ABSTRACT Title of Dissertation: UNDERSTANDING CHANGES IN CONTRACEPTIVE INTENTION, KNOWLEDGE, AND ATTITUDES IN THE CONTEXT OF THE DELCAN INITIATIVE TO REDUCE UNINTENDED PREGNANCY Izidora Skra?i?, Doctor of Philosophy, 2022 Dissertation directed by: Associate Professor Julia R. Steinberg and Associate Professor Amy B. Lewin, Department of Family Science Unintended pregnancies are consistently linked to a higher risk of negative health, social, and economic outcomes for both mother and child. A Delaware public health initiative sought to expand statewide access to all contraceptive methods, particularly IUDs and implants (also known as long-acting reversible contraception (LARC)), due to their high effectiveness, upfront costs, and provision barriers. This study examined changes in self-reported planned contraceptive use, knowledge, and attitudes prior to and following a visit with a medical provider to better understand the provider role in contraceptive outcomes. A diverse group of women (N=474) were recruited at primary care and women?s health Title-X-funded clinics in Delaware. Pre-visit contraceptive use or plan was assessed with two measures: current method use and a composite of current method use and planned method use. Incorporating women?s contraceptive plans in the pre-visit measure resulted in fewer participants being categorized as switching to LARC (2.3%) after a provider visit, compared to the measure that only accounted for current contraceptive use (8.2%). The strongest predictor of changing to a method of higher effectiveness was pre-visit contraceptive choice. On average, women?s knowledge increased; participants with lower pre-visit knowledge were more likely to improve in knowledge post-visit (p<0.001). On average, positive attitudes about LARC decreased, although some individual items changed in the positive direction while others changed in the negative direction. Those with more versus less positive attitudes before the clinic visit had larger decreases in positive attitudes (p<0.001). The full rollout of the DelCAN initiative was associated with an increase in LARC knowledge, while its association with change in LARC attitudes and effectiveness level of planned method use was mixed. These findings suggest that measuring contraceptive plans as opposed to only current method use before a visit is important when applying a pre-post visit design to evaluations of contraceptive use or plans; broadening the conceptualization and measurement of pre-visit contraceptive use or plans could better capture the sources of change that may manifest in post-visit. Additionally, clinic visits may serve as effective education events, particularly for women with lower contraceptive knowledge, and they may provide a more realistic understanding of different contraceptive methods? advantages and disadvantages. It is possible that the lack of increase in positive contraceptive attitudes may be attributable to the negativity bias following changes in knowledge and personal experience, but more research is needed to replicate and understand the phenomenon. UNDERSTANDING CHANGES IN CONTRACEPTIVE INTENTION, KNOWLEDGE, AND ATTITUDES IN THE CONTEXT OF THE DELCAN INITIATIVE TO REDUCE UNINTENDED PREGNANCY by Izidora Skra?i? Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2022 Advisory Committee: Associate Professor Julia R. Steinberg, Co-Chair Associate Professor Amy B. Lewin, Co-Chair Associate Professor Michel H. Boudreaux Professor Kevin M. Roy Associate Professor Marie E. Thoma ?Copyright by Izidora Skra?i? 2022 Acknowledgements This dissertation would not have come to fruition without the many people who mentored, inspired, and supported me along the way. To my mentor and co-chair, Dr. Julia R. Steinberg, whose ceaseless enthusiasm and immense patience kept me afloat as I navigated the perpetually stormy seas of my byzantine dataset. Thank you for persistently teaching by example the value of thorough research and for equally supporting my professional and personal endeavors. To my mentor and co-chair, Dr. Amy Lewin, whose unwavering support served as a much-needed compass throughout the myriad changes to my dissertation aims. Thank you for always expressing your confidence in my ability to do anything I set my mind to. To my committee member, Dr. Marie Thoma, whose wealth of up-to-date statistical skills and literature knowledge was a lifeboat I could always count on. Thank you for always having an unclaimed research topic up your sleeve. To my committee member, Dr. Michel Boudreaux, whose advice was always prompt, constructive, and reasonable. I could not have wished for a more supportive Dean's Representative. To my committee member, Dr. Kevin Roy, whose insistence on grounding my aims in theory has made this dissertation stronger. Thank you for bringing me to UMD and onto the DelCAN evaluation and FMSC 381 teams. To the Maryland Population Research Center for funding my dissertation research and to Jennifer Doiron, in particular, for being such a team player and embracing me as a member of the DelCAN evaluation team. Thank you. To the department of Family Science for supporting me throughout my time at UMD, especially staff members, Nacie Grigsby, Leslie Davis, Kendyl Oliver, and Doris Richardson, each of whom were instrumental in helping me navigate a distinct facet of graduate student life?you are the real MVPs! To Jody Heckman-Bose and Bijoya Chakraborty of International Students and Scholars Services (ISSS), who truly care about students and always went above and beyond to assist me. We are so lucky to have you! To Dr. Linda Macri of the Graduate School Writing Center, who genuinely wants students to succeed, and to Dr. Keisha Allan, who got me to write every Monday morning. Thank you for your commitment to the human layer of the dissertation process. To all of my amazing PhD siblings in arms, who inspired me every day and made this marathon bearable. To my work hubby, Dane De Silva, for his unconditional and non- judgmental support. To my work wife, Debbie Shelef, for her sage advice and unwavering optimism. To my OG study buddy friends, Amara Channell Doig and Sacha St-Onge Ahmad, for ii keeping me accountable in the kindest ways. To my dissertation buddy, Ashley Pantaleao, for being my focus buddy, break buddy, yoga buddy, jigsaw puzzle buddy, ACOTAR buddy, useful resources buddy, and many-titles-yet-to-come buddy. And to everyone else who shared their lives, experiences, resources, time, and wisdom with me, enriching this journey beyond measure (Laura Drew, Ally Pakstis, Yassaman Vafai, Jenni Young, Shawn Kim, Julie Fife, Lauren Ramsey, Laura Golojuch, Kecia Ellick, Jessica Gleason, Yuki Lama, Naheed Ahmed, Andrew Conway, Ronneal Mathews, Matt Rodriguez, Tasha Williams, Gaby Barber, Martha Yumiseva, Jingshuai Du, Lindsay Mallick, Sahra Ibrahimi, Isha Chawla, Rebecca Foss, Krystle McConnell, Sun Cho, Michelle Jasczynski, Juliana Mu?oz)?I am forever grateful <3 To my international friends, Rianna Murray, Nadine Dangerfield, Naette Lee, and Stephanie Cork, who helped me navigate rigid systems and linked me to resources. Thank you for celebrating my every achievement like it was your own and for making me feel understood even when I felt like the world was out to get me. To my CHUM roommates and friends, who played key roles in helping me transition to life in Maryland. Thank you for welcoming me with open arms, exercising equity, and expanding my culinary repertoire. To my roommate, Kalin Schultz, who embraced me with all my quirks and without judgment. Thank you for the endless conversations. To Dr. Yi-Jiun Lin, who introduced me to mindfulness and self-compassion. Thank you for always knowing what to say. To my parents, brother, and partner for cheering me on every step of the way, korak po korak. To the participants of the DelCAN Title X study, whose every experience, knowledge, and feeling about contraception was asked to be reported in the looooooong Title X survey. Thank you for completing the questionnaire prior to *and* following the provider visit. This dissertation would literally have been impossible without their honesty and generosity. iii Table of Contents Acknowledgements ........................................................................................................................ iii Table of Contents ........................................................................................................................... iv List of Tables ............................................................................................................................... viii List of Figures ................................................................................................................................ xi Chapter 1: Introduction ................................................................................................................... 1 1.1. Background and Significance............................................................................................... 1 1.2. Specific Aims and Hypotheses ............................................................................................. 4 1.2.1. Aim 1 ............................................................................................................................. 5 1.2.2. Aim 2 ............................................................................................................................. 6 1.2.3. Aim 3 ............................................................................................................................. 7 1.3. A Note on Gendered Language ............................................................................................ 8 Chapter 2: Literature Review .......................................................................................................... 9 2.1. Unintended Pregnancy ......................................................................................................... 9 2.1.1. Outcomes of Unintended Pregnancy ........................................................................... 11 2.1.2. Disparities in Unintended Pregnancy .......................................................................... 14 2.2. Contraception ..................................................................................................................... 16 2.2.1. Contraceptive Methods ................................................................................................ 18 2.2.1.1. Effectiveness of Contraceptive Methods .............................................................. 19 2.2.1.2. Prevalence of Contraceptive Use and Method Types ........................................... 21 2.3. Contraceptive Behaviors .................................................................................................... 23 2.3.1. Contraceptive Choice .................................................................................................. 24 2.3.1.1. Contraceptive Counseling and Contraceptive Choice .......................................... 26 2.3.2. Contraceptive Use Over Time ..................................................................................... 32 2.3.3. Understanding Disparities in Unintended Pregnancy and Contraceptive Use ............ 33 2.4. Contraceptive Knowledge and Attitudes............................................................................ 36 2.4.1. Contraceptive Counseling and Contraceptive Knowledge .......................................... 37 2.4.2. Contraceptive Counseling and Contraceptive Attitudes .............................................. 45 2.4.3. Summary ...................................................................................................................... 46 2.5. Contraceptive Affordability ............................................................................................... 46 iv 2.6. City and State-wide Interventions ...................................................................................... 48 2.6.1. The Colorado Family Planning Initiative .................................................................... 49 2.6.2. St. Louis Contraceptive CHOICE Project ................................................................... 51 2.6.3. Iowa Initiative to Reduce Unintended Pregnancies ..................................................... 52 2.6.4. HER Salt Lake Contraceptive Initiative ...................................................................... 53 2.6.5. Delaware Contraceptive Access Now (DelCAN) ....................................................... 54 2.6.6. Summary ...................................................................................................................... 57 2.7. Conceptual Framework ...................................................................................................... 58 2.7.1. Integrated Behavioral Model (IBM) ............................................................................ 58 2.7.2. Reproductive Justice .................................................................................................... 62 2.8. Current Study ..................................................................................................................... 65 Chapter 3: Methods ....................................................................................................................... 67 3.1. Data Source and Population ............................................................................................... 67 3.2. Analytic Sample ................................................................................................................. 68 3.3. Variables............................................................................................................................. 69 3.3.1. Aim 1 Outcomes .......................................................................................................... 69 3.3.1.1. Change to Long-Acting Reversible Contraception (LARC) ................................ 73 3.3.1.2. Change to a Higher and Lower Effectiveness Level of Contraceptive Method ... 74 3.3.2. Aim 2 Outcomes .......................................................................................................... 74 3.3.2.1. Change in IUD Knowledge ................................................................................... 75 3.3.2.2. Changes in Implant Knowledge ............................................................................ 77 3.3.2.3. Changes in Effectiveness Knowledge ................................................................... 79 3.3.2.4. Changes in Knowledge about the Benefits of the DelCAN Initiative .................. 80 3.3.3. Aim 3 Outcomes .......................................................................................................... 81 3.3.3.1. Attitude Variables ................................................................................................. 82 3.3.3.2. Scoring Attitudes .................................................................................................. 83 3.3.3.3. Changes in Attitudes ............................................................................................. 83 3.3.4. Predictors ..................................................................................................................... 84 3.4. Analyses ............................................................................................................................. 87 3.4.1. Aim 1 Analyses ........................................................................................................... 87 v 3.4.2. Aim 2 and Aim 3 Analyses.......................................................................................... 88 Chapter 4: Results ......................................................................................................................... 90 4.1. Sample Description ............................................................................................................ 90 4.2. Univariate Analyses ........................................................................................................... 90 4.2.1. Sociodemographic, Reproductive Health, and Provider Visit Measures .................... 90 4.2.2. Current and Planned Contraceptive Use ...................................................................... 91 4.2.3. Contraceptive and DelCAN Knowledge Scores .......................................................... 92 4.2.4. Contraceptive Attitude Scores ..................................................................................... 93 4.3. Bivariate Analyses.............................................................................................................. 94 4.3.1. Differences by Effectiveness Level of Current and Planned Method Use .................. 95 4.3.2. Differences by Contraceptive and DelCAN Knowledge Scores ................................. 96 4.3.3. Differences by Contraceptive Attitude Scores ............................................................ 98 4.4. Change Analyses ? Tests of Hypotheses.......................................................................... 100 4.4.1. Aim 1 ......................................................................................................................... 100 4.4.1.1. Aim 1.1: Overall Changes in Planned Method Use by Effectiveness Levels ..... 100 4.4.1.2. Aim 1.2: Factors that Predict Change in Effectiveness Level of Planned Contraceptive Use ............................................................................................................ 102 4.4.1.3. Aim 1.3: DelCAN Expansion as a Predictor of Change in Effectiveness Level of Planned Contraceptive Use .............................................................................................. 105 4.4.1.4. Aim 1 Summary .................................................................................................. 105 4.4.2. Aim 2 ......................................................................................................................... 106 4.4.2.1. Aim 2.1: Overall Changes in Contraceptive and DelCAN Knowledge .............. 106 4.4.2.2. Aim 2.2: Factors that Predict Change in Contraceptive and DelCAN Knowledge .......................................................................................................................................... 108 4.4.2.3. Aim 2.3: DelCAN Expansion as a Predictor of Change in Contraceptive and DelCAN Knowledge ........................................................................................................ 111 4.4.2.4. Aim 2 Summary .................................................................................................. 111 4.4.3. Aim 3 ......................................................................................................................... 112 4.4.3.1. Aim 3.1: Overall Changes in Contraceptive Attitudes ....................................... 112 4.4.3.2. Aim 3.2: Factors that Predict Change in Contraceptive Attitudes ...................... 113 vi 4.4.3.3. Aim 3.3: DelCAN Expansion as a Predictor of Change in Contraceptive Attitudes .......................................................................................................................................... 116 4.4.3.4. Aim 3 Summary .................................................................................................. 116 Chapter 5: Discussion ................................................................................................................. 118 5.1. Summary of Findings ....................................................................................................... 119 5.1.1. Planned Contraceptive Use ........................................................................................ 119 5.1.2. Contraceptive Knowledge ......................................................................................... 124 5.1.3. Contraceptive Attitudes ............................................................................................. 130 5.1.4. DelCAN Expansion ................................................................................................... 139 5.2. The Integrated Behavioral Model and its Connection to the Findings ............................ 141 5.3. Implications for Practice and Public Health ..................................................................... 145 5.4. Limitations ....................................................................................................................... 152 5.5. Directions for Future Research ........................................................................................ 155 Tables and Figures ...................................................................................................................... 157 Appendix I: Supplementary Analyses ........................................................................................ 224 Supplementary Analyses for Aim 1.2: Factors that Predict Change in Effectiveness Level of Planned Contraceptive Use ..................................................................................................... 224 Supplementary Analyses for Aim 1.3: DelCAN Expansion as a Predictor of Change in Effectiveness Level of Planned Contraceptive Use ................................................................ 224 Appendix II: Supplementary Tables ........................................................................................... 226 References ................................................................................................................................... 250 vii List of Tables Table 1. Baseline characteristics of the sample and their bivariate relationships with effectiveness of contraceptive method(s) currently using (current) and currently using and planning to use (composite) pre-visit. ..................................................................................... 159 Table 2. Baseline characteristics of the sample and their bivariate relationships with effectiveness of contraceptive method(s) planning to use as reported post-visit. ................... 162 Table 3. Contraception and DelCAN knowledge score means and proportions per item prior to and following the provider visit. ............................................................................................. 164 Table 4. Means of contraceptive attitude score prior to and following the provider visit among the portion of the sample who had both pre- and post-visit data. Data for participants who only received attitude questions post-visit are presented separately (rightmost column). .............. 166 Table 5. IUD knowledge means by study variables. ............................................................... 168 Table 6. Implant knowledge means by study variables. ......................................................... 170 Table 7. Contraceptive effectiveness knowledge by study variables. ..................................... 172 Table 8. Knowledge about DelCAN benefits by study variables. .......................................... 174 Table 9. IUD attitude scores by study variables...................................................................... 176 Table 10. Implant attitude scores by study variables .............................................................. 179 Table 11. Hormonal birth control attitude scores by study variables...................................... 182 Table 12. Condom attitude scores by study variables ............................................................. 184 Table 13. Proportions of the sample?s contraceptive methods prior to and following the provider visit by effectiveness levels. ..................................................................................... 189 Table 14. Adjusted odds ratios and 95% confidence intervals for the associations between pre- visit method and study variables and changing to LARC post-visit. ...................................... 190 Table 15. Adjusted ratios and 95% confidence intervals for the associations between changing to a method of higher effectiveness level post-visit and predictor variables, including current and composite pre-visit method. ............................................................................................. 193 Table 16. Adjusted ratios and 95% confidence intervals for the associations between changing to a method of lower effectiveness level post-visit and predictor variables, including current and composite pre-visit method. ............................................................................................. 196 Table 17. Adjusted relationships between change in IUD knowledge and study variables, using a continuous change outcome (percentage point increase in knowledge) and a dichotomous change outcome (increase in knowledge or not) ..................................................................... 203 viii Table 18. Adjusted relationships between change in implant knowledge and study variables, using a continuous change outcome (percentage point increase in knowledge) and a dichotomous change outcome (increase in knowledge or not) ............................................... 206 Table 19. Adjusted relationships between change in effectiveness knowledge and study variables, using two dichotomous change outcomes?any increase in score and increase to perfect score ............................................................................................................................ 208 Table 20. Adjusted relationships between change in knowledge about DelCAN benefits and study variables, using continuous and dichotomous outcomes ............................................... 210 Table 21. Adjusted relationships between change in IUD attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) ... 216 Table 22. Adjusted relationships between change in implant attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) ............................................................................................................................... 218 Table A1. Bivariate relationship between pre-visit current method use and post-visit planned method use............................................................................................................................... 226 Table A2. Bivariate relationship between pre-visit composite current-planned method use and post-visit planned method use ................................................................................................. 227 Table A3. Adjusted relationship between change to LARC and study variables, including current method use, and its corresponding sensitivity analysis for the wave one skip pattern issue. ........................................................................................................................................ 228 Table A4. Adjusted relationship between change to LARC and study variables, including the composite current-planned method use, and its corresponding sensitivity analysis for the wave one skip pattern issue. ............................................................................................................. 231 Table A5. Adjusted relationship of reporting LARC use post-visit compared to any other method and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. .................... 234 Table A6. Adjusted relationship of reporting LARC use post-visit compared to none/low effective methods and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. 237 Table A7. Adjusted relationship between reporting LARC use post-visit compared to moderately effective methods and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. ............................................................................................................................ 240 Table A8. Adjusted relationship between reporting moderately effective methods post-visit compared to none/low effective methods and study variables, including the current and ix composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. .................................................................................................... 243 Table A9. Adjusted relationships between change in effectiveness knowledge and study variables, using one continuous and two dichotomous change outcomes. ............................. 246 Table A10. Proportions of the sample?s reporting of contraceptive decision-making following the provider visit...................................................................................................................... 249 x List of Figures Figure 1. Integrated Behavioral Model. .................................................................................. 157 Figure 2. Diagram of analytic sample selection. ..................................................................... 158 Figure 3. Distribution of contraceptive change by LARC based on pre-visit current and pre- visit composite variables. ........................................................................................................ 187 Figure 4. Distribution of contraceptive change by effectiveness levels based on pre-visit current and pre-visit composite variables................................................................................ 188 Figure 5. Distribution of change in IUD knowledge from pre-visit to post-visit.................... 199 Figure 6. Distribution of change in implant knowledge from pre-visit to post-visit. ............. 200 Figure 7. Distribution of change in effectiveness knowledge from pre-visit to post-visit. ..... 201 Figure 8. Distribution of change in knowledge about DelCAN benefits from pre-visit to post- visit. ......................................................................................................................................... 202 Figure 9. Change in IUD attitude scores from pre-visit to post-visit. ..................................... 212 Figure 10. Change in implant attitude scores from pre-visit to post-visit. .............................. 213 Figure 11. Change in hormonal birth control attitude scores from pre-visit to post-visit. ...... 214 Figure 12. Change in condom attitude scores from pre-visit to post-visit. ............................. 215 xi Chapter 1: Introduction 1.1. Background and Significance Almost half of all pregnancies in the US are unintended; an unintended pregnancy is broadly defined as a pregnancy that is mistimed or unwanted by the pregnant person (Finer & Zolna, 2016). Unintended pregnancies are linked to an increased likelihood of negative health outcomes for both mother and child, as well as to adverse social and economic consequences (Gipson et al., 2008a; Huynh et al., 2013; Sonfield et al., 2013; Tsui et al., 2010; Yazdkhasti et al., 2015). Furthermore, analyses demonstrate substantial savings in overall cost and public spending if unintended pregnancies were to be prevented (Monea & Thomas, 2011; Sonfield et al., 2011; Trussell et al., 2009). Last but not least, a rights-based approach emphasizes the need for reproductive autonomy; women and couples should have the ability to freely choose if and when to have children (Senderowicz, 2020; Sonfield et al., 2013; World Health Organization, 2016). Racial, ethnic, and socioeconomic disparities exist in adverse maternal and child health outcomes (Curtin & Hoyert, 2017; Lang & King, 2008) and in contraceptive use and choice (Daniels & Abma, 2020; Dehlendorf, Park, et al., 2014). Literature consistently demonstrates that women from minority communities and those of lower socioeconomic status have reduced access to contraceptive services and, overall, encounter (sometimes subconscious) biases in their interactions with reproductive health providers that their White counterparts, and those of higher socioeconomic standing, do not (D. Becker & Tsui, 2008; Borrero et al., 2008; Dehlendorf, 1 Rodriguez, et al., 2010; Downing et al., 2007; Harrison & Cooke, 1988), thus, impacting their contraceptive use and preferences. Most pregnancies can be prevented with the effective use of modern contraceptive methods, but not all methods have the same effectiveness level (Trussell et al., 2018). Therefore, contraceptive method type being used is an important determinant of unintended pregnancy (Jaccard et al., 1996; Sonfield et al., 2014; Sundaram et al., 2017). Trussell and Vaughan (1999) estimate that a typical woman using reversible methods will initiate and discontinue a method almost ten times over the course of her 30-year reproductive lifespan, excluding the time she spends trying to get pregnant, pregnant, and post-partum. This suggests that making decisions and implementing related behaviors concerning contraceptive choice, consistency, accuracy, continuity, and switching is an ongoing process that a woman renegotiates repeatedly over her lifespan, and should be approached as such (Downey et al., 2017; Frost et al., 2007b; Grady et al., 2002; R. K. Jones et al., 2015; Marshall et al., 2018; Moreau et al., 2007). Furthermore, contraceptive behaviors do not occur as isolated decisions by individuals and their support systems, but they transpire in the context of health care systems (Dehlendorf, Rodriguez, et al., 2010; Frost & Darroch, 2008; Sonfield et al., 2014). The literature posits that the relatively high rates of unintended pregnancy (Finer & Zolna, 2016) and unmet need for contraception (Frederiksen et al., 2017) in the United States are due to a combination of patient preferences and behaviors, health care system factors, and provider-related factors (Dehlendorf, Rodriguez, et al., 2010). Beyond these three elements, wider systemic factors also play a role; lower socioeconomic status and membership in a minority racial or ethnic group have been 2 found to have independent adverse effects on health (LaVeist, 2005; Williams et al., 2016). Epigenetics and ?weathering? are two models that explain part of the link between social advantage and worse reproductive health outcomes (Frazier et al., 2018; Geronimus, 1992, 1996; Kim et al., 2020; Owen et al., 2013). State- and city-wide initiatives in Colorado, Iowa, Utah, and Missouri have been successful in reducing unintended pregnancy by addressing various barriers pertaining to health care system factors that have impeded access to contraceptive care (Biggs et al., 2015; Kelly et al., 2020; McNicholas et al., 2014; Sanders et al., 2018). Success was measured by reductions in pregnancy, birth, and abortion rates among adults and adolescents, and an increased uptake and continued use of long-acting reversible contraception (Sawhill & Guyot, 2019). In the early 2010s, Delaware had the highest rate of unintended pregnancy in the nation (Kost, 2015; Kost et al., 2018). Consequently, in 2015, it launched its own statewide initiative, the Delaware Contraceptive Access Now (DelCAN) to reduce the rate of unintended pregnancies by increasing access to family planning services for all women of reproductive age. The program focused on (1) policy change and implementation at the state level; (2) a statewide public awareness campaign; and (3) statewide clinician and staff trainings (Choi et al., 2019). This investigation seeks to explore planned contraceptive use, knowledge, and attitudes among Title X patients in Delaware during the time of implementation of the DelCAN initiative. Given that contraceptive intent and choice, i.e., deciding if and which method to use, plays a crucial role in one?s level of protection against unintended pregnancy (Trussell et al., 2018), examining the factors that influence women?s choice and use of contraceptive method is relevant 3 to explore. While contraceptive choice following a clinic visit has been the subject of some prior research (Harper et al., 2015; Secura et al., 2010), according to the author?s knowledge, no studies have used a pre-post-visit design to assess contraceptive choice of all available methods; this study aims to do so using two waves of data from multiple Title X clinics across Delaware at the time of the DelCAN initiative. Secondly, this investigation examines women?s knowledge and attitudes about long-acting reversible contraceptive methods prior to and after visiting with a health care provider in order to assess potential changes in these outcomes, and thus, offer some insight about the benefits of the statewide initiative, including possible effects on patient preferences and behaviors. Thirdly, the conceptualization of this study is based on the Integrated Behavioral Model; the tenets of Reproductive Justice are used to help interpret some of the findings. 1.2. Specific Aims and Hypotheses There are three main aims to this investigation: to investigate changes in pre- and post- clinic visit (1) planned contraceptive use, (2) contraceptive knowledge, and (3) contraceptive attitudes among a diverse group of Delaware patients visiting Title X clinics during the time of the DelCAN initiative. Each of the three aims were carried out by describing changes from pre- visit to post-visit and examining the factors that predicted the change. We also investigate whether the changes differed by the length of time DelCAN had been implemented. 4 1.2.1. Aim 1 1.1. To describe changes in the effectiveness level of contraceptive method: a. women are currently using before, and planning to use after, a Title X provider visit, and b. women are planning to use before and after a Title X provider visit. 1.2. To examine factors that predict changes in effectiveness level of contraceptive method: a. women are currently using before and planning to use after a Title X provider visit, and b. women are planning to use before and after a Title X provider visit. 1.3. To investigate the extent to which the expansion of the DelCAN initiative was associated with change in effectiveness level of contraceptive method: a. women are currently using before and planning to use after a Title X provider visit, and b. women are planning to use before and after a Title X provider visit. Hypothesis 1.1: The percentage of Delaware Title X patients who report planning to use a higher effectiveness contraceptive method after the provider visit will be larger than the percentage who are (a) currently using and (b) planning to use higher effectiveness contraceptive methods before the visit. 5 Hypothesis 1.2: Participants are more likely to change to methods of a higher effectiveness level following a provider visit if pre-visit they report (a) using or (b) planning to use no methods or low effectiveness methods. In both (a) and (b) models, participants will be more likely to change to a higher effectiveness level if they are younger, unmarried, or do not want to become pregnant in the next two years. Hypothesis 1.3: The extent of change in effectiveness level of method choice will be larger in the second wave (2018-2019) than in the first (2017) in both (a) and (b) models. 1.2.2. Aim 2 2.1. To describe changes from before to after a Title X provider visit in self-reported contraceptive knowledge. 2.2. To examine what factors predict changes in contraceptive knowledge from pre to post provider visit. 2.3. To investigate the extent to which the expansion of the DelCAN initiative was associated with changes in contraceptive knowledge from pre to post provider visit. Hypothesis 2.1: After a visit with a provider, Delaware Title X patients will have higher levels of self-reported knowledge about long-acting reversible contraception (LARC) than they did before the visit. Hypothesis 2.2: Those without health insurance and those with less knowledge about LARC pre- visit, will gain more knowledge about LARC 6 Hypothesis 2.3: Changes in knowledge will be larger in the second wave (2018-2019) than in the first (2017). 1.2.3. Aim 3 3.1. To describe changes in self-reported contraceptive attitudes from before to after a Title X provider visit. 3.2. To examine what factors predict changes in contraceptive attitudes from pre to post provider visit. 3.3. To investigate the extent to which the expansion of the DelCAN initiative was associated with changes in contraceptive attitudes from pre to post provider visit. Hypothesis 3.1: Post-visit with a provider, Delaware Title X patients will have more positive attitudes towards long-acting reversible contraception (LARC). Hypothesis 3.2: Those with more negative attitudes about LARC pre-visit will experience more gains in positive attitudes about LARC post-visit. Compared to White women, Black women?s gains in positive attitudes will be smaller. Hypothesis 3.3: Gains in positive will be larger in the second wave (2018-2019) than in the first (2017). 7 1.3. A Note on Gendered Language Unintended pregnancy may be experienced by both women and gender-diverse individuals. As a result, contraceptive care is relevant to all people at risk of unintended pregnancy, regardless of gender identity. In addition to ?woman? and ?she,? I also use ?person,? ?individual,? ?people,? and ?they.? I chose to incorporate both gendered and gender-neutral language in this dissertation because reproductive rights continue to be undermined precisely because they primarily affect women. Issues of reproductive health cannot be divorced from our society?s gendered lens because it informs policy at all levels. The allocation of a society?s resources for reproductive health plays a key role in maternal and child health outcomes. 8 Chapter 2: Literature Review 2.1. Unintended Pregnancy The most recent national data find that 45% of all pregnancies were unintended in 2011, down from 51% in 2008. This amounts to 45 unintended pregnancies per 1000 women aged 15? 44 in 2011, compared to 54 per 1000 in 2008 (Finer & Zolna, 2016). Compared to other developed countries, the United States maintains the highest rate of unintended pregnancy per 1000 women aged 15?44 (Bearak et al., 2018; Sedgh et al., 2014). In literature and in policy, an unintended pregnancy (UIP), also referred to as an unplanned pregnancy, is most frequently defined as a mistimed or an unwanted pregnancy (Guttmacher Institute, 2019). Mistimed pregnancies are operationalized as those occurring at an earlier time than desired, and unwanted pregnancies as those occurring when the person did not want to become pregnant ever or ever again. In contrast, intended pregnancies are those that occur at the time that they are desired or later than desired (Finer & Zolna, 2016; Guttmacher Institute, 2019). In reports based on nationally representative data examining rates of unintended pregnancy, people who state that they are indifferent or unsure about becoming pregnant are usually grouped into the category of their pregnancy being intended (Finer & Henshaw, 2006; Mosher et al., 2012). Unintended pregnancies, like all pregnancies, may result in a live birth, stillbirth, abortion, or fetal loss (Trussell et al., 2009). It should also be noted that the dominant categorization of pregnancy intention status into intended (i.e., on time, later than wanted, indifferent, and unsure) and unintended (i.e., mistimed 9 and unwanted) pregnancies does not represent the full range of an individual?s pregnancy desires despite decades of literature using that categorization (Finer & Zolna, 2016; Guttmacher Institute, 2019; Trussell & Vaughan, 1999). Individuals whose pregnancy intentions are not analogous may be grouped together when examining pregnancy-related behaviors or maternal and child health outcomes. While many studies distinguish between wanted, mistimed and unwanted pregnancies (D?Angelo et al., 2004; Everett et al., 2017; Hartnett & Margolis, 2019; Joyce et al., 2000a), analyses of nationally representative surveys that measure pregnancy intention, such as the Pregnancy Risk Monitoring Assessment System (PRAMS) and the National Survey of Family Growth (NSFG) frequently conflate ?wanted? with ?intended? pregnancies, representing experiences that may not necessarily overlap; this may erroneously influence study conclusions (Kost et al., 2018; Mumford et al., 2016). Additionally, research has been exploring women?s reports of happiness about a pregnancy despite it being unintended (Aiken et al., 2015), as well as experiences of pregnancy ambivalence (Borrero et al., 2015; Higgins, 2017; Patel et al., 2014; Sundstrom et al., 2017; Yoo et al., 2014); it is still unknown whether such phenomena might exhibit more nuanced effects on maternal and child health outcomes that may be lost when grouped in the dominant dichotomy. Although the data used in this dissertation were collected with surveys using the dichotomous intendedness option, and therefore, this work retains the operationalization about prior pregnancy intention as used by nationally representative surveys, we also include predictors such as future pregnancy intention and feelings about a hypothetical pregnancy in the next year in an attempt to capture some of that nuance that is lost with a binary intendedness variable. 10 2.1.1. Outcomes of Unintended Pregnancy The clearest sign of unintended pregnancy (UIP) being recognized as a public health issue is its inclusion in the Healthy People initiative, starting in 2000 (Klerman, 2000) and every decade since. The national health objective is to decrease the proportion of pregnancies that are unintended among women aged 15 to 44 down to 36.5 percent by 2030 (Office of Disease Prevention and Health Promotion, 2020). This addition to the public health agenda was based on copious literature finding that UIPs resulting in live births are linked to higher risk of adverse health outcomes. Many adverse maternal and child health outcomes are associated with unintended pregnancy. First, UIPs are less likely than intended ones to be detected early in pregnancy, which has been associated with delayed uptake of prenatal care (Cheng et al., 2009; Dibaba et al., 2013; Kost & Lindberg, 2015; Mayer, 1997). Second, recent studies find that infants born as a result of UIPs are more likely to be born preterm (Mohllajee et al., 2007; Orr et al., 2000; Shah et al., 2011) and of low birth weight (J. A. Hall et al., 2017; Kost & Lindberg, 2015; Shah et al., 2011), and less likely to be breastfed (Dye et al., 1997; Kost & Lindberg, 2015; Rubin & East, 1999; J. S. Taylor & Cabral, 2002). Third, research examining the association between UIP and maternal risk behaviors, such as smoking and drinking during pregnancy, find positive, mixed, or no effects (Cheng et al., 2009; Dott et al., 2010; Hellerstedt et al., 1998; Joyce et al., 2000a, 2000b; Korenman et al., 2002; Kost et al., 1998; Marsiglio & Mott, 1988). Fourth, some longitudinal studies have found UIP to be associated with increased risk of mental health issues for the 11 children, even as far-reaching as young adulthood (Axinn et al., 1998; Barber et al., 1999). Fifth, Foster et al. (2019) found that children born to women who were denied abortion were more likely to have lower mean child development scores than children born to women who had obtained an abortion prior to the child?s birth. However, a rare study of discordant siblings found little association between pregnancy intentions and child physical and cognitive outcomes after the corresponding siblings were included in the model (Joyce et al., 2000a). Additional adverse health outcomes specific to the pregnant person are as follows. People who give birth and subsequently raise children from an unplanned pregnancy demonstrate higher risks for short and long-term mental health problems, such as postpartum depression, depression, anxiety, and lower levels of happiness (Abajobir et al., 2016; Barber et al., 1999; Cheng et al., 2009; Gipson et al., 2008b; Herd et al., 2016; McCrory & McNally, 2013). Furthermore, any pregnancy, regardless of intention, is an inherently risky biological process, whose potential dangers would be avoided had the person not become pregnant (Campbell & Graham, 2006). First, merely by virtue of occurring, pregnancies involve the potential for hemorrhage, (pre)eclampsia, and infection, which lead to maternal morbidity or mortality (Callaghan, 2012; Creanga et al., 2017; Danel et al., 2003; Goffman et al., 2007). Second, while national level data show that obtaining a legal abortion in the United States entails a much lower likelihood of maternal mortality than childbirth (Raymond & Grimes, 2012), it still entails a health risk compared to not being pregnant at all. Health outcomes may be most salient to public health, but literature consistently demonstrates that social and economic characteristics are closely tied to health outcomes. 12 Therefore, associations between UIP and social and economic outcomes are also relevant to briefly highlight. For instance, studies have found associations between unplanned births and reports of lower relationship happiness and more relationship conflict (Sonfield et al., 2013). In terms of economic consequences, analyses of historical trends link the advent of the pill and women?s corresponding ability to delay childbearing to increased numbers of women pursuing advanced professional degrees; the level of formal educational attainment is associated with women?s future income (Sonfield et al., 2013). Furthermore, children born to women who were denied abortion were more likely to live in poverty than children born to women who had previously sought and obtained an abortion (Foster et al., 2019). The macro-economic perspective is also of concern to public health. Advancing population health occurs within the framework of allocating limited funds to a myriad of public health priorities. Therefore, public spending on UIP is also a direct public health concern. While estimates of costs vary by study methodology and calendar year, two separate investigations using different data found the annual costs of publicly funded births resulting from an UIP to be $11 billion and $12.1 billion, covering an estimated 780,000 and 1 million births, respectively (Monea & Thomas, 2011; Sonfield et al., 2011). In addition to births, some studies also include fetal losses, abortions, and ectopic pregnancies in their cost calculations, and their savings estimates also account for the fact that if mistimed births were prevented, they would still occur in the future (Monea & Thomas, 2011; Trussell et al., 2009). The literature is unanimous that both the overall and the publicly financed medical costs of UIP far eclipse the costs of providing contraceptive care (Frost et al., 2008; Monea & Thomas, 2011; Trussell et al., 2009). It should be 13 noted that this concern about public spending should continue to focus on pregnancies that are unintended, as an unwanted or mistimed health event in a woman?s life that can safely be prevented by modern contraceptive methods. Utilizing contraceptive methods as a policy tool to prevent pregnancies in order to reduce overall public spending or other population-level goals raises issues of contraceptive autonomy, medical ethics, and reproductive justice (Kaneshiro et al., 2020; Roberts, 2015; Senderowicz, 2020). 2.1.2. Disparities in Unintended Pregnancy Neither UIP nor its outcomes are equally distributed among the U.S. population. Despite consistent decreases in most adverse pregnancy-related outcomes across various groups, disparities between populations persist (Finer & Zolna, 2016). People of lower socioeconomic status and lower formal educational attainment, people of color, as well as adolescent girls, young women, and unmarried women have higher rates of UIP (Finer & Zolna, 2016; Koren & Mawn, 2010), although newer findings suggest the association between race and ethnicity and UIP may not be as clear as previously thought (Barber et al., 2021; Everett et al., 2020; Kemet et al., 2018). Furthermore, people of lower socioeconomic status and lower formal educational attainment, Blacks, and Hispanics have higher rates of UIPs ending in births (Finer & Zolna, 2016; Musick, 2002). However, people of lower socioeconomic status and lower formal educational attainment are less likely to have an abortion, while Black and Hispanic women, including adolescents, have higher abortion rates than Whites; unmarried women are also more likely to have an abortion than married women (Finer & Zolna, 2016; Kost et al., 2017). 14 National data indicate that more than one third of all women will experience more than one unintended birth, that rate being the highest among Black women (Wildsmith et al., 2010). Conversely, one in two women having an abortion have already terminated a pregnancy at least once (R. K. Jones et al., 2006). Complementarily, women having an abortion are more likely to become pregnant in the following year compared to women who are not experiencing an abortion (Upadhyay et al., 2012). This pattern of repeat UIP highlights the way that one unintended pregnancy serves as a predictor for future unintended pregnancy, further disadvantaging specific populations. One of the starkest racial disparities in pregnancy-related outcomes in the United States is maternal mortality. Despite maternal mortality being a relatively rare outcome with little change in comparative numbers for the past few decades1 (Hirshberg & Srinivas, 2017; Joseph et al., 2017), women of color are more likely to die of pregnancy-related causes than White women in the United States (Kaunitz et al., 1985; Lang & King, 2008). Furthermore, the data are consistently unequivocal that Black women have the highest maternal mortality ratio of any racial or ethnic group; they have been two to four times more likely to die due to pregnancy- related causes for the past few decades (Centers for Disease Control and Prevention (CDC), 1999; Creanga et al., 2017; Curtin & Hoyert, 2017; Joseph et al., 2017; Mogos et al., 2020). Maternal morbidity is also documented as disproportionately higher for Black people compared to Whites (Curtin & Hoyert, 2017). 1 Recent changes in what constitutes a pregnancy-related death, i.e., maternal mortality, make comparisons between various time periods more challenging. 15 Despite unambiguous racial and ethnic differences in health outcomes, race and ethnicity are not risk factors in and of themselves. Rather, they are likely markers of interrelated social, economic, and cultural factors, as well as of access to and quality of health care, that increase the risk of a negative pregnancy-related outcome for minority, particularly Black, populations (Centers for Disease Control and Prevention (CDC), 1999). In sum, all aforementioned health disparities, which are, by definition, systematic, plausibly avoidable, and based on structural rather than biological differences (Braveman et al., 2011), highlight the need to consider unintended pregnancy as a public health issue. 2.2. Contraception The Centers for Disease Control and Prevention identified family planning as one of the ten greatest public health achievements of the 20th century (Centers for Disease Control and Prevention, 1999). Family planning means that individuals and couples are able to anticipate and attain their desired number of children, even when such number is zero, and freely decide on the spacing and timing of their births (Inter-agency Working Group on Reproductive Health in Crises, 2010). Family planning is achieved through the use of contraceptive methods and the treatment of involuntary infertility, and therefore, the oftentimes synonymous use of the broad concept of ?family planning? with the narrower terms ?contraception? and ?birth control? is incorrect. As the focus of this dissertation is access to contraceptive services, the author will solely use the term ?contraception,? defined as ?any method, medicine, or device used to prevent pregnancy? (Office of Women?s Health, 2017). 16 Contraception is linked to many improvements in health and social outcomes. From a health perspective, access to contraceptive services has reduced the number of infant, child, and maternal deaths. Furthermore, it has contributed to smaller family size and longer interpregnancy intervals between births, as well as increased opportunities for preconception counseling and screening (Centers for Disease Control and Prevention, 1999). From a social and broader societal perspective, the ability to choose if and when to have children has allowed women to complete higher levels of education and increase their status and participation in the labor force; women?s higher earning power has led to a narrowing of the gender pay gap (Sonfield et al., 2013). Educational and economic empowerment has also contributed to increased decision-making power in matters of relationships and family formation, which has been linked to improved mental health indicators (Sonfield et al., 2013). Contraceptive use is a ubiquitous experience in a woman?s life in the United States. According to national data from the National Survey of Family Growth, the percentage of women aged 15-49 who have ever had penile-vaginal intercourse had used at least one contraceptive method at one point in their lifetimes has held steady above 99% since 2006 (Mosher & Jones, 2010; National Center for Health Statistics, 2019a, 2019b). This is expected considering the median ages of women?s key reproductive events the United States; the median age of first penile-vaginal intercourse is 17, first pregnancy is 22.5, first childbirth is 26, intention for no more children is 31, and menopause is 51 (Gold et al., 2009). Therefore, the typical American woman spends about five years trying to get pregnant, pregnant, or post- partum and about thirty years of her reproductive life trying to avoid pregnancy (Gold et al., 17 2009). According to the current paradigm, a woman?s preferences and behavior and her access to contraceptive services are the strongest determinants of whether she will be successful in her attempts to avoid unintended pregnancies (Dehlendorf, Rodriguez, et al., 2010). 2.2.1. Contraceptive Methods Contraception is defined as ?any method, medicine, or device used to prevent pregnancy? (Office of Women?s Health, 2017). However, literature and national data use the term ?contraceptive methods? synonymously with ?contraception.? This dissertation will follow that well-accepted trend, referring to ?contraceptive methods? as an umbrella term that also includes ?medicine? and ?devices.? There are over twenty different types and subtypes of contraceptive methods available today (Mosher & Jones, 2010), most commonly grouped into one of two categories?traditional and modern methods. Traditional methods, also referred to as natural family planning methods, comprise of abstinence, withdrawal, and various fertility-awareness-based methods (incl. calendar, basal body temperature, cervical mucus, and symptothermal methods) (Freundl et al., 2010). Modern contraceptive methods include spermicides, the sponge, external (male) and internal (female) condoms2, the diaphragm, the patch, the vaginal ring, injectables, oral contraceptive pills (incl. emergency contraception3), intrauterine devices (IUDs), subdermal 2 This dissertation will use the terminology of the external and internal condom to refer to the male and female condom, respectively. However, when citing questions and answers from the survey instruments, the exact wording from the questionnaires will be presented. Solely ?condoms? shall refer to external condoms, as the most common barrier method used today. 3 The only emergency contraception approved by the Food and Drug Administration are two types of oral pills (Turok et al., 2021). However, long-standing observational evidence has found the copper IUD to be more effective than pills, leading many organizations to recommend its use as emergency contraception, when appropriate (ACOG 18 implants, as well as female and male sterilization (Trussell et al., 2018). All methods are considered reversible except sterilization. Modern methods can further be categorized into hormonal and non-hormonal contraception, with spermicide, the sponge, condoms, the diaphragm, the copper IUD, and sterilization considered non-hormonal methods, while the others are hormonal. Another relevant categorization is between methods that are coitally dependent, such as spermicides, condoms, the sponge, the diaphragm, the cervical cap, fertility awareness- based methods, and withdrawal, and those that are not (Shoupe, 2020). 2.2.1.1. Effectiveness of Contraceptive Methods Using any contraception has a higher probability of preventing pregnancy than using no method; however, effectiveness rates between contraceptive methods vary widely (Trussell et al., 2018). In measuring effectiveness levels of different contraception, studies distinguish two different rankings: (1) perfect use based on chemical or mechanical properties and assumed correct use of the method at every instance of penile-vaginal intercourse, and (2) typical use based on chemical or mechanical properties and actual use of the method by users during penile- vaginal intercourse, which may be incorrect or inconsistent. The effectiveness rates are measured by the percentage of women experiencing an unintended pregnancy within the first year of use (Trussell et al., 2018). While both rankings are relevant for scientific purposes, typical use is a more accurate representation of women?s risk of unintended pregnancy, and therefore, more relevant in evaluating interventions to reduce unintended pregnancies. Committee on Practice Bulletins?Gynecology, 2015; Cleland et al., 2012; Turok et al., 2021). New evidence suggests that levonorgestrel IUDs may be as effective as the copper IUD for the same purposes (Turok et al., 2021). Unless otherwise noted, in this dissertation ?emergency contraception? refers to its oral pill form only, in line with the usage of the term in the survey instrument used to collect the data for our research. 19 Among women who do not want to become pregnant but engage in penile-vaginal intercourse without using any contraceptive method, 85% will experience an unintended pregnancy within one year; this is also referred to as a failure rate of 85% (Trussell et al., 2018). The data demonstrate that the least effective contraceptive method upon typical use is the sponge, when used by parous women, and the ovulation method, when used by any woman, with failure rates of 27% and 23%, respectively (Trussell et al., 2018). The wide gap between the failure rates of using no method and the least effective methods is further evidence of contraception?s ability to assist women in their family planning. Contraceptive methods are frequently grouped into three effectiveness levels based on the most recent data on their failure rate upon typical use, as presented in Trussell et al. (2018). Methods with a failure rate between 21% and 13% are considered low effective methods, and they include withdrawal (20%) and fertility awareness-based methods (15%), as well as spermicides (21%), external (21%) and internal (13%) condoms, the diaphragm (17%), and the sponge (17%).4 Moderately effective methods, with failure rates between 7% and 4% include oral combined and progestin-only contraceptive pills (7%) , the patch (7%), the ring (7%), and injectables (4%). Highly effective methods, with failure rates below 1%, include intrauterine devices (IUDs), both hormonal (0.1% to 0.7%) and copper (0.8%), the subdermal implant (0.1%), tubal occlusion or female sterilization (0.5%), and vasectomy or male sterilization (0.15%). Implants and IUDs are also referred to as long-acting reversible contraceptives 4 The sponge has a failure rate of 27% for parous women and 14% for nulliparous women; the overall failure rate for the sponge is 17%. Fertility awareness-based methods have an overall failure rate of 15%; however, the rates vary substantially by method, with the ovulation method being the least effective with a failure rate of 23%, while the symptothermal method is the most effective with a failure rate as low as 2% (Trussell et al., 2018). 20 (LARC), whose perfect and typical failure rates are virtually identical because their consistent and correct use is independent of user behavior (Trussell et al., 2018). 2.2.1.2. Prevalence of Contraceptive Use and Method Types The most recent national data from 2017?2019 indicate that 65% of women aged 15?49 are currently using some form of contraception. Data consistently demonstrate that demographic characteristics are associated with contraceptive use. Overall, the youngest age group (15?19) was least likely to use contraception at 38.7%, while the oldest age group (40?49) were most likely to use contraception at 74.8%. Non-Hispanic Whites were more likely to report current use of contraception at 69.2% compared to non-Hispanic Black and Hispanic women, at 61.4% and 60.5% respectively (Daniels & Abma, 2020). Female sterilization is the most common contraceptive method currently used, cited by 18.1% of women. As a non-reversible method, its usage logically increases with age, reaching 39.1% among women aged 40-49. It was almost three times more likely to be used by women without a high school diploma, compared to those with a bachelor?s degree or higher (Daniels & Abma, 2020). The second most cited method are oral contraceptive pills, used by 14% of women. The usage of the pill decreased as women?s age increased, averaging about 20% among women aged 15-29, compared to 6.5% among those 40-49. The pill?s highest uptake continued to be among non-Hispanic Whites at 17.8%, a rate double than among Hispanic and non-Hispanic Black women. Women with a bachelor?s degree or higher used the pill at 18.1%, a percentage more 21 than three times larger than that among women without a high school diploma or equivalent (Daniels & Abma, 2020). LARC was used by 10.4% of women, with IUDs accounting for approximately four fifths of LARC usage (8.4%), and implants only one fifth (2%) (Daniels & Abma, 2020). While IUD uptake has increased moderately but consistently over the past decade, use of the implant appears to be stagnant (Daniels & Abma, 2018; Mosher & Jones, 2010; National Center for Health Statistics, 2019a). However, among adolescents aged 15?19 who have ever had penile-vaginal intercourse, 20% had ever used LARC, whereby the implant accounted for 15% and the IUD for 5% of ever use (Martinez & Abma, 2020). Overall, current LARC usage was highest among women aged 20?39. LARC use did not vary by women?s race or Hispanic ethnicity and only somewhat by educational attainment, where the most educated women were more likely to use LARC (Daniels & Abma, 2020). The external condom was cited by 8.4% of women, with youngest and oldest women being least likely to use it, at 5-6%. Non-Hispanic White women were less likely to use the external condom at 7% compared to Hispanic and non-Hispanic Black women at 10.5% and 11%, respectively. No differences were found between women of different educational attainment (Daniels & Abma, 2020). It is relevant to note that these percentages indicate the most effective method that women use for pregnancy prevention, not the only method. Since external condoms are the most effective method for preventing sexually transmitted infections, overall rates of condom use are higher than 8.7%. 22 2.3. Contraceptive Behaviors High rates of unintended pregnancy and high rates of reported contraceptive use in the same time period (Finer & Zolna, 2016; J. Jones et al., 2012) indicate that, beyond non-use, contraceptive behaviors are important contributors to the occurrence of unintended pregnancy, as evidenced by Trussell et al.'s (2018) differentiation between perfect and typical use of contraceptive methods based on nationally representative data from women of reproductive age. The most recently available population-based estimates find that approximately every other woman of reproductive age will experience an unintended pregnancy at some point in her lifetime (Henshaw, 1998), and one quarter will have an abortion by age 45 (R. K. Jones & Jerman, 2017). Various studies report that contraceptive users account for about half of all unintended pregnancies, of which one in ten is due to method failure and nine out of ten due to inconsistent or incorrect use (Finer & Henshaw, 2006; Frost & Darroch, 2008; Sonfield et al., 2014); this is, again, in line with Trussell et al. (2018). Therefore, beyond increasing access to and uptake of contraceptive methods, as many public health programs aim to do, promoting compliance appears to be just as important for those wanting to avoid pregnancy. Contraceptive behaviors consists of choice, consistency, accuracy, continuation, discontinuation, and switching (Frost et al., 2007b; Jaccard et al., 1996). The data most frequently and easily cited about contraceptive method use are prevalence rates, such as ever use and current use, as presented in earlier sections of this literature review. These data are easy to present to policy makers and they lay the groundwork for other contraceptive research, but they have also obscured the fact that contraceptive behavior is not static. James Trussell & Vaughan 23 (1999) estimate that a typical woman using reversible methods will initiate and discontinue a method almost ten times over the course of her 30-year reproductive lifespan. This suggests that making decisions and effecting related behaviors concerning contraceptive choice, consistency, accuracy, continuity, and switching is an ongoing process that a woman renegotiates repeatedly over her lifetime, and should, therefore, be studied as such (Frost et al., 2007b; Grady et al., 2002; R. K. Jones et al., 2015; Moreau et al., 2007). 2.3.1. Contraceptive Choice Contraceptive choice entails first deciding whether to use a contraceptive method or not, and if yes, choosing what specific method to use (Jaccard et al., 1996). Women?s method choice depends on various factors. While effectiveness in preventing pregnancy is the most frequently cited factor of importance when considering a contraceptive method, safety, side effect profile, ease of use or convenience, protection against sexually transmitted infections, and affordability are also important contraceptive attributes in their decision-making process (Caetano et al., 2020; Donnelly et al., 2014; Downey et al., 2017; Gomez & Clark, 2014; Grady et al., 1999; Harvey et al., 1991; Lamvu et al., 2006; Lessard et al., 2012; Lete et al., 2007; Madden et al., 2015; Mansour, 2014; Marshall et al., 2016; Melo et al., 2015; Weisberg et al., 2013; Wyatt et al., 2014). While most contraceptive studies are based on cross-sectional quantitative data, for many women, contraceptive attribute priorities change over time and life context. Longitudinal studies link changes in contraceptive behavior with changes in pregnancy attitudes (R. K. Jones et al., 2015; Moreau et al., 2013) as do qualitative studies, in which women also cite a change in 24 relationship status as a reason for choosing a different method, suggesting fluctuations in the prioritization of contraceptive attributes over one?s lifetime (Downey et al., 2017; Marshall et al., 2018). Since most modern contraceptive methods require a medical prescription for or medical insertion of the method, including all highly effective methods, providers serve as key intermediaries between women and their contraceptive choice. One study found that more than half of U.S. respondents indicated a healthcare professional as having the most influence on choice of contraception and being the main source of advice on contraceptive methods (Johnson et al., 2013). This is in line with other research that finds American women wanting the provider?s participation in the decision-making process, known as shared decision making, while still having the final say in their selection of contraceptive method (Dehlendorf et al., 2013; Fox et al., 2018). Interestingly, women are far more likely to desire autonomy in contraceptive decision-making than in general health care decisions (Dehlendorf, Diedrich, et al., 2010). A good relationship with family planning providers, or relationship-centered care, has been cited as important by women (Carvajal et al., 2017; Sundstrom et al., 2018) and has been linked to method continuation, method consistency, and use of moderately or highly effective methods (Dehlendorf, Henderson, et al., 2016; Frost & Darroch, 2008). With the advent of new technologies, web-based or other electronic support tools may be used to complement provider counseling by probing patients about the various elements related to the decision-making process, such as educating about different methods, asking about current contraceptive or 25 pregnancy priorities, and providing a summary to the provider prior to the start of the visit (Dehlendorf et al., 2019; Giho et al., 2020; Hebert et al., 2018; Jamin et al., 2017). 2.3.1.1. Contraceptive Counseling and Contraceptive Choice Evidence shows that, despite a heterogeneity of study settings, populations, and counseling scripts and strategies, overall, contraceptive counseling in clinical settings has a positive impact on various contraceptive-related outcomes (Cavallaro et al., 2020; Zapata et al., 2018). The current data suggest that the health care provider can play a key role in assisting women in selecting the method most likely to be a good fit for her life circumstances, by providing accurate information about the range of methods available, their effectiveness, compliance, and side effects profiles, and helping her assess which method best aligns with her individual priorities and contraceptive attribute preferences (Dehlendorf, Krajewski, et al., 2014; Fox et al., 2018). For instance, studies of providers receiving skills-based LARC training have found associations with increased rates of LARC uptake at those sites (Harper et al., 2015; Lewis et al., 2013), and a provider survey from 2010 found that residency training in obstetrics/gynecology or family medicine, compared to internal medicine or pediatrics, predicted LARC provision, in particular for IUDs (Greenberg et al., 2013). This suggests that contraceptive counseling plays an important role in patients? contraceptive choice. Many studies investigating the role of contraceptive counseling in contraceptive-related outcomes has been focused on women seeking abortions and postpartum women as evidenced by two recent systematic reviews (Cavallaro et al., 2020; Zapata et al., 2018); however, this literature review will focus on available research sampling broader populations, of which there are not many. 26 Some studies have focused on evaluating interventions that aimed to increase LARC uptake in part through a contraceptive counseling component. One, a large cluster randomized trial across 15 U.S. states found that young women (18-25) wanting to avoid pregnancy in the next year were significantly more likely to choose an IUD or implant if they received care at an intervention site, where providers had been trained to integrate LARC into their standard contraceptive counseling, compared to those who received care at a control site providing standard care. Women?s autonomy in decision-making was not affected by the intervention? participants in both the intervention and control arm reported the same rates of autonomy (78% chose the method themselves, 14% with the provider, 7% no method, and less than 1% reported the provider choosing the method) (Harper et al., 2015). Two, at a time when the newest generation of the subdermal implant was just approved in the United States (Kaiser Family Foundation, 2019) and there were only two available IUDs (one hormonal and one copper) (Kaiser Family Foundation, 2020), the Contraceptive CHOICE Project found that women looking to initiate a new form of reversible contraception were far more likely to choose LARC following a brief script that introduced them to the IUD and implant as available methods, than was the national average of LARC users among all contraceptive users at the time (67% vs. 10%) (J. Jones et al., 2012; Secura et al., 2010). CHOICE participants were a convenience sample of women not currently using contraception, 60% of whom were under 25, and who, in addition to counseling, were offered all reversible methods at no cost for three years in order to eliminate the financial barrier (Secura et al., 2010), which also contributed to higher uptake levels. 27 As the CHOICE project continued, the researchers developed a structured contraceptive counseling script, modeled after the client-centered GATHER process for counseling (Greet, Ask, Tell, Help, Explain, Return), describing the effectiveness, advantages, and disadvantages of all reversible methods, in order of effectiveness (LARC, injectable, pill, patch, ring, external condom) (Madden et al., 2013); the latter is known as tiered or efficacy-based counseling (A. Hoopes et al., 2020). This counseling was offered pre-enrollment by staff without prior health care experience or clinical training to participants at the university research site, and their LARC uptake was then compared to participants receiving standard counseling, including LARC, at community partner clinics; participants were still able to receive any method at no cost. Researchers found no statistically significant differences between overall LARC uptake after adjusting for baseline participant characteristics, even though uptake of IUDs was much higher at the university research site compared to the implant, with the opposite phenomenon occurring at the partner clinics (Madden et al., 2013). Although CHOICE participants are not nationally representative of women of reproductive age, uptake of LARC compared to other methods at 72% and 78% is drastically higher than the national average, where 10% of all women using reversible contraception were using LARC in 2006-2010 (J. Jones et al., 2012) and 17% in 2011- 2013 (Daniels et al., 2014). Some studies have implemented elements of the CHOICE protocol to increase the uptake of LARC. One, a very small retrospective study in Alaska, mainly including Native American women, found LARC uptake increase after providers received training in the structured contraceptive counseling used by the CHOICE project; access barriers due to contraception cost 28 were minimal among this sample (Tobias & Enriquez, 2018). Two, a cluster randomized trial of family practices in Melbourne, Australia using the structured contraceptive counseling script found uptake of LARC consistently higher across three time points within one year in the intervention than in the control arm, with IUDs being the preferred LARC method (Mazza et al., 2020). Three, one study investigated the effect of reproductive life plan counseling (RLPC) on contraceptive method choice among a diverse group of Title X patients (n=771) in a Midwestern city. The sample included women 16 and older who were at risk of unintended pregnancy; due to the nature of the Title X program their visits fell under, financial barriers to contraceptive use were low for the participants during the study period. Incorporating RLPC into routine family planning visits is recommended by federal and clinical guidelines to improve increase provision of preconception care (Robbins et al., 2017) and serves to assist the patient in considering their pregnancy intentions and what type of contraceptive methods, if any, might help her reach her reproductive goals (Bommaraju et al., 2015). At the time of the study the Title X clinics were not bound to follow a RLPC-standardized tool; the only RLPC measure available and used in the study was whether RLPC occurred or not, although it is known that providers were trained on efficacy-based contraceptive counseling. The study found RLPC not to be associated with contraceptive use, though the group who received RLPC had somewhat higher LARC use and lower no or traditional method use. The counseling did not mediate or moderate the effect of race/ethnicity or educational attainment on contraceptive use (Bommaraju et al., 2015). 29 Four, a prospective cohort study of diverse majority low-income women (n=348) seeking to initiate or change a contraceptive method in the San Francisco Bay Area was conducted to assess quality of communication during the counseling visit based on the dimensions of patient- centered care. Providers in six clinics were asked by the researchers to provide usual contraceptive care. Patients who reported high interpersonal quality of care with their providers immediately post-visit were more likely to continue using their chosen method and to be using highly or moderately effective methods at 6-month follow up (Dehlendorf, Henderson, et al., 2016). Five, one study found women of poorer health being less likely to have used contraception at last penile-vaginal intercourse compared to women reporting better health, and also being less likely to receive contraceptive counseling. However, when women with poorer health received counseling about hormonal contraception, they were more likely to use a hormonal method (Lee et al., 2013). Six, a study in an inner-city Title X-funded family planning clinic investigated the effect of a provider checklist intervention that included a brief structured counseling script about LARC and emergency contraception to patients who were seeking walk-in pregnancy testing. When comparing outcomes with the comparison group (patients before the intervention started), intervention participants were more likely to receive emergency contraception, have a LARC inserted at the visit, or receive a contraceptive prescription. At three-month follow-up, the intervention group were more likely to be using a LARC or injectable than the comparison group (Lee et al., 2015). 30 Seven, an older study in Italy tested an intervention using the Adjusted Contraceptive Score (ACS), a scoring method designed to assist providers and patients choose the best traditional or barrier method (diaphragm, condom, basal temperature, Billings ovulation method, calendar method, and withdrawal). Women who were administered the ACS reported increased use of the diaphragm and basal body temperature and lower use of withdrawal and the calendar method, compared to women who received standard contraception counseling. The women in the intervention arm also had fewer instances of pregnancy in the follow-up period than the controls (4.2% vs. 10.8%) (Custo et al., 1987). However, only two studies were found assessing potential change in contraceptive choice following a clinic visit by measuring pre-visit use or intention and post-visit uptake or continuation. Part of the same data collection project as Dehlendorf et al. (2016), women?s contraceptive preferences prior to a provider visit and method choice following the provider visit were measured among a diverse group of majority low-income women (n=342) seeking to initiate or change a contraceptive method in the San Francisco Bay Area. The goal of the investigation was to study contraceptive counseling about combined hormonal contraceptives (CHC) (pill, patch, ring) and subsequent method choice. Craig et al. (2019) found that, of the women who identified a preference for a specific CHC in the pre-visit survey, 72% had chosen the same method in the post-visit survey. The study found large variability in the consistency of counseling about the three different CHC methods, where women with no pre-visit method preference or with a preference for the ring or the patch were more likely to be comprehensively counseled on all three CHC methods compared to women who indicated a preference for 31 combined oral pills or non-CHC methods. The mixed-methods study was bolstered by qualitative data, which allowed for detailed analyses of all contraceptive counseling encounters (Craig et al., 2019). One industry-funded study spanning ten European countries and Israel investigated the impact of provider?s contraceptive counseling by measuring women?s contraceptive intentions pre-visit and method choice post-visit; only women who were medically eligible for and expressed interest in one of the combined hormonal contraceptives (CHC) (pill, patch, ring) were invited to participate in the study. Although the percentages ranged from 31% to 64% across countries, overall, 47% of participants (n=18,787) chose a different CHC method post-visit compared to the CHC method they intended to use pre-visit (Bitzer et al., 2012). It should be noted that the provider reported the woman?s pre-visit intention in a post-counseling questionnaire and the woman reported her post-visit method choice after the visit. An equivalent investigation was conducted in Portugal where 35% of participants who initially wanted the pill selected the ring or the patch post-visit (Costa et al., 2011). Another equivalent study, in Italy, found that 16.6% of participants who wanted the pill pre-visit selected the ring or the patch post- visit, and overall 38% chose a CHC method different from what they intended pre-visit (Gambera et al., 2015). 2.3.2. Contraceptive Use Over Time Beyond choice, contraceptive behaviors also consist of consistency, accuracy, continuation, discontinuation, and switching (Frost et al., 2007b; Jaccard et al., 1996). While a 32 review of this literature is beyond the scope of our research questions, this context is relevant when considering the difference between LARC and other contraceptive methods. Compared to all other reversible methods, LARC provides the longest time of protection without any action required by the user, i.e., consistency and accuracy do not depend on the user until the devices? expiration date. Furthermore, both discontinuation and switching of methods are common occurrences (Frost et al., 2007a, 2007b). Multiple studies show significantly higher rates of continuation for LARC compared to other hormonal method types (Diedrich et al., 2015; Hubacher et al., 2018; Sanders et al., 2017; Usinger et al., 2016). Detailed studies on switching behaviors are scarce and complex, but overall, the evidence suggests that compared to other methods, there is either no difference or less switching away from LARC and either little difference or more switching to LARC (Hubacher et al., 2018; Simmons et al., 2019; Steinberg et al., 2021). As a result of these features and trends, LARC provides the most effective protection against unintended pregnancy of all reversible methods. This is, arguably, the main reasons why DelCAN and many other initiatives focus on increasing LARC use. 2.3.3. Understanding Disparities in Unintended Pregnancy and Contraceptive Use Demographic characteristics, particularly racial, ethnic, and socioeconomic factors, are predictors of disparities in adverse maternal and child health outcomes (Curtin & Hoyert, 2017; Lang & King, 2008) and in contraceptive prevalence and choice (Daniels & Abma, 2020). Based on Kilbourne et al.?s (2006) conceptualization of health disparities, Dehlendorf, Rodriguez, and colleagues (2010) identify three major factors that contribute to disparities in family planning 33 outcomes: (a) patient preferences and behaviors, (b) health care system factors, and (c) provider- related factors. Patient preferences and behaviors include concerns about the safety and side-effects of contraception, insufficient knowledge about contraception and reproductive health, different levels of ambivalence toward pregnancy, and differing perceptions on the desirability of adolescent pregnancy. Many of these patient-level factors that negatively affect effective contraception use disproportionately impact women from minority communities or those of lower SES; therefore, leading to higher rates of unintended pregnancy and its corresponding outcomes (Dehlendorf, Rodriguez, et al., 2010). However, it is vital to emphasize that patient preferences and behaviors do not exist in a vacuum; the historical and cultural context of the United States, and its treatment of minority and low-income communities, has inevitably influenced women?s and communities? preferences and behaviors in regards to childbearing and contraception (Roberts, 2000; Stern, 2005). Health care system factors refer to women?s differing access to contraceptive services, while provider-related factors pertain to providers? differential treatment of patients according to their demographic characteristics (Dehlendorf, Rodriguez, et al., 2010). Literature consistently finds that women from minority communities and of lower socioeconomic status have reduced access to contraceptive services and, overall, encounter explicit and implicit bias in their interactions with reproductive health providers that their White counterparts and those of higher socioeconomic standing do not (Dehlendorf, Rodriguez, et al., 2010). 34 Although DelCAN offered benefits to all women of reproductive age across the state, the initiative specifically targeted two of the three major factors currently understood to contribute to family planning disparities. As described in more detail in the next section, DelCAN used a multi-prong approach to expand access to all methods of contraception, particularly LARC. Firstly, the initiative trained healthcare providers and support staff to deliver comprehensive, tiered-effectiveness counseling to patients visiting their clinics. This was an effort to modify patient preferences by increasing patient knowledge and positive attitudes about contraception, including LARC. Secondly, DelCAN?s efforts to guarantee that patients can receive All Methods Free, same-day LARC insertion, and transportation subsidies targets health care system factors that often impede women, particularly those of lower socioeconomic status, from accessing the entire array of methods available on the market. Thirdly, the initiative was accompanied by a statewide public awareness campaign aimed to raise awareness about the benefits that DelCAN offers, i.e., to inform Delawarean women about the facilitated access that they should be able to experience if they seek out contraceptive care. By conducting our investigation among a sample of patients at clinics receiving Title X5 funds, most of our participants are likely to be impacted by at least one of the three major factors that contribute to disparities in family planning outcomes. 5 Title X is a federal grant program that provides family planning services and preventive health services, prioritizing serving women and families with low incomes (Schapiro, 2020). 35 2.4. Contraceptive Knowledge and Attitudes Of the estimated 3.4 million unintended pregnancies in 2008 in the United States, only 5% occurred among women consistently using contraception, with 54% attributed to nonuse and 41% to inconsistent use of contraception (Sonfield et al., 2014). A study, based on a data from a nationally representative sample of women at risk of unintended pregnancy in the U.S. in 2004, found that having less than a college degree and feeling that one could not call their contraceptive service provider to answer method-related questions were both associated with nonuse or inconsistent contraceptive use, among other factors. Furthermore, dissatisfaction with one?s method was associated with inconsistent use and switching methods (Frost et al., 2007a). One study using a nationally representative sample found an association between mistrust in the government ensuring contraceptive methods are safe and effectiveness level of current contraceptive method used; the higher the skepticism the less effective the method being used (Rocca & Harper, 2012).6 Considering LARC?s slow uptake among the general population relative to its high rates of efficacy and satisfaction (Daniels & Abma, 2020; Higgins, 2017), many studies have explored women?s perceptions about LARC and LARC provision in order to identify barriers. Across multiple studies, women and adolescent girls have cited various concerns about LARC that fall into four knowledge and attitude categories: lack of objective knowledge (ex., mechanism of action, effect on future fertility), side-effects (correct or perceived), pain during insertion or 6 The same study also found differences in attitudes about effective contraceptive methods, pregnancy, childbearing and fertility among Black, Latina, and White young adult women, but attitudinal differences did not explain differences in contraceptive use by race and ethnicity, which had been the goal of the study (Rocca & Harper, 2012). 36 removal, and lack of personal control once inserted (Asker et al., 2006; Bharadwaj et al., 2012; Foster et al., 2014; Glasier et al., 2008; Kavanaugh et al., 2013; Littlejohn, 2013; Murphy et al., 2017; Spies et al., 2010). A recent review found that peers and family are often sources of negative or inaccurate information about LARC (Mahony et al., 2021) and other types of contraceptive methods; poor scientific evidence and individual cultural beliefs are also sources of contraceptive misconceptions (Belfield, 2009; Binette et al., 2017; Kilbourne et al., 2006; Russo et al., 2013; Yee & Simon, 2010). Therefore, increasing knowledge and positive attitudes about contraception and contraceptive methods has the potential to reduce a key barrier of consistent contraceptive use among those at risk of unintended pregnancy. 2.4.1. Contraceptive Counseling and Contraceptive Knowledge Contraceptive counseling within the scope of a provider visit is an ideal opportunity to increase patients? contraceptive knowledge and positive attitudes due to the provider?s role as content specialist, their expertise on contraceptive contraindications, and the confidentiality of the visit. As such, counseling by clinicians trained on up-to-date contraceptive guidelines may facilitate women in identifying the most appropriate contraceptive method for their current situation and lifestyle (Dehlendorf, Krajewski, et al., 2014; Fox et al., 2018; Glasier et al., 2008; Mahony et al., 2021; Zapata et al., 2018). This section provides an overview of the extant literature on the effect of contraceptive counseling on knowledge. When interventions included additional intervention components or outcome measures, the elements are briefly mentioned and other measure results are omitted. 37 Numerous studies have investigated whether the use of alternative promising audiovisual and interactive tools, such as pre-recorded videos and mobile/tablet applications, increases knowledge and positive attitudes about contraception. While the aids are usually envisioned to supplement contraceptive counseling during a clinic visit, studies commonly (1) test solely the effect of the two, (2) deliver the human component by a volunteer or health promoter, or (3) deliver the human component by a medical professional in a group setting (Delamater et al., 2000; Garbers et al., 2015; Gilliam et al., 2014; Kofinas et al., 2014; O?Donnell et al., 1995; Pedrazzini et al., 2000; Schwarz et al., 2013; Sridhar et al., 2015; Vogt & Schaefer, 2012). This brief review only included studies that conduct a knowledge or attitudes post-test following a consultation or visit with a provider. An inner-city Title X-funded family planning clinic implemented a bundled intervention that included a scripted contraceptive counseling component based on the script used by the aforementioned CHOICE project (Secura et al., 2010) for women presenting for a pregnancy test (Lee et al., 2015). Ten multiple choice and true-false questions about effectiveness, reversibility and duration of IUDs, implants, and injectables were administered immediately post-intervention and at three-month follow up over the phone. The sample was relatively young (mean age 22) and predominantly Black (Lee et al., 2015). The comparison group comprised of patients of similar sociodemographic characteristics who presented in the same way in the eight months prior to the implementation of the intervention; they received only ad lib contraceptive counseling. The intervention group was significantly more likely to report receipt of contraceptive counseling and to have greater knowledge about the IUD and implant 38 regarding effectiveness, reversibility and duration of use; knowledge scores on injectables were non-significant between the groups (Lee et al., 2015). Using a subset of the sample consisting of women seeking emergency contraception (mean age 24 and predominantly Black), the intervention group again had greater knowledge about LARC than the control group. However, immediately post-intervention five out of the seven LARC items were found to be significantly different between the two groups compared to only the out of seven at the three-month follow- up; one of three injectable items showed significantly higher knowledge scores only at three- month follow-up (Schwarz et al., 2014). Two studies assessed knowledge specifically about the emergency contraceptive pill. First, a multi-prong community-based intervention in Sweden consisted of a media campaign, website providing information, and contraceptive counseling specific to emergency contraception. Researchers informed nurse-midwives working in family planning clinics about the intervention and asked them to disseminate information (oral and written) about emergency contraception to women seeking contraception or coming in for postpartum care (the assumption is that this delivery of information would occur within the scope of the visit?s contraceptive counseling). Knowledge was assessed with six multiple-choice items on mechanism of action, effectiveness, and side-effects of emergency contraception. Women aged 16-30 from two equivalent Swedish counties were randomly selected from national tax records and were sent the survey and study invitation letter in the pre-intervention period. After the one-year implementation of the intervention in one of the two counties (experimental group), participants who completed pre-test in both counties were sent the same survey (sample mean age 27). 39 Baseline knowledge about emergency contraception was higher in the intervention group; higher scores were retained following the intervention. No statistical difference in change between baseline and post-intervention between the two groups was observed except for side-effect knowledge, which decreased in the comparison group and remained the same in the experimental group. However, ?there was an overall improvement over time measured by the knowledge index? (Larsson et al., 2004a, p. 823). Second, a pair of sister studies were conducted by a research team of pharmacists investigating the effects of a brief pharmacist-driven counseling session on knowledge retention about the emergency contraceptive pill. A flip chart was utilized during the session, serving both as a visual aid and checklist for the counselor. A 12-item survey assessing mechanism of action, side-effects, effectiveness, administration, and availability was filled out pre-counseling, immediately post-counseling, and at a 1-5 month follow up via phone. The first study took place in 2010 in the waiting room of a women?s clinic at an academic medical center (participants? mean age 25), and the second in 2012 at two grocery store retail pharmacies (participants? mean age 30); both were convenience samples. Both studies found that post-test and follow-up knowledge scores were significantly higher than at baseline, indicating a positive impact on both short-term and long-term knowledge retention (Ragland et al., 2011, 2015). Two international studies focused on knowledge retention about the pill. A complex randomized controlled trial in England recruited patients coming in for a repeat prescription of the pill to one of six groups (three A groups and three B groups). In addition to the standard medical visit, participants in the B groups were asked a brief series of contraceptive questions, 40 with a discussion of the correct answers if the patient answered incorrectly. At the end of the visit, no leaflet, a summary leaflet, or the Family Planning Association leaflet were given to participants in each A-B paired group. The A group that received no leaflet served as the control group. After three months, participants (median ages 26-27) were sent a 32-item questionnaire measuring retention of the twelve basic pill rules. The three types of interventions individually resulted in modest improvement of knowledge; the combination of interactive questions (group B) and leaflet only demonstrated an added benefit in the summary leaflet group (Little et al., 1998). The Oral Contraception Project to Optimise Patient Information (CORALIE) was a multicenter study conducted in France 2009-2013 that assessed knowledge retention among pill initiators. A large group (n=161) of French gynecologists were recruited and randomized into two groups; the intervention group delivered structured contraceptive counseling based on an ?essential information? checklist developed by an elaborate pre-study involving gynecologists across the country, while the control group delivered standard of care. Immediately following the provider visit, patients seen by a provider from either group filled out a 15-item questionnaire about essential information about pill use. Although eligibility for the study was 16-40, the mean age of participants in both groups was 20 years old since any prior use of oral contraceptives disqualified participants from the study. Women in the intervention group reported statistically significant higher knowledge on most questions, with highest significance observed on questions about incorrect use, side-effects, and impact of vomiting on effectiveness (de Reilhac et al., 2016). 41 A rare target population, women seeking sterilization were recruited at two gynecology clinics of two teaching hospitals in England and randomized into a video-intervention followed by standard of care or solely standard of care. The video covered all information that should be provided during a visit for women seeking sterilization; the providers delivering standard of care did not know which group their patients were randomized to (though patients may have self- disclosed having seen the video at any point during the visit). A 15-item sterilization knowledge survey was administered immediately post-visit. The intervention group had significantly higher median scores; among those in the comparison group, women seeking care at the nurse-led dedicated sterilization clinic had more knowledge than those in the general gynecological site (Mason et al., 2003). Two studies assessed contraceptive knowledge among women seeking termination of pregnancy. An Italian randomized controlled trial investigated the effect of a patient-centered contraceptive counseling intervention among women seeking an abortion. Knowledge was measured with a 20 true-false-I-don?t-know questions about six contraceptive methods, including the IUD. Following randomization, women completed the pre-test survey. In the week preceding their abortion, the experimental arm received a 30-minute personalized contraceptive counseling intervention by a psychologist and gynecologist. The counseling started with a 10- minute semi-structured interview by the psychologist asking about ?women?s ?agenda? regarding contraception, including her barriers to use, her perceptions of risk, and her past and present experiences of contraception? (Nobili et al., 2007, p. 363). Subsequently, the gynecologist shared the pros and cons of five contraceptive methods, including the IUD, as well as emergency 42 contraception. The third phase was run by the psychologist and dedicated to the patient choosing a method in line with her previously discussed ?agenda.? Both the psychologist and gynecologist answered any remaining questions and checked for the patient?s understanding. The control group received standard of care, which was to encourage patients to consult the community health center after the abortion. The knowledge questionnaire was administered to both groups at one-month follow-up via phone. More than half of the sample was in their thirties. With no significant difference in pre-test scores between the two groups, the experimental group reported a highly significant increase in knowledge while the control group mean score remained the same (Nobili et al., 2007). A small study in Scotland among women presenting for abortion evaluated the use of a digital video as a source of information about the contraceptive implant. After consulting with a provider and arranging for an abortion, the clinician assessed eligibility for the implant. Patients who were eligible and expressed interest in initiating the implant for the first time were recruited and randomized into the DVD plus provider visit (mean age 24) or standard of care for implant counseling (mean age 23). Immediately following the provider visit, both groups had knowledge recall assessed by four items (side-effects, mechanism of action, duration of use) through a structured interview with the researcher. Information recall was similar in both groups (Michie et al., 2016). Prior to the advent of adolescent-friendly contraceptive counseling, an intervention tested the tailoring of family planning services by shifting from a medical model to a psychosocial model for teenager visits. The components of the intervention emphasized ?in- 43 depth counseling, education geared to an adolescent's level of development, and the provision of reassurance and social support? (Winter & Breckenmaker, 1991, p. 24). The sample was made up of adolescents (98% White) visiting six nonmetropolitan family planning clinics for the first time or for an annual visit. Baseline knowledge about basic reproduction, contraceptive methods, and STIs was measured with eleven true-false items immediately following the index visit that provided standard of care. Subsequently, staff from three clinics, were trained in delivering the treatment, including adolescent-friendly counseling; the other three clinics served as the comparison group. Knowledge scores increased significantly in the treatment group, while those of the comparison group remained (Winter & Breckenmaker, 1991). Also aimed at adolescents, a multisite study at three family planning clinics in Chicago and Madison investigated the impact of a digital decision aid utilized in conjunction with a medical visit for adolescents. The protocol sequence was as follows: recruitment upon arrival to the clinic, pre-test knowledge interview, viewing of computer decision aid and printout, provider visit (standard of contraceptive care), post-test knowledge interview and provision of correct answers, and one-year follow-up. Knowledge about the pill was measured with a five-item scale about side-effects, missed dose, and protective effects. Due to highest volume of patients being interested in oral contraceptives, only those who chose the pill were interviewed post-test and/or follow up. The comparison group was every other adolescent patient who arrived at the clinic and received standard of care. The mean knowledge score of the experimental groups was significantly higher in both cities at post-test, but the difference was only significant among Madison participants at follow-up (Chewning et al., 1999). 44 2.4.2. Contraceptive Counseling and Contraceptive Attitudes I found only two studies that measured contraceptive attitudes following a provider visit. Both studies also assessed knowledge, so study design and sample are presented in the prior section. Larsson and colleagues (2004b) used seven statements with a 5-point Likert-type scale to assess attitudes about emergency contraception?s (EC) prescription-free availability, its similarity to abortion, worries about side-effect, and its benefits. Changes to more favorable attitudes were reported for only three out of seven items; more agreement with over-the counter availability being good, and a less agreement with EC being like abortion and EC use hesitancy due to side-effects. However, the difference observed in change over time between the two groups was non-significant (Larsson et al., 2004a). Nobili and colleagues (2007) ran a study in Italy assessing attitudes about contraception in general and about specific methods with a 10-item scale. Six questions were of a general contraceptive nature such as ?Do you believe contraceptive methods may be damaging to your health?? and ?Do you believe there are some contraceptive methods which are more suitable for casual sex?? Four questions were about specific method types: one about emergency contraception, one about withdrawal, and two about the pill, such as ?Do you believe that the pill causes you to gain weight?? (Nobili et al., 2007, p. 367). Participants chose one of the 4- point Likert likelihood answers (not at all, a little, somewhat, very much so), and difference in baseline attitudes between the two groups was not significant. At one-month follow-up, a 45 significant increase in favorable attitudes in the experimental group was observed; the increase was non-significant for the control group (Nobili et al., 2007). 2.4.3. Summary Overall, most of the studies on contraceptive counseling and knowledge find that counseling does increase patients? knowledge. However, of the few studies that evaluated any LARC knowledge, only one used a pre-post-test design. Very few studies that investigate attitudes about contraception incorporate a provider consultation, so a consensus on evidence cannot be reached. Additionally, none measured LARC attitudes. Due to substantial variability in study settings, counseling procedures and scripts (if any), and outcome measures, it is difficult to identify which types of educational interventions result in the best outcomes. However, evidence suggests that, at minimum, providing additional time for contraceptive counseling or utilizing decision aids or audiovisual content helps patients increase their short-term knowledge on contraception. 2.5. Contraceptive Affordability Despite the existence of Title X funding since 1970 (Schapiro, 2020), women in the United States repeatedly cite affordability as an important attribute of a woman?s contraceptive method (Landau et al., 2006; Lessard et al., 2012; Madden et al., 2015; Marshall et al., 2018). A study using 2002 data from a large nationwide sample of reproductive-aged women at risk of 46 unintended pregnancy found that women without health insurance were 30% less likely to report using contraceptive methods requiring a prescription than those with private or public health insurance (Culwell & Feinglass, 2007). Another nationally representative survey from 2004 found that support for and interest in pharmacy access to self-administered prescription contraception (pills, patch, ring) was higher among uninsured women than insured women, with no cost for doctor?s visit as one of the main benefits of such a service (Landau et al., 2006). Studies have also linked increases in contraceptive uptake, particularly of LARC, to the passage of the Patient Protection and Affordable Care Act (ACA) in 2012, thanks to its elimination of contraceptive cost-sharing for most women with prescription insurance (N. V. Becker et al., 2021; Carlin et al., 2016; Montgomery et al., 2020; Snyder et al., 2018). However, the ACA provision does not help women who are not able to secure private or public insurance, as they still face out-of-pocket costs (August et al., 2016). Clinics nationwide receiving Title X funding provide subsidized or free contraceptive services, but all contraceptive methods may not be available in a timely manner, if at all. For instance, Colorado?s Title X clinics experienced growing waiting lists of patients seeking LARC prior to the launch of the Colorado Family Planning Initiative in 2009 (Colorado Department of Public Health and Environment, 2017). Evidence from Colorado and other city and statewide initiatives (described in more detail in the subsequent section) has linked increases in LARC uptake, in part, to their free provision. DelCAN sought to provide free contraceptive services to all women across the state. A recent analysis by Marthey and colleagues (2021) suggests that the initiative?s public awareness 47 campaign was associated with an increased rate of Title X clinic visits among the intended demographic (18-29). However, the campaign likely did not reach all Delawarean women in need of free contraception. An office manager from a woman?s health clinic participating in the DelCAN initiative shared her regret at not providing a more tailored outreach to uninsured patients when the program began; she explains, ?[Uninsured women] are a very fragile population. And they?re used to going into an office and being charged or being afraid to even go to a walk-in or another doctor because they?ll be charged. So there is a population that you definitely want to make a very strong relationship with. So I wish that we had made these appointments different[ly]? (Skra?i? et al., 2021)p. 214). Therefore, knowledge about the existence of free/affordable and available contraception is just as important for access to contraception as the affordability and availability themselves. 2.6. City and State-wide Interventions Informed by the current literature on contraception-related disparities, a number of city and state-wide initiatives serve as encouraging evidence that reducing or eliminating individual and structural barriers can allow more women to choose more effective contraceptive methods to prevent pregnancy. These initiatives included efforts to make all types of contraceptive methods equally accessible by eliminating out-of- pocket costs, offering same-day contraception initiation for all methods, and increasing the number of health care providers trained in contraceptive counseling and provision, especially for long-acting reversible contraceptive methods (LARC). Success was measured by reductions in pregnancy, birth, and abortion rates among adults and 48 adolescents, and an increased uptake and continued use of long-acting reversible contraception (Sawhill & Guyot, 2019). Interventions that have been researched and evaluated most thoroughly took place in Colorado, Iowa, Utah, and Missouri. 2.6.1. The Colorado Family Planning Initiative Launched in 2008, the Colorado Family Planning Initiative aimed to reduce the number of unintended pregnancies by providing LARC and sterilization at low or no cost to all clients in Title X clinics. In addition to the methods? high levels of effectiveness yet prohibitive costs, unmet need was demonstrated by clinics? waiting lists for these methods among both men and women. With differing capacity needs, community needs, and populations served, each Title X clinic received technical assistance, conferences, and training to address individual needs. Funded by a private donor, the initiative allowed clinics to stock up on LARC, hire and/or train staff, improve clinic infrastructure, and purchase or upgrade needed equipment. Particular attention was paid to the training of health care providers and clinic staff for LARC counseling and managing LARC side-effects (Colorado Department of Public Health and Environment, 2017). In the second year of the initiative, all clinics were offering IUDs, and all but one offered implants (Kelly et al., 2020). Three years into the initiative, a media campaign, Beforeplay, was launched to raise awareness about services at Title X clinics, as well as to normalize reproductive health conversations across the state. The campaign went beyond pregnancy prevention to include information on pregnancy readiness and sexually transmitted diseases (Colorado Department of Public Health and Environment, 2017). In 2015, the private funding ceased, but 49 Colorado?s legislature allowed of the continued funding of the initiative?s services through the state?s ongoing Family Planning Program (Sawhill & Guyot, 2019). Five years into the initiative, Colorado saw notable improvements in multiple maternal and child health indicators. Firstly, there were reductions in the unintended pregnancy rate among adolescents (15-19) and young women (20-24). It also resulted in approximately half the number of births and abortions among adolescents, as well as a 20% and 18% reduction in births and abortions, respectively, among young women. Secondly, the number of repeat births for those two age groups also decreased. Thirdly, births to women without a high school education dropped from 20% to 12%, indicating that some teenagers were able to delay childbirth at least until after high school and that women without a high school diploma were able to avoid unintended pregnancy better than before. Fourthly, the percentage of births following a short interpregnancy interval also dropped (Colorado Department of Public Health and Environment, 2017). The results of various peer-reviewed studies confirm these improving trends (Lindo & Packham, 2017; Ricketts et al., 2014). Perhaps most notably, Kelly et al. (2020), found that the initiative reduced births by about 20% among adolescents living in zip codes within seven miles of Title X clinics, and that after an extensive media coverage five years into the initiative, LARC insertions substantially increased and the birth rate reduction extended to all women in their 20s, as well as adolescents aged 15-17 living up to twelve miles from a Title X clinic. The Colorado Family Planning Initiative is considered as the biggest, but not the only driver, of these improved maternal and child health indicators. For instance, the passage of the Affordable Care Act (ACA) mandated that all contraceptive methods be covered by prescription 50 drug insurance, and it allowed Colorado to expand Medicaid coverage in 2013 (Colorado Department of Public Health and Environment, 2017). However, despite improved access thanks to the ACA, some populations continued to remain outside its coverage, and young people on their parents? health plans worried about confidentiality when seeking contraceptive services? gaps that the statewide initiative model aimed to cover. 2.6.2. St. Louis Contraceptive CHOICE Project Running from 2007 to 2011, the privately funded Contraceptive CHOICE Project recruited 10,000 women of reproductive age in the St. Louis region who wanted to prevent pregnancy for at least one year and were initiating a new form of reversible contraception. Each woman received extensive contraceptive counseling focusing on LARC, and any contraception of her choice for three years at no cost. The project was a prospective cohort study with the goal of increasing LARC uptake and reducing unintended pregnancy in the area (Secura et al., 2010). Using a convenience sample, the study?s participants are not a representative sample of women in St. Louis. However, the results of the CHOICE project suggest improving knowledge and eliminating costs can have transformative effects on women?s reproductive trajectories. When comparing the area of St. Louis to Kansas City and nonmetropolitan Missouri, significant decreases in repeat abortions were observed. Furthermore, within the CHOICE cohort, abortion rates were less than half the regional and national rates, and the teenage birth rate was five times lower than the national rate (Peipert et al., 2012). Overall, study results suggest that when eliminating the barriers of cost, access, and knowledge, women are far more likely to choose 51 LARC as the most effective and least user-dependent methods. The prospective cohort study also showed high continuation and satisfaction rates with LARC (McNicholas et al., 2014). Continuation rates at 12 months for each LARC method were higher than 75% across all age groups, while non-LARC methods had continuation rates of 44% among 14 to 19-year-olds and around 52% among the other age groups. Satisfaction rates closely followed continuation rates where women of all ages were more likely to be satisfied with LARC devices compared to non-LARC methods (Rosenstock et al., 2012). At 24 months, average LARC continuation was 77% compared to 41% for non-LARC methods (Birgisson et al., 2015). Among adolescents, LARC was chosen by 69% of participants aged 14-17, and by 61% of those aged 18-20. Participants in the younger group were more likely to choose the implant (63%), and in the older group the IUD (71%) (Mestad et al., 2011). When comparing contraceptive outcomes for clients who received structured contraceptive counseling at a university research site by CHOICE and standard counseling, including LARC, at community partner clinics, researchers found no statistically significant differences between overall LARC uptake after adjusting for baseline participant characteristics, even though uptake of IUDs was much higher at the university research site compared to the implant, and vice versa at the partner clinics (Madden et al., 2013). 2.6.3. Iowa Initiative to Reduce Unintended Pregnancies Launched in 2007 and concluded in 2013, the Iowa Initiative to Reduce Unintended Pregnancies provided private funding to Title X and other family planning agencies serving low- income women. The initiative?s effort focused on increasing LARC use by subsidizing LARC, 52 training clinicians and staff in LARC insertion and counseling, expanding hours and locations of operation, and raising awareness about LARC (Biggs et al., 2015). A social marketing campaign, Avoid the Stork, complemented the initiative during 2010 (Sawhill & Guyot, 2019). From 2005 to 2012, there was a marked increase, from 1% to 15%, in LARC utilization among clients of clinics participating in the initiative (Biggs et al., 2015; Philliber Research Associates, 2012). In the same time period, abortions among Iowa residents decreased from 8.7 to 6.7 per 1000 women aged 15?44 despite Iowa?s expansion of access to medical abortion through telemedicine in 2008 (Biggs et al., 2015; D. A. Grossman et al., 2012). As LARC usage increased over time, the odds of abortion decreased, suggesting that the initiative contributed to the reduction in abortion rates (Biggs et al., 2015). The percent of unintended pregnancies decreased from 46% in 2007 to 43% in 2010 and then to 33% in 2014 (Kost, 2015; Kost et al., 2018; Philliber Research Associates, 2012). 2.6.4. HER Salt Lake Contraceptive Initiative Running from September 2015 to March 2017, the HER Salt Lake Contraceptive Initiative is a prospective cohort study offering free, same-day reversible contraception to patients in four participating health centers of the Planned Parenthood Association of Utah, the only Title X grantee in the state. The initiative?s first step was to eliminate the cost for patients and ensure the staffing and pharmacy stocking for each clinic. Subsequently, potential clients who may not have heard about the available free service were targeted by an online media campaign (Sawhill & Guyot, 2019). 53 A study investigating changes in method use found that the odds of getting an IUD or implant was 1.6 times higher during the first step of the intervention, and 2.5 times higher during the media campaign. Despite similar upward trends in IUD and implant use prior to the intervention across health centers, a time series analysis found that centers participating in the initiative averaged about 59 more IUD and implant placements per month, compared to non- participating sites (Sanders et al., 2018). 2.6.5. Delaware Contraceptive Access Now (DelCAN) In 2010, Delaware had the highest rate of unintended pregnancy in the nation, at 62 unintended pregnancies per 1000 women aged 15-44. With an intended pregnancy rate of only 47, Delaware?s proportion of 57% of all pregnancies being unintended was also among the highest in the country (Kost, 2015). The next few years saw an overall decrease in pregnancy rates across the nation, including in Delaware (Kost, 2015; Kost et al., 2018). Rates of unintended pregnancy also decreased; in Delaware, the rate dropped to 44 unintended pregnancies per 1000 women aged 15-44, though still remaining among the highest in the country (Kost et al., 2018). The proportion of unintended pregnancies, however, dropped below the midway mark, to 48% of all pregnancies in 2014 (Kost et al., 2018). The pregnancy rate among adolescents aged 15-19 in Delaware followed the same trend between 2010 and 2013 (Kost et al., 2017). In 2014, then-Governor Jack Markell launched the Delaware Contraceptive Access Now (DelCAN) statewide initiative to reduce the rate of unintended pregnancies and increase access 54 to contraceptive services for all women of reproductive age. The program focused on: (a) policy change and implementation at the state level; (b) a statewide public awareness campaign; and (c) statewide clinician and staff trainings and same-day LARC insertions (Choi et al., 2019). Central to the DelCAN initiative was the preparation of primary care providers across the state to offer same-day access to all contraceptive methods, including LARC, at negligible or no cost to women (?All Methods Free?). This was done in a public-private partnership with Upstream USA (Upstream), a non-profit organization that provides health centers with training and technical assistance in order to eliminate barriers to offering the full range of contraceptive options (Upstream.org, 2021). Upstream trained and provided technical assistance to health centers and medical practices that provided general primary care, pediatric care, and women?s health or obstetrics and gynecology care to all Delawareans. Between 2016 and 2019, Upstream trained healthcare providers and support staff on clinical, counseling, and administrative processes necessary to deliver comprehensive contraceptive services, including asking women of reproductive age about their pregnancy intention at every primary care visit (also known as the One Key Question?: ?Would you like to become pregnant in the next year?? (Bellanca & Hunter, 2013)), and offering contraceptive counseling, insertion, removal, and management of LARC and other contraceptive methods (Choi et al., 2019). Furthermore, Upstream?s model encouraged the contraceptive ?conversation? to begin as soon as possible upon arrival at the clinic, including patients reviewing Bedsider educational materials in the waiting room and medical assistants providing contraceptive counseling (Skra?i? et al., 2021). 55 Upstream commissioned Child Trends, a research organization, to investigate changes in contraceptive use based on data from Title X clinics. They found an increase from 13.7% to 31.5% in the use of LARC between 2014 and 2017 among Title X patients aged 20-39. While LARC use also increased nationwide within the same time period, the increase was much smaller. By applying a simulation model, they estimated a 15% decline in unintended pregnancy thanks to the increase in LARC use and decrease in no-method use (Welti & Manlove, 2018). However, significant missing data on Title X patients? use of contraceptive methods is a limitation of the study; providers might have been more likely to report LARC usage post- intervention than pre-intervention, potentially inflating the increases in LARC use over time (Welti & Manlove, 2018); additionally, these results do not encompass the outcomes of the initiative as a whole since DelCAN aimed to train all providers in the state, not just Title X providers. An independent evaluation of DelCAN, assessing effects of the intervention on statewide rates of unintended pregnancy and related outcomes, is currently ongoing, in conjunction with a multiple methods process evaluation. The evaluation is based on data collected through various surveys. This dissertation will use data from one of these datasets for its analyses. Recently published findings from the ongoing independent evaluation by Boudreaux et al. (2020) suggest that LARC use among Title X patients aged 20-44 increased by 3.2- percentage-points when compared to other U.S. states, a 40% increase from baseline. These data support the positive trend identified by Welti & Manlove (2018). On the other hand, a qualitative study based on 86 interviews conducted in 2016-2017 with women living in Delaware found that 56 the patient-provider interaction exacerbated rather than ameliorated the LARC barriers of knowledge and side-effects, as did practice structures that prevented same-day LARC insertion (Berndt & Bell, 2021a). However, the study pertained to women?s lifetime experiences of contraceptive care and does not identify timepoints of the patient-provider interaction; therefore, it is unclear whether the observed and reported patient-provider interactions and practice structure were influenced by DelCAN training to the benefit of contraceptive patients. However, in a related study using the same sample, ?several women described how their clinics have now allotted additional time in appointments for contraceptive counseling with great success? (Manzer & Bell, 2022a, p. 90). Considering the date of interviews and DelCAN?s emphasis on contraceptive counseling, it is plausible that these changes were a result of the DelCAN initiative. 2.6.6. Summary Despite being unable to infer causality in prior research about the rise in LARC use through increasing access, the associations identified by disparities literature and city and state- wide initiatives offer encouraging evidence that reducing or eliminating individual and structural barriers may have a positive impact on choice of more effective contraceptive methods, which, in turn, reduces rates of unintended pregnancies, unintended births, and abortions. Furthermore, evaluations of the HER Salt Lake Contraceptive Initiative and the Colorado Family Planning Initiative suggest that an effective media campaign has a key role to play in raising awareness 57 about available services since in both instances the demand for LARC increased after the launch of the campaign. 2.7. Conceptual Framework Over time, many theoretical models have been used to explain and identify factors that may influence health-related behaviors, including contraceptive behaviors (Hall, 2012; Lopez et al., 2016; Marcinkowski et al., 2021). The mechanisms at play examined by the three research questions are best understood using the Integrated Behavioral Model (Monta?o & Kasprzyk, 2015). However, the tenets of Reproductive Justice (SisterSong Women of Color Reproductive Justice Collective, 2021) are also relevant to the interpretation of findings. 2.7.1. Integrated Behavioral Model (IBM) In the early 1990s, at the behest of the National Institute of Mental Health, an intensive weeklong workshop took place, resulting in a common theoretical model that integrated core constructs from the Theory of Reasoned Action (and the subsequent Theory of Planned Behavior), social learning theory (and the subsequent Social Cognitive Theory), various iterations of the Health Belief Model, and the theory of subjective culture (Fishbein et al., 1992; Jaccard et al., 2002). The general framework serves as the foundation for today?s streamlined Integrated Behavioral Model (IBM), also known as the Integrative Model (IM) (Jaccard et al., 58 2002; Monta?o & Kasprzyk, 2008, 2015). The current understanding of the IBM is described in diagram form in Figure 1. According to the IBM, the strongest predictor of conducting a specific behavior is the intention to perform it, and intent is influenced by an amalgamation of a variety of beliefs and feelings held by an individual. However, the IBM emphasizes that intent to perform a behavior is not enough. Knowledge and skills, in addition to salience of the behavior, environmental constraints, and habit, also influence whether an intended behavior will be undertaken or not (Monta?o & Kasprzyk, 2015). Applying the IBM to contraceptive behavior means that an individual?s intent to use a specific contraceptive method may be the strongest predictor of using that method. Precursors of intent are an individual?s (a) attitudes, (b) perceived norms, and (c) personal agency regarding contraception. According to IBM, this multiplicity of personal beliefs explains much of the variance in behavioral intent (Monta?o & Kasprzyk, 2008). (a) Attitudes are further distinguished as experiential attitudes and instrumental attitudes. Experiential attitudes relate to how using contraception makes one feel, such as feeling relieved that a long-acting reversible contraceptive method requires no effort post-insertion or feeling that it is too much hassle to use a condom during every sexual encounter. Instrumental attitudes pertain to the advantages and disadvantages of using a contraceptive method, such as one?s beliefs regarding the side-effects of a specific method. (b) Perceived norms refer to the perception of others? expectations and behaviors regarding one?s contraceptive use, such as parental disapproval of using hormonal contraception and the expected fallout. (c) Personal agency entails perceived control and self- 59 efficacy. Based on control beliefs, perceived control is one?s perception of environmental facilitators and barriers to contraceptive use, such as the belief that a medical provider might not be willing to remove a long-acting reversible contraceptive upon request. In turn, self-efficacy is the level of confidence in one?s capacity to use or to stop using a contraceptive method even when challenges arise, such as having confidence in one?s self-advocacy abilities to obtain a LARC removal when faced with resistance. Aim 3 explores changes in experiential and instrumental attitudes regarding specific contraceptive methods, as well as perceived control over LARC removal. Changes in attitudes about contraception may prompt women to consider a wider range of contraceptive methods, allowing them to choose the most effective method that suits their needs. In spite of intending to use some form of contraception, individuals may not be able to carry out their intended behavior for a variety of reasons. As the IBM emphasizes, engaging in a specific behavior requires that the individual have knowledge and skills to do so (Monta?o & Kasprzyk, 2008). Therefore, using contraception as a preventative health-seeking behavior is only possible among those who are aware of functionality. Knowing about the range of available contraceptive methods is also important because one?s current health status and medical history, as well as access and patient preferences, play a key role in identifying the most appropriate method for an individual at a specific time in their life (Centers for Disease Control and Prevention, 2020; Dehlendorf, Rodriguez, et al., 2010). Lastly, knowing the correct information about each method is crucial for individuals to be able to consider a method as a feasible option for them. Misinformation about contraception is not uncommon, particularly around LARC. For 60 example, a popular misconception is that IUDs cannot be used by adolescents and nulliparous women (Belfield, 2009; Russo et al., 2013). Aim 2 examines changes in knowledge regarding specific contraceptive methods. An improvement in contraceptive knowledge may increase the likelihood of women choosing more appropriate methods for their current situation, thus influencing their contraceptive behavior. The IBM posits that another factor that may facilitate or hinder behavior are environmental constraints, of which there are many. This is particularly relevant for contraception because most methods available today require a clinician?s prescription and/or financial coverage, thus placing providers and the health care system in the role of intermediaries between the individual?s contraceptive intent and contraceptive use and behavior, as most recently described by Delawarean women (Manzer & Bell, 2022a). Therefore, the clinician and the health care system fit within the scope of IBM?s environmental constraints. Aim 1 investigates changes in participants? contraceptive use and intent prior to and following an interaction with the provider. By utilizing two different measures before the clinic visit, Aim 1 gets partially at changes that occurred due to the provider visit. Such changes may be a result of providing new information or facilitating access to desired methods for patients. As emphasized by the DelCAN initiative, the clinic visit serves as a source of information, potentially modifying the patients? knowledge and attitudes about contraception, which are investigated in Aims 2 and 3 as potential factors that may influence women?s intent and future behavior. Lastly, for all three aims, investigating the effect of the wave of data collection seeks to assess whether having the program in place longer influenced the outcome. 61 2.7.2. Reproductive Justice In June of 1994, a Black caucus of women gathered in Chicago in advance of their attendance at the International Conference on Population and Development in Cairo (Ross & Solinger, 2017). Finding that the mainstream women?s rights movement dominated by middle class and wealthy White women could not adequately represent their needs, they sought to form their own movement, and in doing so, they reconceptualized reproductive rights by combining it with social justice values and rooting it in the United Nation?s human rights framework (Ross & Solinger, 2017; SisterSong Women of Color Reproductive Justice Collective, 2021). Today, reproductive justice is defined as ?the human right to maintain personal bodily autonomy, have children, not have children, and parent the children we have in safe and sustainable communities? (SisterSong Women of Color Reproductive Justice Collective, 2021). Women report that effectiveness of a contraceptive method is repeatedly found to be among the most important features when choosing a contraceptive method (Bracken & Graham, 2014; Grady et al., 1999; Lessard et al., 2012). However, women also indicate taking into consideration a method?s safety, ease of use, impact on sex life, and side-effect profile, including influence on menstruation (Brown et al., 2011; Donnelly et al., 2014; Grady et al., 1999; Higgins et al., 2020; Lessard et al., 2012; Stanwood & Bradley, 2006). Some studies find that sociodemographic characteristics, such as age, race, and socioeconomic status, influence which contraceptive attributes are reported by women to be more or less important (Brown et al., 2011; Lessard et al., 2012). 62 The timeline of a future desired pregnancy is also relevant to consider in contraceptive choice. For instance, as LARC is provider-controlled, patients may encounter the same barriers to LARC-removal that exist for LARC-insertion, such as limited or no access to care and lack of correct knowledge. Additionally, due to a combination of the high cost of LARC devices and institutional and provider bias, evidence is mounting about patients? experiences of delays and denials for ?early? LARC removal, based on reports by both patients and providers (Amico et al., 2016, 2017; Kaneshiro et al., 2020; Manzer & Bell, 2022b). On the other hand, the commonly used injectable, depo-medroxyprogesterone acetate (a.k.a. DMPA, the shot, or injectable), can be discontinued by the patient simply by not attending their three-month follow-up visit. However, return to fertility, i.e., the resumption of ovulation, is almost always delayed and with a very wide range of variability; studies conducted in various countries report an average of five to eight months after the three months since the last shot?arguably the longest timeframe for any reversible contraceptive method currently available (Jain et al., 2004; Pardthaisong et al., 1980; Paulen & Curtis, 2009; Schwallie & Assenzo, 1974; Yland et al., 2020). Thus far, there is no evidence that the lower dosage (104 mg) of subcutaneous DMPA results in a faster return to fertility than the intramuscular shot (150 mg) (Jain et al., 2004), and a very recent clinical trial tested DMPA?s contraceptive effectiveness beyond the currently approved three months of protection against pregnancy with positive results (Halpern et al., 2021; D. J. Taylor et al., 2022). The aforementioned evidence on women?s perspectives and contraceptive side-effects demonstrates that there is more to choosing a contraceptive method than effectiveness alone. One goal of this dissertation is to better understand the provider?s role in women?s contraceptive 63 choices, knowledge, and attitudes within the context of an intervention that recommended tiered- effectiveness contraceptive counseling, a strategy that suggests to patients that LARC is the optimal contraceptive method. Making uncoerced contraceptive choices that are based on correct information is a requisite to maintaining one?s bodily autonomy for women seeking to avoid pregnancy. Therefore, using a reproductive justice lens can help to avoid the pitfall of instrumentalizing women?s bodies in the name of public policy. This will provide for a more holistic interpretation of the evaluation findings of the DelCAN initiative. It should be noted that while Reproductive Justice aids in interpreting our findings, we use it as a lens rather than a full-fledged theoretical framework per se. Although DelCAN allowed for patients to choose any method of contraception free of charge, its provider and clinic staff training focused specifically on LARC, as evidenced by tiered-effectiveness counseling and teaching of device insertion. Indeed, the goal of DelCAN was to prevent unintended pregnancy, not to ensure reproductive justice; in that context, the spotlight on LARC is understandable. Therefore, as much as we endeavored to not give primacy to any specific method in our research questions, our specific aims are circumscribed by the availability of data collected with the DelCAN Title X survey, and the questionnaire focused heavily on LARC. Since neither the program nor the survey were created with Reproductive Justice at its core, we use the framework to aid the interpretation of our results rather than to apply it to the initiative after the fact. 64 2.8. Current Study The reproductive health literature on the effect of a provider visit on contraceptive behaviors, knowledge, and attitudes carries several limitations. First, only two studies have assessed change in contraceptive choice following a clinic visit by measuring use or intention both pre-visit and post-visit, and both focus solely on combined hormonal contraceptives (CHC) (pill, patch, ring) (Bitzer et al., 2012; Dehlendorf, Henderson, et al., 2016). Second, contraceptive counseling that includes a provider visit overwhelmingly has a positive effect on patient contraceptive knowledge (Little et al., 1998; Mason et al., 2003; Ragland et al., 2015; Winter & Breckenmaker, 1991), but knowledge measures vary widely across survey instruments (de Reilhac et al., 2016; Haynes et al., 2017; Lee et al., 2015; Nobili et al., 2007) and of the few studies that evaluated any LARC knowledge, only one used a pre-post clinic visit design (Nobili et al., 2007). Third, very few studies looked at changes in attitudes following contraceptive counseling with a clinician and none measured LARC attitudes (Chewning et al., 1999; Larsson et al., 2004b; Nobili et al., 2007). Lastly, such studies are frequently conducted in one or two, usually urban, sites, with statewide studies more focused on measuring unintended pregnancy and abortion rates as well as general LARC uptake (Biggs et al., 2015; D. A. Grossman et al., 2012; Kelly et al., 2020). Using pre-post test data from a sample of women recruited from multiple clinics statewide can provide us with a better understanding of the effect of a provider visit on patients? contraceptive intentions within the full spectrum of available reversible methods, among both family planning and non-family-planning patients. Additionally, despite increased uptake of 65 LARC in the past decade, there is no evidence on whether LARC knowledge and attitudes improve following a clinic visit, beyond general awareness of LARC?s existence. Examining factors associated with changes in contraceptive intentions, knowledge, and attitudes can offer insight whether the initiative had a differential impact on women based on sociodemographic, reproductive health, or provider visit characteristics. This dissertation addresses these gaps in knowledge. 66 Chapter 3: Methods 3.1. Data Source and Population As part of a larger, multi-component longitudinal evaluation of the DelCAN initiative, a self-report survey about contraceptive knowledge, attitudes, and use was fielded in English and Spanish in two waves in Delaware?s health care facilities receiving Title X funds and participating in DelCAN (Choi et al., 2019). All women of reproductive age (15-44) who were at the clinic to receive care for themselves were eligible to participate. Women were surveyed at two points; the first part of the questionnaire survey was self-administered on tablet computers in the clinic waiting rooms before the woman?s visit with her health provider (pre-visit), and the second part of the survey was self-administered on tablet computers after the visit (post-visit). If a participant did not wish to fill out the survey herself, the recruiter read and input the answers on behalf of the participant. To ensure a probability-based sample representative of the entire Delaware Title X system, sampling was done at three levels: the clinic level, the interview time level, and the patient level. Wave 1 data was collected by a team of 30 recruiters (all women) in Title X clinics in Delaware between June and December 2017. Wave 2 data was collected by a team of 17 recruiters (all women) in Title X clinics in Delaware between November 2018 and October 2019. 67 3.2. Analytic Sample The current study uses data from two waves of the Title X survey, conducted in 2017 and 2018-2019. Between the two waves, a total of 929 eligible participants, who were at the clinic seeking care for themselves, consented to take part in the survey. A diagram describing the rules for inclusion and exclusion in the analytic sample is presented in Figure 2. First, we excluded women who were not at risk of unintended pregnancy (i.e., those who were pregnant or found out they were pregnant during the visit, or those who came to the clinic because they were trying to get pregnant, were seeking prenatal care, or had previously been sterilized or had a hysterectomy), narrowing down the sample to 727 participants. Then, due to a skip pattern gone awry in the first wave, we had to exclude 79 participants for missing on a significant portion of the pre-visit survey in Wave 1, including current and planned contraceptive use questions.7 An additional 24 participants were also missing on those questions due to an intended skip pattern that excluded some participants who said yes to the question, ?Has a doctor ever told you that you are infertile (can?t get pregnant)?? From the remaining sample of 624 individuals, we excluded participants who did not complete the post-visit survey (N=51), were missing on the pre- or post-contraceptive use questions (N=6), or whose answers to those questions were contradictory (N=3) (ex. indicating they are using both the IUD and the implant). Furthermore, we excluded 74 participants for 7 The participant was not shown the pre-visit contraceptive use and intention questions if: (1) she reported ever having had sexual intercourse with a male and she left blank the question ?How many times have you been pregnant in your life?? (18 participants); (2) she reported having sexual intercourse a few times a month in the past 3 months and answered 0 or 00 to ?How many times have you been pregnant in your life?? (44 participants); or (3) she answered, ?How many times have you been pregnant in your life?? with words instead of numbers (17 participants). 68 missing on any of the covariates. Lastly, due to the very small sample size (N=16) and fundamental difference between seeking sterilization as a form of contraception, we excluded all women who reported using or planning to use female or male sterilization as their most effective form of contraception in either the pre- or post-visit. The final analytic sample is 474. One element of the DelCAN initiative was to train providers to ask the One Key Question about pregnancy intention to all incoming female patients of reproductive age, and counsel them on contraception if appropriate. As a result, we included all women at potential risk of unintended pregnancy, regardless of what type of visit she had at the clinic, as long as the visit was for herself. Since the sampling weights available in the dataset pertain only to family planning users, we were not able to conduct a weighted analysis of my results. 3.3. Variables 3.3.1. Aim 1 Outcomes The main outcomes for Aim 1 are: (1) change from pre- to post-visit plan to use LARC; (2) change from pre- to post-visit plan to use a contraceptive method of a higher effectiveness level; (3) change from pre- to post-visit plan to use a contraceptive method of a lower effectiveness level. The outcome variables are based on creating exclusive method type variables, as seen in the prior literature (National Center for Health Statistics, 2019b; Trussell & Vaughan, 1999), where participants who report using more than one method were coded as using the method of 69 highest effectiveness level based on typical failure rates (Trussell et al., 2018). Furthermore, two coding processes for pre-visit and one for post-visit were implemented in order to be able to create the different change measures. The first pre-visit variable captures the effectiveness level of the method currently used by the participants prior to the provider visit. The second pre-visit variable captures the effectiveness level of the method the participants report planning to use prior to the provider visit. Having both of these allows us to gain some perspective regarding the role of the provider in contraceptive plans. Pre-visit Coding Process 1: Pre-visit Current Method Use The pre-visit current contraceptive method use variable was measured by the method of highest effectiveness that each participant reported currently using, based on the question (1) ?Do you currently use any of these birth control methods?? The questionnaire presented a list of fourteen contraceptive methods: (A) Withdrawal (also called ?pulling-out?), (B) Birth control pills, (C) IUD (Mirena, Paragard, Skyla, Kyleena or Liletta), (D) Implant (Nexplanon), (E) Vaginal ring (Nuvaring or other), (F) Depo-Provera (also called ?the shot?), (G) Birth control patch (Evra or other), (H) Male condoms, (I), Natural family planning methods (also called calendar/rhythm method, cycle beads, basal body temperature), (J) Male sterilization (vasectomy), (K) Female sterilization (tubes tied or tubal ligation, Essure), (L) Barrier methods (diaphragm, sponge, cervical cap, female condom), (M) Emergency contraceptives, (N) Other method, please describe (with attached fill-in-the-blank space). Next to each presented method, the participant could choose between the ?yes? and ?no? fields, or skip. 70 Pre-visit Coding Process 2: Pre-visit Composite Current-Planned Method Use The Pre-visit Composite Current-Planned Method Use variable utilized the question used to create the Pre-visit Current Method Use variable, as well as two other questions from the pre- visit questionnaire, as follows: (2) ?Over the next 3 months, do you plan to quit using any of the methods that you're currently using?? Possible survey answers are ?yes,? or ?no.? If they answered ?yes,? they were asked the (2b) question, ?Over the next 3 months, what birth control methods do you plan to QUIT using?? and were presented with the list of methods they had said ?yes? to in the prior (1) question. (3) ?Over the next 3 months, do you plan to start using a new kind of birth control method?? The answer options were ?yes? and ?no.? If they answered ?yes,? they were asked the (3b) question, ?What new forms of birth control are you planning on using?? and presented with a list of methods they had not said ?yes? to in the prior (1) question. Participants who indicated they did not plan to quit a method or planned to quit a method of lower effectiveness than the one they were currently using had their pre-visit contraception use coded as (1) currently used method or (3) new method planned, whichever was of higher effectiveness. Participants who indicated that they planned to quit the most effective method they were currently using were coded according to the next method of highest effectiveness they were currently using but were not planning to quit or the method they planned to use, whichever was 71 of higher effectiveness. If they indicated they were planning to quit the only method they were using and had no plans for another method, they were coded as no method. If participants were currently using no method and did not plan to start a new method, they were also coded as no method. Post-visit Coding Process: Post-visit Planned Method Use The post-visit planned contraception use measure was based on two questions from the post-visit questionnaire: (1) ?During your visit did you start or get a prescription for a new birth control method that you are not currently using?? The survey answers are dichotomized as ?yes? or ?no.? If they answered ?yes,? they received the (1b) question, ?What new method or methods did you start or get a prescription for? Check all that apply.? Possible answers are: (A) Withdrawal (also called ?pulling-out?), (B) Birth control pills, (C) IUD (Mirena, Paragard, Skyla, Kyleena or Liletta), (D) Implant (Nexplanon), (E) Vaginal ring (Nuvaring or other), (F) Depo-Provera (also called ?the shot?), (G) Birth control patch (Evra or other), (H) Male condoms, (I), Natural family planning methods (also called calendar/rhythm method, cycle beads, basal body temperature), (J) Barrier methods (diaphragm, sponge, cervical cap, female condom), (K) Female sterilization (tubes tied or tubal ligation, Essure), (L) Emergency contraceptives, (M) Other method, please describe (with attached fill-in-the-blank space). Participants were shown the complete list, regardless of what their answers in the pre-visit survey. 72 (2) ?Over the next 3 months, do you plan to use any birth control method?? Possible survey answers are ?yes,? ?no,? or ?not sure.? If they answered ?yes,? they were asked the (2b) question, ?Over the next 3 months, which of the following birth control methods do you plan to use?? They were presented with a list of fourteen answers (the same thirteen as for question (1), with the addition of ?male sterilization (vasectomy)?), with ?yes? and ?no? fields next to each method. Women?s planned contraceptive use post-visit was coded based on the method of highest effectiveness level they reported having gotten a new prescription for or were planning to use over the next three months, whichever method was more effective. Participants who indicated that they did not receive a new prescription and that they did not plan to or were ?not sure? about their plans to use a birth control method over the next three months were coded as ?none? for post-visit planned contraception use. 3.3.1.1. Change to Long-Acting Reversible Contraception (LARC) Participants? exclusive classifications of pre-visit and post-visit methods were collapsed into a dichotomous variable: (0) none or non-LARC method, and (1) LARC method. The change measure was calculated using the formula of subtracting the pre-visit variable from the post-visit variable (i.e., post-visit minus pre-visit). A positive difference indicated change from a non- LARC to a LARC method. Zero indicated no change; the participant either reported a LARC method pre- and post-visit or reported a non-LARC method both pre- and post-visit. A negative difference meant a change away from a LARC method post-visit. Change to LARC was 73 calculated by both pre-visit measures: the current method and the composite current-planned method. 3.3.1.2. Change to a Higher and Lower Effectiveness Level of Contraceptive Method Participants? exclusive classifications of pre-visit and post-visit methods were classified by effectiveness level into one of three categories: (0) none or low effective methods; (1) moderately effective methods; and (2) long-acting reversible contraception. The change measure was calculated using the formula of subtracting the pre-visit variable from the post-visit variable (i.e., post-visit minus pre-visit). A negative difference indicated change to a method of a lower effectiveness level, a difference of zero indicated no change in effectiveness level, and a positive difference indicated change to a method of a higher effectiveness level. Changes to a method of higher effectiveness and to a method of lower effectiveness were calculated by both pre-visit measures?the current method and the composite current-planned method. 3.3.2. Aim 2 Outcomes The main outcomes for Aim 2 are changes from pre- to post-visit knowledge about (1) the IUD, (2) the implant, (3) effectiveness of contraceptive methods to prevent pregnancy, and (4) the benefits of the DelCAN initiative. Each of the four changes was operationalized into two outcomes. Both outcomes were based on the same set of survey questions, as explained in 74 subsequent sections. The survey questions are based on up-to-date facts about contraceptive methods and service provision. The first question in the knowledge section of the survey was: ?Which of these birth control methods have you heard of? Check all that apply,? followed by these multiple-choice options: ?(1) Implant (Nexplanon); (2) IUD or Intrauterine Device (ParaGard, Mirena, Liletta, Skyla); (3) The shot (Depo-Provera); (4) Birth Control Pills; (5) I have not heard of any of these.? This was a key question because it served as a gateway to multiple other survey questions; every item not checked triggered a different skip pattern. For instance, participants who left the implant option unchecked were skipped on all true-false knowledge questions that pertained solely to the implant; the same occurred with not checking the IUD. In the post-visit survey, participants were asked the same question again, but, of the four options, were only shown methods that they left unchecked in the pre-visit survey. As a result, some participants received questions about the IUD and/or implant only in the post-visit survey. The scoring for participants who did not receive all questions used in our outcomes is explained in the following sections, by outcome. 3.3.2.1. Change in IUD Knowledge The IUD knowledge variable was created from six survey questions: the aforementioned ?heard of? question and five true-false questions. The true-false questions also offered a third option, ?I don?t know,? and were each phrased in two different ways in order to reduce acquiescence bias, as follows: 75 (1) Women can use an IUD even if they?ve never had children (group A; correct answer: True); Only women who have had children can use an IUD (group B, correct answer: False) (2) There is a type of IUD without hormones (group A; correct answer: True); Every type of IUD has hormones (group B; correct answer: False) (3) If you use an IUD it might affect your ability to have a child in the future, even if you have the IUD taken out (group A; correct answer: False); Using an IUD will not affect your ability to have a child in the future, after you have the IUD taken out (group B; correct answer: True) (4) Teenagers can use an IUD (group A; correct answer: True); Teenagers cannot use an IUD (group B; correct answer: False) (5) The IUD cannot be removed early, even if a woman changes her mind about wanting to get pregnant (group A; correct answer: False); The IUD can be removed early if a woman changes her mind about wanting to get pregnant (group B; correct answer: True). Once participants reported having heard of an IUD or implant, the survey instrument randomly assigned them to group A or group B and showed them either questions from the A block or the B block. If assigned group A pre-visit, the participant also received the A questions post-visit.8 Our IUD knowledge variable was a score ranging from 0 to 6. Having heard of the IUD was scored with one point, and each correct response to the true-false questions was one point, allowing for a range of 0 to 6 points. Those who had not heard of the IUD received zero points for the ?heard of IUD? question, as well as all IUD knowledge questions (they did not receive the IUD questions because they had not heard of the IUD). Those who answered incorrectly or ?I 8 Seventeen participants in our sample had data for both the A and B group in post-visit IUD and implant knowledge questions; the survey team is still investigating why this occurred. These participants were included in our sample by using the post-survey data that matches their pre-survey group assignment, as occurred with the rest of the sample. 76 don?t know? to any of the true-false questions were scored zero on that question, as seen in other studies (de Reilhac et al., 2016; Nobili et al., 2007; O?Donnell et al., 1995; Ragland et al., 2015). Change in IUD knowledge was evaluated with two outcome variables. The first IUD knowledge change measure captured participants? individual percentage change in knowledge. Participants? pre-score was subtracted from their post-score and then divided by the number of points they could have increased, i.e., the difference between the highest possible score (6) and their pre-score. A positive percentage indicated how many percentage points the participants increased in knowledge, and a negative percentage indicated a decrease in knowledge. The second IUD knowledge change measure was calculated by coding any individual increase as a success, i.e., whenever the post-score was higher than the pre-score. A post-score that was lower or the same as its corresponding pre-score was coded as a failure. 3.3.2.2. Changes in Implant Knowledge The implant knowledge variable was created from four survey questions: the aforementioned ?heard of? question and three true-false questions. The true-false questions also offered a third option, ?I don?t know,? and were each phrased in two different ways in order to reduce acquiescence bias, as follows: (1) If you use an implant it might affect your ability to have a child in the future, even if you have your implant taken out (group A; correct answer: False); Using an implant will not affect your ability to have a child in the future, after you have the implant taken out (group B; correct answer: True) 77 (2) Teenagers can use an implant (group A; correct answer: True); Teenagers cannot use an implant (group B; correct answer: False) (3) The implant cannot be removed early, even if a woman changes her mind about wanting to get pregnant (group A; correct answer: False); The implant can be removed early if a woman changes her mind about wanting to get pregnant (group B; correct answer: True) The implant knowledge variable was a score between 0 to 4. Having heard of the implant was scored as one point, and each correct response to the true-false questions was one point each, allowing for a range of 0 to 4 points. Those who had not heard of the implant received zero points for the ?heard of implant? question, as well as all implant knowledge questions (since they did not receive them, as explained earlier). Incorrect and ?I don?t know? answers were scored zero. The randomization to the A and B group occurred in the same manner as for the IUD knowledge questions. Changes in implant knowledge was evaluated with two outcome variables. The first implant knowledge change measure captured participants? individual percentage change in knowledge. Participants? pre-score was subtracted from their post-score and then divided by the number of points they could have increased, i.e., the difference between the highest possible score (4) and their pre-score. A positive percentage indicated how many percentage points the participants increased in knowledge, and a negative percentage indicated a decrease in knowledge. The second implant knowledge change measure was calculated by coding any individual increase as a success, i.e., whenever the post-score was higher than the pre-score. A post-score that was lower or the same as its corresponding pre-score was coded as a failure. 78 3.3.2.3. Changes in Effectiveness Knowledge Knowledge about contraceptive effectiveness was measured with the question, ?Which birth control method is the most effective in preventing pregnancy? Please rank these methods from 1st (most effective) to 4th (least effective),? followed by a list of ?The pill; IUD or implant; Male condoms; The shot (Depo-Provera).? Unlike for the prior two outcomes, all participants received this question, even if they had never heard of the IUD and implant. Participants who ranked all methods in the correct order (LARC, the shot, the pill, external condom) scored four points on our effectiveness knowledge variable. Participants scored three points if they ranked LARC first and condoms last, two points if they ranked LARC first, one point if they ranked condoms last, and zero points for any other ranking combination. The scoring categories were exclusive one from another. Changes in effectiveness knowledge was evaluated with two outcome variables. The first measure captured improvement in knowledge, where any increase in score from pre- to post-visit was coded as a success, indicated by any positive difference resulting from the subtraction of the pre-score from the post-score. The second measure of change coded as success only an increase to a perfect score of 4 post-visit, indicating that a participant with less than 4 pre-visit ranked all methods correctly post-visit. All other participants? change scores were coded as failure. 79 3.3.2.4. Changes in Knowledge about the Benefits of the DelCAN Initiative The DelCAN benefits knowledge variable, ranging from 0 to 7 points, was created from two survey questions. One question stated, ?True or False: You can get any birth control method that you want with only one office visit to a clinic or health care provider,? and offered multiple choice answers of ?True,? ?False,? ?It depends on the type of birth control method you want,? and ?I Don?t Know.? Since one element of the DelCAN initiative was to ensure that all reversible contraceptive methods could be obtained in the same visit they were requested, ?True? was scored with one point and all other options with zero points. The second question was, ?Do you know how YOU can get any of the following birth control methods for free (by FREE we mean you don?t have to pay anything out-of-pocket)?? followed by a list of six contraceptive methods (?Male condoms; IUD (Mirena, Paragard, Skyla, Kyleena, or Liletta); Implant (Nexplanon); Depo-Provera (also called ?the shot?); NuvaRing (vaginal birth control ring); Birth control pills?). For each method, participants could answer ?Yes,? ?No,? or ?I?m not sure.? ?Yes? answers were scored as one point each; the other answers were scored as zero. Changes in DelCAN benefits knowledge was evaluated with two outcome variables. The first measure captured participants? individual percentage change in knowledge. Participants? pre-score was subtracted from their post-score and then divided by the number of points they could have increased, i.e., the difference between the highest possible score (7) and their pre-score. A positive percentage indicated how many percentage points the participants increased in knowledge, and a negative percentage indicated a decrease in knowledge. The second DelCAN benefits knowledge change measure was calculated by coding any individual 80 increase as a success, i.e., whenever the post-score was higher than the pre-score. A post-score that was lower or the same as its corresponding pre-score was coded as a failure. 3.3.3. Aim 3 Outcomes The main outcomes for Aim 3 are changes from pre- to post-visit attitudes about (1) the IUD, (2) the implant, (3) hormonal birth control, and (4) condoms. Each of the four changes was operationalized into two outcomes, one continuous and one dichotomous. Both outcomes were based on the same set of statements, as explained in the following paragraphs. Each statement presented a positive or negative feature or belief about a type of contraceptive method. As explained in the Aim 2 Outcomes section, the first knowledge question asking participants which birth control methods they had heard of served as a crossroads to future skip patterns. Participants who indicated they had not heard about the IUD or the implant did not receive attitude questions about the specific method(s). As a result, some participants received questions about the IUD and/or implant only in the post-visit survey and others never received the questions. For our change analyses, we only included participants who had received and answered the attitude questions both pre- and post-visit; participants who had not heard about a method could not have an attitude or change in attitude about it. Descriptive results from participants who only received the questions post-visit are reported separately (Whitaker et al., 2010). 81 3.3.3.1. Attitude Variables The IUD attitude variable was created from four survey questions. Participants were asked to rate their agreement with the following statements: (1) It?s a relief to have an IUD or implant that works without having to do anything. (2) My provider will remove my IUD whenever I decide I?d like it removed. (3) IUDs are uncomfortable. [we reverse coded this item] (4) IUDs have bad side effects such as weight gain and irregular bleeding. [we reverse coded this item] The implant attitude variable was created from two survey questions. Participants were asked to rate their agreement with the following statements: (1) It?s a relief to have an IUD or implant that works without having to do anything. (2) My provider will remove my implant whenever I decide I?d like it removed. Hormonal birth control attitude was measured by participants? level of agreement with one statement: ?Mood swings become worse with hormonal birth control? [we reverse coded this item]. This question was asked of all participants regardless of any prior answer about familiarity with a specific contraceptive method. Condom attitude was measured by participants? level of agreement with one statement, ?It is too much hassle to use a condom every time I have sex? [we reverse coded this item]. The question was asked of all participants regardless of any prior answer about familiarity with a specific contraceptive method. 82 3.3.3.2. Scoring Attitudes The response options for the survey attitude questions came in the form of a 5-point Likert agreement scale consisting of Strongly Agree, Agree, Neither Agree nor Disagree, Disagree, Strongly Disagree, and Don?t Know as a sixth option. Statements highlighting a positive feature about a contraceptive method were coded as follows: 4 points for ?strongly agree,? 3 for ?agree,? 2 for ?neither agree nor disagree? and ?don?t know,? 1 for ?disagree,? and 0 for ?strongly disagree.? Statements that identify a negative feature about a contraceptive method were reverse coded as follows: 4 points for ?strongly disagree,? 3 for ?disagree,? 2 for ?neither agree nor disagree? and ?don?t know,? 1 for ?agree,? and 0 for ?strongly agree.? Consequently, a higher score on an attitude measure indicated more agreement with positive aspects of the method or more disagreement with negative aspects of a method, regardless of the wording of each question (Delamater et al., 2000; Nobili et al., 2007). For short, we refer to participants? agreement with positive aspects and disagreement with negative aspects as positive attitudes, and thus, we discuss changes in positive attitudes. IUD attitude scores ranged from 0 to 16. Implant attitude scores ranged from 0 to 8. Hormonal birth control attitude scores ranged from 0 to 4. Condom attitude scores ranged from 0 to 4. 3.3.3.3. Changes in Attitudes We calculated a continuous and a dichotomous variable for changes in attitudes for each of our 4 outcomes (IUD attitudes, implant attitudes, hormonal birth control attitudes, and 83 condom attitudes). The first continuous change measure was calculated by subtracting participants? pre-visit score from their post-visit score. A positive number indicated that participants moved towards holding a more positive attitude, and a negative number indicated a shift towards a less positive (or more negative) attitude; a difference of zero indicated no change in overall attitude score. The second attitude change measure assessed whether participants had any change towards more positive attitudes and was calculated by coding any individual increase as 1, i.e., whenever the post-visit score was higher than the pre-visit score. A post-visit score that was lower or the same as its corresponding pre-score was coded as 0. 3.3.4. Predictors Various sociodemographic, reproductive health, and provider visit characteristics served as the main predictor variables for Aims 1.2, 2.2, and 3.2; their operationalization is explained in the following paragraphs. Additional predictors were also included for each aim: (1) the effectiveness level of contraceptive method reported pre-visit for Aims 1.2 and 1.3; (2) the pre- visit score of the knowledge outcomes for Aims 2.2 and 2.3, as well as the assigned randomized group (A or B) for the IUD and implant knowledge outcomes in Aims 2.2 and 2.3; and (3) the pre-visit score of the attitude outcomes for Aims 3.2 and 3.3. The operationalization of these additional predictors is explained in the preceding Outcome sections. The main predictor variable for Aims 1.3, 2.3, and 3.3, wave of data collection, is also explained below. Age was analyzed as a categorical variable in the following groups: 15-19, 20-29, and 30- 45. 84 Race or Hispanic ethnicity is a composite variable based on the Hispanic ethnicity and race questions in the survey: those who reported being of Hispanic, Latina, or Spanish origin were categorized as Hispanic, regardless of race. Those who were not Hispanic, but chose Black or African American were categorized as non-Hispanic Black, even if they also checked another race. Participants who were not Hispanic or Black, and checked the Asian or Asian American, the Native American, Alaska Native, or American Indian, or the Native Hawaiian or Pacific Islander box, were categorized as non-Hispanic Other. Participants were categorized as White if they were non-Hispanic and only checked the White box. One participant answered only ?Other? and wrote in ?Egyptian.? She was categorized as White, in line with Census and national statistics that group people of Middle Eastern and North African descent as White. Current relationship status was grouped into three exclusionary categories: currently single, currently married, and currently cohabitating. Insurance type was analyzed categorically, with three exclusive categories of private, public, and none. Emergency Medicaid was categorized as ?none,? while student and military insurance was coded as ?private.? Education level was also examined categorically, with participants grouped by ?high school or less,? ?some college/vocational,? and ?bachelor?s degree or higher.? Nativity was dichotomized into U.S.-born and foreign-born. Future pregnancy desire was based on the survey question, ?How do you feel about having a child now or sometime in the future??; women?s answers were recoded into four 85 categories: (a) wanting a child within the next two years, (b) wanting a child in 2 years or more, (c) wanting a child but not sure when or unsure about wanting a child, and (d) not wanting a(nother) child. Feelings about a hypothetical pregnancy in the next year were based on the survey question, ?In the next year, how happy would you be if you got pregnant?? Answers were collapsed into three categories: (a) very happy and happy; (b) very unhappy or unhappy; and (c) unsure. Lifetime experience of unintended pregnancy was dichotomized as ?yes? or ?no.? Types of lifetime experience of reproductive coercion, measured by the short-form reproductive coercion scale (McCauley et al., 2017), was analyzed with two categories of experience: no reproductive coercion and any type of reproductive coercion (verbal only and/or behavioral). Reason for clinic visit was dichotomized as a visit for the purposes of family planning or not. The clinic focus variable was dichotomized as primary care or specializing in women?s health. The clinic location covariate was based on the geographical location of the clinic where the visit took place: urban, suburban, or rural. The wave variable pertains to whether the data was collected as part of the first or second wave. Wave indicated the length of time since the launch of the DelCAN initiative?wave one 86 occurred during the earlier part of the initiative (2017) and wave two when the initiative was well under way (2018-2019). 3.4. Analyses All analyses were conducted using IBM SPSS Statistics 28 and Stata/MP 17.0 for Windows. 3.4.1. Aim 1 Analyses We described changes in planned contraceptive use from pre-visit to post-visit (Aim 1.1) by calculating participant shifts (or lack thereof) to and from LARC, as well as shifts (or lack thereof) to and from methods of different effectiveness levels (higher and lower). We further examined changes in planned contraceptive use from pre- to post-visit (Aim 1.1) by conducting McNemar paired tests within each method effectiveness level. We examined what sociodemographic, reproductive health, and provider visit factors were associated with changes to LARC, to methods of higher effectiveness levels, and to methods of lower effectiveness levels (Aim 1.2) using binomial logistic regression. We investigated whether these changes vary by when DelCAN was carried out by examining the effect of wave of data collection on changes to LARC, to methods of higher effectiveness levels, and to methods of lower effectiveness levels (Aim 1.3). Sample sizes varied by each regression analysis because models were restricted to participants who were able to change to the outcome in question (e.g., participants coded as LARC pre-visit were excluded from the sample used in the models testing change to LARC). 87 Owing to the skip pattern gone awry in the first wave (as explained in the Analytic Sample section), we conducted sensitivity analyses by excluding participants in the second wave who would have been skipped had the skip pattern from the first wave not been corrected9; a total of 78 Wave 2 participants fit the criteria. We also conducted the following supplemental analyses. Due to empty cells between some of the predictors and the outcomes, we ran robust linear probability models, in order to supplement the change to LARC regression results. Furthermore, due to our modest sample sizes in analyzing change, we also examined the effect of pre-visit method effectiveness level on post- visit intention to use LARC and to use moderately effective methods by running adjusted binomial and multinomial logistic regression models. The results from these analyses are not part of the main results; they are presented in the Appendix instead. 3.4.2. Aim 2 and Aim 3 Analyses We described changes in contraceptive knowledge and attitudes from pre-visit to post- visit (Aims 2.1 and 3.1) by calculating participant shifts (or lack thereof) towards more knowledge or more positive attitudes and towards less knowledge or less positive attitudes for each of our eight outcomes. We further examined changes in knowledge and attitudes from pre- 9 The sensitivity sample was obtained imperfectly as only two of the three skip pattern issues were able to be mimicked in the second wave. Namely, the sensitivity samples exclude participants in the first wave who (1) report ever having had sexual intercourse with a male and left blank the question ?How many times have you been pregnant in your life?? (31 participants) and (2) reported having sexual intercourse a few times a month in the past 3 months and answered 0 to ?How many times have you been pregnant in your life?? (47 participants). The skip pattern issue that in the first wave skipped participants who answered with words instead of digits to the question, ?How many times have you been pregnant in your life??, was unable to be applied because the second wave questionnaire restricted answers only to digits. 88 to post-visit (Aims 2.1 and 3.1) by conducting paired t-tests for all continuous outcomes and individual questions as well as McNemar paired tests for all dichotomous knowledge questions with identical samples. The bivariate relationships between the continuous knowledge and attitude change scores and sociodemographic, reproductive health, and provider visit characteristics were examined using one-way ANOVA tests (Aim 2.1 and 3.1). We examined what sociodemographic, reproductive health, and provider visit factors were associated with (a) changes in knowledge about the IUD, the implant, effectiveness of contraceptive methods to prevent pregnancy, and the benefits of the DelCAN initiative (Aim 2.2) and (b) changes in attitudes about the IUD, implant, hormonal birth control, and condoms (Aim 3.2). We used linear regression for continuous outcomes and binomial logistic regression for dichotomous outcomes. We further investigated whether changes in knowledge and attitudes varied by the length of time DelCAN had been carried out (Aim 2.3 and Aim 3.3). Our knowledge models were restricted to participants who were able to increase in score post-visit, including those who had not heard of the IUD or implant; participants who attained the highest score pre-visit were excluded from regression analyses (Whitaker et al., 2010). Our models with the continuous attitude outcomes were restricted to participants who had both pre- and post-visit scores for the attitude outcome(s); those who reported not having heard of the IUD or implant pre-visit were excluded from the regression analyses. For models with the dichotomous attitude outcomes, analyses also excluded participants who already had the most positive attitude as they could not change. Therefore, knowledge and attitude samples vary by each method type outcome. 89 Chapter 4: Results 4.1. Sample Description As described in the Methods section, the participants of this study were recruited as part of a probability-based sample representative of the entire Delaware Title X system. However, beyond our study?s eligibility criteria (being at risk for unintended pregnancy N=727), we had to exclude additional participants for missing data due to skip pattern issues (N=103), declining the post-visit survey (N=51) and non-response to relevant study variables (N=83) (see Figure 2). We also excluded participants who were using or planning to use male or female sterilization as a contraceptive method (N=16). Our final analytic sample included N=474 female participants at risk of unintended pregnancy who were seeking care for themselves in Title-X-funded clinics across the state of Delaware. 4.2. Univariate Analyses 4.2.1. Sociodemographic, Reproductive Health, and Provider Visit Measures As presented in Table 1 (first column), more than half of the participants were in their 20s. The sample is made up of similar proportions of Black and White participants who comprise three-quarters of the sample with Hispanics accounting for approximately one fifth. More than half of the participants were currently single; similar proportions had either public or private insurance, which made up about three quarters of the sample. Forty-two percent of the sample 90 had an associate degree, some college, or vocational training. The vast majority (83.5%) of participants were born in the United States. Reproductive health variables were as follows. Only 8.9% of the sample wanted a child within the next two years; 42.8% wanted a child in two years or more, 28.5% were unsure of their pregnancy desire, and almost one-fifth did not want any (more) children. Yet, 38.4% of the sample reported that they would feel very happy or happy if they were to become pregnant in the next year. Forty percent of the sample had ever experienced an unintended pregnancy, and 15.2% reported having experienced any reproductive coercion by a romantic partner. In terms of provider visit factors, two thirds of the sample was meeting with a provider for family planning, and 67.1% were attending a clinic dedicated to women?s health. The majority of the clinics were located in urban areas (65.2%), and just over half of the sample had been recruited in the second wave of data collection (Table 1). 4.2.2. Current and Planned Contraceptive Use Also presented in Table 1 were participants? pre-visit reports of contraceptive use and intention. At pre-visit, one quarter of the sample was using LARC, and 43.2% moderately effective methods. Almost one third of the participants reported using no method or low effective methods, which includes 8.6% using no methods, 17.7% using external condoms, and 5.5% using emergency contraception, withdrawal, or natural family planning methods (data shown in Table A1 in the Appendix). However, as the pre-visit composite current-planned method variable 91 indicates in Table 1, less than one fifth of participants (18.8%) planned to use no method or low effective methods, half of which were attributed to external condoms (Table A2 in Appendix). In fact, half of the sample (50.4%) intended to use moderately effective methods, while 30.8% of participants were currently using or intending to use LARC. Table 2 presents participant reports of post-visit prescriptions, methods received or planned contraceptive use. Half of the sample (50.8%) planned to use moderately effective methods. LARC was reported by 27% of the participants, while 22.2% reported planning to use no method or a low effective method, which includes 20% who planned to use no method or were unsure if they would use any method (Table A1 in the Appendix). 4.2.3. Contraceptive and DelCAN Knowledge Scores Out of the sample?s 474 participants, N=469 make up the sample size for the three knowledge measures?IUD, implant, and effectiveness knowledge. Five participants were excluded for not providing a single answer on at least one pre- or post-visit question set that comprises the three knowledge measures. These were participants who were shown the question set, but skipped answering every single question of the set. Table 3 describes the sample?s contraceptive knowledge. Out of range from 0 to 6, participants? mean IUD knowledge score was 3.2 pre-visit, and 3.5 post-visit. The sample?s knowledge about IUD use varied starkly both pre- and post-visit, with two-thirds or more knowing that nulliparous women can use the IUD and that IUDs can be removed before the 92 expiration date, yet less than half knowing that an IUD without hormones exists and that IUDs will not affect future fertility. Participants? mean implant knowledge was 2.2 pre-visit and 2.5 post-visit, out of a range from 0 to 4. Lastly, from a range of 0 to 4 points, the average score for effectiveness knowledge was 1.5 pre-visit and 1.8 post-visit. Fully correct ranking was performed by 25.2% of the participants pre-visit and 30.9% post-visit (Table 3). All participants were shown the two questions that make up the knowledge about DelCAN benefits measure. After excluding three participants who were missing on at least one element of the two questions in either the pre- or post-visit survey, our sample for the DelCAN knowledge measure was N=471. From a range of 0 to 7 points, the average score was 3.2 pre- visit and 3.6 points post-visit (Table 3). 4.2.4. Contraceptive Attitude Scores As previously mentioned, participants needed to have reported hearing about the IUD or implant to be asked about their attitudes regarding the corresponding method. The results presented in Table 4 stratify the sample in two groups, those who received the attitude questions pre- and post-visit, and those who only reported hearing the about the method post-visit and therefore only received the questions post-visit (column on the far right). Among the sample of 399 women who received and answered the IUD attitude questions both pre- and post-visit10, their pre-visit mean total score was 9.9, and their post-visit mean total score was 9.7. As 10 Of the 399 women included in our sample, two participants received and left unanswered one of four IUD attitude questions?one pre-visit and another post-visit. A value that equals the mean score of the three answered questions was assigned to the blank item. Participants with missing data on any other attitude measure were excluded because it would constitute 50% or more of the total score. 93 explained in the Methods section, IUD attitude scoring ranged from 0 to 16, with a higher score indicating a more positive attitude. Among participants who only received the IUD attitude questions post-visit (N=40), their mean score was 9.0. Among those who had heard of the implant prior to visiting with a provider and answered all the implant attitude questions (N=363), mean total scores for both pre- and post-visit were 6.0, out of a range of 0 to 8. Participants with only post-visit implant attitude data had a mean total score of 6.1 (Table 4). The mean attitude score about hormonal birth control based on one statement was 1.5 in both pre- and post-visit, while the mean attitude score about external condoms was 2.5 pre-visit and 2.3 post-visit. The range for both measures was 0 to 4 (Table 4). 4.3. Bivariate Analyses We examined the bivariate relationships between each of our aims? pre-visit variable, post-visit variable, and all study predictors. Since our study predictors were all categorical variables, we used Pearson chi-square tests to examine their relationship with all categorical pre- visit and post-visit variables, i.e., contraceptive use, planned contraceptive use, and contraceptive effectiveness knowledge. We used one-way ANOVA F-tests to examine the relationship between study predictors and continuous pre-visit and post-visit variables, i.e., knowledge about IUDs, implants, and DelCAN benefits, as well as all contraceptive attitude measures. Results of these bivariate analyses are presented in Tables 1-2 and 5-12. 94 4.3.1. Differences by Effectiveness Level of Current and Planned Method Use Women who were currently using or planning to use LARC pre-visit compared to other methods were more likely to be married, to not want any (more) children, to be unsure about their feelings if they were to become pregnant in the next year, and to be coming in for a non- family-planning visit. Being foreign-born was associated only with current LARC use, while participants visiting clinics in a rural location were more likely to be planning to use LARC pre- visit than those visiting urban or suburban clinics (Table 1). The effectiveness level of post-visit planned method use presented in Table 2 retains all previously mentioned significant associations from Table 1 in relation to planned LARC, except reason for clinic visit, where the proportion of participants reporting LARC as their post-visit planned method was similar in those having family-planning and non-family planning visits, about one-quarter. However, participants at the clinic for a non-family-planning visit were most likely to report plans to use no or low effective methods post-visit (42.1%), while those visiting for family planning were most likely to report plans to use moderately effective methods (61.3%). Although moderately effective methods account for the largest proportion of users overall and by race, ranging approximately 40-60%, race or Hispanic ethnicity was associated with the effectiveness level of post-visit planned method use; Hispanic women were more likely to report plans to use LARC post-visit compared to White and Black women (31.7% vs. 28.7% and 22.3%, respectively). Clinic focus was not significantly associated with participants? plans to use LARC (26.3% vs. 27.4%), but participants visiting women?s health clinics were far more 95 likely to report plans to use moderately effective methods compared to their counterparts visiting primary care clinics (56.3% vs. 39.7%) and much less likely to report plans to use no or low effective methods (16.4% vs. 34.0%) (Table 2). 4.3.2. Differences by Contraceptive and DelCAN Knowledge Scores The first two columns of Table 5 present the results of bivariate analyses of total IUD knowledge scores and study variables. Black and Hispanic women were more likely to have lower mean scores in IUD knowledge compared to White women (Tukey?s test p<0.001 and p= 0.010, respectively). Women with private health insurance had higher IUD knowledge both pre- and post-visit compared to women with no insurance (Tukey?s test p-values<0.01). Having any level of post-high-school education was found to be correlated with higher IUD knowledge scores in both pre- and post-visit surveys compared to having a high school diploma or less (Tukey?s test p-values<0.001). Patients visiting clinics specializing in women?s health were more knowledgeable about IUDs than those visiting primary care clinics in both pre- and post-visit surveys (p<0.001 and p=0.016, respectively). Current relationship status, nativity, future pregnancy desire, clinic location, and wave of data collection were statistically significant (p- values<0.05) only at one of the two time points. Additionally, in the pre-visit survey, those randomized to version B IUD knowledge questions had a 0.3 points higher mean than those who received version A questions (p=0.017); the significance was not retained post-visit (Table 5). Means of total implant knowledge scores by study variables are shown in Table 6. Age was significantly related to both pre-visit and post-visit knowledge scores. Compared to women 96 aged 20-29 and 30-44, adolescents had the highest implant scores (pre-visit Tukey?s test p=0.006 and p<0.001; post-visit Tukey?s test p=0.235 and p=0.012), and pre-visit women in their 20s had more knowledge than women thirty and above (Tukey?s test p=0.048). Single and cohabitating women were more knowledgeable about the implant both pre- and post-visit compared to their married counterparts (Tukey?s test p-values<0.05), as were U.S.-born women compared to foreign-born women (p-values<0.05). Women who want to have a child in two years or more scored higher on implant knowledge than women who were unsure if or when they wanted to have (more) children (Tukey?s test p-values<0.04), as did women recruited in wave two of data collection (pre- and post-visit p-values<0.02). Patients visiting a women?s health clinic was only significant pre-visit (p=0.003), but post-visit results were only marginally significant (p=0.070), following the same trend as IUD knowledge?higher knowledge among women?s health patients. Randomization to version B resulted in statistically significant higher total implant knowledge scores both pre- and post-visit (p-values<0.03) (Table 6). As shown in Table 7, the only statistically significant factor associated with effectiveness knowledge both pre- and post-visit was age. More adolescents, compared to adult women, obtained a perfect score when ranking contraceptive methods by effectiveness pre-visit (31.7% vs. 23%, p=0.072, respectively) and post-visit (41.5% vs. 29%, p=0.035, respectively). Education level was associated with pre-visit effectiveness knowledge (p<0.003), with almost half of all women with a high school education or less (49.2%) scoring zero on effectiveness knowledge compared to approximately one third among women with a Bachelor?s degree or higher degree. Post-visit, more White women ranked all methods correctly than Black and 97 Hispanic women (41.4% vs. 23.6% and 23.3%, respectively, p=0.002). Approximately one third of women attending women?s health clinics obtained a perfect score post-visit compared to 23.4% of patients visiting primary care sites (p=0.021) (Table 7). Table 8 presents the results of bivariate analyses of knowledge about DelCAN benefits and study variables. Identifying as Black was associated with higher pre-visit and post-visit knowledge about the benefits of the DelCAN initiative, compared to White women and those identifying as neither White, Black, nor Hispanic (Tukey?s test p-values<0.03), in addition to Hispanic women having higher post-visit DelCAN knowledge compared to those identifying as neither White, Black, nor Hispanic (Tukey?s test p=0.026). Having public insurance compared to private insurance and no insurance was associated with higher DelCAN knowledge in both pre- and post-visit surveys (Tukey?s test p-values<0.01). Women who were U.S. born and those with experience of unintended pregnancy had higher DelCAN knowledge both pre- and post-visit (p- values?0.02) (Table 8). 4.3.3. Differences by Contraceptive Attitude Scores Few study variables were associated with contraceptive attitude scores. As presented in Table 9, age and race/Hispanic ethnicity were associated with pre-visit IUD attitude scores. Those who were 30 and older had more positive pre- and post-attitudes than those who were in their 20s (Tukey?s test p-value=0.033). White and Hispanic women had more positive attitudes about IUDs than Black women pre-visit (Tukey?s test p-values=0.048 and p=0.019). Table 10 presents bivariate relationships between implant attitudes and study variables. While we found 98 that no variable was associated with both pre- and post-visit attitudes, married women compared to single women (Tukey?s test p=0.003) and those who had experienced an unintended pregnancy (p=0.012) held more positive implant attitudes pre-visit. Similarly, associations with hormonal birth control attitudes presented in Table 11 indicated that no study variable was statistically significant with both pre- and post-visit attitudes. Single women held more positive attitudes towards hormonal birth control pre-visit than cohabitating women (Tukey?s test p=0.031), and those with experiences of reproductive coercion held less positive attitudes post- visit (p=0.048). Table 12 presents bivariate results showing that women who were single compared to married (Tukey?s test p-value<0.001) and cohabitating (Tukey?s test p=0.008) held more positive attitudes about condom use pre-visit, although the significant difference was retained post-visit only between single and married women (Tukey?s test p<0.001); nevertheless, cohabitating women held more positive attitudes about condoms than married women post-visit (Tukey?s test p=0.035). U.S.-born women held more positive attitudes about condom use in both pre- and post-visit surveys (p-values?0.001), while those with any prior experience of reproductive coercion had less positive attitudes towards using condoms pre- and post-visit (p=0.021 and p=0.006). Lastly, Black women compared to Hispanic women had more positive pre-visit attitudes towards condom use (Tukey?s test p=0.004). 99 4.4. Change Analyses ? Tests of Hypotheses 4.4.1. Aim 1 4.4.1.1. Aim 1.1: Overall Changes in Planned Method Use by Effectiveness Levels We calculated individual participant shifts (or lack thereof) to and from methods of different effectiveness levels, and we present these changes as stacked bar charts in Figures 3 and 4. Figure 3 demonstrates the distribution of participants changing to and from planned LARC use. At post-visit, the majority of the sample remained without a LARC (67.1% when using pre- visit current method use, and 66.9% when using pre-visit composite current-planned method use). Approximately 6% changed away from LARC, by both pre-visit measures (data inferred from Tables A1-A2. The main difference in the results (presented in the two stacked bar charts of Figure 3) lies between the pre-visit measures that account for the ?remained with LARC? and ?changed to LARC? categories. When considering only pre-visit current LARC use, 18.8% of the sample remained with LARC and 8.2% changed to LARC post-visit. However, when taking into account what women plan to use in conjunction with what they were currently using, as captured by the pre-visit composite current-planned method use measure, the results demonstrate that 24.7% participants remained with LARC and only 2.3% of participants changed to LARC. This would indicate that only 2.3% of our sample, i.e., 11 participants, changed their minds in favor of LARC after visiting with a provider. Figure 4 presents the distribution of pre-post-visit contraceptive change by three effectiveness levels. At post-visit, the largest proportion of the sample remained with a moderately effective method (38.2% when using the pre-visit current method variable and 46.4% 100 when using the pre-visit composite current-planned method variable). The proportions of the sample who changed to and remained at a less effective or no method were virtually the same between the pre-visit current and composite variables (9.5% changed from both pre-visit versions to a less effective or no method and approximately 15% in both pre-visit versions remained with a low effectiveness method or no method). However, the right-hand side of the figure shows the same trend seen in Figure 3; when the pre-visit variable accounts for women?s pre-visit intentions, the proportion of women who changed to a method of higher effectiveness level was only 4%, compared to 18.6% based on the pre-visit current method variable. This variation was also visible in the difference between the proportions of participants who remained at the highest effective methods, i.e., LARC (18.8% at pre-visit current method; 24.7% at pre- visit composite current-planned method). We further examined participant shifts (or lack thereof) within each method effectiveness level by conducting McNemar paired tests (Table 13). When comparing pre-visit current use and post-visit planned use, we found a statistically significant decrease in no method or low effective method (p<0.001) and a statistically significant increase in moderately effective methods (p<0.001). The increase in LARC was not statistically significant. However, when comparing the pre-visit composite current-planned use and the post-visit planned use, we found a 3.4% statistically significant increase in planned use of no method or low effective method (p=0.030) and a 3.8% statistically significant decrease in planned use of LARC. The slight increase in moderately effective methods was not statistically significant. Again, this would indicate that a higher percentage of the sample was planning to use no or low methods and a lower percent of 101 the sample was planning to use LARC after visiting with a provider 4.4.1.2. Aim 1.2: Factors that Predict Change in Effectiveness Level of Planned Contraceptive Use We examined factors associated with changes (a) to LARC, (b) to methods of higher effectiveness levels, and (c) to methods of lower effectiveness levels using binomial logistic regression (Tables 14-16). As presented in Table 14, we found that reporting moderately effective methods compared to no or low effective methods pre-visit was significantly associated with not changing to LARC post-visit (current: OR=0.08, 95% CI: 0.03-0.20; composite current- planned OR=0.05, 95% CI: 0.01-0.36). No other predictor was significant in both the current and composite models. Predictors that were significantly associated with change to LARC in the model using the current method variable were: racial identity (being non-Hispanic Black (OR=0.35, 95% CI: 0.13-0.98) and non-Hispanic other racial category (OR=15.4, 95% CI: 1.89- 125.53), vs. being White); current relationship status (cohabitating OR=0.28, 95% CI: 0.08-0.95) vs. being single), and having experienced any reproductive coercion (OR=3.00, 95% CI: 1.09- 8.24) compared to none. No other predictors were significant in the model using the composite current-planned method variable. The results from the sensitivity analyses were very similar to the main results (also presented in Table 14). As explained in the Analysis Aim 1 section, the binomial regression models omitted observations that predicted failure perfectly (hence the presence of non-estimable cells in Table 14). As a result, the regression models use smaller samples compared to the total eligible sample, i.e., those who did not report LARC pre-visit. Therefore, we ran robust linear probability 102 models, presented in the supplemental analyses in Tables A3 and A4 in the Appendix, in order to supplement the logistic regression results. The majority of the results from the robust linear probability model reflect the results from the main analyses. When investigating factors associated with change to methods of higher effectiveness levels, we found that reporting moderately effective methods pre-visit compared to no method or low effectiveness methods was strongly associated with not changing to a method of higher effectiveness, i.e., LARC, according to both pre-visit measures (current: OR= 0.01, 95% CI: 0.003-0.03; composite current-planned OR=0.02, 95% CI: 0.004-0.10; first and third column of Table 15). No other predictor was significant in either pre-visit model. Significant associations in the pre-visit current method models include future pregnancy desire and reason for clinic visit, where wanting a to become pregnant in two years or more was associated with changing to a method of higher effectiveness level (OR=5.40, 95% CI: 1.18-24.68) and coming in for a family planning visit was strongly associated with changing to a method of higher effectiveness level (OR=4.74, 95% CI: 1.95-11.48). In the pre-visit composite current-planned method models, no independent variable was significantly associated with the outcome. The sensitivity models followed the same patterns as the full models, with one exception: identifying as Black was associated with not changing to a method of higher effectiveness level in the pre-visit current model (OR=0.36, 95% CI: 0.14-0.95) (Table 15). Table 16 presents the adjusted odds ratios for planning to use to a method of a lower effectiveness level post-visit. The reported pre-visit method remains the strongest predictor of post-visit outcome. Reporting LARC compared to moderately effective methods pre-visit was 103 strongly associated with changing to a method of a lower effectiveness level in both the pre-visit current and pre-visit composite current-planned method models (OR=3.06, 95% CI: 1.42-6.59 vs. OR=3.72, 95% CI: 1.72-8.04). Furthermore, in both models, having private health insurance was associated with changing to a method of lower effectiveness (current: OR=3.86, 95% CI: 1.49-10.04 vs. composite: OR=4.49, 95% CI: 1.72-11.72). No other predictor was significant in both the current and composite models. In the pre-visit current method model, wanting a child in two years or more and not wanting any more children was associated with not changing to a method of lower effectiveness level (OR=0.22, 95% CI: 0.05-0.92; OR= 0.18, 95% CI: 0.04-0.89, respectively). In the pre-visit composite method model, identifying as Black vs. White was associated with changing to a method of a lower effectiveness level (OR=2.67, 95% CI: 1.10-6.49), while attending a family planning visit was associated with not changing to a method of lower effectiveness level (OR=0.41, 95% CI: 0.19-0.87). The sensitivity models followed the same patterns as the full models, with one exception: participants with no insurance had statistically significant higher odds of changing to a method of a lower effectiveness level compared to those with public insurance in the pre-visit composite method model (OR=3.04, 95% CI: 1.06-8.72) (last column in Table 16). 104 4.4.1.3. Aim 1.3: DelCAN Expansion as a Predictor of Change in Effectiveness Level of Planned Contraceptive Use In most of our models, wave of data collection, which indicates how long the DelCAN initiative had been in place, was not significantly associated with changes to methods of different effectiveness levels. When investigating change to LARC, only the sensitivity composite model finds wave two to be significantly associated with reporting LARC post-visit (OR=10.94, 95% CI: 1.13-105.49); women visiting clinics in 2018-19, when the DelCAN initiative was more established, were more likely to change to LARC than those visiting clinics in 2017, even after taking account women?s intent with the composite model. The results from the main analysis were not significant but followed the same direction (OR=7.29, 95% CI: 0.85-62.31) (Table 14). This is supported by the results from the aforementioned robust linear probability models; we found the second wave to be associated with changing to LARC in both the composite main and sensitivity models (B=0.05, 95% CI: 0.01-0.09 and B=0.06, 95% CI: 0.01-0.11, respectively) (Table A4). As seen in the last rows of Tables 15 and 16, wave was not significantly associated with change to a method of higher effectiveness level or change to a method of lower effectiveness level. 4.4.1.4. Aim 1 Summary Changes in planning to use a method of higher effectiveness levels (including LARC) were relatively small. Of those who changed pre- to post-visit to a method of higher effectiveness level, most had arrived at the clinic already planning to initiate or switch to a more 105 effective method. Effectiveness level of method reported pre-visit was the strongest and only consistent predictor of changing to a method of a different effectiveness level?both higher and lower. Further expansion of the DelCAN initiative was not associated with changing to a method of a higher or lower effectiveness level following a visit with a Title X provider, although findings suggest DelCAN had an effect on changing to LARC methods. 4.4.2. Aim 2 4.4.2.1. Aim 2.1: Overall Changes in Contraceptive and DelCAN Knowledge We calculated individual participant shifts (or lack thereof) in knowledge about the IUD, implant, contraceptive effectiveness, and DelCAN benefits from pre-visit to post-visit, and we present these changes as stacked bar charts in Figures 5-8. The distribution of pre- to post-visit changes in IUD knowledge among the sample is presented in Figure 5; almost half maintained the same IUD knowledge score from pre-to post, while 21.3% decreased and 31.1% increased in IUD knowledge. Similarly, 19.4% of participants decreased in implant knowledge while 24.9% increased (Figure 6). Figure 7 describes the change in relative contraceptive effectiveness knowledge; 8.1% decreased in effectiveness knowledge, while 17.3% increased. Lastly, more than twice as many participants increased in DelCAN knowledge than decreased (Figure 8). We further examined changes in contraceptive knowledge with two-sided paired t-tests. The increases in mean knowledge scores were statistically significant for all four knowledge outcomes (p-values<0.001), as presented in the far-right column of Table 3. Additionally, we found that increases of 8.7% and 10.6% in having heard about the IUD and implant, respectively, 106 were statistically significant based on the McNemar tests (p-values<0.001). The question about contraceptive effectiveness allowed for multiple different ranking combinations; the proportion of participants who ranked LARC as most effective increased, while the proportion of participants who ranked non-LARC as most effective decreased (Table 3). Furthermore, participant knowledge about the benefits of DelCAN improved across all individual questions. Using McNemar tests, we found statistically significant increases in knowledge on how to obtain for free male condoms (7.4%, p<0.001), the IUD (4.9%, p=0.021), the NuvaRing (4.7%, p=0.024), and birth control pills (4.7%, p=0.016); the implant change of 3.6% approached significance (p=0.065). Knowledge that all contraceptive methods can be accessed on the same day they are requested to a provider increased by 12.1% (p<0.001) (Table 3). We also examined the correlations between our study variables and the change means of IUD, implant, and DelCAN knowledge by conducting one-way ANOVA tests and post-hoc Tukey tests if indicated. As presented in Table 5, the increase in IUD knowledge scores from pre- to post-visit survey was statistically significant by race/Hispanic ethnicity, where Black women?s knowledge scores increased more than White women?s (Tukey?s test p=0.005), and by clinic focus, where change in knowledge was higher among women visiting primary care sites (p=0.036). Increase in implant knowledge was only significantly associated with age, where the oldest age group (30-44) had a larger increase compared to the adolescent group (Tukey?s test p=0.024) (Table 6). Increase in knowledge about DelCAN benefits was significantly associated with reason for clinic visit and with clinic focus. Women coming in for family planning had a larger increase in knowledge than those coming in for other healthcare services (p=0.025), and 107 those visiting women?s health clinics compared to primary care sites had larger increases in knowledge (p=0.032) (Table 8). 4.4.2.2. Aim 2.2: Factors that Predict Change in Contraceptive and DelCAN Knowledge We examined factors associated with changes in knowledge about (1) the IUD, (2) the implant, (3) effectiveness of contraceptive methods to prevent pregnancy, and (4) the benefits of the DelCAN initiative using adjusted linear and logistic regressions (Tables 17-20). As presented in Table 17, among participants who did not already have the highest IUD knowledge pre-visit (N=418), we found pre-visit knowledge score to be the strongest predictor of change in both models of IUD knowledge change; for each unit of pre-survey knowledge increase, the positive change was 8% smaller (B=-0.08, 95% CI: -0.11 - -0.05, p<0.001), i.e., those with higher scores pre-visit had lower odds of positive change (OR: 0.69, 95% CI: 0.60-0.79, p<0.001). Additionally, compared with having a high school education or less, holding a bachelor?s degree or higher was associated with a 27% higher probability of increasing in IUD knowledge from pre- to post-visit (B=0.27, 95% CI: 0.10-0.43, p=0.002). Similarly, holding a Bachelor?s degree or higher was associated with twice the odds (OR=2.14, 95% CI: 1.04-4.42, p=0.039) of any increased IUD knowledge from pre- to post-visit compared to those with a high school education or less. We found pre-visit implant knowledge to be the strongest predictor of change in implant knowledge when fitting models for both the continuous and dichotomous outcomes (Table 18); every point of increase in pre-visit implant knowledge was associated with a 14% smaller 108 increase in one?s implant knowledge (B= -0.14, 95% CI: -0.19 - -0.09, p<0.001) as well as lower odds of any increase (OR=0.61, 95% CI: 0.49-0.76, p<0.001). The logistic regression model also indicates that women visiting rural sites had a higher odds of any increase in implant knowledge (OR=1.91, 95% CI: 1.10-3.31, p=0.021) compared to those attending urban sites; those at suburban sites had a three times higher odds of knowledge increase compared to those at urban sites but the result was only marginally significant (OR= 3.16, 95% CI: 0.97-10.32, p=0.056). Since eleven participants of the N=469 sample (2.3%) were never assigned a randomized group for the IUD and implant knowledge questions because they reported never having heard of the IUD or implant post-visit, we ran a sensitivity model excluding them so that we could control for random group assignment. Receiving version B questions compared to A questions did not predict change in IUD knowledge nor in implant knowledge (data shown in the last rows of Tables 17 and 18; full model results are not shown but the inclusion of the random group assignment did not materially alter the values of the other predictors). We used two dichotomous outcome models to examine factors associated with changes in contraceptive effectiveness knowledge, as shown in Table 19. Having scored 1 to 3 compared to zero pre-visit was associated with more than twice the odds of ranking all methods correctly by effectiveness post-visit (OR=2.37, 95% CI: 1.16-4.84, p=0.018); the same association was not statistically significant when testing for any increase (OR=0.98, 95% CI: 0.57-1.66, p=0.933). When testing only for change to perfect effectiveness knowledge score, we observed that identifying as Black compared to White was strongly associated with a lower odds of increasing to a perfect score (OR=0.25, 95% CI: 0.10-0.60, p=0.002); the same direction was observed in 109 the testing for any increase model among identifying as Black and Hispanic and the results were marginally significant (OR=0.56, 95% CI: 0.30-1.04, p=0.064 and OR=0.44, 95% CI: 0.19-1.02, p=0.056). Additionally, when testing for any increase in effectiveness knowledge from pre-to post-visit, we observed that women who had experienced any reproductive coercion by a partner had twice the odds of improving their effectiveness score (OR=2.17, 95% CI 1.07-4.39, p=0.032). We also conducted supplemental analyses to match the models of our other three outcomes (percentage increase with linear regression and using the continuous pre-survey score compared to dichotomous score in the logistic regression models), as shown in Table A9. Notably, we found that a continuous pre-visit score predicted the same direction as in all other knowledge outcome results from linear regression analyses (the higher the pre-visit score, the lower the probability of increasing post-visit, B=-0.10, 95% CI: -0.15 - -0.15, p<0.001), while results from the supplementary logistic effectiveness knowledge models were very similar to those from the main effectiveness knowledge analyses. As in the IUD and implant knowledge change models, in linear regression we found the continuous pre-visit knowledge score to be the strongest predictor of percent change in knowledge about DelCAN benefits (Table 20). For every unit increase in pre-visit knowledge, the post-visit knowledge score decreased by 15% (B=-0.15, 95% CI: -0.20- -0.10, p<0.001). Similarly, in logistic regression, we observed that higher pre-visit scores were associated with a lower odds of any increase in DelCAN knowledge post-visit (OR: 0.85, 95% CI: 0.76-0.94, p<0.001). Identifying as a race/ethnicity other than White, Black, or Hispanic was associated with decrease in knowledge compared to White women, but our sample size for that group was 110 very small (B=-0.64, 95% CI: -1.24- -0.04, p=0.037). Being older was associated with a lower odds of any increase in knowledge compared to adolescents (OR=0.45, 95% CI: 0.21-0.96, p=0.039). 4.4.2.3. Aim 2.3: DelCAN Expansion as a Predictor of Change in Contraceptive and DelCAN Knowledge Being recruited for the study during wave two compared to wave one was not associated with a percentage increase in IUD knowledge, but we observed increased odds of any improvement in IUD knowledge among wave two participants (OR=1.59, 95% CI: 1.00-2.52, p=0.048) (Table 17). For change in implant knowledge, we found that wave two was associated with a 15% higher change in implant knowledge (B=0.15, 95% CI: 0.03-0.27, p=0.017), while in the logistic regression, wave two trended in the same direction approaching significance (OR=1.59, 95% CI: 0.94-2.69, p=0.081). In contrast, when testing DelCAN expansion as a predictor of change in effectiveness knowledge, wave two participants had half the odds of increasing their effectiveness score (OR: 0.56, 95% CI 0.33-0.96, p=0.036) and to reach a perfect score (OR: 0.40, 95% CI: 0.19-0.84, p=0.015). Lastly, wave was not significant in either the linear or logistic regression DelCAN knowledge models, although the direction was negative like the relationship with change in effectiveness knowledge. 4.4.2.4. Aim 2 Summary Following a visit with a provider, mean knowledge scores increased across all four knowledge outcomes (p-values<0.001). Pre-survey knowledge scores were the strongest and 111 most consistent predictors of change; a positive change in knowledge was more likely among those who came to the clinic with less knowledge. Overall, a few factors were associated with change in knowledge (age, education, race/ethnicity, reproductive coercion, clinic location), but no sociodemographic, reproductive health, or provider visit characteristics were associated with more than one change in knowledge outcome. Further expansion of the DelCAN initiative was positively associated with increases in IUD and implant knowledge, but negatively associated with increases in effectiveness knowledge. 4.4.3. Aim 3 4.4.3.1. Aim 3.1: Overall Changes in Contraceptive Attitudes We calculated individual participant shifts (or lack thereof) in attitudes about the IUD, implant, hormonal birth control, and condoms and visually described them as stacked bar charts in Figures 9-12. Figure 9 presents the distribution of pre- to post-visit change in IUD attitudes; more participants decreased in positive attitudes towards IUDs between pre- and post-visit (39.8%) compared to those whose attitudes increased positively (28.6%). Conversely, the proportions of participants decreasing and increasing in positive attitudes about the implant were similar (27.0% and 25.3%, respectively) (Figure 10). Figure 11 shows that more participants decreased in positive attitudes toward hormonal birth control methods than increased (21.2% vs. 17.4%). When it comes to condom attitudes, one quarter decreased in positive condom attitudes and 16.3% increased (Figure 12). 112 We further examined changes in contraceptive attitudes with two-sided paired t-tests, as presented in Table 4. For changes in IUD attitudes, the mean score decline of 0.2 points was statistically significant (p=0.009), even though only change in one of the four items that make up the measure (?IUDs are uncomfortable?) was significant (p<0.001). In contrast, the mean total implant attitude change score was not statistically significant, but the mean change of each implant attitude question was significant. Positive attitudes declined for ?It is a relief to have an IUD or implant? by 0.1 point (p=0.020) and rose by 0.1 point for ?Provider will remove my implant when I want? (p=0.033). On average, attitudes about hormonal birth control attitudes remained similar, but the positive attitude about condom use declined by 0.14 points (p=0.003). We used one-way ANOVA tests, and post-hoc Tukey tests where indicated, to examine associations between changes in attitude outcomes and study variables; results are presented in the two rightmost columns of Tables 9-12. Changes in attitudes about the IUD, implant, hormonal birth control, and condom were not significantly associated with any study variable, except wave two was associated with changes in condom attitudes (p=0.032) (Table 12). 4.4.3.2. Aim 3.2: Factors that Predict Change in Contraceptive Attitudes We examined factors that may predict changes in attitudes about (1) the IUD, (2) the implant, (3) hormonal birth control, and (4) condoms using adjusted linear and logistic regressions (Tables 21-24). As presented in Table 21, for every unit increase in pre-visit positive IUD attitude score, we observed a 0.3-point decrease in post-visit positive IUD attitudes (B=-0.30, 95% CI: -0.37- -0.22, p<0.001). The results from the logistic regression were 113 similar; a one unit increase in pre-visit IUD attitudes was associated with a lower odds of any increase in positive post-visit IUD attitudes (OR=0.82, 95% CI: 0.72-0.92, p=0.001). Having a bachelor's degree or higher compared to a high school diploma or less was marginally associated with 0.5-point increase in post-visit positive attitudes (B=0.50, 95% CI: -0.03-1.03, p=0.067). Similarly, having at least a bachelor?s degree compared to a high school degree at most was associated with higher odds of reporting any positive change in positive IUD attitudes (OR=2.32, 95% CI: 1.10-4.88, p=0.027). Table 22 presents results from the adjusted relationships between changes in positive implant attitudes and the study variables. Again, pre-visit attitude score was the strongest predictor of change, where every unit increase in pre-visit positive implant attitudes was associated with a 0.34-point decrease in post-visit positive implant attitudes (B=-0.34, 95% CI: - 0.43- -0.25, p<0.001). Similarly, a one unit increase in pre-visit positive attitudes was associated with a lower odds of experiencing any positive change in positive implant attitudes post-visit (OR=0.75, 95% CI: 0.59-0.94, p=0.013). Additionally, results from the logistic model indicate that visiting a clinic in rural Delaware compared to urban locations was associated with higher odds of changing to more positive attitudes (OR: 1.90, 95% CI: 1.06-3.40, p=0.032). Having more positive attitudes about hormonal birth control pre-visit was associated with decrease in positive attitude changes post-visit. As presented in Table 23, for every unit increase in positive attitudes about hormonal birth control pre-visit, there was a 0.41-point decrease in positive attitudes post-visit (B=-0.41, 95% CI: -0.48- -0.35, p<0.001). This was similar to the results from our logistic regression model (OR=0.34, 95% CI: 0.24-0.48, p<0.001). 114 Not wanting any (more) children compared to women who want children within the next two years was associated with a decrease in positive attitudes about hormonal birth control (B=-0.34, 95% CI -0.65- -0.04, p=0.029), and we observed a similar association in our logistic model (OR=0.26, 95% CI: 0.07-0.95, p=0.042). Additionally, having any experience of reproductive coercion was negatively associated with positive attitude change (B=-0.20, 95% CI: -0.39- -0.01, p=0.037). In terms of condom attitude change, we observed from our linear model that every unit increase in pre-visit positive condom attitudes is associated with a 0.34-point decrease in post- visit positive attitudes (B=-0.34, 95% CI: -0.40- -0.27, p<0.001). A similar association is apparent in our logistic regression results (OR=0.63, 95% CI: 0.48-0.83, p=0.001). Additionally, feeling unhappy or very unhappy about a hypothetical pregnancy in the next year compared to those who would be happy or very happy had a marginally significant positive effect on change in positive condom attitudes (B=0.24, 95% CI: 0.00-0.48, p=0.053); the same association was statistically significant in our logistic model (OR=3.03, 95% CI: 1.32-6.98, p=0.009) (Table 24). Although our linear regression results found an association between identifying as a race or ethnicity other than White, Black, and Hispanic (B=-0.57, 95% CI: -1.12- -0.03, p=0.040), it should be noted that this is based on a very small sample and our logistic regression results point to no difference in condom attitude change among those identifying as non-White, non-Black, and non-Hispanic (as denoted by the non-estimable (NE) designation in Table 24). 115 4.4.3.3. Aim 3.3: DelCAN Expansion as a Predictor of Change in Contraceptive Attitudes Being recruited for the study during wave two compared to wave one was associated with positive attitude change for the implant and the condom, but not for the IUD and hormonal birth control (Tables 21-24). Specifically, the effect of the DelCAN expansion is a 0.29-point increase in positive implant attitudes (B=0.29, 95% CI: 0.02-0.57, p=0.038) compared to the start of the initiative during wave one; results from the logistic model trend in the same direction but are non-significant (Table 22). In contrast, wave two had a negative effect on change in positive condom attitudes (B=-0.22, 95% CI: -0.40- -0.04, p=0.017); the same trend was observed in the logistic model but the result was non-significant (Table 24). 4.4.3.4. Aim 3 Summary Following a visit with a provider, mean scores of positive attitudes across all four contraceptive methods decreased to less positive; the changes varied in magnitude and only two out of four were significant (p<0.01). Pre-visit attitude scores were the strongest and only consistent predictors of change when testing both continuous and dichotomous change outcomes; a more positive attitude pre-visit was associated with a larger magnitude of negative change post- visit, and vice-versa. Similarly, a larger decrease in positive attitudes was more likely among those who came to the clinic with more positive attitudes. Overall, a few factors were associated with changes in attitudes (education, clinic location, reproductive coercion, pregnancy desire, feelings about hypothetical pregnancy), but none were associated with more than one method attitude change. Further expansion of the DelCAN initiative was not consistently associated with 116 change in attitudes following a visit with a Title X provider, but wave two was positively associated with changes in positive implant attitudes and negatively associated with changes in positive condom attitudes. 117 Chapter 5: Discussion Continuing to increase the proportion of pregnancies that are intended remains a public health priority. Despite the availability of many effective contraceptive methods and women?s desire to avoid pregnancy during much of their reproductive lifespan, almost half of all pregnancies in the United States are unintended. Reasons for inconsistent contraceptive use and non-use are many and complex. Using data collected between 2017 and 2019 in Delaware?s Title-X-funded clinics, the overarching objective of this dissertation was to contribute to the larger evaluation of the DelCAN initiative by investigating the role of a health care provider visit in women?s planned contraceptive use, contraceptive knowledge, and contraceptive attitudes. Serving simultaneously as gatekeepers and counselors, providers have both institutional power and clout to influence women?s decisions about what method to use. They can also be sources of knowledge about factual information and the range of experiences women report about contraception, particularly lesser-known methods. Information gained regarding possible experiences may influence patients? attitudes. Since DelCAN placed an emphasis on LARC due to their high effectiveness, upfront costs, provision barriers, and relatively recent rollout, the available evaluation data skewed in LARC?s favor. Therefore, this dissertation addressed various gaps in literature on the aforementioned topics, mainly focusing on LARC as an outcome. This chapter provides a summary of the main findings of this dissertation, situating them in the context of existing evidence on contraceptive counseling and contraceptive choice, 118 knowledge, and attitudes. Notable findings are related to DelCAN goals. Furthermore, dissertation findings are examined within the framework of the Integrated Behavioral Model. Since DelCAN was not designed with reproductive justice in mind, the framework is merely used as a lens to interpret findings and present implications for practice and public health. Lastly, limitations of the study are explained, and suggestions for future research are highlighted at the end of the chapter. 5.1. Summary of Findings 5.1.1. Planned Contraceptive Use Overall, changes in patients? plans to use a method of higher effectiveness level as a result of the provider visit were relatively small. By using two different pre-visit measures that assessed both current and intended contraceptive use and method type, we found that pre-visit intent accounted for much of the change to a method of higher effectiveness level. Of the women who changed to a method of higher effectiveness level (18.6%), most had arrived at the clinic already planning to initiate or switch to a more effective method, and only 4% had not planned to do so. The proportion of women who changed to a method of lower effectiveness or to no method (9.5%) did not differ by pre-visit measure. The strongest and only consistent predictor of changing to a method of a different effectiveness level, higher and lower, was effectiveness level of method reported in pre-visit. Reporting moderately effective methods in pre-visit was associated with lower likelihood of changing to higher effectiveness level methods (compared to reporting no methods or low effective methods) and of changing to lower effectiveness level 119 methods (compared to reporting LARC). This was the case for all models, regardless of the pre- visit measure that was used. Switching to and from methods of different effectiveness levels and no method is not unusual in and of itself. Quantitative and qualitative data consistently demonstrate that making contraceptive decisions and implementing related behaviors is an ongoing process that women renegotiate repeatedly over their reproductive lifespans based on multiple factors, not just the prevention or desire for pregnancy (Downey et al., 2017; Frost et al., 2007b; Grady et al., 2002; R. K. Jones et al., 2015; Marshall et al., 2018; Moreau et al., 2007). Although we cannot directly compare our results to national averages since discontinuation and switching rates pertain to switching between method types rather than effectiveness levels, our numbers are not out of the ordinary. What is notable about our results is the aforementioned difference between the changes to a method of a higher effectiveness level when using different metrics (18.6% vs. 4%). Women?s contraceptive intentions as reported in pre-survey explained most of the shift to higher effective methods, including LARC. This would suggest that the provider visit played a supportive role in women?s contraceptive decision-making, either by facilitating the obtainment of a new desired contraceptive method and/or by respecting the effectiveness level of their current method. In spite of the LARC-first approach advocated for by DelCAN trainers, there is no evidence to suggest that providers in Delaware?s Title-X-funded clinics were persuading women to initiate LARC. Indeed the rates of change to moderately effective methods were higher by both pre-visit measures compared to rates of change to LARC (data inferred from Tables A1-A2). 120 This is, furthermore, supported by participants? reporting of the decision-maker in the context of contraceptive counseling. Our data indicates that making the decision to initiate or not initiate a new method was overwhelmingly in the hands of the patients. At least 97% of women whose patient-provider interaction included any contraceptive talk11 reported that they made the decision individually or together with the provider (see Table A10). Of the 1% of participants (N=4) who reported that the contraceptive decision was made solely by the provider, one participant initiated a new method during the visit and three did not.12 Although having even one participant report that the decision was made by the healthcare provider is problematic13, the presumed coercion does not appear to favor LARC in any of the four cases.14 In spite of likely experiencing disempowerment during the provider visit, the three participants who did not obtain a new method (or prescription for a new method) did not appear to be dissuaded from their future contraceptive plans?their pre-visit contraceptive intention matched their post-visit contraceptive plan. In sum, although we do not know whether the interaction between the providers and the women who participated in our study met all the standards of patient-centered contraceptive care, almost all the women reported that the shared decision-making provision was met during 11 In our N=474 sample, 83.1% of women reported mention of at least one contraceptive method or conversation about trying to get pregnant during their visit (data inferred from Table A10). 12 Participants who reported starting or obtaining a prescription for a new method were asked, ?Who would you say made this decision (to start a new method today)?? and those who reported not starting a new method were asked, ?Who would you say made this decision (not to start a new method today)?? 13 The proportion of participants who reported that the healthcare provider decided on initiating or not initiating a new method was also 1% among the larger sample, i.e., all participants who completed the post-survey questionnaire and received the decision-making questions (N=744) (data not shown). 14 The participant who initiated a new method reported seeking the pill and obtaining the pill; it is possible that the provider decided on a different type or brand of pill than the patient wanted. Of the three participants who did not obtain a new contraceptive method, two were using the shot and the IUD and they reported not wanting a new method pre-visit; their post-visit plan remained the same as their pre-visit use. Lastly, one participant was using low effective methods and wanted to switch to an IUD pre-visit; however, she reported not receiving any new method during her visit with the provider at a women?s health clinic; she reported still wanting the IUD as her post-visit contraceptive plan. 121 their visit. This suggests that the vast majority of our participants did not experience the decision-making process as coercive. We would be remiss not to note that approximately 6% of the total sample changed away from LARC, by both pre-visit measures. Slightly lower rates are observed for moderately effective methods (5.3% by the current measure and 4.0% by the composite measure) (data inferred from Tables A1-A2). However, the rates of changing away from LARC are twice as high as those of changing away from moderately effective methods when observing within effectiveness level changes by both pre-visit measures. To illustrate, 12.1% of the 206 participants who reported currently using moderately effective methods reported a different effectiveness level post-visit, compared to 23.9% of 117 participants who were currently using LARC. The rates for the composite measure were 7.9% of 239 and 19.9% of 146, respectively (data inferred from Tables A1-A2). While this difference in rates may be explained by reasons such as impending plans to become pregnant, a perceived lack of control to quit LARC, and the side-effect profiles, the overall shifts to and from different effectiveness levels by both pre-visit measures likely reflects the dynamic nature of the contraceptive decision-making process as well as the provider as gatekeeper with the ability to both facilitate and hinder access to desired methods. Despite the tiered effectiveness counseling approach endorsed by the trainers of DelCAN, some Delawarean providers were reticent to embrace the new LARC provision standards. Qualitative interviews conducted as part of the DelCAN process evaluation by Skra?i? and colleagues (2021) also revealed that some primary care and women?s health providers in 122 Delaware preferred to follow longstanding and outdated requirements (pregnancy test, waiting period, no IUDs for nulliparous women) prior to LARC insertion. This directly impeded same- day LARC insertion among their patients and went against the goals of the initiative. According to the DelCAN site leaders, the reluctance was explained, in part, by providers? beliefs that immediate LARC insertions would increase the likelihood of ?early? LARC removals due to undesired side-effects. Indeed, research by Bell and colleagues find that virtually all providers in their study engage in delay tactics to avoid performing a LARC removal among patients who want theirs removed due to side effects (Manzer & Bell, 2022b). In an effort to give patients until the next visit to think about whether they truly want a LARC, a minority of reticent providers in Delaware might have reduced the number of LARC-seekers as some patients would have wanted a method they could start immediately, while others would have wanted to return for a second appointment to get their LARC but never did. Despite DelCAN, some providers likely hindered some patients from obtaining LARC by not following updated recommendations for same-day LARC for patients of all ages who want them. To our knowledge, our Aim 1 is the first study to investigate the role of contraceptive counseling by a provider by measuring pre-visit and post-visit contraceptive use and intentions, as reported by patients, that includes all contraceptive methods. Two prior studies utilizing a pre- post design similar to ours focused only on combined hormonal contraceptive methods (CHC; pill, patch, ring). Again, we cannot make direct comparisons of results because the two prior studies differ from ours in terms of both sample selection and contraceptive outcomes, but both studies found that less than half of the participants chose a different method in post-survey 123 compared to the method they had been interested in prior to the contraceptive counseling (Bitzer et al., 2012; Costa et al., 2011; Craig et al., 2019; Gambera et al., 2015). This aligns with our finding that pre-visit contraceptive intention is the strongest predictor of post-visit contraceptive plans. In terms of methodology, the novelty of our Aim 1 study lies in measuring both current and planned contraceptive use prior to engaging in contraceptive counseling with a provider. Aside from the two aforementioned studies that focused on CHCs, past research on contraceptive counseling and contraceptive choice does not consistently account for pre-visit method use or interest. In some studies, eligibility guidelines require women to be seeking to initiate or switch contraceptive methods to participate, but their interest in a specific method is not measured (Harper et al., 2015; Mazza et al., 2020; Secura et al., 2010), while other study designs account for neither current method use nor method interest prior to contraceptive counseling (Bommaraju et al., 2015; Lee et al., 2013). By assessing women?s current and intended method use pre-visit and intended method use post-visit, we were able to better hone in on the role of the clinic visit on women?s contraceptive plans across all effectiveness levels. It appears that, by and large, the clinic visit facilitated women?s contraceptive wishes rather than changing their minds, in line with the reproductive justice tenet of maintaining personal bodily autonomy. 5.1.2. Contraceptive Knowledge Across each of our four knowledge measures about the IUD, implant, contraceptive effectiveness, and DelCAN benefits, participants overall left the clinic with more knowledge 124 than upon arrival (p<0.001). Pre-visit knowledge scores were the strongest and most consistent predictors of changes, i.e., a positive change in knowledge was more likely among those who came to the clinic with less knowledge. No sociodemographic, reproductive health, or provider visit factors were associated with more than one knowledge outcome. It should be noted that, although many participants at the individual level increased in knowledge across all four of our knowledge scales, some participants decreased in knowledge (Figures 5-8). As a result, the overall increase in knowledge among our sample was modest. This further raises the question whether these relatively small yet statistically significant changes are clinically significant. We posit two reasons why they might be. First, an overall 0.3-point positive difference in knowledge should be considered along with the extent to which participants? knowledge could increase. For instance, a pre-visit mean score of 3.2 in IUD knowledge that has a range from 0 to 6, indicates that the sample?s mean score could increase up to 2.8 points; a 0.3-point change amounts to a change of 10.7%.15 This perspective allows us to appropriately account for participants? prior knowledge. Second, our findings show that women coming to the clinic with lower levels of knowledge experienced the highest knowledge gains. They are, arguably, the group of patients most in need of the increased knowledge. We also observe a distinctive result from our increase to total score in the effectiveness knowledge logistic models (rightmost model in Table 19 and rightmost model in Table A9), where participants with a higher score in pre-visit had higher odds of ranking all four methods correctly according to their efficacy rates. Although this may ostensibly contradict the results 15 Dividing a change of 0.3 by the total available change of 2.8 results in 10.7%. 125 from the linear regression analyses (leftmost model in Table A9), where participants with higher pre-visit scores had a lower probability of increasing in knowledge, this is not the case. Indeed, both findings are logical in the real world?participants coming in with less knowledge have a larger leeway for knowledge gain, and it is precisely because their knowledge gap is bigger that they are less likely to bridge that gap in totality. These findings support our methodological approach to test both continuous and dichotomous outcomes with both continuous and dichotomous variables. Our knowledge results are consistent with prior literature that commonly finds that a contraceptive counseling component beyond the standard of care has a positive effect on contraceptive knowledge (de Reilhac et al., 2016; Larsson et al., 2004a; Lee et al., 2015; Mason et al., 2003; Nobili et al., 2007). However, only two small European studies have specifically investigated LARC knowledge in the context of contraceptive counseling by a provider, both among patients seeking abortion services. One study utilizing a pre-post-provider visit design included IUD questions among many other contraceptive knowledge questions; results found personalized counseling to have an overall positive effect on knowledge, but individual scores by method were not reported (Nobili et al., 2007). The second study focused only on implant knowledge, examining the effect of an implant-focused provider visit enhanced by a video in the intervention group relative to the standard of care; knowledge recall in the intervention group after seeing a provider was not significantly higher than in the standard of care group (Michie et al., 2016). It should be noted that, unlike our study, all aforementioned intervention evaluations 126 used a comparison group in their analyses, and most used a study design where the contraceptive counseling component was standardized in the experimental group. As new methods and brands are introduced onto the market and contraceptive guidelines are updated, facts about contraceptive methods and people?s experiences using them also change. Despite radically different histories (Callahan & Caughey, 2013; French & Darney, 2015), modern versions of the IUD and implant are the most effective reversible contraceptive methods currently available on the market, appropriately termed LARC (long-acting reversible contraception). In the past decade, there has been a strong push to expand access by attempting to address the multiple barriers that have impeded access to these (and, at times, other) methods. At the large-scale policy level, two decisions paved the way for LARC access. Firstly, in 2007, ACOG endorsed the use of IUDs as appropriate for most women, including adolescents and nulliparous women, and expanded that endorsement to both LARC methods in 2012 (ACOG Committee on Adolescent Health Care Long-Acting Reversible Contraception Working Group, 2012; American College of Obstetricians and Gynecologists, 2007). Secondly, the passage of the ACA in 2012 has allowed millions of women with prescription insurance to obtain any of the FDA-approved method types with a prescription, including LARC, without copayment (Montgomery et al., 2020), although millions of women still cannot benefit from this provision, by law or in practice.16 At state and local levels, initiatives such as DelCAN have provided 16 This ACA provision guaranteeing contraceptive choice from the financial aspect does not cover women who are uninsured nor those living in states that have not implemented the Medicaid expansion, although they might be able to obtain some kind of contraceptive care through other pathways (Ranji et al., 2022). Despite financial coverage many women may not be able to access specific methods due to lack of provider training or method availability. Furthermore, women who work for employers and educational institutions that have religious or moral objections to some or all contraception do not have contraceptive coverage as part of their employer-sponsored health plan, thanks 127 resources and training to clinic teams in order to equip administrators and providers to offer same-day LARC insertions, which is in stark contrast to prior clinical guidance and often poses logistical and workflow challenges to existing schedules and stock inventory (Skra?i? et al., 2021). These relatively recent changes provide more women with a wider choice of contraceptive options, a key feature of comprehensive contraceptive care. However, these policies and practices are helpful only insofar as women know about them. Considering that myths and misconceptions about contraception are not uncommon, and that many stem from partially factual or outdated information, the clinic visit is not only an opportunity to educate patients, but, in fact, a responsibility. The latest guidance by ACOG calls for patient-centered contraceptive care by ?counseling about an individual's values, preferences, and goals? in order to ?help patients obtain contraceptive methods that best suit their values, needs, and priorities? while adhering to a shared decision-making process (ACOG Committee on Health Care for Underserved Women and Committee on Ethics, 2022). Bearing in mind the vast array of contraceptive methods available and the relative frequency of new iterations, shared decision- making during contraceptive counseling should entail an educational component in all patient- provider interactions. Indeed, we can think of few scenarios where an educational component would be misplaced or redundant within a contraceptive care context (e.g., patients who express to a 2020 Supreme Court decision that upheld sweeping exemptions granted to such employers by the Trump administration in 2018 (Sonfield, 2021). Lastly, a more recent federal provision was implemented to allow patients, with endorsement from a provider, to seek an expedient waiver to the insurance plan?s list of covered contraceptives in order to obtain the contraceptive brand that the patient needs without cost sharing; however, many insurance companies, patients, and providers are not knowledgeable about this requirement, and states do not enforce this policy (Andrews, 2021). 128 that they do not have the time nor interest in obtaining more information, patients long-term users of the same method coming in for a renewal of prescription or replacement of a LARC device). Improvements in contraceptive knowledge are in direct service to the reproductive justice tenet of maintaining personal bodily autonomy. Since virtually all women will use contraception at some point in their lives and many throughout their reproductive lifespan, factual contraceptive information allows women to make informed choices about their bodies and health. To our knowledge, our study is the first to evaluate knowledge about both the IUD and the implant following a visit with a provider. Although we cannot know whether the patient- provider interaction in our study met all the standards of patient-centered contraceptive care, results find that, on average, patients left the clinic visit with more knowledge than when they arrived?knowledge about LARC, about how to get free contraceptive methods, including LARC, and about how effective contraceptive methods are relative to each other, including LARC. Women who benefitted the most had lower levels of knowledge coming in, arguably the group of patients most in need of the increased knowledge. Although clinic type (women?s health or primary care) and reason for visit (family planning or not) were associated with some changes in knowledge in bivariate analyses, neither variable was predictive of change in knowledge for any of our four knowledge outcomes in adjusted analyses. That there were no effects of clinic type or reason for visit may suggest that DelCAN had a positive effect on facilitating education about contraception in primary care visits. Providers may have utilized scripts recommended by the DelCAN initiative, such as the One Key Question, to inquire about 129 women?s pregnancy intentions in order to initiate a conversation about contraceptive counseling. Although this is pure speculation as we do not have any information about how contraceptive topics were broached, we know from our data that 65.4% of our participants came in for a family planning visit (Table 1), but 83% reported mention of any family planning topic (Table A10). Again, with no baseline to compare to, we cannot claim a direct link with the DelCAN initiative, but we note that these are encouraging findings. 5.1.3. Contraceptive Attitudes In contrast to modest increases in knowledge outcomes observed in our study, changes in mean attitude scores were mostly negative and relatively small, as discussed in the subsequent paragraphs. However, similar to the results from our knowledge outcomes, pre-survey attitude scores were the strongest and most consistent predictors of change in attitudes when testing both continuous and dichotomous change outcomes; those with more positive attitudes before the visit had decreases of larger magnitude after the visit. Of all our sociodemographic, reproductive health, and provider visit measures, only feelings about a hypothetical pregnancy were associated with more than one method type attitude change, as discussed further on in this section. In terms of individual shifts, although more participants decreased in positive contraceptive attitudes, a substantial proportion also increased in positive attitudes across all four contraceptive types (Figures 9-12). Adjusted logistic models suggest these increases come from individuals with lower positive attitudes pre-visit (Tables 21-24). Furthermore, our findings allude to a link between contraceptive attitudes and pregnancy desires and feelings, which 130 change over the course of woman?s reproductive lifespan, possibly explaining the wide range of change in contraceptive attitudes observed in this study. Although changes in knowledge and positive attitudes are distinct phenomena, conceptually speaking, our study finds that the pattern of individual shifts in contraceptive attitudes is analogous to the individual shifts in contraceptive knowledge, but trends in the opposite direction. This suggests that may be a link between knowledge and attitudes, as explored further in this section. On average, participants? positive IUD attitude statements decreased following a visit with the provider, while positive implant attitudes remained relatively stable. Similarly, positive attitudes about hormonal birth control and condom use decreased post-visit, although the former change was small and non-significant. Despite differences in magnitude and significance, changes in the four groups of attitude statements all trend in the same direction (denoted by the negative values in the third column of Table 4), indicating that attitudes moved in the direction of holding less favorable perceptions about these methods. For example, the fact that the level of agreement increased with the statement, ?It is too much of a hassle to use a condom every time I have sex,? indicates that participants? belief that condoms are more trouble than they are worth became stronger following the provider visit. Participants with only post-visit LARC attitude scores, overall reported less agreement with positive IUD statements compared to women with pre- and post-visit scores, while implant attitudes were similar. Significance levels also varied when considering the individual items that make up the IUD and implant belief composite scores. Perhaps most interesting about these results is that only one item (and a similar one) per composite scale increased in agreement; following a 131 provider visit, participant agreement was higher for the statement, ?My provider will remove my IUD/implant whenever I decide I?d like it removed? (as denoted by positive values in the change column of Table 4). Although the change is only significant for the implant attitude, the fact that only these two items trend in the opposite direction compared to all other items suggests that the patient-provider interaction played a positive role in modestly nurturing patient trust in providers? future willingness to carry out ?early? LARC removals17. While provider intentions are not the same as provider behaviors and evidence finds that providers nationwide are generally resistant to ?early? LARC removals (Manzer & Bell, 2022b), our results suggest the Title X provider visit in Delaware served to somewhat ameliorate one of the concerns, i.e., barriers, for LARC uptake?provider-controlled removal. To our knowledge, our Aim 3 study is the first to evaluate attitudes about LARC following a visit with a provider. Two prior studies utilizing a pre-post design similar to ours focused on attitudes and beliefs about non-LARC types of contraception (Larsson et al., 2004a; Nobili et al., 2007). In a Swedish study about emergency contraception, positive attitudes increased in only three of seven survey statements in the intervention group and one in the comparison group; however, the difference in change between the two groups was not significant (Larsson et al., 2004a). In an Italian study, positive attitudes as measured by a ten-item scale increased significantly in the experimental group, but not in the control group (Nobili et al., 2007). Interestingly, both studies also measure knowledge in addition to attitudes; Larsson and 17 There is no clear definition of what constitutes ?early? LARC removal, but other studies have defined it anytime from three to twelve months (Amico et al., 2016, 2017; Grunloh et al., 2013; Sznajder et al., 2017). For context, according to the FDA, the current indicated duration of product use for LARC is three to ten years, depending on device, dosage, and brand (U.S. Food and Drug Administration, 2021). The Nexplanon implant and the Skyla IUD have the shortest duration of product use of any LARC device in the U.S.?three years. 132 colleagues (2004a) found non-significant knowledge increases in the intervention group across all items and in the comparison group across most items, while Nobili and colleagues (2007) found significant knowledge increases in the combined knowledge score of the experimental group and no change in the control group. Lastly, the two studies use very different study designs; one is a population-level study with little information on the informational content that was delivered (Nobili et al., 2007), while the other uses an experimental design with a highly standardized contraceptive counseling encounter (Larsson et al., 2004a). While both studies differ from ours, our study design is more like the Swedish study. In the absence of comparable research on attitude changes, cross-sectional studies using measures similar to ours may provide some context. Harper and colleagues (2010) constructed a 5-point Likert attitudes scale about hormonal contraception from items about the pill, patch, ring, and shot that denote skepticism about the methods, referred to as negative attitudes. Adolescent and young women of low-income completed the scale only once, following a provider visit to initiate one of the four contraceptive methods. Negative attitudes towards each of the four methods were found to be common, ranging from 30% to 68%, depending on each individual item (Harper et al., 2010). When considering the scale as a whole, made up of pill, patch, and ring items, 53% endorsed negative attitudes about these methods (Raine et al., 2011). Furthermore, a nationally representative survey of unmarried young men and women aged 18-29 conducted by the Guttmacher Institute found that 30% thought it was likely that using an IUD would cause an infection, 36% that the pill would likely cause weight gain, and 40% that the pill was likely to cause serious mood swings (Kaye et al., 2009). Although this dissertation uses 133 continuous rather than binary measures for attitudes, when dichotomizing relevant attitude items, we find our results are similar; ?Mood swings become worse with hormonal birth control? was endorsed (agree and strongly agree) by 47.4% pre-visit and 45.0% post-visit, and ?IUDs have bad side effects such as weight gain and irregular bleeding? by 29.5% pre-visit and 33.3% post- visit (data not shown). Negative attitudes about contraception, particularly hormonal methods, is common in prior literature and our findings. This scarcity in literature along with our own mixed results, does not lend itself to straightforward conclusions. Rather, situating our findings in the context of the two aforementioned studies raises the question of what kind of link exists between gaining more knowledge and changing one?s attitudes. Our participants were exposed to more information about contraceptive methods as evidenced by overall increases in knowledge levels across all items, but this was not accompanied by more favorable perceptions of contraceptive methods, similar to findings from Larsson and colleagues (2004a). Furthermore, participants who reported only becoming aware of the IUD in the post-visit survey, had lower scores, i.e., lower levels of agreement on IUD positive attitude items (see fifth column in Table 4) than those who had heard of the IUD before seeing a provider. In contrast, the implant scores were similar among the two groups of participants. Lastly, while the Italian study found a significant increase in favorable attitudes following an extended counseling visit, the increase was far smaller than the increase in knowledge (Nobili et al., 2007). These results suggest that increases in people?s contraceptive knowledge may be accompanied by a stronger focus on the negative aspects of contraceptive methods. 134 In psychology, the human tendency to give more weight to negative information, events, or emotions compared to positive ones is known as negativity bias (Rozin & Royzman, 2001). Evidence finds this to be true across all kinds of contexts, including ?everyday events, major life events (e.g., trauma), close relationship outcomes, social network patterns, interpersonal interactions, and learning processes? (Baumeister et al., 2001). Although the impact of contraceptive knowledge on contraceptive attitudes has not been studied, a patchwork of evidence aligns with our theory that contraceptive knowledge about negative features is associated with contraceptive attitudes. For instance, women report that few or no contraceptive methods have all the features that they want (Dixon et al., 2014; Lessard et al., 2012) and studies find women?s social networks to be particularly influential sources of negative attitudes and inaccurate information about contraception (A. J. Hoopes et al., 2018; Mahony et al., 2021; Yee & Simon, 2010). Women?s prior negative experiences with side-effects, the health care system, and a history of reproductive injustice has been found to influence their future contraceptive use (Gilliam et al., 2004; Gomez & Wapman, 2017; Guendelman et al., 2000; Thorburn & Bogart, 2005). Additionally, beyond simply the medical aspect, women?s concerns about contraceptive side-effects are also gendered and cultural; specific side-effects (ex., weight gain, emotionality) are widely infused with a negative social meaning that cannot but inform women?s attitudes about contraception (Guendelman et al., 2000; Littlejohn, 2013). In order for patients to be able to make an informed decision about contraception, providers must be upfront about the possibility and likelihood of having less favorable experiences with contraception (ex., possible short-term and long-term side-effects, coitus- 135 dependent use, daily use), and what to do if undesirable effects occur. However, as the gatekeepers of both contraceptive knowledge and access, providers are called upon to provide factual information without undermining nor overselling contraceptive attributes, as well as not steering the patient towards a predetermined method. Examples of educational patient-provider interactions where young adults are counseled on either the IUD, implant, shot, or combined hormonal methods (pill, patch, ring) are presented in comic format in The Birth Control Tales18 (Sridhar & Roher, 2017). It is perfectly valid to refuse a method due to its side-effect profile or any other undesirable attribute. However, these legitimate concerns should be weighed against the reason for seeking contraception?most frequently, the prevention of unintended pregnancy. It is possible that due to negativity bias, contraceptive counseling, such as the kind presented in The Birth Control Tales, may have an unanticipated, negative effect on contraceptive attitudes, potentially to the point of an unfavorable contraceptive attribute(s) (temporarily) eclipsing a method?s benefits. The statistically significant increase in agreement with the statement, ?It is too much hassle to use a condom every time I have sex,? is worth discussing separately. Although condoms are currently the best source of protection against sexually transmitted infections, as a method of pregnancy prevention they are among the less effective options. Contraceptive 18 The scenarios are focused on counseling about each method individually after the patient expressed interest in that specific method, so choice between various methods is not modelled in this source. The data provided in the comics are presumed to be correct as of 2017 when they were created; it is possible that newer versions of existing method types compare differently by hormone levels. 136 counseling by a provider typically focuses on more effective methods?those not available over the counter, as patients would not typically seek out a contraceptive visit for condoms. While we can only speculate as to the cause behind the change in condom use attitudes, recent research reports many examples of paternalistic clinicians who believe that LARC is the best option for their patients, as they call into question patients? ability, willingness, and motivation to engage in consistent non-LARC contraceptive behaviors (Mann, 2022; Manzer & Bell, 2021). Indeed, the oft-touted benefit of LARC is in its ?set it and forget it? nature, requiring neither daily, weekly, nor quarterly adherence, nor use at every instance of sexual intercourse. During contraceptive counseling, a provider might benevolently inquire about the patient?s current contraceptive use, including satisfaction, problems, consistency of use, and plans for future contraceptive use. It is in such a context that the topic of condoms as ?too much hassle to use? may have arisen. When such a contraceptive counseling interaction leads to queries about the patient?s interest in other forms of contraceptive methods (e.g., those that do not require frequent compliance behavior), the provider is implicitly signaling that the patient?s current or requested method (ex., condom, pill) is not a valid choice (Manzer & Bell, 2022a). In fact, administrative leaders championing DelCAN in their own clinics expressed unease about ?the absence of condoms and sexually transmitted infections from DelCAN training, funding, and mission,? and they reported the concern was shared by both providers and adolescent parents (Skra?i? et al., 2021, p. 214). Interestingly, despite the overall significant trend in our results of more agreement with the condom statement (indicating more less positive attitudes about condom use post-visit), we 137 found that women who reported that they would feel unhappy about a hypothetical pregnancy in the next year, compared to those who would be happy, had more positive attitude changes about using condoms. This would suggest that among women who are more wary of an unintended pregnancy, the provider visit had a positive influence on the patients? cost-benefit analysis in favor of condom usage, whether intended or not. Indeed, despite DelCAN?s apparent disregard for condoms, among participants who discussed any type of family planning during their provider visit, just about half reported that condoms were talked about, second in frequency only to the pill (data not shown).19 Nevertheless, the lack of data on the quality of the condom conversation prevents us from understanding the mechanism behind the change in condom attitudes. Perhaps the simplest explanation might be that condom counseling (or lack thereof) differed by both patient and provider, particularly in view of DelCAN?s lack of guidance in that respect. Or once women learn about LARC options, condoms seem less worth the hassle to some women. Be it as it may, our speculations about the cost-benefit analyses and differences in contraceptive counseling can be extended to the finding that patients who were unsure how they would feel about a hypothetical pregnancy, compared to those who would be happy, had more positive attitude changes regarding hormonal birth control. 19 The IUD, implant, and the shot were mentioned in just over one third of visits among women who discussed any type of family planning; the ring, the patch, withdrawal, and emergency contraception were discussed in approximately one fifth of such visits, and natural family planning and other barrier methods in one tenth of such visits (data not shown). As expected, mention of female and male sterilizations was the most infrequent since our sample excluded patients who were currently using or intending to use pre- or post-visit sterilization as a form of contraceptive method. 138 5.1.4. DelCAN Expansion Despite DelCAN?s launch in 2014, implementation of the multipronged initiative was staggered as some facets served as prerequisites to other initiative features. For instance, the provision of same-day LARC insertion only became possible after (1) the adoption of a policy to provide clinics with funding to stock up on LARC, (2) Upstream training, (3) enough trained providers became LARC-precepted, and (4) clinic workflow was adapted to facilitate longer appointments when LARC was requested on the spot, among others. Upstream training of providers began in 2016, and by the end of 2018, the vast majority, if not all, of the targeted sites had been trained by Upstream and were in the implementation phase (Skra?i? et al., 2021). The first wave of data collection between 2016 and 2017 was before the full program was rolled out, and the second wave of data collection between 2018-and 2019 was after all providers were trained but may have been before clinic workflow was adapted. Thus, the second wave of Title X data collection is the closest to the full rollout of the program and the first wave is the closest to before the program was rolled out, though the timing of measurement is imperfect. When investigating changes in effectiveness level of planned contraceptive method use, our findings suggest that the full expansion of the DelCAN was associated with changing to LARC for participants who arrived to the clinic not planning to use LARC. Although the result of the main adjusted model (third column in Table 14) is not significant, the large odds ratio indicates the non-significance is due to lack of power in the sample. This is supported by the significant results in the same direction that we obtain from the sensitivity model (also Table 14) and the supplementary models (Tables A4 and A7). However, since our overall data appear to 139 indicate that interaction with the provider did not play a large role in patients? plans to switch to a more effective method, it is possible that over time more women became interested in switching, perhaps through word of mouth and DelCAN?s media campaign. As we had hypothesized, increases in IUD and implant knowledge were more likely for women visiting clinics in 2018-19 compared to those visiting in 2017. We also found the DelCAN expansion to be associated with more positive implant attitudes, even with the overall sample trending towards less positive implant attitudes following a provider visit. In contrast, wave two was associated with lower effectiveness knowledge compared to wave one participants. Wave one participants had twice the likelihood of reporting both any increase in effectiveness knowledge and increase to highest effectiveness knowledge compared to women interviewed in wave two. Similarly, wave two was associated with a larger drop in positive attitudes about condom use compared to wave one. The lack of association between the DelCAN expansion and change to methods of higher effectiveness as well as the association between the DelCAN expansion and decreases in effectiveness knowledge and condom attitudes go against our overarching hypothesis of higher effectiveness level methods, more knowledge, and more positive attitudes in wave two. This contradiction cannot be attributed to differential pre-visit scores as we controlled for this in analyses. Interpreting these unexpected findings in the context of our other knowledge outcomes is difficult considering the sparse data available about the content of the provider visit, but we may speculate that as the DelCAN initiative continued to expand and providers and clinics adapted to the program and new recommendations, they prioritized increasing patients? LARC 140 knowledge over comparative effectiveness and the benefits of DelCAN. Or, perhaps, other components of the DelCAN initiative, such as the media campaign, were more successful in that respect during the early part the initiative. 5.2. The Integrated Behavioral Model and its Connection to the Findings The study?s main research questions were grounded in the Integrated Behavioral Model (IBM). While our hypotheses do not directly test the relationships presented in the IBM, the model allows for the contextualization of contraceptive intent, knowledge, and attitudes within the sphere of women?s contraceptive planning and future behavior. Consequently, this section examines the extent to which our findings are consistent with the IBM. Our contraceptive choice findings mostly align with the main premise of the IBM?intent as the primary predictor of future behavior. When observing individual shifts in contraceptive choice by comparing the two pre-visit measures (current and composite) (Figures 3-4), we see that participants? post-visit initiation and plans are, by and large, strongly informed by their pre- visit intentions; i.e., when taking into account women?s pre-visit intent, far fewer women report changing to a method of higher effectiveness, compared to only considering women?s current contraceptive use (4% vs. 18.6%). Although this distinction appears to be missing in the opposite direction based on numbers alone (9.5% vs. 9.5%), upon closer inspection, the participants who 141 make up the two groups of 9.5% (N=45) were not the same.20 This difference suggests that, while intent also plays a role in shifting to a method of lower effectiveness level or no method, the connection is not as clear as when seeking out a method of higher effectiveness. This discrepancy may be explained by the differential effort that is required to obtain different methods of contraception; no method and withdrawal are free, other low effective methods can be obtained over the counter at almost any store, most moderately effective methods require a provider's prescription and pharmacy visit, and highly effective methods require a (minimally) invasive procedure. Our findings suggest that when it comes to contraceptive choice and planning, intent plays a clearer role when method initiation requires more effort and elaborate long-term decision-making on the part of the patient. To be more specific, users remain in control and are able to change their minds throughout the processes of obtaining and using a condom (assuming the sexual act is consensual) and combined hormonal contraceptives. In contrast, using LARC entails accepting the idea of a device being inserted into one?s body, accepting that the device should only be removed by a trained provider, attending a clinic visit, and submitting to the procedure; by nature, the involvement of the provider and the larger health care system necessitates a degree of relinquishing one?s control over one?s body. It would be logical, therefore, that the larger the mental and logistical burden of initially obtaining and using a specific method, the clearer one?s intent needs to be for the behavior to occur. This 20 Of the 45 participants who decreased, 32 (71.1%) individuals decrease in effectiveness level by both pre-visit measures (current and composite), and 13 (28.9%) decrease only by one of the two pre-visit measures (data inferred from Tables A1-A2). 142 differentiation of effort as it relates to an individual?s intent is reflected in the elements of perceived control and self-efficacy in the current IBM. According to the IBM, ?knowledge and skills to perform the behavior? are an influencing factor on carrying out a specific behavior, but they are separate from the intention to perform the behavior. Although we agree that both elements are crucial to carrying out contraceptive behaviors, the parallel positioning of knowledge and intent do not aptly reflect the contraceptive decision-making process. Setting aside ?skills to perform the behavior,? we propose that women cannot have the intention to prevent pregnancy if they do not have the basic knowledge on how to do so. Similarly, women cannot intend to use LARC if they do not know that LARC exists. Similarly, they may not develop the intention to use LARC if they do not know that they are medically eligible to use it, that their insurance will cover it, and that it is the most highly effective form of reversible birth control. Although our research questions do not specifically test this proposed pathway as an element of contraceptive decision-making, the inclusion of knowledge in IBM, nevertheless, supports our rationale for investigating knowledge about LARC, effectiveness, and DelCAN benefits. This study?s attitude outcomes map fairly well onto the IBM components of experiential attitude, instrumental attitude, perceived control, and self-efficacy. Although this dissertation does not examine the relationship between these four components and intent itself, our findings suggest that a change in knowledge may play a role in shaping and changing attitudes, as mentioned earlier. Should this, indeed, be the case, it is not in line with the places of knowledge 143 and attitudes in the IBM framework. Further research is needed to confirm this link between attitudes and knowledge. It should be noted, however, that, in light of the health care system?s role as gatekeeper in obtaining most of the currently available contraceptive methods, including all moderately and highly effective ones, the IBM does not adequately describe the influence of environmental constraints on contraceptive behavior, nor the larger scale cumulative disadvantages experienced by racial and ethnic minorities and those of lower socioeconomic status in our society. Indeed, many contraceptive barriers and facilitators may be better explained by the Ecological Model for Health Promotion (McLeroy et al., 1988) and the Behavioral Model of Health Services Use (Andersen, 1995), as two distinct but complementary approaches for examining the individual?s behavior within their environmental context. In addition to highlighting the myriad of ways that external factors can impact individual health-seeking behaviors, both theoretical frameworks support the idea that the individual is in constant interaction with the world and the systems around them, emphasizing that behavior is an iterative, everchanging process. Studies about contraceptive behaviors and decision-making confirm this theory as reality (Downey et al., 2017; Frost et al., 2007b; Grady et al., 2002; R. K. Jones et al., 2015; Marshall et al., 2018; Moreau et al., 2007; Trussell & Vaughan, 1999). Nevertheless, since our hypotheses did not test external factors, the two models are beyond the scope of our research. 144 5.3. Implications for Practice and Public Health Both anecdotal and peer-reviewed evidence has found that family planning providers engage in limited conversation with their patients during counseling appointments, missing the opportunity to discuss preferences relevant to the patient (Dehlendorf et al., 2017; Dehlendorf, Kimport, et al., 2014; Manzer & Bell, 2022a; Yee & Simon, 2011). Not unique to the U.S., evidence of a similar phenomenon was found in Europe through an innovative online study that simulated the contraceptive counseling encounter. Primary care and women?s health practitioners across ten European countries engaged with three distinct patient profiles, each with a ?hidden? need (poor compliance, headaches, and desire for hormone-free contraception). Less than half of the practitioners uncovered the hidden needs of the poorly compliant and headaches patients, and about one third did not uncover the hormone-free desire. ?Clinicians who uncovered their patients? hidden needs or preferences asked significantly more questions than those who did not? (Nappi et al., 2022, p. 85). DelCAN provider trainings recommended tiered-contraceptive counseling as a way for patients to learn about LARC and to encourage them to consider LARC as a valid contraceptive method. This counseling technique?s meteoric rise in popularity among providers has faced a backlash from reproductive justice advocates for its potential to coerce women?s choices. It has since been disavowed by a number of organizations, who now emphasize the need for patient- centered care in contraceptive counseling. Before describing the latest recommendations for counseling techniques, we must first understand the rejection of the old. Its fall from grace is described in the following paragraphs. 145 The increased availability of LARC in the late 2000s occurred concurrently with the rise of tiered-effectiveness contraceptive counseling, whose script held the promise of a contraceptive counseling encounter that was comprehensive within a shorter timeframe while emphasizing effectiveness as the most important feature, in conjunction with the possibility of transferring the counseling task to a trained counselor thus freeing up the clinician?s time (Madden et al., 2013). Endorsed by major (inter)governmental and professional organizations, such as the World Health Organization, the Centers for Disease Control and Prevention (CDC), the American College of Obstetricians and Gynecologists (ACOG), and the American Academy of Pediatrics (AAP) (Brandi & Fuentes, 2020; Klein et al., 2015; Ott et al., 2014; World Health Organization, 2018), some scholars considered the tiered approach to be aligned with a rights-based approach (Stanback et al., 2015; Stevens & Berlan, 2014), while others cautioned about likely risks to women?s autonomous decision-making when efficacy is emphasized above all else (Gomez et al., 2014; Higgins, 2014). It is perhaps difficult to fully grasp the risks of a ?LARC first? counseling approach without contextualizing it within the legacy of coercive reproductive practices experienced by marginalized populations in the United States and abroad (Gold, 2014; Gomez et al., 2014). In the U.S., compulsory state-sanctioned sterilization began in the early 1900s and continued into the current century (Luna & Luker, 2013; Stern, 2005). Informed by a eugenicist perspective, at least 60,000 people, including immigrant women, women of color, women of low income, women with disabilities and mental illness, and women in the criminal justice system, have been involuntarily sterilized across the country in the name of population control or public health 146 (Luna & Luker, 2013; Stern, 2016). It was not uncommon for sterilizations to be conducted covertly and to include nulliparous women and adolescent girls, as well as men (Stern, 2005). In the 1990s, the focus shifted from sterilization to other strategies for controlling the fertility of specific populations. Following the FDA-approval of Norplant, a subdermal contraceptive implant that can be used for up to five years, more than a dozen states introduced legislation for financially incentivizing Norplant insertion for women receiving public assistance (Gold, 2014). While no such legislation ever went into effect, welfare reform incorporated a parallel coercive tactic?so-called ?family caps.? Instead of calculating the dollar amount of public assistance per number of children in the household, the family cap policy limited public assistance payments for women whose family size was over the designated ?cap.? Although research proved this policy of ?fertility prevention as poverty prevention? as unsuccessful in both reducing poverty and fertility, many states still maintain family caps on public assistance (Romero & Ag?nor, 2009; U.S. Government Accountability Office, 2001; Wiltz, 2019). Other examples of coercive practices include judges offering reductions in jail or prison sentences in exchange for LARC insertions or male and female sterilization, which have continued well into the 21st century (Gold, 2014). Lastly, compelling evidence exists that clinician prejudice results in biased contraceptive counseling. Younger women, women of color, and women of lower socioeconomic status (SES) are more often directed towards LARC compared to their older, white, and higher SES counterparts (Dehlendorf, Ruskin, et al., 2010; Manzer & Bell, 2021). In light of increasing criticism for the ?LARC first? counseling approach, the language of tiered-effectiveness contraceptive counseling has been scrubbed from recently published 147 recommendations by well-regarded organizations, emphasizing counseling in line with the principles of reproductive justice. In their 2020 publication regarding adolescents and LARC, the American Academy of Pediatrics recommend that clinicians ?[p]rovide LARC counseling within the reproductive justice framework to prevent directive and potentially coercive counseling? and ?[f]ocus on an end goal of improving the availability of LARC services to adolescents and not on increasing adolescent use of LARC methods? (Menon et al., 2020). Similarly, the newest recommendations by ACOG, published in February 2022, categorize and criticize the tiered- effectiveness model as directive counseling, advocating instead for a ?shared decision-making model, [where] each party?counselor and patient?is recognized as having valued expertise. [. . . ] Information can be shared, priorities can be explored, and, ultimately, the patient is the final arbiter of their decision, arriving at a choice that best meets their needs informed by the clinician's expertise? (ACOG Committee on Health Care for Underserved Women and Committee on Ethics, 2022). Promoting patient-centered care, these new recommendations are based on evidence showing that shared decision-making fosters provider trust and contraceptive satisfaction (Dehlendorf, Fox, et al., 2016; Dehlendorf, Krajewski, et al., 2014; Fox et al., 2018; Morse et al., 2017; Shay & Lafata, 2015). Regardless of the specific type of contraceptive counseling, an approach that allows space for patients to express their needs and ask questions while offering comprehensive care often takes more time than the system?s allotted time for a clinician visit. Indeed, when patient- centered contraceptive care and shared decision-making requires the prioritization of ?patients? values, preferences, and lived experiences in the selection or discontinuation of a contraceptive 148 method? (ACOG Committee on Health Care for Underserved Women and Committee on Ethics, 2022, p. 350), we inevitably wonder how much time one might reasonably schedule for such a visit. One model example of this was examined in an Italian study whose experimental patient- centered counseling session script started with a discussion of women?s desires about and prior experiences with contraception, followed by an offer of contraceptive information and education. Patients who accepted were educated about usage, pros, and cons of six methods, and the patient chose one of the presented methods in consultation with the provider. Following the intervention, not only did patients? knowledge increase, but so did their positive attitudes about contraception. Notable about this study is its allotment of a full 30 minutes for the extended counseling encounter that was carried out jointly by a gynecologist and a psychologist (Nobili et al., 2007). Although only a single study, it is worth investigating further whether a non-rushed patient- centered contraceptive care encounter can lead to more positive attitudes. One alternative to the perceived time shortage has been to shift a part of the onus onto patients. It is not uncommon for patients to be asked to review informational materials ahead of time; in this way the provider saves time by not having to go over the options, and they, instead, ask if the patients has any questions. The transition to online platforms and data management systems in health care has facilitated this trend. For instance, the University of Maryland health center?s online appointment platform asks patients to review Bedsider educational materials about contraceptive methods prior to coming in for a first-time appointment. The form also requires an expression of interest for at least one contraceptive method. While this is certainly an effective way to encourage their, primarily young, patients to seek out relevant information prior 149 to the visit with the provider, it also shifts much of the responsibility onto the patient and shapes the contraceptive counseling encounter before the provider and patient have even met. While obtaining more information prior to a provider visit is not undesirable in and of itself, it is concerning if its sole goal is to shorten the contraceptive counseling encounter. When appointment time slots become predicated on both the reviewal and understanding of pre-visit documents, it is difficult to imagine a patient-centered contraceptive care visit that prioritizes the ?values, preferences, and lived experiences in the selection or discontinuation of a contraceptive method? for all patients. Ultimately, it must be recognized that, to be executed appropriately, this type of counseling technique necessitates more time than a tiered-effectiveness counseling script or a delivery of information. It necessitates a conversation. This relates to our hypothesis that increased contraceptive knowledge may be linked to change towards more negative contraceptive attitudes due to humans? negativity bias upon learning about the possibility of less favorable features about certain contraceptive methods, such as irregular bleeding or headaches. While it may be tempting to gloss over side-effects precisely to avoid this negativity bias, particularly in instances where the patient clearly wants to avoid unintended pregnancy, research repeatedly shows that (dis)satisfaction with a method is associated with its (dis)continuation or intention to switch methods (Frost et al., 2007a; Steinberg et al., 2021), that patients want to be informed about side-effects (Cicerchia et al., 2022), and that misrepresentation of side-effects may discourage women from using the method or related methods again as well as result in a rejection of the entire healthcare system (Berndt & Bell, 2021a). Therefore, within the contraceptive counseling context, as a provider, it may be better to 150 err on the side of caution and present commonly reported side-effects, including less desirable features. This may create realistic expectations or if such side effects do not occur patients may be pleasantly surprised. If less desirable side effects are not discussed and they occur, then patients will be unpleasantly surprised. While this approach may not be aligned with the aims of programs like DelCAN, who may measure success by LARC uptake rates, it is a bare minimum requirement for facilitating women?s ability to make informed decisions over their bodies and from an overall perspective of reproductive justice. A model counseling interaction on contraceptive side-effects was described by a provider in The Birth Control Tales (Sridhar & Roher, 2017). The depicted encounter once again highlights the need for time to conduct comprehensive contraceptive counseling. Recent trends to incorporate the tenets of reproductive justice into contraceptive counseling and care are a step in the right direction. However, it is yet to be seen whether and to what extent providers will be willing and able to heed the new guidelines. Recent qualitative research by Bell and colleagues offers evidence of contraceptive providers nationwide utilizing the tenets of patient-centered care strategically in order to subvert their patient?s autonomy, particularly when it comes to LARC (Manzer & Bell, 2021, 2022b). Although clinicians believe that their patient interactions occur in the context of empowering equal partnerships, the biomedicalization of contraceptive care results in providers? preferences superseding women?s decision-making (Berndt & Bell, 2021b). The journey to achieving true reproductive justice in contraceptive counseling requires continued vigilance. Moving the pill, patch, and ring from 151 prescription-only to over-the-counter may be one way to shift that power dynamic in favor of the patient.21 5.4. Limitations This dissertation is not without limitations. Our results may not generalize to reproductive-aged women seeking services at Delaware?s Title-X-funded clinics, although the sampling strategy was designed to do so by being conducted across three levels and by adding sampling weights to each participant. First, our sample size is relatively small (N=474) when considering its diversity across sociodemographic, reproductive health, and provider visit measures, causing large confidence intervals around our relative risk ratios and the inability of many coefficients to attain significance. Our relatively small sample size also created the need for collapsing categories within effectiveness levels. Due to the small proportion of participants planning to use low effective methods as their highest method of effectiveness pre-visit, we combined the no method and low effective method groups into one effectiveness level to allow for comparison between our two pre-visit measures. It should also be noted that while our sample?s pre-visit contraceptive choices do not match those of the wider population by either of our two pre-visit measures (Daniels & Abma, 2020), this is not unusual considering that Title-X-funded clinics provide highly and 21 In more than one hundred countries oral contraception is available without a prescription, either formally or informally (Grindlay et al., 2013). Evidence from the U.S. finds that having the pill available over the counter is a viable option (D. Grossman et al., 2013; D. Grossman & Fuentes, 2013; Wollum et al., 2020). The patch and the ring are combined hormonal contraceptives like the pill, with similar hormone profiles. 152 moderately effective methods at higher rates than clinics not funded by Title X (Darney et al., 2022). Nevertheless, the merging of the two categories prevents us from gaining insight into differences between changes in planning to use a low effective method compared to no method. Insofar as the wide-ranging diversity of our sample is a limitation due to our sample size, it is also a strength to be able to test an array of characteristics that may have bearing on women?s intended contraceptive choice. For instance, the inclusion of all women at risk of unintended pregnancy in our studies, not just family planning patients, offers a more comprehensive perspective on the role of a range of provider visits on contraceptive choice, knowledge, and attitudes, and a better insight into the benefits of the statewide initiative. Indeed, DelCAN intended the initiative to help all reproductive-aged women, not only those explicitly interested in contraception or at high risk for unintended pregnancy, as is the case for many studies about contraception. Second, missing data and multiple non-random errors due to the survey instruments? skip patterns resulted in the exclusion of approximately 33% of participants who were missing on key variables and sociodemographic characteristics, such as contraceptive method use and age, and who might have been eligible for our study. Although the original skip patterns intended to reduce participant burden by omitting questions that were irrelevant following a specific participant response, the complexity of the skip pattern rules resulted in many unintended omissions. Most of the errors relevant to our study sample were corrected for the second wave of data collection, but this left us with inequivalent groups of participants by wave, which we were 153 able to imperfectly address by conducting sensitivity analyses for all our changes in planned contraceptive use outcomes. All data used in our studies were collected by women?s self-report before and after the provider visit, with no verification from medical charts or clinic staff to complement and confirm women?s responses. While self-reported data is commonly accepted as trustworthy, the issue lies in the broad and future-oriented phrasing of the survey?s contraceptive questions. As a result, we do not have any data to confirm whether women carried out any of their intended contraceptive behaviors. For instance, some participants reported in pre-visit that they wanted to quit LARC; apart from the contraceptive plan that they report at post-visit, women are not asked whether their LARC was removed during the visit. Similarly, we do not know whether patients who reported initiating or obtaining a prescription for contraception during the provider visit received the method in that visit or whether they indeed picked up their prescription. Consequently, our study is only able to make conclusions about contraceptive intent both pre- and post-visit, and not contraceptive choice. Lastly, the aims of this dissertation form part of a larger multiple-component longitudinal evaluation of the DelCAN initiative. The pre-post visit survey design aimed to capture changes in women?s self-reported contraceptive choice, knowledge, and attitudes that may be attributable to the initiative?s training of medical staff and subsidy of contraceptive methods. Ascertaining a causal link between changes in participants? responses between pre- and post-visit is impossible without a control group. Furthermore, we cannot know for certain whether the providers seen by our participants were DelCAN trained, and if they were, whether they subscribed to the tenets of 154 DelCAN and implemented its recommendations in their daily work?the latter was cited as one of the challenges in qualitative interviews with clinic leaders (Skra?i? et al., 2021). More specifically, beyond identifying which methods were discussed and who made the contraceptive decision, we do not know much else about the contraceptive counseling that women received. We also do not know if they utilized the tiered-effectiveness approach. Therefore, we had to engage in some speculation in order to interpret some of our findings. 5.5. Directions for Future Research This section briefly summarizes directions for future research, as suggested throughout this chapter. As relates to this dissertation?s first aim, the main novelty of our study was the use of two different pre-visit measures in order to account for patients? intent prior to visiting with a provider in order to gain a better understanding of the visit?s influence on women?s contraceptive planning. This method of measurement should be further explored and examined in other populations. Second, including more participants who use solely low effective methods may provide insight as to any potential differences between women using low effective methods and those using no methods. Third, the incorporation of a larger sample size would allow for analyses to have more power to examine what characteristics are significant predictors of change. The vast majority of studies that examine the connection between contraceptive education, including contraceptive counseling, and contraceptive knowledge find positive results 155 (Pazol et al., 2018), in line with this dissertation?s second aim results. The effect of counseling on contraceptive attitudes and contraceptive behaviors is not nearly as extensively studied as knowledge, and results have ranged from mixed to positive (Larsson et al., 2004b; Nobili et al., 2007; Zapata et al., 2018); findings from this dissertation?s third aim contribute to this small body of literature. In particular, there is a dearth of literature on changes in contraceptive attitudes, as well as lack of clarity between the terms beliefs and attitudes. Attempts to discern the two might elucidate the mixed findings of current studies. Additionally, the relationship between knowledge and attitudes requires further examination. Contraceptive intervention research aims to provide evidence of innovations that improve contraceptive outcomes, and ultimately, maternal and child health outcomes. However, as the review by Zapata and colleagues (2018) so aptly calls for, improved documentation of counseling content and processes is needed. This will facilitate the linking of observed changes to specific content and quality of interaction with the patient, which is needed in order to replicate the benefits of successful interventions for future patients. In the context of our study and available datapoints, our next step will be to investigate the extent to which changes in contraceptive knowledge and attitudes are associated with changes in intended contraceptive use and choice. Additionally, our unexpected attitude findings highlight the need for further research into the relationship between changes in contraceptive knowledge and changes in contraceptive attitudes. The possibility of knowledge mediating attitudes and vice versa needs to be explored. 156 Tables and Figures Figure 1. Integrated Behavioral Model. Diagram extracted from Monta?o and Kasprzyk (2008, p. 77). 157 Figure 2. Diagram of analytic sample selection. 158 Table 1. Baseline characteristics of the sample and their bivariate relationships with effectiveness of contraceptive method(s) currently using (current) and currently using and planning to use (composite) pre-visit. Current Pre-visit Composite Current-Planned Pre-visit NONE Chi- NONE Chi- % MODERATELY LARC MODERATELY LARC & LOW square & LOW square TOTAL 100% 31.9% 43.5% 24.7% 18.8% 50.4% 30.8% Baseline N=474 N=151 N=206 N=117 N=89 N=239 N=146 characteristics Age 0.664 0.086 15-19 17.3% 31.7% 48.8% 19.5% 9.8% 57.3% 32.9% 20-29 55.1% 32.2% 43.3% 24.5% 18.4% 51.0% 30.7% 30-44 27.6% 31.3% 40.5% 28.2% 25.2% 45.0% 29.8% Race/Hispanic 0.463 0.056 Ethnicity White NH 36.1% 27.5% 47.4% 25.1% 11.7% 55.6% 32.7% Black NH 38.8% 35.3% 41.3% 23.4% 25.0% 47.8% 27.2% Hispanic 21.9% 33.7% 38.5% 27.9% 20.2% 45.2% 34.6% Other NH 3.2% 26.7% 60.0% 13.3% 13.3% 60.0% 26.7% Current Relationship 0.016 0.014 Status Single 54.4% 34.5% 45.7% 19.8% 15.1% 54.3% 30.6% Married 20.0% 29.5% 33.7% 36.8% 27.4% 35.8% 36.8% Cohabitating 25.5% 28.1% 46.3% 25.6% 19.8% 53.7% 26.4% Insurance Type 0.892 0.376 Public 36.9% 33.7% 41.7% 24.6% 21.1% 46.3% 32.6% Private 36.3% 32.6% 43.6% 23.8% 16.3% 50.6% 33.1% None 26.8% 28.3% 45.7% 26.0% 18.9% 55.9% 25.2% 159 Education Level 0.523 0.513 High school or 38.4% 33.0% 39.0% 28.0% 17.0% 48.9% 34.1% less Some 42.2% 30.5% 47.5% 22.0% 20.0% 53.5% 26.5% college/vocational Bachelor?s degree 19.4% 32.6% 43.5% 23.9% 19.6% 46.7% 33.7% or more Nativity 0.030 0.108 U.S.-born 83.5% 32.1% 45.5% 22.5% 17.7% 52.5% 29.8% Foreign-born 16.5% 30.8% 33.3% 35.9% 24.4% 39.7% 35.9% Future Pregnancy 0.003 <0.001 Desire Within the next 2 8.9% 52.4% 21.4% 26.2% 52.4% 26.2% 21.4% years In 2 years or more 42.8% 33.0% 47.3% 19.7% 16.3% 52.7% 31.0% Unsure & Yes, 28.5% 31.1% 43.7% 25.2% 19.3% 50.4% 30.4% but not sure when No (more) 19.8% 21.3% 44.7% 34.0% 8.5% 56.4% 35.1% children Hypothetical 0.019 <0.001 Pregnancy Happy 38.4% 39.6% 39.0% 21.4% 29.7% 45.1% 25.3% Unhappy 34.0% 28.6% 49.1% 22.4% 9.9% 60.2% 29.8% Unsure 27.6% 25.2% 42.7% 32.1% 14.5% 45.8% 39.7% Unintended 0.831 0.817 Pregnancy No 59.3% 31.7% 44.5% 23.8% 18.1% 51.6% 30.2% Yes 40.7% 32.1% 42.0% 25.9% 19.7% 48.7% 31.6% Reproductive 0.083 0.342 Coercion by a 160 Partner - Ever None 84.8% 29.9% 44.5% 25.6% 17.9% 51.7% 30.3% Any 15.2% 43.1% 37.5% 19.4% 23.6% 43.1% 33.3% Reason for Clinic <0.001 <0.001 Visit Not Family 34.6% 42.7% 27.4% 29.9% 34.1% 32.9% 32.9% Planning Family Planning 65.4% 26.1% 51.9% 21.9% 10.6% 59.7% 29.7% Clinic Focus 0.102 0.018 Primary Care 32.9% 35.3% 36.5% 28.2% 25.6% 43.6% 30.8% Women's Health 67.1% 30.2% 46.9% 23.0% 15.4% 53.8% 30.8% Clinic Location 0.187 0.019 Urban 65.2% 31.7% 45.0% 23.3% 18.1% 52.4% 29.4% Rural 30.0% 28.9% 43.0% 28.2% 16.2% 47.9% 35.9% Suburban 4.9% 52.2% 26.1% 21.7% 43.5% 39.1% 17.4% Wave 0.292 0.530 One 46.4% 35.5% 40.9% 23.6% 20.9% 48.6% 30.5% Two 53.6% 28.7% 45.7% 25.6% 16.9% 52.0% 31.1% Notes. Due to rounding, the percentages may not always add up to 100%. 161 Table 2. Baseline characteristics of the sample and their bivariate relationships with effectiveness of contraceptive method(s) planning to use as reported post- visit. Planning to use Post-Visit NONE & Chi- % MODERATELY LARC LOW square TOTAL 100% 22.2% 50.8% 27.0% Baseline characteristics N=474 N=105 N=241 N=128 Age 0.162 15-19 17.3% 13.4% 57.3% 29.3% 20-29 55.1% 21.8% 50.6% 27.6% 30-44 27.6% 28.2% 47.3% 24.4% Race/Hispanic 0.026 Ethnicity White NH 36.1% 14.0% 57.3% 28.7% Black NH 38.8% 27.2% 50.5% 22.3% Hispanic 21.9% 26.9% 41.3% 31.7% Other NH 3.2% 20.0% 46.7% 33.3% Current Relationship 0.016 Status Single 54.4% 20.5% 51.9% 27.5% Married 20.0% 31.6% 36.8% 31.6% Cohabitating 25.5% 18.2% 59.5% 22.3% Insurance Type 0.426 Public 36.9% 21.1% 49.7% 29.1% Private 36.3% 24.4% 47.1% 28.5% None 26.8% 20.5% 57.5% 22.0% Education Level 0.761 High school or less 38.4% 23.6% 47.8% 28.6% Some 42.2% 21.5% 54.0% 24.5% college/vocational Bachelor?s degree 19.4% 20.7% 50.0% 29.3% or more Nativity 0.006 U.S.-born 83.5% 20.2% 54.0% 25.8% Foreign-born 16.5% 32.1% 34.6% 33.3% Future Pregnancy <0.001 Desire 162 Within the next 2 8.9% 57.1% 23.8% 19.0% years In 2 years or more 42.8% 16.7% 57.1% 26.1% Unsure & Yes, but 28.5% 25.9% 46.7% 27.4% not sure when No (more) children 19.8% 12.8% 55.3% 31.9% Hypothetical <0.001 Pregnancy 38.4% Happy 32.4% 45.1% 22.5% 34.0% Unhappy 12.4% 59.0% 28.6% Unsure 27.6% 19.8% 48.9% 31.3% Unintended Pregnancy 0.959 No 59.3% 21.7% 51.2% 27.0% Yes 40.7% 22.8% 50.3% 26.9% Reproductive Coercion 0.465 by a Partner - Ever None 84.8% 21.9% 52.0% 26.1% Any 15.2% 23.6% 44.4% 31.9% Reason for Clinic Visit <0.001 Not Family 34.6% 42.1% 31.1% 26.8% Planning Family Planning 65.4% 11.6% 61.3% 27.1% Clinic Focus <0.001 Primary Care 32.9% 34.0% 39.7% 26.3% Women's Health 67.1% 16.4% 56.3% 27.4% Clinic Location 0.008 Urban 65.2% 19.1% 54.7% 26.2% Rural 30.0% 24.6% 44.4% 31.0% Suburban 4.9% 47.8% 39.1% 13.0% Wave 0.394 One 46.4% 23.6% 52.3% 24.1% Two 53.6% 20.9% 49.6% 29.5% Notes. Due to rounding, the percentages may not always add up to 100%. 163 Table 3. Contraception and DelCAN knowledge score means and proportions per item prior to and following the provider visit. Knowledge measures (out of N=469) Pre-visit Post-visit Change Paired tests Mean Mean Mean t-test p-value (2- (SD) (SD) (SD) sided) IUD Knowledge score (range 0 to 6) 3.2 (1.9) 3.5 (1.8) 0.3 (1.4) <0.001 IUD knowledge questions (% having heard of IUD McNemar test p- % % % or with a correct response): value Heard of the IUD 85.1% 93.8% 8.7% <0.001 Nulliparous women can use IUDs 70.9% 66.4% Version A wording 68.4% 68.2% Version B wording 73.0% 64.9% An IUD without hormones exists 35.6% 39.1% Version A wording 37.3% 47.5% Version B wording 34.2% 32.2% IUDs will not affect future fertility 45.1% 49.1% Version A wording 36.7% 40.9% Version B wording 51.8% 55.8% Adolescents can use IUDs 54.4% 53.6% Version A wording 53.7% 54.5% Version B wording 55.0% 52.9% IUD can be removed before expiry date 71.7% 66.6% Version A wording 62.1% 56.1% Version B wording 79.3% 75.2% Mean Mean Mean t-test p-value (2- (SD) (SD) (SD) sided) Implant Knowledge score (range 0 to 4) 2.2 (1.5) 2.5 (1.3) 0.2 (1.1) <0.001 Implant knowledge questions (% having heard of McNemar test p- % % % IUD or with a correct response): value Heard of the implant 78.5% 89.1% 10.6% <0.001 Implant will not affect future fertility 50.8% 49.8% Version A wording 42.9% 40.9% Version B wording 57.1% 56.9% Adolescents can use implants 61.4% 57.7% Version A wording 62.6% 61.8% Version B wording 60.5% 54.3% Implant can be removed before expiry date 73.9% 69.4% Version A wording 66.9% 59.1% Version B wording 79.5% 77.6% 164 Mean Mean Mean t-test p-value (2- (SD) (SD) (SD) sided) Effectiveness Ranking score (range 0 to 4) 1.5 (1.7) 1.8 (1.7) 0.3 (1.3) <0.001 (% whose rankings are:) % % % All 4 methods ranked correctly (4 pts) 25.2% 30.9% 5.7% LARC first, condoms last (3pts) 5.8% 7.0% 1.2% LARC first (2 pts) 11.7% 14.3% 2.6% Condoms last (1 pts) 13.4% 10.7% -2.7% All other combinations (0 pts) 43.9% 37.1% -6.8% DelCAN Knowledge measure (out of N=471) Pre-visit Post-visit Change Paired tests Mean Mean Mean t-test p-value (2- (SD) (SD) (SD) sided) Knowledge about DelCAN score (range 0 to 7) 3.2 (2.3) 3.6 (2.4) 0.4 (1.9) <0.001 McNemar test p- Know how to obtain free (% who checked method): % % % value Male Condoms 77.1% 84.5% 7.4% <0.001 IUD (any brand) 38.6% 43.5% 4.9% 0.021 Implant 37.4% 41.0% 3.6% 0.065 Depo-Provera 44.4% 47.1% 2.7% 0.188 NuvaRing 35.9% 40.6% 4.7% 0.024 Birth Control Pills 62.8% 67.5% 4.7% 0.016 Same-day contraceptive access 24.8% 36.9% 12.1% <0.001 Notes. Due to rounding, the change mean presented in this table may not match the difference between the pre-visit and post-visit means. For the total knowledge scores and effectiveness ranking, the sample consists of 469 women. Participants who reported they had not heard of the method (IUD or implant, respectively) pre-visit or post-visit were given zero on the total knowledge score about the method. To calculate the percentage of correct answers for each individual item of each method, we used different sample sizes for pre- and post-visit knowledge items because those who had not heard of the method (IUD or implant) were not given the question and so could not get a correct or incorrect response. Sample sizes used were: Pre-visit heard of IUD: N=399, where 177 received version A and 222 version B; Post-visit heard of IUD: N=440, where 198 received version A and 242 version B; Pre-visit heard of implant: N=368, where 163 received version A and 205 version B; Post-visit heard of implant: N=418, where 186 received version A and 232 version B. For the DelCAN knowledge measure and corresponding individual items, the sample consists of 471 women. 165 Table 4. Means of contraceptive attitude score prior to and following the provider visit among the portion of the sample who had both pre- and post-visit data. Data for participants who only received attitude questions post-visit are presented separately (rightmost column). Post-visit Pre-visit Post-visit Change Paired tests participants only IUD Attitudes (N=399) Mean (SD) Mean (SD) Mean (SD) t-test p-value (2-s) Mean (SD) (N=40) IUD Attitude score (range 0 to 16) 9.9 (2.2) 9.7 (2.2) -0.2 (1.7) 0.009 9.0 (2.0) IUD attitude questions (range 0-4) It is a relief to have an IUD or implant 3.0 (0.9) 2.9 (0.9) -0.1 (0.9) 0.120 2.6 (1.0) My provider will remove my IUD when I want 3.0 (0.9) 3.1 (0.9) 0.1 (0.04) 0.356 2.8 (0.9) IUDs are uncomfortable? 2.1 (0.9) 2.0 (0.9) -0.1 (0.8) <0.001 2.0 (0.8) IUDs have bad side effects? 1.8 (0.9) 1.7 (0.9) -0.1 (0.9) 0.212 1.6 (0.7) Implant Attitudes (N=363) Mean (SD) Mean (SD) Mean (SD) t-test p-value (2-s) Mean (SD) (N=50) Implant Attitude score (range 0 to 8) 6.0 (1.5) 6.0 (1.6) -0.01 (1.3) 0.845 6.1 (1.7) Implant attitude questions (range 0-4) It is a relief to have an IUD or implant 3.0 (0.9) 2.9 (1.0) -0.1 (0.9) 0.020 3.0 (1.0) Provider will remove my implant when I want 3.0 (0.9) 3.1 (0.9) 0.1 (0.9) 0.033 3.1 (0.9) Hormonal Birth Control Attitude (N=466) Mean (SD) Mean (SD) Mean (SD) t-test p-value (2-s) Mood swings become worse with hormonal birth 1.5 (1.0) 1.5 (0.9) -0.05 (0.8) 0.237 control? (range 0-4) External Condom Attitude (N=467) Mean (SD) Mean (SD) Mean (SD) t-test p-value (2-sided) It is too much hassle to use a condom every time I 2.5 (1.3) 2.3 (1.3) -0.1 (1.0) 0.003 have sex? (range 0-4) 166 Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. Response options were scored as follows: 4 points for ?strongly agree,? 3 for ?agree,? 2 for ?neither agree nor disagree? and ?don?t know,? 1 for ?disagree,? and 0 for ?strongly disagree.? ? Items were reverse coded as follows: 4 points for ?strongly disagree,? 3 for ?disagree,? 2 for ?neither agree nor disagree? and ?don?t know,? 1 for ?agree,? and 0 for ?strongly agree.? 167 Table 5. IUD knowledge means by study variables. N=469 Pre-visit Post-visit Change p-value Mean p-value Mean p-value Mean (SD) ANOVA (SD) ANOVA (SD) ANOVA Total Sample (range 0-6) 3.2 (1.9) 3.5 (1.8) 0.3 (1.4) Age 0.335 0.131 0.542 15-19 3.2 (2.0) 3.4 (1.9) 0.2 (1.4) 20-29 3.3 (1.9) 3.7 (1.7) 0.4 (1.4) 30-44 3.0 (1.9) 3.3 (1.9) 0.3 (1.4) Race/Hispanic Ethnicity <0.001 0.086 0.006 White NH 3.7a (1.7) 3.8 (1.7) 0.1a (1.1) Black NH 2.8b (2.0) 3.3 (1.7) 0.5b (1.6) Hispanic 3.0b (2.0) 3.4 (1.9) 0.4ab (1.4) Other NH 3.6ab (1.7) 3.5 (2.1) -0.1ab (1.0) Current Relationship 0.180 0.006 0.267 Status Single 3.2 (1.9) 3.5ab (1.7) 0.3 (1.5) Married 2.9 (1.9) 3.1a (1.8) 0.1 (1.2) Cohabitating 3.4 (1.8) 3.8b (1.8) 0.4 (1.3) Insurance Type 0.008 0.006 0.337 Public 3.2ab (2.0) 3.6ab (1.8) 0.4 (1.5) Private 3.5a (1.8) 3.8a (1.7) 0.2 (1.3) None 2.8b (1.9) 3.1b (1.9) 0.3 (1.3) Education Level <0.001 <0.001 0.740 High school or less 2.7a (2.0) 3.1 a (1.7) 0.3 (1.5) Some b 3.7 b (1.8) 0.3 (1.2) college/vocational 3.4 (1.9) Bachelor?s degree or b 4.0 b (1.7) 0.4 (1.4) more 3.7 (1.7) Nativity 0.098 0.019 0.466 U.S.-born 3.3 (1.9) 3.6 (1.8) 0.3 (1.4) Foreign-born 2.9 (2.0) 3.1 (1.9) 0.2 (1.3) Future Pregnancy Desire 0.044 0.221 0.672 Within the next 2 years 3.1ab (2.0) 3.5 (1.9) 0.4 (1.2) In 2 years or more 3.5a (1.8) 3.7 (1.7) 0.2 (1.3) 168 Unsure & Yes, but not 2.9b (2.0) 3.3 (1.8) 0.4 (1.5) sure when No (more) children 3.2ab (1.9) 3.5 (1.8) 0.3 (1.4) Hypothetical Pregnancy 0.390 0.249 0.604 Happy 3.1 (1.9) 3.4 (1.8) 0.3 (1.2) Unhappy 3.3 (1.9) 3.7 (1.7) 0.4 (1.5) Unsure 3.3 (1.9) 3.5 (1.8) 0.2 (1.4) Unintended Pregnancy 0.278 0.464 0.579 No 3.1 (1.9) 3.5 (1.8) 0.3 (1.4) Yes 3.3 (1.9) 3.6 (1.7) 0.3 (1.3) Reproductive Coercion by 0.656 0.154 0.219 Partner ? Ever None 3.2 (1.9) 3.5 (1.8) 0.3 (1.4) Any 3.3 (1.9) 3.8 (1.8) 0.5 (1.4) Reason for Clinic Visit 0.372 0.529 0.675 Not Family Planning 3.1 (1.9) 3.4 (1.8) 0.3 (1.3) Family Planning 3.3 (1.9) 3.6 (1.8) 0.3 (1.4) Clinic Focus <0.001 0.016 0.036 Primary Care 2.7 (2.0) 3.2 (1.9) 0.5 (1.4) Women's Health 3.4 (1.8) 3.7 (1.7) 0.2 (1.3) Clinic Location 0.020 0.375 0.119 Urban 3.3a (1.9) 3.6 (1.8) 0.2 (1.4) Rural 3.1ab (1.9) 3.5 (1.7) 0.4 (1.4) Suburban 2.2b (1.9) 3.0 (1.9) 0.8 (1.7) Wave 0.243 0.016 0.131 One 3.1 (1.9) 3.3 (1.8) 0.2 (1.2) Two 3.3 (1.9) 3.7 (1.7) 0.4 (1.5) Randomized Group 0.017 0.364 different pre- and post- A 3.6 (1.5) 3.7 (1.6) samples B 3.9 (1.4) 3.8 (1.6) Notes. Due to rounding, the change mean presented in this table may not match the difference between the pre-visit and post-visit means. The means presented in the randomized knowledge group cells are based only on participants who received the questions and, consequently, were assigned a group. Pre-visit 399 participants had heard of the IUD (177 received version A and 222 version B), and post-visit 440 (198 received version A and 242 version B). When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 169 Table 6. Implant knowledge means by study variables. N=469 Pre-visit Post-visit Change p-value Mean p-value Mean p-value Mean (SD) ANOVA (SD) ANOVA (SD) ANOVA Total Sample (range 0-4) 2.2 (1.5) 2.5 (1.3) 0.2 (1.1) Age <0.001 0.016 0.030 15-19 2.8a (1.2) 2.8a (1.1) -0.1a (0.9) 20-29 2.2b (1.5) 2.5ab (1.3) 0.2ab (1.1) 30-44 1.9c (1.6) 2.2b (1.4) 0.4b (1.3) Race/Hispanic Ethnicity 0.766 0.775 0.689 White NH 2.3 (1.4) 2.6 (1.3) 0.3 (1.2) Black NH 2.3 (1.5) 2.4 (1.2) 0.1 (1.1) Hispanic 2.1 (1.6) 2.4 (1.4) 0.3 (1.1) Other NH 2.1 (1.7) 2.4 (1.6) 0.3 (1.0) Current Relationship 0.013 0.006 0.764 Status Single 2.4a (1.4) 2.5a (1.2) 0.2 (1.2) Married 1.8b (1.6) 2.1b (1.5) 0.2 (1.1) Cohabitating 2.3a (1.4) 2.6a (1.2) 0.3 (1.1) Insurance Type 0.206 0.066 0.793 Public 2.3 (1.5) 2.5 (1.3) 0.3 (1.2) Private 2.4 (1.5) 2.6 (1.3) 0.2 (1.1) None 2.1 (1.5) 2.2 (1.3) 0.2 (1.1) Education Level 0.261 0.059 0.834 High school or less 2.1 (1.5) 2.3 (1.3) 0.2 (1.1) Some 2.4 (1.5) 2.6 (1.3) 0.3 (1.1) college/vocational Bachelor?s degree or 2.2 (1.6) 2.4 (1.4) 0.2 (1.2) more Nativity 0.046 0.006 0.594 U.S.-born 2.3 (1.4) 2.5 (1.2) 0.2 (1.2) Foreign-born 1.9 (1.7) 2.1 (1.5) 0.2 (1.0) Future Pregnancy Desire 0.014 0.013 0.133 Within the next 2 years 2.0ab (1.7) 2.5ab (1.3) 0.5 (1.3) In 2 years or more 2.5a (1.4) 2.7a (1.2) 0.2 (1.0) 170 Unsure & Yes, but not 2.1b (1.5) 2.2b (1.3) 0.1 (1.2) sure when No (more) children 2.1ab (1.6) 2.4ab (1.3) 0.4 (1.3) Hypothetical Pregnancy 0.177 0.557 0.123 Happy 2.1 (1.5) 2.4 (1.4) 0.3 (1.2) Unhappy 2.3 (1.5) 2.6 (1.2) 0.3 (1.2) Unsure 2.4 (1.4) 2.5 (1.3) 0.1 (1.0) Unintended Pregnancy - 0.756 0.630 0.340 Ever No 2.2 (1.5) 2.5 (1.3) 0.3 (1.1) Yes 2.3 (1.5) 2.4 (1.3) 0.2 (1.2) Reproductive Coercion by 0.955 0.596 0.498 Partner None 2.2 (1.5) 2.5 (1.3) 0.2 (1.1) Any 2.2 (1.6) 2.5 (1.2) 0.3 (1.2) Reason for Clinic Visit 0.197 0.090 0.803 Not Family Planning 2.1 (1.5) 2.3 (1.4) 0.2 (1.1) Family Planning 2.3 (1.5) 2.5 (1.3) 0.2 (1.1) Clinic Focus 0.003 0.070 0.072 Primary Care 2.0 (1.5) 2.3 (1.3) 0.4 (1.3) Women's Health 2.4 (1.5) 2.5 (1.3) 0.2 (1.0) Clinic Location 0.123 0.478 0.146 Urban 2.3 (1.5) 2.5 (1.3) 0.2 (1.1) Rural 2.2 (1.4) 2.5 (1.2) 0.3 (1.2) Suburban 1.6 (1.5) 2.2 (1.3) 0.5 (1.4) Wave 0.012 0.001 0.691 One 2.1 (1.5) 2.3 (1.4) 0.2 (1.2) Two 2.4 (1.4) 2.6 (1.2) 0.2 (1.1) Randomized Group 0.021 0.007 different pre- and A 2.7 (1.0) 2.6 (1.0) post-samples B 3.0 (1.0) 2.9 (1.0) Notes. Due to rounding, the change means presented in this table may not match the difference between the corresponding pre-visit and post-visit means. The means presented in the randomized knowledge group cells are based only on participants who received the questions and, consequently, were assigned a group. Pre-visit 368 participants had heard of the implant (163 received version A and 205 version B), and post-visit 418 (186 received version A and 232 received version B). When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 171 Table 7. Contraceptive effectiveness knowledge by study variables. N=469 Pre-visit Post-visit All All Score 1- Score 1-3 Score 0 Chi- Score 0 Correct Correct 3 Chi- square square % of Total Sample 25.2% 30.9% 43.9% 30.9% 32.0% 37.1% Age 0.072 0.035 15-19 31.7% 26.8% 41.5% 41.5% 29.3% 29.3% 20-29 24.3% 27.8% 47.9% 28.6% 29.3% 42.1% 30-44 22.7% 39.8% 37.5% 28.9% 39.1% 32.0% Race/Hispanic 0.424 0.002 Ethnicity White NH 30.2% 30.2% 39.6% 41.4% 24.9% 33.7% Black NH 21.4% 29.7% 48.9% 23.6% 36.3% 40.1% Hispanic 22.3% 35.0% 42.7% 23.3% 36.9% 39.8% Other NH 33.3% 26.7% 40.0% 53.3% 26.7% 20.0% Current 0.212 0.830 Relationship Status Single 25.0% 28.9% 46.1% 30.9% 31.6% 37.5% Married 23.7% 40.9% 35.5% 26.9% 35.5% 37.6% Cohabitating 26.7% 27.5% 45.8% 34.2% 30.0% 35.8% Insurance Type 0.190 0.126 Public 25.0% 29.7% 45.3% 29.1% 34.9% 36.0% Private 30.4% 28.7% 40.9% 37.4% 26.3% 36.3% None 18.3% 35.7% 46.0% 24.6% 35.7% 39.7% Education Level 0.003 0.500 High school or 23.5% 27.4% 49.2% 28.5% 32.4% 39.1% less Some 30.5% 27.0% 42.5% 34.5% 29.0% 36.5% college/voc. Bachelor?s 16.7% 46.7% 36.7% 27.8% 37.8% 34.4% degree+ Nativity 0.468 0.180 U.S.-born 25.4% 29.8% 44.8% 32.1% 30.3% 37.7% Foreign-born 23.7% 36.8% 39.5% 25.0% 40.8% 34.2% Future Pregnancy 0.784 0.603 Desire Within next 2 21.4% 40.5% 38.1% 33.3% 38.1% 28.6% years In 2 years or 25.7% 27.7% 46.5% 31.2% 29.7% 39.1% more 172 Unsure & Yes, 25.6% 30.8% 43.6% 30.1% 29.3% 40.6% but not sure when No (more) 25.0% 33.7% 41.3% 30.4% 38.0% 31.5% children Hypothetical 0.468 0.426 Pregnancy Happy 21.0% 33.7% 45.3% 26.0% 34.3% 39.8% Unhappy 26.3% 28.8% 45.0% 32.5% 31.9% 35.6% Unsure 29.7% 29.7% 40.6% 35.9% 28.9% 35.2% Unintended 0.437 0.204 Pregnancy - Ever No 23.5% 30.3% 46.2% 28.2% 34.7% 37.2% Yes 27.6% 31.8% 40.6% 34.9% 28.1% 37.0% Reproductive Coercion by a 0.669 0.248 Partner - Ever None 24.4% 31.0% 44.6% 29.5% 32.2% 38.3% Any 29.2% 30.6% 40.3% 38.9% 30.6% 30.6% Reason for Clinic 0.954 0.565 Visit Not Family 25.9% 30.2% 43.8% 27.8% 33.3% 38.9% Planning Family Planning 24.8% 31.3% 44.0% 32.6% 31.3% 36.2% Clinic Focus 0.616 0.021 Primary Care 22.7% 30.5% 46.8% 23.4% 39.0% 37.7% Women's Health 26.3% 31.1% 42.5% 34.6% 28.6% 36.8% Clinic Location 0.250 0.136 Urban 22.6% 32.1% 45.2% 27.9% 32.5% 39.7% Rural 31.7% 28.9% 39.4% 38.7% 31.0% 30.3% Suburban 18.2% 27.3% 54.5% 22.7% 31.8% 45.5% Wave 0.950 0.107 One 24.5% 31.5% 44.0% 35.6% 28.7% 35.6% Two 25.7% 30.4% 43.9% 26.9% 34.8% 38.3% Notes. Due to rounding, the percentages presented in this table may not always add up to 100%. 173 Table 8. Knowledge about DelCAN benefits by study variables. Pre-visit Post-visit Change N=471 p-value p-value Mean p-value Mean (SD) Mean (SD) ANOVA ANOVA (SD) ANOVA Total Sample (range 0-7) 3.2 (2.3) 3.6 (2.4) 0.4 (1.9) Age 0.758 0.509 0.294 15-19 3.3 (2.4) 3.8 (2.3) 0.5 (1.8) 20-29 3.1 (2.3) 3.6 (2.4) 0.5 (1.9) 30-44 3.3 (2.3) 3.5 (2.4) 0.2 (2.2) Race/Hispanic Ethnicity 0.003 <0.001 0.411 White NH 2.9a (2.2) 3.2ac (2.3) 0.3 (1.7) Black NH 3.6b (2.4) 4.2b (2.4) 0.6 (2.1) Hispanic 3.2ab (2.3) 3.6ab (2.4) 0.3 (2.1) Other NH 1.8a (2.0) 1.7c (1.8) -0.1 (1.8) Current Relationship 0.082 0.288 0.601 Status Single 3.0 (2.2) 3.5 (2.3) 0.5 (1.8) Married 3.2 (2.4) 3.5 (2.4) 0.3 (1.9) Cohabitating 3.6 (2.5) 3.9 (2.5) 0.3 (2.2) Insurance Type <0.001 <0.001 0.489 Public 3.9a (2.3) 4.3a (2.4) 0.4 (2.0) Private 2.8b (2.2) 3.1b (2.2) 0.3 (1.7) None 2.8b (2.3) 3.4b (2.4) 0.6 (2.2) Education Level 0.074 0.094 0.198 High school or less 3.3 (2.3) 3.5 (2.4) 0.2 (1.9) Some 3.4 (2.4) 3.9 (2.4) 0.5 (2.0) college/vocational Bachelor?s degree or 2.7 (2.2) 3.3 (2.3) 0.6 (1.9) more Nativity 0.002 0.010 0.650 U.S.-born 3.4 (2.4) 3.7 (2.4) 0.4 (1.9) Foreign-born 2.5 (2.0) 3.0 (2.2) 0.5 (2.2) Future Pregnancy Desire 0.607 0.585 0.147 Within the next 2 years 3.3 (2.4) 3.6 (2.4) 0.3 (1.4) In 2 years or more 3.1 (2.2) 3.7 (2.5) 0.6 (2.0) Unsure & Yes, but not 3.4 (2.4) 3.7 (2.3) 0.3 (1.8) sure when 174 No (more) children 3.2 (2.4) 3.3 (2.3) 0.1 (2.1) Hypothetical Pregnancy 0.091 0.031 0.548 Happy 3.3 (2.4) 3.5ab (2.5) 0.3 (1.7) Unhappy 2.9 (2.2) 3.3a (2.3) 0.4 (2.2) Unsure 3.5 (2.4) 4.1b (2.3) 0.5 (1.9) Unintended Pregnancy 0.001 0.016 0.255 No 2.9 (2.3) 3.4 (2.4) 0.5 (2.0) Yes 3.7 (2.4) 3.9 (2.4) 0.3 (1.9) Reproductive Coercion by 0.832 0.455 0.507 Partner None 3.2 (2.3) 3.6 (2.4) 0.4 (1.9) Any 3.3 (2.3) 3.8 (2.2) 0.5 (2.1) Reason for Clinic Visit 0.769 0.035 0.025 Not Family Planning 3.2 (2.3) 3.3 (2.3) 0.1 (1.7) Family Planning 3.2 (2.3) 3.8 (2.4) 0.5 (2.1) Clinic Focus 0.313 0.446 0.032 Primary Care 3.4 (2.3) 3.5 (2.4) 0.1 (2.0) Women's Health 3.1 (2.3) 3.7 (2.4) 0.5 (1.9) Clinic Location 0.823 0.827 0.342 Urban 3.2 (2.3) 3.7 (2.4) 0.5 (1.9) Rural 3.3 (2.3) 3.5 (2.3) 0.2 (2.1) Suburban 3.3 (2.6) 3.6 (2.6) 0.3 (1.8) Wave 0.073 0.760 0.078 One 3.0 (2.3) 3.6 (2.4) 0.6 (1.8) Two 3.4 (2.3) 3.6 (2.4) 0.3 (2.0) Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 175 Table 9. IUD attitude scores by study variables Post-visit Pre-visit Post-visit Change participants only N=40 N=399 p- p- p- p- value Mean value value Mean value Mean (SD) Mean (SD) ANO (SD) ANO ANO (SD) ANO VA VA VA VA Total Sample 9.9 (2.2) 9.7 (2.2) -0.2 (1.7) 9.0 (2.0) Age 0.042 0.036? 0.077 0.766 15-19 10.0ab (1.8) 9.3 (1.7) -0.7 (1.7) 9.5 (2.2) 20-29 9.7a (2.2) 9.6 (2.2) -0.1 (1.7) 8.8 (2.0) 30-44 10.4b (2.2) 10.2 (2.4) -0.2 (1.8) 9.1 (2.2) Race/Hispanic 0.004 0.053 0.427 0.004 Ethnicity White, NH 10.2a (2.2) 9.9 (2.4) -0.3 (1.4) 9.3ab (1.7) Black, NH 9.5b (2.1) 9.3 (2.1) -0.2 (1.9) 9.6a (2.0) Hispanic 10.4a (2.1) 10.0 (2.2) -0.3 (2.0) 7.1b (1.2) Other, NH 9.1ab (1.9) 9.5 (1.9) 0.5 (2.0) NE ? Current Relationship 0.183 0.277 0.971 0.515 Status Single 9.8 (2.2) 9.6 (2.2) -0.2 (1.6) 9.2 (2.0) Married 10.3 (1.9) 10.1 (1.9) -0.3 (2.0) 8.2 (2.3) Cohabitating 10.0 (2.3) 9.7 (2.5) -0.2 (1.8) 8.9 (1.9) Insurance Type 0.194 0.642 0.308 0.081 Public 10.2 (2.3) 9.8 (2.3) -0.4 (1.8) 9.2 (2.0) Private 9.9 (2.0) 9.6 (2.1) -0.3 (1.6) 9.8 (1.8) None 9.7 (2.2) 9.7 (2.3) -0.01 (1.9) 7.9 (2.0) Education Level 0.605 0.699 0.082 0.778 High school 10.1 (2.0) 9.6 (2.1) -0.5 (1.7) 8.9 (2.2) or less Some 9.9 (2.2) 9.7 (2.3) -0.2 (1.8) 9.0 (1.9) college/vocational Bachelor?s 9.8 (2.2) 9.9 (2.2) 0.1 (1.7) 9.6 (1.8) degree or more Nativity 0.974 0.566 0.440 0.153 U.S.-born 9.9 (2.2) 9.7 (2.2) -0.3 (1.7) 9.2 (2.0) Foreign-born 9.9 (2.2) 9.9 (2.1) -0.1 (1.9) 8.0 (2.2) Future 0.610 0.629 0.350 0.629 176 Pregnancy Desire Within the 9.7 (2.5) 9.8 (2.1) 0.1 (1.7) 9.8 (2.1) next 2 years In 2 years or 10.0 (2.0) 9.6 (2.0) -0.4 (1.5) 9.6 (2.0) more Unsure & Yes, but not sure 9.8 (2.2) 9.7 (2.3) -0.1 (2.0) 9.7 (2.3) when No (more) 10.2 (2.2) 10.0 (2.7) -0.2 (1.9) 10.0 (2.7) children Hypothetical 0.625 0.657 0.703 0.501 Pregnancy Happy 9.8 (2.2) 9.6 (2.1) -0.3 (1.9) 8.4 (1.8) Unhappy 9.9 (2.2) 9.8 (2.5) -0.1 (1.8) 9.2 (2.1) Unsure 10.1 (2.1) 9.8 (2.0) -0.3 (1.4) 9.3 (2.2) Unintended 0.400 0.202 0.559 0.108 Pregnancy No 9.9 (2.1) 9.6 (2.1) -0.3 (1.7) 8.6 (1.9) Yes 10.0 (2.2) 9.9 (2.4) -0.2 (1.8) 9.7 (2.1) Reproductive Coercion by a 0.327 0.212 0.706 0.442 Partner - Ever None 10.0 (2.2) 9.8 (2.3) -0.2 (1.8) 8.9 (1.9) Any 9.7 (2.0) 9.4 (2.0) -0.3 (1.6) 9.8 (2.9) Reason for Clinic 0.593 0.296 0.502 0.748 Visit Not Family 9.9 (2.1) 9.5 (2.1) -0.3 (1.6) 8.9 (1.9) Planning Family 10.0 (2.2) 9.8 (2.3) -0.2 (1.8) 9.1 (2.1) Planning Clinic Focus 0.919 0.797 0.649 0.057 Primary Care 10.0 (2.2) 9.7 (2.3) -0.3 (1.6) 8.3 (1.9) Women's Health 9.9 (2.1) 9.7 (2.2) -0.2 (1.8) 9.5 (2.0) Clinic Location 0.854 0.954 0.590 0.054 Urban 10.0 (2.2) 9.7 (2.3) -0.3 (1.8) 9.3 (2.0) Rural 9.9 (2.1) 9.8 (2.1) -0.1 (1.6) 9.1 (1.9) Suburban 10.2 (2.4) 9.7 (1.9) -0.5 (1.7) 6.3 (1.5) Wave 0.063 0.064 0.952 0.427 One 9.7 (2.1) 9.5 (2.1) -0.2 (1.7) 8.7 (2.0) Two 10.1 (2.2) 9.9 (2.3) -0.2 (1.8) 9.2 (2.1) ?Non-estimable; the cell was empty. ?The omnibus F test was significant, but the post-hoc Tukey test did not detect any significant differences (the 177 difference between the thirty and over group and adolescents as well as the difference between the thirty and over group and participants in their twenties were marginally significant p<0.1). Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 178 Table 10. Implant attitude scores by study variables Post-visit Pre-visit Post-visit Change participants only N=50 N=363 p- p- p- p-value Mean value Mean value Mean value Mean ANOV (SD) ANO (SD) ANO (SD) ANO (SD) A VA VA VA Total Sample 6.0 (1.5) 6.0 (1.6) -0.01 (1.3) 6.1 (1.7) Age 0.839 0.850 1.000 0.353 15-19 6.0 (1.5) 6.0 (1.5) -0.01 (1.4) 4.5 (0.7) 20-29 6.0 (1.5) 6.0 (1.6) -0.01 (1.4) 6.1 (1.7) 30-44 6.1 (1.4) 6.1 (1.5) -0.01 (1.3) 6.3 (1.7) Race/Hispanic 0.472 0.399 0.499 0.924 Ethnicity White, NH 6.0 (1.5) 6.1 (1.6) 0.02 (1.1) 5.9 (1.7) Black, NH 5.9 (1.6) 5.9 (1.6) -0.02 (1.5) 6.2 (1.7) Hispanic 6.3 (1.3) 6.1 (1.5) -0.2 (1.4) 6.3 (1.6) Other, NH 6.2 (0.8) 6.7 (1.3) 0.5 (1.0) 6.5 (2.1) Current Relationship 0.003 0.071 0.486 0.092 Status Single 5.8a (1.5) 5.9 (1.6) 0 (1.3) 5.8 (1.6) Married 6.5b (1.3) 6.3 (1.5) -0.2 (1.4) 7.1 (1.3) Cohabitating 6.2ab (1.4) 6.2 (1.6) 0 (1.4) 6.0 (1.8) Insurance Type 0.826 0.605 0.854 0.962 Public 6.1 (1.6) 6.1 (1.6) 0.04 (1.4) 6.2 (1.5) Private 6.0 (1.4) 6.0 (1.6) -0.1 (1.2) 6.1 (1.7) None 6.0 (1.5) 5.9 (1.7) -0.03 (1.5) 6.2 (1.9) Education Level 0.662 0.991 0.695 0.721 High school 6.0 (1.4) 6.0 (1.5) 0.01 (1.3) 6.0 (1.7) or less Some 6.1 (1.5) 6.0 (1.7) -0.1 (1.4) 6.4 (1.6) college/vocational Bachelor?s 5.9 (1.6) 6.0 (1.6) 0.1 (1.4) 5.9 (1.8) degree or more Nativity 0.088 0.142 0.877 0.086 U.S.-born 6.0 (1.5) 6.0 (1.6) -0.01 (1.4) 6.0 (1.7) Foreign-born 6.4 (1.3) 6.3 (1.3) -0.04 (1.3) 7.1 (0.9) Future 0.959 0.314 0.452 0.014 179 Pregnancy Desire Within the 6.2 (1.7) 6.6 (1.4) 0.4 (1.2) 6.3ab (1.5) next 2 years In 2 years or 6.0 (1.5) 6.0 (1.6) -0.05 (1.3) 6.7b (1.7) more Unsure & Yes, but not sure 6.0 (1.5) 6.0 (1.6) 0.0 (1.4) 4.8a (1.3) when No (more) 6.0 (1.5) 5.9 (1.7) -0.1 (1.4) 6.5b (1.6) children Hypothetical 0.819 0.587 0.855 0.913 Pregnancy Happy 6.1 (1.4) 6.1 (1.6) 0.03 (1.3) 6.1 (1.6) Unhappy 6.0 (1.5) 5.9 (1.6) -0.1 (1.4) 6.1 (1.7) Unsure 6.0 (1.6) 6.0 (1.6) -0.1 (1.4) 6.4 (1.8) Unintended 0.012 0.225 0.179 0.953 Pregnancy No 5.9 (1.5) 5.9 (1.6) 0.1 (1.4) 6.1 (1.7) Yes 6.3 (1.4) 6.1 (1.6) -0.1 (1.2) 6.2 (1.6) Reproductive Coercion by a 0.156 0.192 0.976 0.770 Partner - Ever None 6.0 (1.5) 6.0 (1.6) -0.01 (1.4) 6.2 (1.7) Any 6.3 (1.5) 6.3 (1.6) -0.02 (1.1) 6.0 (1.6) Reason for 0.996 0.609 0.541 0.522 Clinic Visit Not Family 6.0 (1.5) 6.0 (1.6) -0.1 (1.6) 6.4 (1.7) Planning Family 6.0 (1.5) 6.0 (1.6) 0.02 (1.2) 6.0 (1.7) Planning Clinic Focus 0.561 0.310 0.578 0.727 Primary Care 6.0 (1.5) 5.9 (1.6) -0.1 (1.5) 6.2 (1.6) Women's 6.1 (1.5) 6.1 (1.6) 0.01 (1.3) 6.1 (1.7) Health Clinic Location 0.493 0.870 0.151 0.226 Urban 6.1 (1.5) 6.0 (1.5) -0.1 (1.3) 5.9 (1.7) Rural 5.9 (1.4) 6.1 (1.7) 0.1 (1.4) 6.2 (1.7) Suburban 6.4 (1.4) 5.9 (1.7) -0.5 (1.4) 7.7 (0.6) Wave 0.658 0.095 0.139 0.125 One 6.0 (1.3) 5.9 (1.7) -0.1 (1.3) 5.7 (1.6) Two 6.1 (1.6) 6.1 (1.5) 0.1 (1.4) 6.4 (1.7) 180 Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 181 Table 11. Hormonal birth control attitude scores by study variables Pre-visit Post-visit Change N=466 p-value p-value p-value Mean (SD) Mean (SD) Mean (SD) ANOVA ANOVA ANOVA Total Sample 1.5 (1.0) 1.5 (0.9) -0.05 (0.8) Age 0.872 0.997 0.837 15-19 1.5 (0.9) 1.5 (0.9) -0.01 (0.8) 20-29 1.6 (1.0) 1.5 (0.9) -0.04 (0.8) 30-44 1.6 (1.0) 1.5 (1.0) -0.1 (0.8) Race/Hispanic Ethnicity 0.324 0.258 0.678 White NH 1.6 (1.0) 1.5 (0.9) -0.1 (0.8) Black NH 1.5 (0.9) 1.5 (0.9) -0.03 (0.8) Hispanic 1.5 (1.0) 1.5 (0.9) 0.01 (0.9) Other NH 1.9 (1.1) 1.9 (1.2) 0.1 (0.5) Current Relationship 0.040 0.232 0.292 Status Single 1.6a (1.0) 1.6 (0.9) -0.1 (0.8) Married 1.6ab (1.1) 1.5 (1.0) -0.1 (0.9) Cohabitating 1.4b (0.9) 1.4 (0.9) 0.1 (0.8) Insurance Type 0.847 0.713 0.345 Public 1.5 (1.0) 1.5 (1.0) -0.04 (0.9) Private 1.5 (1.0) 1.6 (0.8) 0.01 (0.8) None 1.6 (0.9) 1.5 (0.9) -0.1 (0.7) Education Level 0.045 0.568 0.187 High school or less 1.7a (1.0) 1.6 (0.9) -0.1 (0.8) Some 1.4b (0.9) 1.5 (0.9) 0.03 (0.9) college/vocational Bachelor?s degree or 1.6ab (1.0) 1.5 (0.9) -0.03 (0.7) more Nativity 0.319 0.480 0.681 U.S.-born 1.5 (1.0) 1.5 (0.9) -0.04 (0.8) Foreign-born 1.7 (1.0) 1.6 (0.9) -0.1 (0.7) Future Pregnancy Desire 0.370 0.891 0.174 Within the next 2 1.3 (0.9) 1.5 (1.0) 0.2 (0.7) years In 2 years or more 1.6 (1.0) 1.5 (0.9) -0.1 (0.8) Unsure & Yes, but 1.5 (1.0) 1.5 (0.8) -0.02 (0.8) not sure when 182 No (more) children 1.6 (1.1) 1.5 (0.9) -0.2 (0.9) Hypothetical Pregnancy 0.929 0.828 0.694 Happy 1.6 (1.0) 1.5 (0.9) -0.1 (0.8) Unhappy 1.6 (1.0) 1.5 (0.9) -0.1 (0.9) Unsure 1.5 (0.9) 1.5 (0.9) 0.01 (0.8) Unintended Pregnancy 0.083 0.277 0.386 No 1.6 (1.0) 1.6 (0.9) -0.1 (0.8) Yes 1.5 (0.9) 1.5 (0.9) -0.01 (0.8) Reproductive Coercion 0.349 0.048 0.293 by a Partner - Ever None 1.6 (1.0) 1.5 (0.9) 0 (0.8) Any 1.5 (1.1) 1.3 (0.9) -0.1 (0.9) Reason for Clinic Visit 0.942 0.919 0.980 Not Family Planning 1.6 (1.0) 1.5 (0.9) -0.04 (0.9) Family Planning 1.6 (1.0) 1.5 (0.9) -0.05 (0.8) Clinic Focus 0.668 0.619 0.289 Primary Care 1.5 (1.0) 1.5 (0.9) 0.01 (0.9) Women's Health 1.6 (1.0) 1.5 (0.9) -0.1 (0.8) Clinic Location 0.524 0.524 0.398 Urban 1.6 (1.0) 1.5 (0.9) -0.03 (0.8) Rural 1.5 (1.0) 1.4 (0.9) -0.05 (0.9) Suburban 1.7 (1.2) 1.5 (1.0) -0.3 (0.9) Wave 0.739 0.524 0.272 One 1.5 (1.0) 1.5 (0.9) 0.00 (0.8) Two 1.6 (0.9) 1.5 (0.9) -0.1 (0.9) Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 183 Table 12. Condom attitude scores by study variables Pre-visit Post-visit Change N=467 Mean p-value Mean p-value p-value Mean (SD) (SD) ANOVA (SD) ANOVA ANOVA Total Sample 2.5 (1.3) 2.3 (1.3) -0.14 (1.0) Age 0.997 0.890 0.800 15-19 2.5 (1.3) 2.5 (1.3) -0.2 (1.2) 20-29 2.5 (1.3) 2.3 (1.3) -0.1 (0.9) 30-44 2.5 (1.5) 2.4 (1.3) -0.2 (1.1) Race/Hispanic 0.006 0.099 0.016? Ethnicity White NH 2.4ab (1.4) 2.4 (1.4) -0.04 (0.1) Black NH 2.7a (1.3) 2.5 (1.3) -0.3 (1.1) Hispanic 2.2b (1.3) 2.1 (1.3) -0.03 (1.1) Other NH 2.6ab (1.4) 1.9 (1.4) -0.7 (1.1) Current <0.001 <0.001 0.238 Relationship Status Single 2.7a (1.2) 2.5a (1.3) -0.2 (1.1) Married 2.0b (1.4) 1.9b (1.3) -0.2 (1.1) Cohabitating 2.3b (1.4) 2.3a (1.3) -0.01 (0.93) Insurance Type 0.818 0.281 0.495 Public 2.5 (1.4) 2.3 (1.3) -0.2 (1.2) Private 2.5 (1.3) 2.5 (1.3) -0.1 (0.9) None 2.4 (1.3) 2.2 (1.3) -0.2 (1.0) Education Level 0.789 0.761 0.705 High school or 2.4 (1.3) 2.4 (1.3) -0.1 (1.1) less Some 2.5 (1.3) 2.4 (1.3) -0.2 (1.1) college/vocational Bachelor?s 2.5 (1.4) 2.3 (1.3) -0.2 (0.9) degree or more Nativity 0.001 0.001 0.786 U.S.-born 2.6 (1.3) 2.4 (1.3) -0.1 (1.0) Foreign-born 2.0 (1.4) 1.9 (1.3) -0.2 (1.3) Future Pregnancy 0.094 0.109 0.695 Desire Within the next 2.1 (1.4) 2.0 (1.3) -0.1 (1.0) 184 2 years In 2 years or 2.6 (1.3) 2.4 (1.3) -0.2 (1.1) more Unsure & Yes, 2.6 (1.3) 2.5 (1.3) -0.1 (0.9) but not sure when No (more) 2.3 (1.5) 2.2 (1.3) -0.1 (1.2) children Hypothetical 0.277 0.098 0.335 Pregnancy Happy 2.4 (1.4) 2.2 (1.4) -0.2 (1.0) Unhappy 2.6 (1.3) 2.5 (1.2) -0.1 (1.0) Unsure 2.4 (1.4) 2.3 (1.3) -0.1 (1.0) Unintended 0.055 0.153 0.503 Pregnancy No 2.6 (1.3) 2.4 (1.3) -0.2 (1.1) Yes 2.3 (1.4) 2.2 (1.3) -0.1 (1.0) Reproductive Coercion by a 0.021 0.006 0.651 Partner - Ever None 2.5 (1.3) 2.4 (1.3) -0.1 (1.0) Any 2.2 (1.4) 2.0 (1.3) -0.2 (1.0) Reason for Clinic 0.591 0.971 0.519 Visit Not Family 2.5 (1.3) 2.3 (1.3) -0.2 (1.0) Planning Family 2.5 (1.4) 2.3 (1.3) -0.1 (1.0) Planning Clinic Focus 0.350 0.612 0.573 Primary Care 2.4 (1.3) 2.3 (1.3) -0.1 (1.1) Women's 2.5 (1.3) 2.4 (1.3) -0.2 (1.0) Health Clinic Location 0.096 0.038? 0.761 Urban 2.6 (1.3) 2.4 (1.3) -0.1 (1.1) Rural 2.4 (1.4) 2.2 (1.4) -0.2 (1.0) Suburban 2.0 (1.2) 1.9 (1.2) -0.1 (1.1) Wave 0.137 0.858 0.032 One 2.4 (1.4) 2.4 (1.3) -0.03 (0.1) Two 2.6 (1.3) 2.3 (1.3) -0.2 (1.1) ?The omnibus F test was significant, but the post-hoc Tukey test did not detect any significant differences (the differences between White NH participants and Other NH as well as between Hispanic and Other NH were marginally significant at p<0.1; no differences were marginally significant for clinic location). Notes. Due to rounding, the change means presented in this table may not match the difference between the pre-visit and post-visit means. 185 When omnibus F-tests found a significant relationship (p<0.05), we conducted post-hoc Tukey tests. Means with different superscripts are significantly different at p<0.05, according to Tukey's post-hoc test. 186 Figure 3. Distribution of contraceptive change by LARC based on pre-visit current and pre-visit composite variables. Contraceptive Change by LARC Current 67.1% 5.9% 18.8% 8.2% 6.1% Composite 66.9% 24.7% 2.3% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained without LARC Changed away from LARC Changed to LARC Remained with LARC 187 Figure 4. Distribution of contraceptive change by effectiveness levels based on pre-visit current and pre-visit composite variables. Contraceptive Change by Effectiveness Level current 15.0% 9.5% 38.2% 18.6% 18.8% composite 15.4% 9.5% 46.4% 4.0% 24.7% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at low effectiveness level or no method Changed to lower effectiveness level or to no method Remained at moderate effectiveness level Changed to higher effectiveness level Remained at high effectiveness level (LARC) 188 Table 13. Proportions of the sample?s contraceptive methods prior to and following the provider visit by effectiveness levels. % % % asymptotic p-value Method Effectiveness Level Pre-visit (current) Post-visit Change McNemar paired test None or Low 31.9% 22.2% -9.7% <0.001 Moderately 43.5% 50.8% 7.3% <0.001 LARC 24.7% 27.0% 2.3% 0.222 Method Effectiveness Level Pre-visit (composite) Post-visit Change McNemar paired test None or Low 18.8% 22.2% 3.4% 0.030 Moderately 50.4% 50.8% 0.4% 0.874 LARC 30.8% 27.0% -3.8% 0.007 189 Table 14. Adjusted odds ratios and 95% confidence intervals for the associations between pre-visit method and study variables and changing to LARC post-visit. Change to LARC - Change to LARC - Outcome of Interest Change to LARC Change to LARC sensitivity sensitivity Pre-visit measure used Current method Current method Composite current-planned Composite current-planned Eligible sample 357 323 328 299 Omitted observations 18 18 33 47 Sample used for Analysis 339 305 280 252 Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 0.08*** (0.03-0.20) 0.05*** (0.02-0.16) 0.05*** (0.01-0.36) 0.04*** (0-0.34) LARC Age 15-19 1.00 1.00 1.00 1.00 20-29 0.59 (0.19-1.83) 0.74 (0.2-2.7) 0.85 (0.09-8.52) 1.5 (0.12-18) 30-44 0.46 (0.09-2.27) 0.87 (0.15-4.9) 0.16 (0.01-2.85) 0.14 (0.01-3.49) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.35* (0.13-0.98) 0.32* (0.1-0.97) 1.27 (0.12-13.46) 0.96 (0.08-11.53) Hispanic 0.99 (0.3-3.3) 1.19 (0.32-4.43) 2.45 (0.19-31.16) 2.38 (0.14-39.9) Other NH 15.4* (1.89-125.53) 10.82* (1.08-107.87) 17.61 (0.2-1571.85) 6.49 (0.07-578.59) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 1.53 (0.4-5.9) 2.36 (0.56-9.88) 2.63 (0.27-25.45) 5.33 (0.47-61.03) Cohabitating 0.28* (0.08-0.95) 0.34 (0.1-1.17) 1.34 (0.2-9.03) 1.01 (0.13-7.65) 190 Insurance Type Public 1.00 1.00 1.00 1.00 Private 0.71 (0.25-1.97) 0.72 (0.23-2.24) 1.15 (0.07-17.89) 2.68 (0.13-55.35) None 0.42 (0.12-1.43) 0.54 (0.14-2.03) 5.51 (0.59-51.55) 7.71 (0.66-90.42) Education Level High school or less 1.00 1.00 1.00 1.00 Some 1.04 (0.36-2.98) 1.39 (0.44-4.34) 2.44 (0.24-24.35) 3.45 (0.23-51.02) college/vocational Bachelor?s degree or 2.31 (0.60-8.84) 2.17 (0.47-10.06) 0.74 (0.03-17.62) 1.04 (0.03-38.85) higher Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 0.45 (0.08-2.38) 0.37 (0.06-2.27) 1.66 (0.16-17.49) 3.04 (0.21-43.44) Future Pregnancy Desire Within the next 2 0.28 (0.03-2.27) 0.37 (0.05-2.97) NE? NE? years In 2 years or more 1.00 1.00 1.00 1.00 Unsure & Yes, but 1.26 (0.47-3.38) 1.02 (0.34-3.07) 5.66 (0.55-58.15) 7.00 (0.57-85.6) not sure when No (more) children 0.34 (0.08-1.55) 0.19 (0.03-1.05) 1.15 (0.02-53.66) 0.87 (0.01-51.19) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.63 (0.53-5.01) 3.52 (0.92-13.44) 4.6 (0.4-53.01) 7.27 (0.53-98.85) Unsure 1.64 (0.5-5.32) 2.13 (0.57-7.92) 2.24 (0.2-25.25) 2.8 (0.23-34.37) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 1.02 (0.39-2.69) 0.77 (0.27-2.16) 2.67 (0.35-20.13) 1.57 (0.19-13) 191 Reproductive Coercion by a Partner - Ever None 1.00 1.00 1.00 1.00 Any 3.00* (1.09-8.24) 3.24* (1.03-10.16) 1.28 (0.13-12.49) 3.30 (0.28-38.96) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 2.78 (0.92-8.36) 2.57 (0.81-8.17) 0.84 (0.11-6.28) 1.20 (0.13-11.32) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 1.48 (0.48-4.57) 1.13 (0.34 -3.73) 0.93 (0.13-6.65) 1.15 (0.15-8.79) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 1.65 (0.67-4.07) 1.65 (0.62-4.41) 0.42 (0.05-3.66) 0.46 (0.04-5.49) Suburban NE? NE? NE? NE? Wave One 1.00 1.00 1.00 1.00 Two 1.81 (0.76-4.34) 1.84 (0.7-4.88) 7.29 (0.85-62.31) 10.94* (1.13-105.49) ?Non-estimable because the observations in the cell predict failure perfectly (100%). *p?0.05. **p?0.01. ***p?0.001. 192 Table 15. Adjusted ratios and 95% confidence intervals for the associations between changing to a method of higher effectiveness level post-visit and predictor variables, including current and composite pre-visit method. Outcome of Interest Change to a method of higher effectiveness level Type of analysis Main analysis Sensitivity analysis Main analysis Sensitivity analysis Composite current- Composite current- Pre-visit measure used Current method Current method planned planned Eligible sample 340 308 312 284 Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 0.01*** (0.003-0.03) 0.01*** (0-0.02) 0.02*** (0.004-0.10) 0.02*** (0.003-0.09) LARC Age 15-19 1.00 1.00 1.00 1.00 20-29 0.51 (0.17-1.55) 0.52 (0.15-1.78) 0.56 (0.11-2.92) 0.62 (0.11-3.42) 30-44 0.31 (0.08-1.22) 0.40 (0.09-1.8) 0.32 (0.04-2.45) 0.31 (0.04-2.71) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.44 (0.18-1.08) 0.36* (0.14-0.95) 1.1 (0.24-4.99) 0.72 (0.15-3.46) Hispanic 0.68 (0.22-2.14) 0.57 (0.16-2.02) 0.59 (0.09-3.88) 0.42 (0.05-3.3) Other NH 6.78 (0.85-53.91) 2.97 (0.28-31.36) 4.34 (0.16-116.69) 2.09 (0.07-65.15) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 1.15 (0.37-3.62) 1.89 (0.56-6.39) 2.59 (0.44-15.13) 4.06 (0.68-24.05) Cohabitating 0.68 (0.27-1.74) 0.81 (0.3-2.17) 2.14 (0.51-9.04) 1.89 (0.43-8.27) Insurance Type 193 Public 1.00 1.00 1.00 1.00 Private 0.77 (0.28-2.06) 1.01 (0.34-2.94) 0.96 (0.18-5.19) 1.27 (0.22-7.29) None 0.82 (0.3-2.23) 1.09 (0.35-3.34) 3.52 (0.76-16.24) 3.92 (0.79-19.6) Education Level High school or less 1.00 1.00 1.00 1.00 Some college/vocational 0.38 (0.14-1.01) 0.49 (0.17-1.41) 0.54 (0.12-2.39) 0.59 (0.13-2.68) Bachelor?s degree or 0.83 (0.27-2.57) 0.71 (0.2-2.5) 0.46 (0.07-2.95) 0.54 (0.08-3.66) higher Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 0.82 (0.22-3.06) 0.83 (0.2-3.48) 1.4 (0.21-9.41) 1.69 (0.22-13.03) Future Pregnancy Desire Within the next 2 years 1.00 1.00 1.00 1.00 In 2 years or more 5.4* (1.18-24.68) 4.91* (1.03-23.47) 9.11 (0.71-117.31) 7.67 (0.56-104.71) Unsure & Yes, but not 3.8 (0.79-18.28) 2.11 (0.41-10.72) 8.6 (0.66-111.99) 7.21 (0.5-103.93) sure when No (more) children 2.59 (0.38-17.68) 1.17 (0.16-8.86) 1.25 (0.04-41.61) 0.53 (0.01-21.79) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.74 (0.63-4.8) 3.05 (0.95-9.79) 2.27 (0.44-11.76) 4.03 (0.65-25.01) Unsure 1.19 (0.44-3.21) 1.29 (0.44-3.73) 0.93 (0.18-4.86) 1.09 (0.2-5.99) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 1.25 (0.55-2.84) 1.11 (0.46-2.67) 2.15 (0.59-7.83) 1.69 (0.45-6.32) Reproductive Coercion by a Partner - Ever None 1.00 1.00 1.00 1.00 Any 2.52 (0.95-6.7) 2.8 (0.95-8.3) 0.94 (0.17-5.13) 1.58 (0.26-9.56) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 4.74*** (1.95-11.48) 5.15*** (1.96-13.52) 2.42 (0.57-10.25) 2.88 (0.64-12.89) 194 Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 2.67 (0.98-7.26) 2.11 (0.73-6.15) 0.79 (0.17-3.64) 0.76 (0.15-3.78) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 0.71 (0.31-1.65) 0.66 (0.26-1.64) 0.75 (0.19-2.98) 0.75 (0.16-3.45) Suburban 0.59 (0.1-3.61) 0.49 (0.08-3.12) 0.20 (0.01-3.44) 0.13 (0.01-2.7) Wave One 1.00 1.00 1.00 1.00 Two 0.90 (0.42-1.92) 0.84 (0.37-1.93) 1.26 (0.36-4.46) 1.82 (0.49-6.78) *p?0.05. **p?0.01. ***p?0.001. 195 Table 16. Adjusted ratios and 95% confidence intervals for the associations between changing to a method of lower effectiveness level post-visit and predictor variables, including current and composite pre-visit method. Outcome of Interest Change to a method of lower effectiveness level Type of analysis Main analysis Sensitivity analysis Main analysis Sensitivity analysis Pre-visit measure used Current method Current method Composite current-planned Composite current-planned Eligible sample 315 286 382 346 Pre-Visit Method None or Low Moderately 1.00 1.00 1.00 1.00 LARC 3.06** (1.42-6.59) 3.12** (1.38-7.05) 3.72*** (1.72-8.04) 4.00*** (1.78-8.96) Age 15-19 1.00 1.00 1.00 1.00 20-29 2.05 (0.59-7.06) 1.92 (0.49-7.53) 0.80 (0.28-2.32) 0.72 (0.24-2.17) 30-44 3.84 (0.93-15.89) 3.19 (0.68-15.01) 1.30 (0.37-4.62) 1.10 (0.29-4.12) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 2.06 (0.85-4.99) 2.12 (0.8-5.64) 2.67* (1.10-6.49) 2.72* (1.06-6.98) Hispanic 1.60 (0.41-6.25) 1.68 (0.39-7.27) 1.99 (0.57-6.97) 2.12 (0.56-8.01) Other NH 2.28 (0.31-16.79) 2.65 (0.33-21.17) 2.23 (0.3-16.58) 3.03 (0.38-24.24) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 1.37 (0.47-3.94) 1.41 (0.47-4.25) 1.33 (0.46-3.87) 1.21 (0.4-3.61) Cohabitating 1.38 (0.55-3.43) 1.71 (0.64-4.56) 1.35 (0.54-3.39) 1.42 (0.55-3.68) Insurance Type Public 1.00 1.00 1.00 1.00 196 Private 3.86** (1.49-10.04) 4.07** (1.49-11.12) 4.49** (1.72-11.72) 4.47** (1.67-11.92) None 1.7 (0.61-4.76) 2 (0.69-5.76) 2.58 (0.92-7.26) 3.04* (1.06-8.72) Education Level High school or less 1.00 1.00 1.00 1.00 Some college/vocational 0.63 (0.25-1.59) 0.62 (0.24-1.6) 0.94 (0.38-2.32) 0.89 (0.36-2.22) Bachelor?s degree or 0.48 (0.15-1.54) 0.45 (0.13-1.53) 0.67 (0.21-2.14) 0.59 (0.17-2.01) higher Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 1.06 (0.29-3.86) 1.06 (0.26-4.25) 1.33 (0.39-4.57) 1.4 (0.38-5.16) Future Pregnancy Desire Within the next 2 years 1.00 1.00 1.00 1.00 In 2 years or more 0.22* (0.05-0.92) 0.22* (0.05-0.95) 1.43 (0.27-7.69) 1.49 (0.27-8.28) Unsure & Yes, but not 0.57 (0.14-2.26) 0.68 (0.17-2.8) 2.75 (0.51-14.77) 2.99 (0.54-16.59) sure when No (more) children 0.18* (0.04-0.89) 0.19* (0.04-0.99) 0.6 (0.09-3.87) 0.58 (0.08-3.92) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.39 (0.47-4.1) 1.52 (0.48-4.81) 0.79 (0.28-2.21) 0.84 (0.29-2.43) Unsure 1.57 (0.59-4.15) 1.4 (0.5-3.9) 1.12 (0.46-2.73) 1.01 (0.4-2.58) Unintended Pregnancy - Ever No 1.00 1.00 1.00 1.00 Yes 0.59 (0.26-1.34) 0.55 (0.24-1.26) 1.13 (0.52-2.48) 1.03 (0.46-2.28) Reproductive Coercion by a Partner - Ever None 1.00 1.00 1.00 1.00 Any 0.44 (0.1-1.9) 0.5 (0.11-2.18) 0.39 (0.11-1.33) 0.42 (0.12-1.47) Reason for Clinic Visit 197 Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 0.57 (0.26-1.27) 0.69 (0.29-1.63) 0.41* (0.19-0.87) 0.45* (0.21-0.97) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 1.22 ( 0.48-3.16) 1.09 (0.41-2.89) 1.01 (0.40-2.59) 1.02 (0.39-2.70) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 1.97 (0.89-4.33) 2.31 (0.99-5.41) 2.16 (0.99-4.7) 2.26 (1-5.14) Suburban 1.88 (0.28-12.57) 1.67 (0.24-11.38) 2.61 (0.41-16.55) 2.51 (0.39-16.13) Wave One 1.00 1.00 1.00 1.00 Two 1.03 (0.48-2.23) 1.14 (0.5-2.58) 0.74 (0.35-1.54) 0.90 (0.41-1.96) *p?0.05. **p?0.01. ***p?0.001. 198 Figure 5. Distribution of change in IUD knowledge from pre-visit to post-visit. Change in IUD Knowledge 6.2% 21.3% 33.9% 31.1% 7.5% 0% 20% 40% 60% 80% 100% Remained at lowest score (0) Decreased in score Remained at same middle score (1-5) Increased in score Remained at highest score (6) 199 Figure 6. Distribution of change in implant knowledge from pre-visit to post-visit. Change in Implant Knowledge 10.9% 19.4% 27.7% 24.9% 17.1% 0% 20% 40% 60% 80% 100% Remained at lowest score (0) Decreased in score Remained at same middle score (1-3) Increased in score Remained at highest score (4) 200 Figure 7. Distribution of change in effectiveness knowledge from pre-visit to post- visit. Change in Effectiveness Knowledge 32.4% 8.1% 21.1% 17.3% 21.1% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at 0 Decreased in score Remained at same middle score (1-3) Increased in score Remained at All Correct 201 Figure 8. Distribution of change in knowledge about DelCAN benefits from pre- visit to post-visit. Change in Knowledge of DelCAN Benefits 4.0% 15.1% 39.3% 34.6% 7.0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at lowest score (0) Decreased in score Remained at same middle score (1-6) Increased in score Remained at maximum score (7) 202 Table 17. Adjusted relationships between change in IUD knowledge and study variables, using a continuous change outcome (percentage point increase in knowledge) and a dichotomous change outcome (increase in knowledge or not) Increase in IUD knowledge Increase in IUD Knowledge % (Yes vs. No) N=418 B SE p-value 95% CI OR p-value 95% CI Pre-visit IUD Knowledge Score -0.08 0.02 <0.001 -0.11 -0.05 0.69 <0.001 0.60 0.79 (cont.) Age 15-19 [ref] 1.00 20-29 0.05 0.08 0.525 -0.11 0.21 1.19 0.617 0.60 2.38 30-44 -0.03 0.10 0.763 -0.22 0.16 0.82 0.643 0.36 1.87 Race/Hispanic Ethnicity White, NH [ref] 1.00 Black, NH 0.00 0.07 0.952 -0.12 0.13 0.69 0.190 0.40 1.20 Hispanic 0.11 0.09 0.201 -0.06 0.28 1.56 0.223 0.76 3.16 Other, NH -0.04 0.17 0.793 -0.38 0.29 0.66 0.581 0.15 2.88 Current Relationship Status Single [ref] 1.00 Married -0.01 0.08 0.864 -0.17 0.14 0.75 0.395 0.39 1.45 Cohabitating 0.12 0.07 0.073 -0.01 0.25 1.26 0.401 0.74 2.16 Insurance Type Public [ref] 1.00 Private -0.04 0.07 0.537 -0.17 0.09 0.78 0.386 0.45 1.36 None -0.07 0.07 0.329 -0.20 0.07 0.80 0.450 0.45 1.43 Education Level High school [ref] 1.00 or less Some 0.11 0.07 0.103 -0.02 0.24 1.50 0.162 0.85 2.64 college/vocational Bachelor?s 0.27 0.09 0.002 0.10 0.43 2.14 0.039 1.04 4.42 degree or higher 203 Nativity U.S.-born [ref] 1.00 Foreign-born -0.09 0.09 0.334 -0.28 0.09 0.78 0.540 0.35 1.74 Future Pregnancy Desire Within the [ref] 1.00 next 2 years In 2 years or -0.06 0.10 0.586 -0.26 0.15 1.06 0.892 0.44 2.54 more Unsure & Yes, but not sure -0.07 0.11 0.543 -0.28 0.15 1.25 0.630 0.51 3.07 when No (more) -0.04 0.12 0.748 -0.28 0.20 1.07 0.900 0.38 2.97 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy -0.03 0.07 0.721 -0.17 0.12 1.34 0.343 0.73 2.43 Unsure -0.04 0.07 0.523 -0.18 0.09 0.95 0.859 0.53 1.70 Unintended Pregnancy No [ref] 1.00 Yes 0.01 0.06 0.860 -0.10 0.12 1.26 0.344 0.78 2.04 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any 0.07 0.08 0.357 -0.08 0.22 1.09 0.792 0.59 2.00 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family 0.02 0.06 0.704 -0.09 0.14 0.70 0.145 0.43 1.13 Planning Clinic Focus Primary Care [ref] 1.00 Women's -0.07 0.07 0.287 -0.20 0.06 1.63 0.094 0.92 2.90 Health Clinic Location Urban [ref] 1.00 Rural 0.05 0.06 0.381 -0.06 0.17 1.06 0.815 0.65 1.74 Suburban 0.12 0.13 0.349 -0.13 0.38 2.15 0.151 0.76 6.12 Wave One [ref] 1.00 204 Two 0.03 0.05 0.622 -0.08 0.13 1.59 0.048 1.00 2.52 Randomized Group (N=407) ? sensitivity model A [ref] 1.00 B -0.05 0.05 0.318 -0.16 0.05 0.92 0.709 0.58 1.44 OR = odds ratios. 205 Table 18. Adjusted relationships between change in implant knowledge and study variables, using a continuous change outcome (percentage point increase in knowledge) and a dichotomous change outcome (increase in knowledge or not) Increase in implant knowledge Increase in implant knowledge % (Yes vs. No) N=345 B SE p-value 95% CI OR p-value 95% CI Pre-visit Implant -0.14 0.02 <0.001 -0.19 -0.09 0.61 <0.001 0.49 0.76 Knowledge Score (cont.) Age 15-19 [ref] 1.00 20-29 -0.05 0.09 0.578 -0.24 0.13 1.33 0.491 0.59 3.03 30-44 0.06 0.11 0.586 -0.16 0.28 1.42 0.476 0.54 3.70 Race/Hispanic Ethnicity White NH [ref] 1.00 Black NH -0.14 0.07 0.056 -0.28 0.00 0.66 0.189 0.36 1.22 Hispanic 0.03 0.10 0.724 -0.16 0.23 0.95 0.909 0.42 2.16 Other NH -0.08 0.19 0.684 -0.44 0.29 1.37 0.695 0.28 6.62 Current Relationship Status Single [ref] 1.00 Married -0.05 0.08 0.526 -0.22 0.11 0.76 0.465 0.36 1.60 Cohabitating -0.02 0.07 0.826 -0.16 0.13 1.28 0.426 0.70 2.34 Insurance Type Public [ref] 1.00 Private -0.06 0.07 0.418 -0.21 0.09 0.69 0.253 0.37 1.30 None -0.02 0.08 0.812 -0.17 0.13 0.58 0.105 0.30 1.12 Education Level High school or less [ref] 1.00 Some 0.10 0.07 0.183 -0.05 0.24 1.17 0.625 0.62 2.20 college/vocational Bachelor?s degree or 0.01 0.09 0.886 -0.17 0.20 0.97 0.945 0.43 2.19 higher Nativity U.S.-born [ref] 1.00 206 Foreign-born -0.18 0.10 0.082 -0.38 0.02 0.49 0.146 0.19 1.28 Future Pregnancy Desire Within the next 2 [ref] 1.00 years In 2 years or more 0.03 0.11 0.779 -0.19 0.25 0.98 0.975 0.37 2.59 Unsure & Yes, but -0.13 0.12 0.269 -0.36 0.10 0.67 0.436 0.24 1.83 not sure when No (more) children -0.01 0.13 0.965 -0.27 0.26 0.91 0.877 0.29 2.84 Hypothetical Pregnancy Happy [ref] 1.00 Unhappy 0.00 0.08 0.971 -0.16 0.16 1.90 0.074 0.94 3.84 Unsure -0.12 0.08 0.125 -0.27 0.03 1.45 0.283 0.74 2.85 Unintended Pregnancy No [ref] 1.00 Yes -0.04 0.06 0.503 -0.17 0.08 0.75 0.309 0.44 1.30 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any 0.03 0.08 0.694 -0.13 0.20 1.20 0.612 0.60 2.40 Reason for Clinic Visit Not Family Planning [ref] 1.00 Family Planning -0.02 0.06 0.744 -0.15 0.11 0.93 0.790 0.53 1.61 Clinic Focus Primary Care [ref] 1.00 Women's Health -0.05 0.07 0.517 -0.19 0.10 0.83 0.568 0.45 1.56 Clinic Location Urban [ref] 1.00 Rural 0.04 0.06 0.556 -0.09 0.17 1.91 0.021 1.10 3.31 Suburban 0.19 0.14 0.189 -0.09 0.47 3.16 0.056 0.97 10.32 Wave One [ref] 1.00 Two 0.15 0.06 0.017 0.03 0.27 1.59 0.081 0.94 2.69 Randomized Group (N=334) ? sensitivity model A [ref] 1.00 B 0.03 0.06 0.603 -0.09 0.15 1.18 0.529 0.71 1.96 207 Table 19. Adjusted relationships between change in effectiveness knowledge and study variables, using two dichotomous change outcomes?any increase in score and increase to perfect score Any increase in Score Increase to Total Correct N=351 OR p-value 95% CI OR p-value 95% CI Pre-visit Effectiveness Knowledge Zero 1.00 1.00 One to Three 0.98 0.933 0.57 1.66 2.37 0.018 1.16 4.84 Age 15-19 1.00 1.00 20-29 0.86 0.704 0.39 1.89 0.54 0.218 0.20 1.44 30-44 1.59 0.336 0.62 4.11 0.67 0.521 0.19 2.31 Race/Hispanic Ethnicity White NH 1.00 1.00 Black NH 0.56 0.064 0.30 1.04 0.25 0.002 0.10 0.60 Hispanic 0.44 0.056 0.19 1.02 0.53 0.238 0.19 1.52 Other NH 1.29 0.744 0.28 5.80 1.94 0.446 0.35 10.76 Current Relationship Status Single 1.00 1.00 Married 0.68 0.350 0.31 1.52 0.59 0.388 0.18 1.95 Cohabitating 1.13 0.691 0.61 2.11 1.23 0.615 0.54 2.80 Insurance Type Public 1.00 1.00 Private 0.80 0.518 0.42 1.55 0.94 0.885 0.38 2.30 None 0.79 0.478 0.40 1.53 0.65 0.359 0.26 1.63 Education Level High school or 1.00 1.00 less Some 0.68 0.250 0.35 1.31 0.95 0.910 0.39 2.31 college/vocational Bachelor?s degree 0.76 0.499 0.33 1.70 0.78 0.670 0.26 2.39 or higher Nativity 208 U.S.-born 1.00 1.00 Foreign-born 1.31 0.562 0.53 3.23 1.04 0.956 0.28 3.79 Future Pregnancy Desire Within the next 2 1.00 1.00 years In 2 years or more 1.35 0.550 0.50 3.61 0.95 0.940 0.27 3.41 Unsure & Yes, 0.89 0.828 0.32 2.48 0.89 0.859 0.24 3.34 but not sure when No (more) 0.86 0.804 0.27 2.78 0.49 0.381 0.10 2.41 children Hypothetical Pregnancy Happy 1.00 1.00 Unhappy 1.34 0.426 0.65 2.74 1.05 0.922 0.40 2.76 Unsure 1.58 0.182 0.81 3.09 1.44 0.419 0.59 3.50 Unintended Pregnancy No 1.00 1.00 Yes 0.69 0.214 0.38 1.24 1.34 0.463 0.61 2.93 Reproductive Coercion by a Partner - Ever None 1.00 1.00 Any 2.17 0.032 1.07 4.39 1.89 0.195 0.72 4.97 Reason for Clinic Visit Not Family 1.00 1.00 Planning Family Planning 0.84 0.557 0.48 1.49 1.10 0.815 0.49 2.47 Clinic Focus Primary Care 1.00 1.00 Women's Health 1.15 0.673 0.60 2.22 2.00 0.175 0.74 5.41 Clinic Location Urban 1.00 1.00 Rural 1.75 0.064 0.97 3.15 1.67 0.213 0.75 3.71 Suburban 2.21 0.213 0.63 7.71 3.33 0.222 0.48 22.91 Wave One 1.00 1.00 Two 0.56 0.036 0.33 0.96 0.40 0.015 0.19 0.84 209 Table 20. Adjusted relationships between change in knowledge about DelCAN benefits and study variables, using continuous and dichotomous outcomes Increase in DelCAN knowledge Increase in DelCAN % knowledge (Yes vs. No) N=426 B SE p-value 95% CI OR p-value 95% CI Pre-visit Knowledge Score -0.15 0.02 <0.001 -0.20 -0.10 0.85 <0.001 0.76 0.94 (cont.) Age 15-19 [ref] 1.00 20-29 -0.08 0.15 0.601 -0.37 0.22 0.58 0.095 0.31 1.10 30-44 -0.04 0.18 0.815 -0.40 0.31 0.45 0.039 0.21 0.96 Race/Hispanic Ethnicity White, NH [ref] 1.00 Black, NH 0.14 0.12 0.241 -0.09 0.37 1.60 0.065 0.97 2.64 Hispanic 0.14 0.16 0.393 -0.18 0.45 1.10 0.793 0.55 2.18 Other, NH -0.64 0.31 0.037 -1.24 -0.04 0.61 0.477 0.16 2.37 Current Relationship Status Single [ref] 1.00 Married -0.15 0.15 0.313 -0.43 0.14 0.93 0.819 0.50 1.74 Cohabitating -0.23 0.12 0.061 -0.47 0.01 1.21 0.474 0.72 2.02 Insurance Type Public [ref] 1.00 Private -0.11 0.12 0.392 -0.35 0.14 0.93 0.784 0.55 1.57 None -0.09 0.13 0.474 -0.35 0.16 0.85 0.574 0.49 1.49 Education Level High school or [ref] 1.00 less Some 0.16 0.12 0.207 -0.09 0.40 1.55 0.104 0.91 2.63 college/vocational Bachelor?s 0.10 0.16 0.516 -0.20 0.41 1.46 0.268 0.75 2.83 degree or higher Nativity U.S.-born [ref] 1.00 Foreign-born 0.05 0.17 0.792 -0.29 0.38 1.12 0.770 0.54 2.32 210 Future Pregnancy Desire Within the [ref] 1.00 next 2 years In 2 years or -0.10 0.20 0.612 -0.49 0.29 1.59 0.308 0.65 3.91 more Unsure & Yes, but not sure -0.24 0.20 0.230 -0.64 0.15 1.72 0.252 0.68 4.35 when No (more) -0.35 0.23 0.128 -0.80 0.10 1.97 0.198 0.70 5.51 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy -0.09 0.13 0.515 -0.35 0.18 1.02 0.958 0.58 1.79 Unsure 0.03 0.13 0.827 -0.23 0.28 1.20 0.515 0.69 2.08 Unintended Pregnancy No [ref] 1.00 Yes 0.02 0.11 0.832 -0.19 0.24 1.03 0.888 0.65 1.64 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any -0.09 0.14 0.498 -0.36 0.18 1.02 0.957 0.57 1.81 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family 0.09 0.11 0.432 -0.13 0.30 1.30 0.277 0.81 2.08 Planning Clinic Focus Primary Care [ref] 1.00 Women's 0.24 0.13 0.059 -0.01 0.49 1.29 0.356 0.75 2.23 Health Clinic Location Urban [ref] 1.00 Rural -0.16 0.11 0.153 -0.37 0.06 1.07 0.775 0.67 1.70 Suburban 0.01 0.26 0.961 -0.50 0.53 0.88 0.837 0.27 2.87 Wave One [ref] 1.00 Two -0.13 0.10 0.194 -0.33 0.07 0.86 0.497 0.56 1.32 211 Figure 9. Change in IUD attitude scores from pre-visit to post-visit. Change in IUD attitudes 1.0% 39.8% 13.0% 28.6% 17.5% 0% 20% 40% 60% 80% 100% Remained at the same negative leaning score <8 Decreased in positive attitude Remained at a neutral score (8) Increased in positive attitude Remained at the same positive leaning score >8 212 Figure 10. Change in implant attitude scores from pre-visit to post-visit. Change in Implant Attitudes 0.3% 11.6% 25.3% 35.8% 27.0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at the same negative leaning score <4 Decreased in positive attitude Remained at a neutral score (4) Increased in positive attitude Remained at the same positive leaning score >4 213 Figure 11. Change in hormonal birth control attitude scores from pre-visit to post- visit. Change in Attitudes about Side Effects in Hormonal Birth Control Methods 28.3% 21.2% 27.9% 17.4% 5.2% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at the same negative leaning score <2 Decreased in positive attitude Remained at a neutral score (2) Increased in positive attitude Remained at the same positive leaning score >2 214 Figure 12. Change in condom attitude scores from pre-visit to post-visit. Change in Condom Attitudes 16.1% 24.4% 9.6% 16.3% 33.6% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Remained at the same negative leaning score <2 Decreased in positive attitude Remained at a neutral score (2) Increased in positive attitude Remained at the same positive leaning score >2 215 Table 21. Adjusted relationships between change in IUD attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) Any increase in positive IUD Increase in positive IUD attitudes attitudes Yes vs. No N=399 N=396 B SE p-value 95% CI OR p-value 95% CI Pre-visit Attitude -0.30 0.04 <0.001 -0.37 -0.22 0.82 0.001 0.72 0.92 Score (continuous) Age 15-19 [ref] 1.00 20-29 0.38 0.26 0.151 -0.14 0.89 0.80 0.546 0.39 1.66 30-44 0.36 0.32 0.251 -0.26 0.99 0.67 0.369 0.28 1.60 Race/Hispanic Ethnicity White NH [ref] 1.00 Black NH -0.10 0.20 0.630 -0.49 0.30 0.96 0.884 0.54 1.69 Hispanic 0.09 0.27 0.753 -0.45 0.62 1.42 0.362 0.67 2.99 Other NH 0.39 0.50 0.436 -0.59 1.37 1.61 0.467 0.44 5.85 Current Relationship Status Single [ref] 1.00 Married -0.07 0.25 0.780 -0.56 0.42 1.10 0.791 0.55 2.18 Cohabitating -0.05 0.20 0.788 -0.46 0.35 0.96 0.898 0.55 1.70 Insurance Type Public [ref] 1.00 Private -0.11 0.21 0.593 -0.53 0.31 0.85 0.591 0.46 1.55 None 0.12 0.23 0.583 -0.32 0.57 1.07 0.829 0.58 1.99 Education Level High school or [ref] 1.00 less Some 0.22 0.21 0.310 -0.20 0.64 1.51 0.184 0.82 2.77 college/vocational Bachelor?s 0.50 0.27 0.067 -0.03 1.03 2.32 0.027 1.10 4.88 degree or higher Nativity U.S.-born [ref] 1.00 Foreign-born 0.06 0.31 0.842 -0.55 0.67 1.53 0.312 0.67 3.47 216 Future Pregnancy Desire Within the next [ref] 1.00 2 years In 2 years or -0.52 0.33 0.118 -1.16 0.13 0.66 0.363 0.27 1.62 more Unsure & Yes, -0.22 0.35 0.527 -0.90 0.46 1.19 0.718 0.47 3.00 but not sure when No (more) -0.35 0.40 0.372 -1.13 0.42 1.47 0.483 0.50 4.30 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy 0.19 0.24 0.428 -0.28 0.65 0.72 0.336 0.37 1.40 Unsure 0.10 0.22 0.659 -0.34 0.53 0.78 0.430 0.42 1.45 Unintended Pregnancy No [ref] 1.00 Yes 0.12 0.19 0.504 -0.24 0.49 0.96 0.870 0.57 1.61 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any -0.24 0.24 0.320 -0.71 0.23 0.81 0.527 0.41 1.57 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family Planning 0.13 0.19 0.485 -0.24 0.50 1.01 0.961 0.60 1.71 Clinic Focus Primary Care [ref] 1.00 Women's Health -0.01 0.22 0.958 -0.45 0.42 1.30 0.408 0.69 2.45 Clinic Location Urban [ref] 1.00 Rural 0.07 0.18 0.688 -0.29 0.44 0.98 0.934 0.59 1.63 Suburban -0.19 0.46 0.678 -1.09 0.71 0.96 0.947 0.26 3.53 Wave One [ref] 1.00 Two 0.23 0.17 0.181 -0.11 0.57 1.26 0.345 0.78 2.04 217 Table 22. Adjusted relationships between change in implant attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) Any increase in positive implant Increase in positive implant attitudes attitudes Yes vs. No. N=363 N=281 B SE p-value 95% CI OR p-value 95% CI Pre-visit Attitude -0.34 0.05 <0.001 -0.43 -0.25 0.75 0.013 0.59 0.94 Score (continuous) Age 15-19 [ref] 1.00 20-29 -0.05 0.20 0.790 -0.45 0.34 0.72 0.411 0.32 1.59 30-44 -0.01 0.25 0.958 -0.51 0.48 0.71 0.499 0.27 1.90 Race/Hispanic Ethnicity White NH [ref] 1.00 Black NH -0.11 0.16 0.509 -0.42 0.21 0.83 0.563 0.44 1.57 Hispanic -0.18 0.22 0.419 -0.60 0.25 1.05 0.915 0.45 2.46 Other NH 0.47 0.46 0.305 -0.43 1.38 1.33 0.730 0.26 6.82 Current Relationship Status Single [ref] 1.00 Married -0.02 0.22 0.919 -0.45 0.40 0.61 0.299 0.24 1.56 Cohabitating 0.05 0.17 0.769 -0.28 0.38 1.46 0.250 0.77 2.78 Insurance Type Public [ref] 1.00 Private -0.20 0.17 0.237 -0.54 0.13 0.89 0.752 0.45 1.79 None -0.13 0.18 0.456 -0.49 0.22 0.90 0.783 0.44 1.86 Education Level High school or [ref] 1.00 less Some -0.03 0.17 0.849 -0.37 0.31 1.21 0.587 0.61 2.41 college/vocational Bachelor?s 0.08 0.22 0.731 -0.36 0.52 1.79 0.181 0.76 4.22 degree or higher Nativity U.S.-born [ref] 1.00 218 Foreign-born 0.17 0.25 0.508 -0.33 0.66 1.89 0.225 0.68 5.26 Future Pregnancy Desire Within the [ref] 1.00 next 2 years In 2 years or -0.38 0.28 0.183 -0.94 0.18 0.43 0.134 0.14 1.30 more Unsure & Yes, but not sure -0.28 0.30 0.341 -0.87 0.30 0.67 0.492 0.21 2.11 when No (more) -0.33 0.34 0.334 -0.99 0.34 0.65 0.539 0.17 2.54 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy -0.08 0.19 0.683 -0.44 0.29 0.83 0.625 0.38 1.78 Unsure 0.01 0.18 0.939 -0.34 0.37 1.23 0.578 0.59 2.54 Unintended Pregnancy No [ref] 1.00 Yes -0.03 0.15 0.841 -0.33 0.27 0.87 0.654 0.47 1.60 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any 0.11 0.20 0.591 -0.28 0.50 1.56 0.272 0.71 3.44 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family 0.02 0.15 0.884 -0.27 0.32 0.60 0.097 0.32 1.10 Planning Clinic Focus Primary Care [ref] 1.00 Women's 0.10 0.18 0.568 -0.25 0.45 1.39 0.361 0.69 2.79 Health Clinic Location Urban [ref] 1.00 Rural 0.12 0.15 0.422 -0.18 0.42 1.90 0.032 1.06 3.40 Suburban -0.36 0.37 0.330 -1.09 0.37 0.73 0.732 0.12 4.49 Wave One [ref] 1.00 Two 0.29 0.14 0.038 0.02 0.57 1.44 0.203 0.82 2.54 219 Table 23. Adjusted relationships between change in hormonal birth control attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) Any increase in positive attitudes Increase in positive attitudes about hormonal about hormonal birth control birth control Yes vs. No. N=466 N=448 B SE p-value 95% CI OR p-value 95% CI Pre-visit Attitude -0.41 0.03 <0.001 -0.48 -0.35 0.34 <0.001 0.24 0.48 Score (continuous) Age 15-19 [ref] 1.00 20-29 -0.05 0.10 0.636 -0.25 0.15 0.92 0.832 0.41 2.05 30-44 -0.02 0.12 0.857 -0.27 0.22 1.00 0.995 0.36 2.79 Race/Hispanic Ethnicity White NH [ref] 1.00 Black NH 0.01 0.08 0.891 -0.15 0.17 1.02 0.944 0.53 1.97 Hispanic 0.12 0.11 0.281 -0.10 0.33 1.55 0.309 0.67 3.61 Other NH 0.24 0.21 0.266 -0.18 0.65 0.97 0.974 0.16 5.99 Current Relationship Status Single [ref] 1.00 Married -0.09 0.10 0.385 -0.28 0.11 0.85 0.716 0.36 2.02 Cohabitating -0.02 0.08 0.858 -0.18 0.15 1.14 0.688 0.59 2.20 Insurance Type Public [ref] 1.00 Private 0.08 0.09 0.375 -0.09 0.24 1.32 0.407 0.68 2.55 None -0.06 0.09 0.534 -0.23 0.12 0.75 0.460 0.35 1.61 Education Level High school or [ref] 1.00 less Some 0.10 0.09 0.265 -0.07 0.26 1.35 0.405 0.67 2.71 college/vocational Bachelor?s 0.08 0.11 0.450 -0.13 0.30 1.31 0.569 0.52 3.29 degree or higher Nativity U.S.-born [ref] 1.00 Foreign-born -0.01 0.12 0.952 -0.24 0.23 0.71 0.517 0.26 1.97 220 Future Pregnancy Desire Within the [ref] 1.00 next 2 years In 2 years or -0.21 0.13 0.124 -0.47 0.06 0.63 0.372 0.23 1.73 more Unsure & Yes, but not sure -0.20 0.14 0.138 -0.47 0.07 0.66 0.431 0.24 1.85 when No (more) -0.34 0.16 0.029 -0.65 -0.04 0.26 0.042 0.07 0.95 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy 0.11 0.09 0.256 -0.08 0.29 1.46 0.340 0.67 3.20 Unsure 0.11 0.09 0.217 -0.06 0.28 1.71 0.145 0.83 3.50 Unintended Pregnancy No [ref] 1.00 Yes 0.04 0.07 0.611 -0.11 0.18 0.99 0.983 0.55 1.80 Reproductive Coercion by a Partner - Ever None [ref] 1.00 Any -0.20 0.09 0.037 -0.39 -0.01 1.08 0.842 0.52 2.24 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family 0.04 0.07 0.637 -0.11 0.18 1.42 0.267 0.76 2.63 Planning Clinic Focus Primary Care [ref] 1.00 Women's -0.16 0.09 0.063 -0.33 0.01 0.56 0.094 0.29 1.10 Health Clinic Location Urban [ref] 1.00 Rural -0.07 0.08 0.384 -0.21 0.08 1.06 0.847 0.58 1.96 Suburban -0.23 0.17 0.182 -0.57 0.11 0.48 0.388 0.09 2.54 Wave One [ref] 1.00 Two -0.09 0.07 0.213 -0.22 0.05 0.73 0.270 0.42 1.28 221 Table 24. Adjusted relationships between change in condom attitudes and study variables, using a continuous outcome (linear regression) and a dichotomous outcome (logistic regression) Any increase in positive condom Increase in positive condom attitudes attitudes Yes vs. No. N=467 N=326 B SE p-value 95% CI OR p-value 95% CI Pre-visit Attitude -0.34 0.03 <0.001 -0.40 -0.27 0.63 <0.001 0.48 0.83 Score (continuous) Age 15-19 [ref] 1.00 20-29 0.18 0.14 0.177 -0.08 0.45 1.54 0.320 0.66 3.62 30-44 0.19 0.16 0.251 -0.13 0.51 2.16 0.150 0.76 6.18 Race/Hispanic Ethnicity White NH [ref] 1.00 Black NH -0.06 0.11 0.573 -0.27 0.15 1.50 0.246 0.76 2.98 Hispanic 0.10 0.14 0.468 -0.18 0.38 0.98 0.972 0.39 2.47 Other NH -0.57 0.28 0.040 -1.12 -0.03 NE? Current Relationship Status Single [ref] 1.00 Married -0.11 0.13 0.395 -0.38 0.15 1.32 0.507 0.58 3.02 Cohabitating 0.07 0.11 0.553 -0.15 0.28 1.24 0.542 0.62 2.49 Insurance Type Public [ref] 1.00 Private 0.18 0.11 0.116 -0.04 0.40 1.03 0.934 0.51 2.10 None -0.05 0.12 0.682 -0.28 0.18 0.78 0.511 0.37 1.64 Education Level High school or [ref] 1.00 less Some -0.10 0.11 0.360 -0.32 0.12 0.64 0.213 0.32 1.29 college/vocational Bachelor?s -0.19 0.14 0.186 -0.47 0.09 0.54 0.190 0.21 1.36 degree or higher Nativity U.S.-born [ref] 1.00 Foreign-born -0.18 0.16 0.253 -0.49 0.13 0.99 0.980 0.37 2.64 222 Future Pregnancy Desire Within the next [ref] 1.00 2 years In 2 years or -0.13 0.18 0.468 -0.47 0.22 0.74 0.602 0.24 2.27 more Unsure & Yes, -0.03 0.18 0.871 -0.39 0.33 0.61 0.390 0.20 1.89 but not sure when No (more) -0.28 0.21 0.173 -0.69 0.12 0.52 0.311 0.15 1.83 children Hypothetical Pregnancy Happy [ref] 1.00 Unhappy 0.24 0.12 0.053 0.00 0.48 3.03 0.009 1.32 6.98 Unsure 0.15 0.12 0.202 -0.08 0.38 2.03 0.074 0.93 4.42 Unintended Pregnancy No [ref] 1.00 Yes -0.05 0.10 0.636 -0.24 0.15 0.79 0.438 0.43 1.44 Reproductive Coercion by a Partner - Ever None 1.00 Any -0.19 0.13 0.129 -0.44 0.06 0.83 0.634 0.39 1.78 Reason for Clinic Visit Not Family [ref] 1.00 Planning Family Planning 0.09 0.10 0.386 -0.11 0.28 0.95 0.884 0.51 1.78 Clinic Focus Primary Care [ref] 1.00 Women's Health -0.09 0.11 0.414 -0.32 0.13 1.07 0.849 0.51 2.25 Clinic Location Urban [ref] 1.00 Rural -0.14 0.10 0.170 -0.33 0.06 1.03 0.929 0.54 1.95 Suburban -0.08 0.22 0.718 -0.52 0.36 0.68 0.569 0.18 2.58 Wave One [ref] 1.00 Two -0.22 0.09 0.017 -0.40 -0.04 0.70 0.217 0.40 1.23 ?Non-estimable because the observations in the cell predict failure perfectly (100%). 223 Appendix I: Supplementary Analyses Supplementary Analyses for Aim 1.2: Factors that Predict Change in Effectiveness Level of Planned Contraceptive Use When comparing the results from the main analyses (Table 14) to those of the robust linear probability models, we find many similarities. Exceptions are as follows. In the robust linear probability model using the current method variable (Table A3), the main differences that appear are that the association with reproductive coercion was no longer significant, while the association with being a family planning user was significant?the probability of changing to LARC was 10% if participants came in for a family planning visit compared to any other type of visit (B=0.1, 95% CI: 0.03-0.18). The association with racial identity was no longer statistically significant, but the same racial categories were marginally significant (p<0.1). The suburban cell that was unable to be estimated in the logistic regression models shows a marginally significant association with not changing to LARC in the robust linear probability model. The same differences appear in the sensitivity models of the logistic regression and linear probability regression models. Comparing the robust linear regression probability model using the composite current- planned method variable with its bivariate logistic regression counterpart, the strongest association remains with the pre-visit report of moderately effective methods (B=-0.11, 95% CI: -0.19- -0.03 vs. OR=0.05, 95% CI: 0.01-0.36) (Table A4). However, the linear regression probability model also identifies that having some college/vocational education compared to a high school diploma or less and seeking care during the second wave as significantly associated with changing to LARC, as well as wanting a child within the next two years, compared to after two years or more, and visiting a suburban clinic, compared to an urban or rural one, as significantly associated with not changing to LARC. The latter two results were in line with the logistic regression, as they predicted failure perfectly in those models. Comparisons of the equivalent sensitivity analyses demonstrate the same pattern of results, except that the second wave was significantly associated with changing to LARC in both the logistic and linear regression models (B=0.06, 95% CI: 0.01-0.11 vs. OR=10.94, 95% CI: 1.13-105.5) (Table A4 and Table 14). Supplementary Analyses for Aim 1.3: DelCAN Expansion as a Predictor of Change in Effectiveness Level of Planned Contraceptive Use In additional supplementary analyses investigating what predicts post-visit LARC use after controlling for pre-visit LARC use, we found that the second wave was significantly associated with planning to use LARC compared to moderately effective methods post-visit (OR: 3.82, 95% CI: 1.43-10.17) in the composite pre-visit model (Table A7). No other models examining LARC as the post-visit outcome found wave to be significant (Tables A5-A7). Conversely, when investigating what predicts post-visit effectiveness level controlling for the effect of pre-visit method effectiveness level, we found participants in the second wave had a lower odds of planning to use moderately effective methods compared to no method or low 224 effective methods in the main composite model (OR: 0.38; 95% CI: 0.15-0.95); results from the other models followed the same direction but were not significant (Table A8). 225 : Supplementary Tables Bivariate relationship between pre-visit current method use and post-visit planned method use Pre-visit current method use EC, NFP, Patch, None Condoms Pills Shot IUD Implant TOTAL visit planned method use withdr, barr Ring 41 26 84 106 15 85 62 55 474 TOTAL counts and % 8.6% 5.5% 17.7% 22.4% 3.2% 17.9% 13.1% 11.6% 100.0% 27 10 26 7 1 9 6 9 95 None or Unsure 65.9% 38.5% 31.0% 6.6% 6.7% 10.6% 9.7% 16.4% 20.0% 0 1 0 0 0 0 0 0 1 EC, NFP, withdr., barrier 0.0% 3.8% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.2% 0 1 6 0 0 0 1 1 9 External Condoms 0.0% 3.8% 7.1% 0.0% 0.0% 0.0% 1.6% 1.8% 1.9% 5 6 26 93 3 3 4 3 143 12.2% 23.1% 31.0% 87.7% 20.0% 3.5% 6.5% 5.5% 30.2% 1 3 3 1 9 1 0 1 19 2.4% 11.5% 3.6% 0.9% 60.0% 1.2% 0.0% 1.8% 4.0% 0 0 5 0 0 71 3 0 79 0.0% 0.0% 6.0% 0.0% 0.0% 83.5% 4.8% 0.0% 16.7% 5 2 7 4 1 1 46 0 66 12.2% 7.7% 8.3% 3.8% 6.7% 1.2% 74.2% 0.0% 13.9% 3 3 11 1 1 0 2 41 62 7.3% 11.5% 13.1% 0.9% 6.7% 0.0% 3.2% 74.5% 13.1% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% EC stands for emergency contraception, NFP for natural family planning methods, withdr for withdrawal, and barr for barrier methods such as the diaphragm and spermicide. 226 Table A2. Bivariate relationship between pre-visit composite current-planned method use and post-visit planned method use Pre-visit composite current-planned method use EC, NFP, Patch, None Condoms Pills Shot IUD Implant TOTAL withdr., barr Ring Post-visit planned method use 34 11 44 134 19 86 74 72 474 TOTAL counts and % 7.2% 2.3% 9.3% 28.3% 4.0% 18.1% 15.6% 15.2% 100.0% 28 8 28 8 1 7 4 11 95 None or Unsure 82.4% 72.7% 63.6% 6.0% 5.3% 8.1% 5.4% 15.3% 20.0% 0 1 0 0 0 0 0 0 1 EC, NFP, withdr., barrier 0.0% 9.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.2% 1 1 6 0 0 0 1 0 9 External Condoms 2.9% 9.1% 13.6% 0.0% 0.0% 0.0% 1.4% 0.0% 1.9% 1 0 4 124 4 3 4 3 143 Pills 2.9% 0.0% 9.1% 92.5% 21.1% 3.5% 5.4% 4.2% 30.2% 1 0 0 1 13 1 2 1 19 Patch, Ring 2.9% 0.0% 0.0% 0.7% 68.4% 1.2% 2.7% 1.4% 4.0% 0 0 2 0 0 74 2 1 79 Shot 0.0% 0.0% 4.5% 0.0% 0.0% 86.0% 2.7% 1.4% 16.7% 1 1 1 1 0 1 61 0 66 IUD 2.9% 9.1% 2.3% 0.7% 0.0% 1.2% 82.4% 0.0% 13.9% 2 0 3 0 1 0 0 56 62 Implant 5.9% 0.0% 6.8% 0.0% 5.3% 0.0% 0.0% 77.8% 13.1% TOTAL 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Notes: EC stands for emergency contraception, NFP for natural family planning methods, withdr for withdr., and barr for barrier methods such as the diaphragm and spermicide. 227 Table A3. Adjusted relationship between change to LARC and study variables, including current method use, and its corresponding sensitivity analysis for the wave one skip pattern issue. Outcome of Interest Change to LARC Change to LARC (sensitivity) Pre-visit measure used Current method Current method Sample size 357 323 ROBUST B 95% CI p-value 95% CI ROBUST SE p-value SE B Pre-Visit Method None or Low [ref] [ref] Moderately -0.21 -0.29 -0.13 0.04 <0.001 -0.22 -0.30 -0.14 0.04 <0.001 LARC Age 15-19 [ref] [ref] 20-29 -0.03 -0.13 0.08 0.05 0.616 0.00 -0.11 0.11 0.06 0.996 30-44 -0.04 -0.15 0.06 0.05 0.409 -0.01 -0.13 0.10 0.06 0.815 Race/Hispanic Ethnicity White NH [ref] [ref] Black NH -0.06 -0.13 0.01 0.04 0.072 -0.06 -0.13 0.01 0.04 0.112 Hispanic 0.02 -0.09 0.13 0.05 0.712 0.03 -0.08 0.14 0.06 0.625 Other NH 0.24 0.00 0.48 0.12 0.051 0.19 -0.04 0.43 0.12 0.107 Current Relationship Status Single [ref] [ref] Married 0.01 -0.09 0.10 0.05 0.902 0.02 -0.08 0.12 0.05 0.635 Cohabitating -0.07 -0.13 0.00 0.03 0.049 -0.06 -0.13 0.01 0.04 0.100 228 Insurance Type Public [ref] [ref] Private -0.02 -0.10 0.06 0.04 0.647 -0.02 -0.10 0.07 0.04 0.726 None -0.06 -0.14 0.02 0.04 0.157 -0.04 -0.12 0.04 0.04 0.357 Education Level High school or less [ref] [ref] Some 0.01 -0.07 0.08 0.04 0.875 0.01 -0.07 0.09 0.04 0.772 college/vocational Bachelor?s degree or 0.04 -0.05 0.14 0.05 0.352 0.03 -0.06 0.12 0.05 0.512 more Nativity U.S.-born [ref] [ref] Foreign-born -0.04 -0.14 0.06 0.05 0.403 -0.07 -0.17 0.03 0.05 0.197 Future Pregnancy Desire Within the next 2 years -0.06 -0.17 0.04 0.05 0.252 -0.05 -0.16 0.06 0.06 0.412 In 2 years or more [ref] [ref] Unsure & Yes, but not 0.02 -0.06 0.10 0.04 0.627 0.02 -0.06 0.11 0.04 0.583 sure when No (more) children -0.06 -0.16 0.04 0.05 0.255 -0.08 -0.18 0.03 0.05 0.156 Hypothetical Pregnancy Happy [ref] [ref] Unhappy 0.05 -0.04 0.14 0.05 0.251 0.08 -0.02 0.18 0.05 0.101 Unsure 0.03 -0.05 0.12 0.04 0.426 0.04 -0.04 0.13 0.04 0.337 Unintended Pregnancy No [ref] [ref] Yes -0.02 -0.08 0.05 0.03 0.600 -0.03 -0.09 0.04 0.03 0.387 229 Reproductive Coercion by a Partner - Ever None [ref] [ref] Any 0.08 -0.03 0.19 0.06 0.145 0.08 -0.03 0.19 0.06 0.160 Reason for Clinic Visit Not Family Planning [ref] [ref] Family Planning 0.10 0.03 0.18 0.04 0.005 0.11 0.03 0.19 0.04 0.007 Clinic Focus Primary Care [ref] [ref] Women's Health 0.01 -0.06 0.08 0.04 0.739 -0.01 -0.08 0.07 0.04 0.827 Clinic Location Urban [ref] [ref] Rural 0.01 -0.06 0.08 0.04 0.726 0.01 -0.07 0.08 0.04 0.814 Suburban -0.09 -0.20 0.01 0.05 0.088 -0.11 -0.21 0.00 0.05 0.044 Wave One [ref] [ref] Two 0.05 -0.02 0.11 0.03 0.136 0.05 -0.02 0.11 0.03 0.179 230 Table A4. Adjusted relationship between change to LARC and study variables, including the composite current- planned method use, and its corresponding sensitivity analysis for the wave one skip pattern issue. Outcome of Interest Change to LARC Change to LARC (sensitivity) Pre-visit measure used Composite current-planned method Composite current-planned method Sample size 328 299 B 95% CI ROBUST SE p-value B 95% CI ROBUST SE p-value Pre-Visit Method None or Low [ref] [ref] Moderately -0.11 -0.19 -0.03 0.04 0.009 -0.11 -0.20 -0.03 0.04 0.008 LARC Age 15-19 [ref] [ref] 20-29 -0.03 -0.09 0.04 0.03 0.430 -0.03 -0.10 0.05 0.04 0.452 30-44 -0.07 -0.14 0.01 0.04 0.084 -0.07 -0.16 0.01 0.04 0.088 Race/Hispanic Ethnicity White NH [ref] [ref] Black NH <0.0001 -0.04 0.04 0.02 0.949 0.00 -0.04 0.05 0.02 0.836 Hispanic 0.01 -0.05 0.08 0.03 0.663 0.01 -0.06 0.08 0.04 0.695 Other NH 0.12 -0.03 0.28 0.08 0.116 0.12 -0.04 0.27 0.08 0.136 Current Relationship Status Single [ref] [ref] Married 0.01 -0.06 0.07 0.03 0.838 0.01 -0.06 0.08 0.04 0.728 Cohabitating <0.0001 -0.04 0.04 0.02 0.976 0.00 -0.05 0.05 0.02 0.976 Insurance Type Public [ref] [ref] 231 Private <0.001 -0.05 0.04 0.02 0.885 0.00 -0.05 0.05 0.03 0.926 None 0.05 0.00 0.11 0.03 0.069 0.06 0.00 0.12 0.03 0.060 Education Level High school or less [ref] [ref] Some college/vocational 0.04 0.01 0.08 0.02 0.026 0.05 0.01 0.09 0.02 0.027 Bachelor?s degree or 0.01 -0.04 0.06 0.03 0.645 0.01 -0.04 0.07 0.03 0.685 more Nativity U.S.-born [ref] [ref] Foreign-born 0.04 -0.03 0.10 0.03 0.299 0.04 -0.03 0.11 0.04 0.302 Future Pregnancy Desire Within the next 2 years -0.06 -0.12 0.00 0.03 0.048 -0.06 -0.12 0.00 0.03 0.067 In 2 years or more [ref] [ref] Unsure & Yes, but not 0.04 -0.02 0.10 0.03 0.174 0.04 -0.02 0.11 0.03 0.187 sure when No (more) children 0.02 -0.03 0.06 0.02 0.449 0.01 -0.04 0.07 0.03 0.644 Hypothetical Pregnancy Happy [ref] [ref] Unhappy 0.02 -0.04 0.07 0.03 0.540 0.03 -0.04 0.09 0.03 0.403 Unsure 0.02 -0.04 0.07 0.03 0.531 0.02 -0.04 0.07 0.03 0.533 Unintended Pregnancy No [ref] [ref] Yes 0.04 -0.01 0.08 0.02 0.106 0.03 -0.02 0.07 0.02 0.215 Reproductive Coercion by a Partner - Ever None [ref] [ref] Any -0.0003 -0.06 0.06 0.03 0.992 0.01 -0.06 0.08 0.04 0.811 Reason for Clinic Visit 232 Not Family Planning [ref] [ref] Family Planning <0.0001 -0.06 0.05 0.03 0.882 0.00 -0.06 0.06 0.03 0.981 Clinic Focus Primary Care [ref] [ref] Women's Health -0.01 -0.06 0.04 0.03 0.769 -0.01 -0.06 0.05 0.03 0.850 Clinic Location Urban [ref] [ref] Rural -0.03 -0.07 0.02 0.02 0.262 -0.02 -0.07 0.02 0.03 0.326 Suburban -0.08 -0.15 0.00 0.04 0.041 -0.08 -0.16 -0.01 0.04 0.037 Wave One [ref] [ref] Two 0.05 0.01 0.09 0.02 0.025 0.06 0.01 0.11 0.02 0.016 233 Table A5. Adjusted relationship of reporting LARC use post-visit compared to any other method and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. Outcome of Interest LARC vs. Non-LARC Composite current- Composite current- Pre-visit measure used Current Current planned planned Sample size N=474 N=432 (sensitivity) N=474 N=432 (sensitivity) Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 0.11*** (0.05-0.26) 0.08*** (0.03-0.22) 0.07*** (0.02-0.31) 0.07*** (0.02-0.31) LARC 21.69*** (10.32-45.59) 18.63*** (8.81-39.38) 41.45*** (15.69-109.53) 34.46*** (13.03-91.13) Age 15-19 1.00 1.00 1.00 1.00 20-29 0.58 (0.23-1.43) 0.75 (0.29-1.99) 0.79 (0.24-2.56) 0.96 (0.29-3.2) 30-44 0.34 (0.1-1.12) 0.49 (0.14-1.69) 0.4 (0.1-1.71) 0.51 (0.12-2.22) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.55 (0.26-1.13) 0.52 (0.24-1.14) 0.64 (0.26-1.6) 0.58 (0.22-1.53) Hispanic 1.16 (0.45-2.95) 1.14 (0.41-3.15) 1.05 (0.31-3.62) 0.94 (0.25-3.5) Other NH 5.38 (0.94-30.71) 3.65 (0.51-25.95) 5.12 (0.45-58.31) 3.73 (0.29-47.58) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 0.67 (0.26-1.7) 0.82 (0.31-2.2) 0.73 (0.26-2.07) 0.8 (0.27-2.35) Cohabitating 0.44* (0.2-0.98) 0.46 (0.2-1.05) 0.78 (0.31-1.97) 0.75 (0.29-1.91) Insurance Type 234 Public 1.00 1.00 1.00 1.00 Private 0.64 (0.3-1.35) 0.66 (0.3-1.46) 0.66 (0.26-1.67) 0.73 (0.29-1.85) None 0.43* (0.19-0.99) 0.46 (0.19-1.09) 0.85 (0.31-2.31) 0.87 (0.31-2.44) Education Level High school or less 1.00 1.00 1.00 1.00 Some 0.95 (0.44-2.05) 1.02 (0.46-2.27) 1.19 (0.45-3.15) 1.27 (0.48-3.39) college/vocational Bachelor?s degree or 1.71 (0.63-4.62) 1.38 (0.48-3.93) 1.44 (0.43-4.81) 1.36 (0.4-4.67) more Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 1.49 (0.51-4.36) 1.33 (0.42-4.22) 1.71 (0.49-6.01) 1.7 (0.44-6.52) Future Pregnancy Desire Within the next 2 1.00 1.00 1.00 1.00 years In 2 years or more 2.45 (0.72-8.39) 2.23 (0.64-7.71) 0.77 (0.19-3.11) 0.76 (0.19-3.08) Unsure & Yes, but 2.39 (0.67-8.44) 1.98 (0.56-7.04) 1.33 (0.33-5.45) 1.2 (0.29-4.88) not sure when No (more) children 2.45 (0.59-10.13) 1.93 (0.46-8.11) 1.98 (0.39-10.03) 1.66 (0.33-8.37) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.35 (0.58-3.11) 1.74 (0.7-4.35) 1.53 (0.52-4.47) 1.76 (0.57-5.45) Unsure 1.27 (0.57-2.83) 1.46 (0.63-3.38) 0.85 (0.33-2.21) 0.96 (0.36-2.54) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 1.35 (0.7-2.62) 1.23 (0.62-2.43) 1.10 (0.49-2.47) 1.00 (0.45-2.25) Reproductive Coercion by a Partner - Ever None 1.00 1.00 1.00 1.00 235 Any 2.78* (1.25-6.21) 2.89* (1.19-7.03) 1.93 (0.7-5.36) 2.07 (0.71-6.07) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 1.56 (0.79-3.08) 1.42 (0.7-2.86) 1.61 (0.72-3.6) 1.53 (0.67-3.49) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 1.29 (0.60-2.82) 1.20 (0.53-2.69) 0.99 (0.39-2.53) 0.94 (0.36-2.40) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 1.09 (0.55-2.14) 0.98 (0.48-1.99) 0.79 (0.34-1.84) 0.75 (0.31-1.79) Suburban 0.14* (0.02-0.84) 0.13* (0.02-0.79) 0.24 (0.03-1.97) 0.22 (0.03-1.75) Wave One 1.00 1.00 1.00 1.00 Two 1.51 (0.82-2.8) 1.44 (0.75-2.75) 2.07 (0.96-4.48) 2.13 (0.97-4.7) *p?0.05. **p?0.01. ***p?0.001. 236 Table A6. Adjusted relationship of reporting LARC use post-visit compared to none/low effective methods and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. Outcome of Interest LARC vs. NONE and LOW Composite current- Composite current- Pre-visit measure used Current Current planned planned Sample size N=474 N=432 (sensitivity) N=474 N=432 (sensitivity) Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 1.00 (0.35-2.87) 0.72 (0.22-2.36) 1.54 (0.3-7.89) 1.41 (0.27-7.34) LARC 22.04*** (9.15-53.08) 19.59*** (8.00-48.00) 81.81*** (27.28-245.31) 67.53*** (22.37-203.88) Age 15-19 1.00 1.00 1.00 1.00 20-29 0.59 (0.2-1.77) 0.72 (0.22-2.3) 0.73 (0.19-2.77) 0.83 (0.21-3.36) 30-44 0.31 (0.08-1.2) 0.42 (0.1-1.73) 0.48 (0.1-2.44) 0.54 (0.1-2.9) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.45 (0.19-1.07) 0.41 (0.16-1.01) 0.65 (0.23-1.84) 0.55 (0.18-1.66) Hispanic 0.77 (0.25-2.36) 0.71 (0.21-2.37) 0.7 (0.18-2.7) 0.55 (0.13-2.34) Other NH 4.21 (0.48-37.05) 2.53 (0.25-25.45) 3.75 (0.26-54.08) 2.26 (0.14-36.46) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 0.64 (0.22-1.86) 0.82 (0.27-2.49) 0.97 (0.3-3.11) 1.1 (0.33-3.69) Cohabitating 0.63 (0.24-1.62) 0.62 (0.24-1.64) 1.69 (0.55-5.25) 1.46 (0.47-4.58) 237 Insurance Type Public 1.00 1.00 1.00 1.00 Private 0.43 (0.18-1.03) 0.46 (0.19-1.16) 0.39 (0.13-1.13) 0.43 (0.15-1.29) None 0.57 (0.22-1.52) 0.56 (0.2-1.55) 1.07 (0.34-3.4) 1.06 (0.32-3.49) Education Level High school or less 1.00 1.00 1.00 1.00 Some 0.7 (0.29-1.71) 0.84 (0.33-2.09) 1.27 (0.42-3.83) 1.37 (0.45-4.15) college/vocational Bachelor?s degree or 1.43 (0.46-4.46) 1.41 (0.42-4.7) 1.62 (0.42-6.34) 1.75 (0.42-7.23) more Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 1.21 (0.36-4.04) 1.24 (0.34-4.51) 1.27 (0.33-4.96) 1.42 (0.32-6.19) Future Pregnancy Desire Within the next 2 1.00 1.00 1.00 1.00 years In 2 years or more 6.35** (1.62-24.91) 5.53* (1.39-22.06) 2.61 (0.53-12.94) 2.47 (0.48-12.61) Unsure & Yes, but 3.78 (0.96-14.88) 2.72 (0.69-10.78) 2.37 (0.49-11.41) 1.96 (0.4-9.54) not sure when No (more) children 7.16* (1.42-36.03) 4.69 (0.91-24.13) 4.58 (0.72-29.2) 3.72 (0.57-24.22) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.65 (0.61-4.45) 2.39 (0.81-7.07) 1.46 (0.43-5.01) 1.58 (0.43-5.81) Unsure 1.14 (0.46-2.81) 1.44 (0.57-3.67) 0.68 (0.23-1.98) 0.82 (0.27-2.46) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 1.27 (0.59-2.71) 1.11 (0.51-2.42) 0.97 (0.39-2.42) 0.86 (0.34-2.15) Reproductive Coercion by a Partner - Ever 238 None 1.00 1.00 1.00 1.00 Any 2.79* (1.07-7.26) 2.92* (1.03-8.25) 1.76 (0.56-5.49) 1.99 (0.59-6.67) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 3.39** (1.57-7.33) 3.01** (1.36-6.67) 2.98* (1.2-7.38) 2.82* (1.11-7.18) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 1.78 (0.75-4.25) 1.65 (0.67-4.06) 1.48 (0.54-4.12) 1.43 (0.51-4.02) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 0.63 (0.29-1.36) 0.56 (0.25-1.26) 0.52 (0.2-1.33) 0.47 (0.18-1.24) Suburban 0.12* (0.02-0.8) 0.10* (0.01-0.73) 0.23 (0.02-2.27) 0.21 (0.02-1.99) Wave One 1.00 1.00 1.00 1.00 Two 1.05 (0.51-2.17) 1.11 (0.52-2.37) 1.44 (0.6-3.45) 1.56 (0.64-3.85) *p?0.05. **p?0.01. ***p?0.001. 239 Table A7. Adjusted relationship between reporting LARC use post-visit compared to moderately effective methods and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. Outcome of Interest LARC vs. MODERATELY Composite current- Composite current- Pre-visit measure used Current Current planned planned Sample size N=474 N=432 (sensitivity) N=474 N=432 (sensitivity) Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 0.05*** (0.02-0.13) 0.04*** (0.01-0.12) 0.003*** (0.001-0.02) 0.004*** (0.0007-0.03) LARC 24.55*** (9.59-62.86) 20.04*** (7.82-51.38) 9.75*** (2.63-36.15) 8.65** (2.33-32.07) Age 15-19 1.00 1.00 1.00 1.00 20-29 0.53 (0.19-1.45) 0.69 (0.23-2.1) 1 (0.24-4.18) 1.38 (0.32-6.01) 30-44 0.3 (0.08-1.13) 0.42 (0.1-1.71) 0.25 (0.04-1.5) 0.36 (0.06-2.23) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.6 (0.27-1.37) 0.61 (0.26-1.45) 0.67 (0.22-2.02) 0.61 (0.19-1.93) Hispanic 1.48 (0.53-4.15) 1.54 (0.5-4.76) 1.38 (0.31-6.17) 1.24 (0.26-5.92) Other NH 6.96 (0.99-49.07) 4.81 (0.53-43.43) 15.49 (0.9-267.29) 11.22 (0.61-205.79) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 0.77 (0.26-2.27) 0.93 (0.3-2.86) 0.38 (0.1-1.48) 0.43 (0.11-1.7) Cohabitating 0.31* (0.13-0.76) 0.34* (0.14-0.86) 0.35 (0.11-1.09) 0.35 (0.11-1.11) 240 Insurance Type Public 1.00 1.00 1.00 1.00 Private 0.93 (0.4-2.19) 0.95 (0.38-2.35) 1.65 (0.52-5.3) 1.79 (0.54-5.87) None 0.36* (0.14-0.9) 0.39 (0.15-1.03) 0.89 (0.26-3) 0.93 (0.27-3.14) Education Level High school or less 1.00 1.00 1.00 1.00 Some 1.26 (0.54-2.99) 1.27 (0.52-3.13) 1.18 (0.36-3.82) 1.21 (0.37-3.93) college/vocational Bachelor?s degree or 1.98 (0.66-5.95) 1.39 (0.44-4.46) 1.21 (0.29-5.12) 0.96 (0.22-4.12) more Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 2.1 (0.61-7.17) 1.74 (0.47-6.51) 3.19 (0.64-15.83) 2.82 (0.55-14.56) Future Pregnancy Desire Within the next 2 1.00 1.00 1.00 1.00 years In 2 years or more 1.00 (0.24-4.19) 0.96 (0.22-4.14) 0.19 (0.03-1.21) 0.20 (0.03-1.27) Unsure & Yes, but not 1.42 (0.32-6.23) 1.37 (0.3-6.19) 0.87 (0.13-5.68) 0.82 (0.13-5.31) sure when No (more) children 0.93 (0.18-4.92) 0.84 (0.15-4.61) 1.05 (0.12-8.9) 0.89 (0.11-7.4) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.04 (0.41-2.66) 1.2 (0.43-3.37) 1.08 (0.3-3.83) 1.18 (0.31-4.47) Unsure 1.31 (0.53-3.23) 1.41 (0.54-3.63) 0.82 (0.25-2.63) 0.83 (0.25-2.72) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 1.52 (0.72-3.2) 1.36 (0.63-2.95) 1.6 (0.58-4.42) 1.36 (0.49-3.74) Reproductive Coercion by a Partner - Ever 241 None 1.00 1.00 1.00 1.00 Any 2.77* (1.11-6.89) 2.85* (1.04-7.82) 1.48 (0.42-5.22) 1.6 (0.44-5.86) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 0.69 (0.31-1.5) 0.63 (0.28-1.41) 0.67 (0.25-1.81) 0.64 (0.23-1.76) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 0.90 (0.37-2.17) 0.82 (0.32-2.08) 0.47 (0.15-1.52) 0.47 (0.14-1.50) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 1.82 (0.84-3.96) 1.71 (0.75-3.93) 1.54 (0.53-4.44) 1.52 (0.51-4.55) Suburban 0.12* (0.02-0.91) 0.12* (0.01-0.89) 0.29 (0.02-3.66) 0.27 (0.02-3.32) Wave One 1.00 1.00 1.00 1.00 Two 1.75 (0.87-3.51) 1.60 (0.77-3.34) 3.82** (1.43-10.17) 3.84** (1.41-10.48) *p?0.05. **p?0.01. ***p?0.001. 242 Table A8. Adjusted relationship between reporting moderately effective methods post-visit compared to none/low effective methods and study variables, including the current and composite current-planned method use, and the corresponding sensitivity analyses for the wave one skip pattern issue. Outcome of Interest MODERATELY vs. NONE and LOW Composite current- Composite current- Pre-visit measure used Current Current planned planned Sample size N=474 N=432 (sensitivity) N=474 N=432 (sensitivity) Pre-Visit Method None or Low 1.00 1.00 1.00 1.00 Moderately 18.60*** (8.56-40.42) 17.86*** (7.92-40.27) 423.35*** (109.28-1640.1) 347.86*** (88.0-1374.6) LARC 0.90 (0.33-2.44) 0.98 (0.35-2.71) 8.39** (2.32-30.28) 7.81** (2.13-28.64) Age 15-19 1.00 1.00 1.00 1.00 20-29 1.13 (0.4-3.18) 1.04 (0.34-3.14) 0.73 (0.19-2.74) 0.6 (0.15-2.48) 30-44 1.03 (0.31-3.46) 1 (0.28-3.57) 1.93 (0.39-9.48) 1.49 (0.29-7.76) Race/Hispanic Ethnicity White NH 1.00 1.00 1.00 1.00 Black NH 0.75 (0.35-1.62) 0.66 (0.29-1.52) 0.98 (0.35-2.77) 0.91 (0.3-2.73) Hispanic 0.52 (0.19-1.44) 0.46 (0.15-1.37) 0.50 (0.14-1.89) 0.44 (0.11-1.78) Other NH 0.61 (0.07-5.19) 0.53 (0.06-4.68) 0.24 (0.02-3.15) 0.2 (0.02-2.62) Current Relationship Status Single 1.00 1.00 1.00 1.00 Married 0.83 (0.32-2.2) 0.88 (0.32-2.42) 2.53 (0.72-8.89) 2.56 (0.71-9.24) 243 Cohabitating 2.02 (0.89-4.58) 1.82 (0.78-4.28) 4.85** (1.52-15.41) 4.23* (1.28-13.9) Insurance Type Public 1.00 1.00 1.00 1.00 Private 0.46 (0.2-1.04) 0.49 (0.21-1.15) 0.24** (0.08-0.71) 0.24* (0.08-0.75) None 1.61 (0.67-3.9) 1.45 (0.57-3.66) 1.2 (0.37-3.87) 1.14 (0.35-3.73) Education Level High school or less 1.00 1.00 1.00 1.00 Some 0.56 (0.25-1.24) 0.66 (0.29-1.52) 1.08 (0.37-3.13) 1.13 (0.39-3.31) college/vocational Bachelor?s degree or 0.72 (0.26-1.99) 1.01 (0.34-2.95) 1.34 (0.36-5.07) 1.83 (0.46-7.26) more Nativity U.S.-born 1.00 1.00 1.00 1.00 Foreign-born 0.58 (0.19-1.8) 0.71 (0.22-2.33) 0.4 (0.1-1.66) 0.5 (0.12-2.17) Future Pregnancy Desire Within the next 2 1.00 1.00 1.00 1.00 years In 2 years or more 6.33** (1.95-20.56) 5.76** (1.7-19.46) 14.03** (2.62-75.13) 12.64** (2.31-69.26) Unsure & Yes, but 2.66 (0.81-8.74) 1.99 (0.58-6.77) 2.72 (0.53-14.11) 2.39 (0.45-12.65) not sure when No (more) children 7.69** (1.74-34.01) 5.56* (1.21-25.52) 4.37 (0.6-32.03) 4.19 (0.56-31.11) Hypothetical Pregnancy Happy 1.00 1.00 1.00 1.00 Unhappy 1.59 (0.64-3.95) 1.99 (0.73-5.41) 1.36 (0.41-4.53) 1.33 (0.37-4.8) Unsure 0.87 (0.38-1.99) 1.02 (0.43-2.42) 0.83 (0.27-2.51) 0.99 (0.32-3.06) Unintended Pregnancy No 1.00 1.00 1.00 1.00 Yes 0.84 (0.41-1.69) 0.81 (0.39-1.68) 0.61 (0.23-1.61) 0.63 (0.24-1.68) 244 Reproductive Coercion by a Partner - Ever None 1.00 1.00 1.00 1.00 Any 1.01 (0.41-2.49) 1.03 (0.39-2.7) 1.19 (0.35-4.02) 1.24 (0.35-4.4) Reason for Clinic Visit Not Family Planning 1.00 1.00 1.00 1.00 Family Planning 4.95*** (2.48-9.88) 4.77*** (2.31-9.85) 4.43** (1.81-10.88) 4.41** (1.75-11.12) Clinic Focus Primary Care 1.00 1.00 1.00 1.00 Women's Health 1.99 (0.90-4.39) 2.02 (0.87-4.70) 3.14* (1.14-8.70) 3.08* (1.09-8.67) Clinic Location Urban 1.00 1.00 1.00 1.00 Rural 0.35** (0.17-0.72) 0.33** (0.15-0.71) 0.34* (0.13-0.85) 0.31* (0.12-0.82) Suburban 0.97 (0.24-3.95) 0.90 (0.22-3.74) 0.81 (0.11-5.82) 0.76 (0.11-5.17) Wave One 1.00 1.00 1.00 1.00 Two 0.6 (0.31-1.17) 0.69 (0.35-1.39) 0.38* (0.15-0.95) 0.41 (0.16-1.06) *p?0.05. **p?0.01. ***p?0.001. 245 Table A9. Adjusted relationships between change in effectiveness knowledge and study variables, using one continuous and two dichotomous change outcomes. Increase in effectiveness Increase in effectiveness knowledge % increase in effectiveness knowledge knowledge (Yes vs. No) (to Total Score only) N=351 B SE p-value 95% CI OR p-value 95% CI OR p-value 95% CI Pre-visit Effectiveness -0.10 0.03 <0.001 -0.15 -0.05 0.98 0.867 0.75 1.28 1.61 0.004 1.16 2.25 Knowledge (cont.) Age 15-19 [ref] 1.00 1.00 20-29 -0.02 0.08 0.850 -0.18 0.15 0.85 0.696 0.39 1.89 0.58 0.276 0.21 1.55 30-44 0.06 0.10 0.537 -0.14 0.26 1.59 0.338 0.62 4.10 0.66 0.521 0.19 2.32 Race/Hispanic Ethnicity White NH [ref] 1.00 1.00 Black NH -0.21 0.06 0.002 -0.33 -0.08 0.56 0.064 0.30 1.03 0.26 0.003 0.11 0.63 Hispanic -0.12 0.08 0.174 -0.28 0.05 0.44 0.057 0.19 1.02 0.51 0.212 0.18 1.47 Other NH 0.09 0.17 0.609 -0.25 0.43 1.28 0.752 0.28 5.77 2.29 0.343 0.41 12.77 Current Relationship Status Single [ref] 1.00 1.00 Married -0.09 0.08 0.242 -0.24 0.06 0.69 0.354 0.31 1.52 0.61 0.414 0.19 1.99 Cohabitating 0.00 0.07 0.957 -0.13 0.13 1.14 0.689 0.61 2.11 1.20 0.662 0.53 2.74 Insurance Type Public [ref] 1.00 1.00 Private -0.06 0.07 0.383 -0.20 0.08 0.81 0.525 0.42 1.56 0.87 0.763 0.35 2.16 None -0.03 0.07 0.654 -0.17 0.10 0.79 0.482 0.41 1.53 0.64 0.335 0.25 1.59 Education Level 246 High school or less [ref] 1.00 1.00 Some -0.05 0.07 0.499 -0.18 0.09 0.68 0.256 0.35 1.32 0.88 0.771 0.36 2.14 college/vocational Bachelor?s degree or 0.00 0.09 0.962 -0.16 0.17 0.76 0.512 0.34 1.72 0.74 0.599 0.24 2.26 higher Nativity U.S.-born [ref] 1.00 1.00 Foreign-born -0.02 0.09 0.857 -0.20 0.17 1.31 0.562 0.53 3.23 1.07 0.917 0.30 3.85 Future Pregnancy Desire Within the next 2 [ref] 1.00 1.00 years In 2 years or more -0.07 0.10 0.479 -0.27 0.13 1.35 0.551 0.50 3.60 0.92 0.900 0.26 3.30 Unsure & Yes, but -0.08 0.11 0.436 -0.29 0.13 0.89 0.825 0.32 2.48 0.91 0.885 0.24 3.42 not sure when No (more) children -0.13 0.12 0.291 -0.37 0.11 0.86 0.801 0.27 2.78 0.50 0.393 0.10 2.45 Feeling about a Hypothetical Pregnancy in Next Year Very Happy & Happy [ref] 1.00 1.00 Very Unhappy & 0.04 0.07 0.631 -0.11 0.18 1.34 0.429 0.65 2.73 1.05 0.927 0.40 2.76 Unhappy Unsure 0.07 0.07 0.352 -0.07 0.20 1.58 0.183 0.81 3.09 1.47 0.396 0.60 3.59 Unintended Pregnancy No [ref] 1.00 1.00 Yes -0.01 0.06 0.813 -0.13 0.10 0.69 0.214 0.38 1.24 1.33 0.482 0.60 2.92 Reproductive Coercion by a Partner - Ever None [ref] 1.00 247 Any 0.03 0.08 0.688 -0.12 0.18 2.16 0.032 1.07 4.38 2.12 0.129 0.80 5.59 Reason for Clinic Visit Not Family Planning [ref] 1.00 1.00 Family Planning 0.00 0.06 0.975 -0.12 0.12 0.84 0.561 0.48 1.50 1.06 0.887 0.47 2.40 Clinic Focus Primary Care [ref] 1.00 1.00 Women's Health 0.06 0.07 0.378 -0.07 0.19 1.15 0.678 0.60 2.21 2.12 0.144 0.77 5.83 Clinic Location Urban [ref] 1.00 1.00 Rural 0.04 0.06 0.486 -0.08 0.17 1.75 0.064 0.97 3.15 1.71 0.194 0.76 3.86 Suburban 0.14 0.13 0.297 -0.12 0.41 2.19 0.221 0.62 7.68 3.91 0.165 0.57 26.77 Wave One [ref] 1.00 1.00 Two -0.16 0.06 0.004 -0.27 -0.05 0.56 0.037 0.33 0.97 0.40 0.015 0.19 0.84 248 Table A10. Proportions of the sample?s reporting of contraceptive decision- making following the provider visit. N=473? % Count Started a new method 24.9% 118 Patient chose 80.5% 95 Health care provider and patient chose 16.9% 20 Health care provider chose 0.8% 1 Other? or Missing data 1.7% 2 Did not start a new method 58.1% 275 Patient chose 86.9% 239 Health care provider and patient chose 8.7% 24 Health care provider chose 1.1% 3 Other? or Missing data 3.3% 9 No mention of contraception or pregnancy intention during visit? 16.9% 80 Notes. ? One participant indicated they had talked about various forms of contraception with the provider, but left blank the question if they had started a new contraceptive method, so they were not asked the decision-making question. ? Participants who chose "Other" were able to write-in their answers. Write-ins were reclassified into the three main categories when possible. The Other category here refers to participants whose write-ins left the decision-maker unclear or clicked Other but left the write-in box blank. ? Patients who did not discuss contraception with the provider were not asked the contraceptive decision-making question in wave two. They were not supposed to be asked the question in wave one, but the skip pattern was not implemented as designed. For consistency, those who reported not discussing contraception or pregnancy intention during their visit were placed in the ?No mention? category even if they answered the new method and decision- making question (for context, of such 41 participants in wave one, none initiated a new method (as is expected), and only one indicated that the provider made that choice). 249 References Abajobir, A. A., Maravilla, J. C., Alati, R., & Najman, J. M. (2016). A systematic review and meta-analysis of the association between unintended pregnancy and perinatal depression. Journal of Affective Disorders, 192, 56?63. https://doi.org/10.1016/j.jad.2015.12.008 ACOG Committee on Adolescent Health Care Long-Acting Reversible Contraception Working Group. (2012). Committee opinion no. 539: Adolescents and long-acting reversible contraception: implants and intrauterine devices. Obstetrics and Gynecology, 120(4), 983?988. https://doi.org/10.1097/AOG.0b013e3182723b7d ACOG Committee on Health Care for Underserved Women and Committee on Ethics. (2022, February). Patient- Centered Contraceptive Counseling. American College of Obstetricians and Gynecologists. https://www.acog.org/en/clinical/clinical-guidance/committee-statement/articles/2022/02/patient-centered- contraceptive-counseling ACOG Committee on Practice Bulletins?Gynecology. (2015). Emergency Contraception. Practice Bulletins? Gynecology, 152. https://www.acog.org/en/clinical/clinical-guidance/practice- bulletin/articles/2015/09/emergency-contraception Aiken, A. R. A., Dillaway, C., & Mevs-Korff, N. (2015). A blessing I can?t afford: Factors underlying the paradox of happiness about unintended pregnancy. Social Science & Medicine, 132, 149?155. https://doi.org/10.1016/j.socscimed.2015.03.038 American College of Obstetricians and Gynecologists. (2007). ACOG Committee Opinion No. 392, December 2007. Intrauterine device and adolescents. Obstetrics and Gynecology, 110(6), 1493?1495. https://doi.org/10.1097/01.AOG.0000291575.93944.1a Amico, J. R., Bennett, A. H., Karasz, A., & Gold, M. (2016). ?She just told me to leave it?: Women?s experiences discussing early elective IUD removal. Contraception, 94(4), 357?361. https://doi.org/10.1016/j.contraception.2016.04.012 250 Amico, J. R., Bennett, A. H., Karasz, A., & Gold, M. (2017). ?I wish they could hold on a little longer?: Physicians? experiences with requests for early IUD removal. Contraception, 96(2), 106?110. https://doi.org/10.1016/j.contraception.2017.05.007 Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36(1), 1?10. Andrews, M. (2021, July 21). Contraception Is Free To Women, Except When It?s Not. NPR. https://www.npr.org/sections/health-shots/2021/07/21/1018483557/contraception-is-free-to-women-except- when-its-not Asker, C., Stokes-Lampard, H., Wilson, S., & Beavan, J. (2006). What is it about intrauterine devices that women find unacceptable? Factors that make women non-users: a qualitative study. BMJ Sexual & Reproductive Health, 32(2), 89?94. https://doi.org/10.1783/147118906776276170 August, E. M., Steinmetz, E., Gavin, L., Rivera, M. I., Pazol, K., Moskosky, S., Weik, T., & Ku, L. (2016). Projecting the Unmet Need and Costs for Contraception Services After the Affordable Care Act. American Journal of Public Health, 106(2), 334?341. https://doi.org/10.2105/AJPH.2015.302928 Axinn, W. G., Barber, J. S., & Thornton, A. (1998). The long-term impact of parents? childbearing decisions on children?s self-esteem. Demography, 35(4), 435?443. Barber, J. S., Axinn, W. G., & Thornton, A. (1999). Unwanted childbearing, health, and mother-child relationships. Journal of Health and Social Behavior, 40(3), 231?257. Barber, J. S., Guzzo, K. B., Budnick, J., Kusunoki, Y., Hayford, S. R., & Miller, W. (2021). Black-White Differences in Pregnancy Desire During the Transition to Adulthood. Demography, 58(2), 603?630. https://doi.org/10.1215/00703370-8993840 Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is Stronger than Good. Review of General Psychology, 5(4), 323?370. https://doi.org/10.1037/1089-2680.5.4.323 Bearak, J., Popinchalk, A., Alkema, L., & Sedgh, G. (2018). Global, regional, and subregional trends in unintended pregnancy and its outcomes from 1990 to 2014: Estimates from a Bayesian hierarchical model. The Lancet Global Health, 6(4), e380?e389. https://doi.org/10.1016/S2214-109X(18)30029-9 251 Becker, D., & Tsui, A. O. (2008). Reproductive Health Service Preferences And Perceptions of Quality Among Low-Income Women: Racial, Ethnic and Language Group Differences. Perspectives on Sexual and Reproductive Health, 40(4), 202?211. https://doi.org/10.1363/4020208 Becker, N. V., Keating, N. L., & Pace, L. E. (2021). ACA Mandate Led To Substantial Increase In Contraceptive Use Among Women Enrolled In High-Deductible Health Plans. Health Affairs, 40(4), 579?586. https://doi.org/10.1377/hlthaff.2020.01710 Belfield, T. (2009). Principles of contraceptive care: Choice, acceptability and access. Best Practice & Research Clinical Obstetrics & Gynaecology, 23(2), 177?185. https://doi.org/10.1016/j.bpobgyn.2008.11.006 Bellanca, H. K., & Hunter, M. S. (2013). One Key Question: Preventive reproductive health is part of high quality primary care. Contraception, 88(1), 3?6. https://doi.org/10.1016/j.contraception.2013.05.003 Berndt, V. K., & Bell, A. V. (2021a). Contextualizing barriers to long-acting reversible contraception in Delaware. Contraception, 103(6), 439?443. https://doi.org/10.1016/j.contraception.2021.02.007 Berndt, V. K., & Bell, A. V. (2021b). ?This is what the truth is?: Provider-patient interactions serving as barriers to contraception. Health, 25(5), 613?629. https://doi.org/10.1177/1363459320969775 Bharadwaj, P., Akintomide, H., Brima, N., Copas, A., & D?Souza, R. (2012). Determinants of long-acting reversible contraceptive (LARC) use by adolescent girls and young women. The European Journal of Contraception & Reproductive Health Care, 17(4), 298?306. Biggs, M. A., Rocca, C. H., Brindis, C. D., Hirsch, H., & Grossman, D. (2015). Did increasing use of highly effective contraception contribute to declining abortions in Iowa? Contraception, 91(2), 167?173. https://doi.org/10.1016/j.contraception.2014.10.009 Binette, A., Howatt, K., Waddington, A., & Reid, R. L. (2017). Ten Challenges in Contraception. Journal of Women?s Health, 26(1), 44?49. https://doi.org/10.1089/jwh.2016.5854 Birgisson, N. E., Zhao, Q., Secura, G. M., Madden, T., & Peipert, J. F. (2015). Preventing Unintended Pregnancy: The Contraceptive CHOICE Project in Review. Journal of Women?s Health, 24(5), 349?353. https://doi.org/10.1089/jwh.2015.5191 252 Bitzer, J., Gemzell-Danielsson, K., Roumen, F., Marintcheva-Petrova, M., van Bakel, B., & Oddens, B. J. (2012). The CHOICE study: Effect of counselling on the selection of combined hormonal contraceptive methods in 11 countries. The European Journal of Contraception & Reproductive Health Care, 17(1), 65?78. https://doi.org/10.3109/13625187.2011.637586 Bommaraju, A., Malat, J., & Mooney, J. L. (2015). Reproductive Life Plan Counseling and Effective Contraceptive Use among Urban Women Utilizing Title X Services. Women?s Health Issues, 25(3), 209?215. https://doi.org/10.1016/j.whi.2015.02.005 Borrero, S., Nikolajski, C., Steinberg, J. R., Freedman, L., Akers, A. Y., Ibrahim, S., & Schwarz, E. B. (2015). ?It just happens?: A qualitative study exploring low-income women?s perspectives on pregnancy intention and planning. Contraception, 91(2), 150?156. https://doi.org/10.1016/j.contraception.2014.09.014 Borrero, S., Schwarz, E. B., Creinin, M., & Ibrahim, S. (2008). The Impact of Race and Ethnicity on Receipt of Family Planning Services in the United States. Journal of Women?s Health, 18(1), 91?96. https://doi.org/10.1089/jwh.2008.0976 Boudreaux, M., Xie, L., Choi, Y. S., Roby, D. H., & Rendall, M. S. (2020). Changes to Contraceptive Method Use at Title X Clinics Following Delaware Contraceptive Access Now, 2008?2017. American Journal of Public Health, 110(8), 1214?1220. https://doi.org/10.2105/AJPH.2020.305666 Bracken, J., & Graham, C. A. (2014). Young women?s attitudes towards, and experiences of, long-acting reversible contraceptives. The European Journal of Contraception & Reproductive Health Care, 19(4), 276?284. https://doi.org/10.3109/13625187.2014.917623 Brandi, K., & Fuentes, L. (2020). The history of tiered-effectiveness contraceptive counseling and the importance of patient-centered family planning care. American Journal of Obstetrics and Gynecology, 222(4, Supplement), S873?S877. https://doi.org/10.1016/j.ajog.2019.11.1271 Braveman, P. A., Kumanyika, S., Fielding, J., LaVeist, T., Borrell, L. N., Manderscheid, R., & Troutman, A. (2011). Health Disparities and Health Equity: The Issue Is Justice. American Journal of Public Health, 101(Suppl 1), S149?S155. https://doi.org/10.2105/AJPH.2010.300062 253 Brown, W., Ottney, A., & Nguyen, S. (2011). Breaking the barrier: The Health Belief Model and patient perceptions regarding contraception. Contraception, 83(5), 453?458. https://doi.org/10.1016/j.contraception.2010.09.010 Caetano, C., Bliekendaal, S., Engler, Y., & Lombardo, M. (2020). From awareness to usage of long-acting reversible contraceptives: Results of a large European survey. International Journal of Gynecology & Obstetrics, 151(3), 366?376. https://doi.org/10.1002/ijgo.13363 Callaghan, W. M. (2012). Overview of Maternal Mortality in the United States. Seminars in Perinatology, 36(1), 2? 6. https://doi.org/10.1053/j.semperi.2011.09.002 Callahan, T., & Caughey, A. B. (2013). Blueprints Obstetrics and Gynecology. Lippincott Williams & Wilkins. Campbell, O. M., & Graham, W. J. (2006). Strategies for reducing maternal mortality: Getting on with what works. The Lancet, 368(9543), 1284?1299. https://doi.org/10.1016/S0140-6736(06)69381-1 Carlin, C. S., Fertig, A. R., & Dowd, B. E. (2016). Affordable Care Act?s Mandate Eliminating Contraceptive Cost Sharing Influenced Choices Of Women With Employer Coverage. Health Affairs, 35(9), 1608?1615. https://doi.org/10.1377/hlthaff.2015.1457 Carvajal, D. N., Gioia, D., Mudafort, E. R., Brown, P. B., & Barnet, B. (2017). How can Primary Care Physicians Best Support Contraceptive Decision Making? A Qualitative Study Exploring the Perspectives of Baltimore Latinas. Women?s Health Issues, 27(2), 158?166. https://doi.org/10.1016/j.whi.2016.09.015 Cavallaro, F. L., Benova, L., Owolabi, O. O., & Ali, M. (2020). A systematic review of the effectiveness of counselling strategies for modern contraceptive methods: What works and what doesn?t? BMJ Sexual & Reproductive Health, 46(4), 254?269. https://doi.org/10.1136/bmjsrh-2019-200377 Centers for Disease Control and Prevention. (1999). Ten Great Public Health Achievements?United States, 1900- 1999. https://www.cdc.gov/mmwr/preview/mmwrhtml/00056796.htm Centers for Disease Control and Prevention. (2020). Summary of Classifications for Hormonal Contraceptive Methods and Intrauterine Devices | CDC. Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion. https://www.cdc.gov/reproductivehealth/contraception/mmwr/mec/appendixk.html 254 Centers for Disease Control and Prevention (CDC). (1999). State-specific maternal mortality among Black and White women?United States, 1987-1996. MMWR. Morbidity and Mortality Weekly Report, 48(23), 492? 496. Cheng, D., Schwarz, E. B., Douglas, E., & Horon, I. (2009). Unintended pregnancy and associated maternal preconception, prenatal and postpartum behaviors. Contraception, 79(3), 194?198. https://doi.org/10.1016/j.contraception.2008.09.009 Chewning, B., Mosena, P., Wilson, D., Erdman, H., Potthoff, S., Murphy, A., & Kennedy Kuhnen, K. (1999). Evaluation of a computerized contraceptive decision aid for adolescent patients. Patient Education and Counseling, 38(3), 227?239. https://doi.org/10.1016/S0738-3991(99)00014-2 Choi, Y., Rendall, M., Boudreaux, M., & Roby, D. (2019). Summary of the Delaware Contraceptive Access Now (DelCAN) Initiative. DelCAN Initiative Evaluation. https://popcenter.umd.edu/delcaneval/summary-init Cicerchia, G., Reid, L. D., & Carvajal, D. N. (2022). Contraceptive Decision-Making and the Importance of Side Effect Information Among a Sample of Latinas. Women?s Health Reports, 3(1), 78?84. https://doi.org/10.1089/whr.2021.0115 Cleland, K., Zhu, H., Goldstuck, N., Cheng, L., & Trussell, J. (2012). The efficacy of intrauterine devices for emergency contraception: A systematic review of 35 years of experience. Human Reproduction, 27(7), 1994?2000. https://doi.org/10.1093/humrep/des140 Colorado Department of Public Health and Environment,. (2017). Colorado?s success with long-acting reversible contraception (LARC) | Department of Public Health & Environment. https://cdphe.colorado.gov/fpp/about-us/colorados-success-long-acting-reversible-contraception-larc Costa, A. R. R., Palma, F., S?, J. L., Vicente, L., Bombas, T., Nogueira, A. M., & Rocha, P. (2011). Impact of a women?s counselling programme on combined hormonal contraception in Portugal ? The IMAGINE Study. The European Journal of Contraception & Reproductive Health Care, 16(6), 409?417. https://doi.org/10.3109/13625187.2011.608441 255 Craig, A. D., Steinauer, J., Kuppermann, M., Schmittdiel, J. A., & Dehlendorf, C. (2019). Pill, patch or ring? A mixed methods analysis of provider counseling about combined hormonal contraception. Contraception, 99(2), 104?110. https://doi.org/10.1016/j.contraception.2018.09.001 Creanga, A. A., Syverson, C., Seed, K., & Callaghan, W. M. (2017). Pregnancy-Related Mortality in the United States, 2011?2013. Obstetrics and Gynecology, 130(2), 366?373. https://doi.org/10.1097/AOG.0000000000002114 Culwell, K. R., & Feinglass, J. (2007). The Association of Health Insurance with Use of Prescription Contraceptives. Perspectives on Sexual and Reproductive Health, 39(4), 226?230. https://doi.org/10.1363/3922607 Curtin, S. C., & Hoyert, D. L. (2017). Maternal Morbidity and Mortality: Exploring Racial/Ethnic Differences Using New Data from Birth and Death Certificates. In M. N. Hoque, B. Pecotte, & M. A. McGehee (Eds.), Applied Demography and Public Health in the 21st Century (pp. 95?113). Springer International Publishing. https://doi.org/10.1007/978-3-319-43688-3_7 Custo, G., Saitto, C., Cerza, S., & Sertoli, G. (1987). The adjusted contraceptive score (ACS) improves the overall performance of behavioural and barrier contraceptive methods. Advances in Contraceptive Delivery Systems: CDS, 3(4), 367?373. Danel, I., Berg, C., Johnson, C. H., & Atrash, H. (2003). Magnitude of Maternal Morbidity During Labor and Delivery: United States, 1993?1997. American Journal of Public Health, 93(4), 8. D?Angelo, D. V., Gilbert, B. C., Rochat, R. W., Santelli, J. S., & Herold, J. M. (2004). Differences Between Mistimed and Unwanted Pregnancies Among Women Who Have Live Births. Perspectives on Sexual and Reproductive Health, 36(5), 192?197. https://doi.org/10.1363/3619204 Daniels, K., & Abma, J. C. (2018). Current Contraceptive Status Among Women Aged 15?49: United States, 2015? 2017. NCHS Data Brief, 327, 8. Daniels, K., & Abma, J. C. (2020). Current Contraceptive Status Among Women Aged 15?49: United States, 2017? 2019. NCHS Data Brief, 388, 8. 256 Daniels, K., Daugherty, J., & Jones, J. (2014). Current Contraceptive Status Among Women Aged 15?44: United States, 2011?2013. NCHS Data Brief, 173, 8. Darney, B. G., Biel, F. M., Hoopes, M., Rodriguez, M. I., Hatch, B., Marino, M., Templeton, A., Oakley, J., Schmidt, T., & Cottrell, E. K. (2022). Title X Improved Access To Most Effective And Moderately Effective Contraception In US Safety-Net Clinics, 2016?18. Health Affairs, 41(4), 497?506. https://doi.org/10.1377/hlthaff.2021.01483 de Reilhac, P., Plu-Bureau, G., Serfaty, D., Letombe, B., Gondry, J., & Christin-Maitre, S. (2016). The CORALIE study: Improving patient education to help new users better understand their oral contraceptive. The European Journal of Contraception & Reproductive Health Care, 21(5), 388?394. https://doi.org/10.1080/13625187.2016.1217323 Dehlendorf, C., Diedrich, J., Drey, E., Postone, A., & Steinauer, J. (2010). Preferences for decision-making about contraception and general health care among reproductive age women at an abortion clinic. Patient Education and Counseling, 81(3), 343?348. https://doi.org/10.1016/j.pec.2010.06.021 Dehlendorf, C., Fox, E., Sobel, L., & Borrero, S. (2016). Patient-Centered Contraceptive Counseling: Evidence to Inform Practice. Current Obstetrics and Gynecology Reports, 5(1), 55?63. https://doi.org/10.1007/s13669- 016-0139-1 Dehlendorf, C., Grumbach, K., Schmittdiel, J. A., & Steinauer, J. (2017). Shared decision making in contraceptive counseling. Contraception, 95(5), 452?455. https://doi.org/10.1016/j.contraception.2016.12.010 Dehlendorf, C., Henderson, J. T., Vittinghoff, E., Grumbach, K., Levy, K., Schmittdiel, J., Lee, J., Schillinger, D., & Steinauer, J. (2016). Association of the quality of interpersonal care during family planning counseling with contraceptive use. American Journal of Obstetrics and Gynecology, 215(1), 78.e1-78.e9. https://doi.org/10.1016/j.ajog.2016.01.173 Dehlendorf, C., Kimport, K., Levy, K., & Steinauer, J. (2014). A Qualitative Analysis of Approaches To Contraceptive Counseling. Perspectives on Sexual and Reproductive Health, 46(4), 233?240. https://doi.org/10.1363/46e2114 257 Dehlendorf, C., Krajewski, C., & Borrero, S. (2014). Contraceptive Counseling: Best Practices to Ensure Quality Communication and Enable Effective Contraceptive Use. Clinical Obstetrics and Gynecology, 57(4), 659? 673. https://doi.org/10.1097/GRF.0000000000000059 Dehlendorf, C., Levy, K., Kelley, A., Grumbach, K., & Steinauer, J. (2013). Women?s preferences for contraceptive counseling and decision making. Contraception, 88(2), 250?256. https://doi.org/10.1016/j.contraception.2012.10.012 Dehlendorf, C., Park, S. Y., Emeremni, C. A., Comer, D., Vincett, K., & Borrero, S. (2014). Racial/ethnic disparities in contraceptive use: Variation by age and women?s reproductive experiences. American Journal of Obstetrics and Gynecology, 210(6), 526.e1-9. https://doi.org/10.1016/j.ajog.2014.01.037 Dehlendorf, C., Reed, R., Fitzpatrick, J., Kuppermann, M., Steinauer, J., & Kimport, K. (2019). A mixed-methods study of provider perspectives on My Birth Control: A contraceptive decision support tool designed to facilitate shared decision making. Contraception, 100(5), 420?423. https://doi.org/10.1016/j.contraception.2019.08.001 Dehlendorf, C., Rodriguez, M. I., Levy, K., Borrero, S., & Steinauer, J. (2010). Disparities in Family Planning. American Journal of Obstetrics and Gynecology, 202(3), 214?220. https://doi.org/10.1016/j.ajog.2009.08.022 Dehlendorf, C., Ruskin, R., Grumbach, K., Vittinghoff, E., Bibbins-Domingo, K., Schillinger, D., & Steinauer, J. (2010). Recommendations for intrauterine contraception: A randomized trial of the effects of patients? race/ethnicity and socioeconomic status. American Journal of Obstetrics and Gynecology, 203(4), 319.e1- 319.e8. https://doi.org/10.1016/j.ajog.2010.05.009 Delamater, J., Wagstaff, D. A., & Havens, K. K. (2000). The Impact of a Culturally Appropriate STD/AIDS Education Intervention on Black Male Adolescents? Sexual and Condom Use Behavior. Health Education & Behavior, 27(4), 454?470. https://doi.org/10.1177/109019810002700408 Dibaba, Y., Fantahun, M., & Hindin, M. J. (2013). The effects of pregnancy intention on the use of antenatal care services: Systematic review and meta-analysis. Reproductive Health, 10(1), 50. https://doi.org/10.1186/1742-4755-10-50 258 Diedrich, J. T., Zhao, Q., Madden, T., Secura, G. M., & Peipert, J. F. (2015). Three-year continuation of reversible contraception. American Journal of Obstetrics and Gynecology, 213(5), 662.e1-662.e8. https://doi.org/10.1016/j.ajog.2015.08.001 Dixon, S. C., Herbert, D. L., Loxton, D., & Lucke, J. C. (2014). ?As many options as there are, there are just not enough for me?: Contraceptive use and barriers to access among Australian women. The European Journal of Contraception & Reproductive Health Care, 19(5), 340?351. https://doi.org/10.3109/13625187.2014.919380 Donnelly, K. Z., Foster, T. C., & Thompson, R. (2014). What matters most? The content and concordance of patients? and providers? information priorities for contraceptive decision making. Contraception, 90(3), 280?287. https://doi.org/10.1016/j.contraception.2014.04.012 Dott, M., Rasmussen, S. A., Hogue, C. J., & Reefhuis, J. (2010). Association Between Pregnancy Intention and Reproductive-health Related Behaviors Before and After Pregnancy Recognition, National Birth Defects Prevention Study, 1997?2002. Maternal & Child Health Journal, 14(3), 373?381. https://doi.org/10.1007/s10995-009-0458-1 Downey, M. M., Arteaga, S., Villase?or, E., & Gomez, A. M. (2017). More Than a Destination: Contraceptive Decision Making as a Journey. Women?s Health Issues, 27(5), 539?545. https://doi.org/10.1016/j.whi.2017.03.004 Downing, R. A., LaVeist, T. A., & Bullock, H. E. (2007). Intersections of Ethnicity and Social Class in Provider Advice Regarding Reproductive Health. American Journal of Public Health, 97(10), 1803?1807. https://doi.org/10.2105/AJPH.2006.092585 Dye, T. D., Wojtowycz, M. A., Aubry, R. H., Quade, J., & Kilburn, H. (1997). Unintended pregnancy and breast- feeding behavior. American Journal of Public Health, 87(10), 1709?1711. https://doi.org/10.2105/ajph.87.10.1709 Everett, B. G., McCabe, K. F., & Hughes, T. L. (2017). Sexual Orientation Disparities in Mistimed and Unwanted Pregnancy Among Adult Women. Perspectives on Sexual and Reproductive Health, 49(3), 157?165. https://doi.org/10.1363/psrh.12032 259 Everett, B. G., Mollborn, S., Jenkins, V., Limburg, A., & Diamond, L. M. (2020). Racial/Ethnic Differences in Unwanted Births: Moderation by Sexual Orientation. Journal of Marriage and Family, 82(4), 1234?1249. https://doi.org/10.1111/jomf.12656 Finer, L. B., & Henshaw, S. K. (2006). Disparities in Rates of Unintended Pregnancy In the United States, 1994 and 2001. Perspectives on Sexual and Reproductive Health, 38(2), 90?96. https://doi.org/10.1363/3809006 Finer, L. B., & Zolna, M. R. (2016). Declines in Unintended Pregnancy in the United States, 2008-2011. The New England Journal of Medicine, 374(9), 843?852. https://doi.org/10.1056/NEJMsa1506575 Fishbein, M., Bandura, A., Triandis, H., Kanfer, F., Becker, M., Middlestadt, S., & m, A. (1992). Factors influencing behavior and behavior change: Final report-Theorist?s Workshop. 1992. National Institute of Mental Health. Foster, D. G., Grossman, D., Turok, D. K., Peipert, J. F., Prine, L., Schreiber, C. A., Jackson, A. V., Barar, R. E., & Schwarz, E. B. (2014). Interest in and experience with IUD self-removal. Contraception, 90(1), 54?59. https://doi.org/10.1016/j.contraception.2014.01.025 Foster, D. G., Raifman, S. E., Gipson, J. D., Rocca, C. H., & Biggs, M. A. (2019). Effects of Carrying an Unwanted Pregnancy to Term on Women?s Existing Children. The Journal of Pediatrics, 205, 183-189.e1. https://doi.org/10.1016/j.jpeds.2018.09.026 Fox, E., Reyna, A., Malcolm, N. M., Rosmarin, R. B., Zapata, L. B., Frederiksen, B. N., Moskosky, S. B., & Dehlendorf, C. (2018). Client Preferences for Contraceptive Counseling: A Systematic Review. American Journal of Preventive Medicine, 55(5), 691?702. https://doi.org/10.1016/j.amepre.2018.06.006 Frazier, T., Hogue, C. J. R., Bonney, E. A., Yount, K. M., & Pearce, B. D. (2018). Weathering the storm; a review of pre-pregnancy stress and risk of spontaneous abortion. Psychoneuroendocrinology, 92, 142?154. https://doi.org/10.1016/j.psyneuen.2018.03.001 Frederiksen, B. N., Ahrens, K. A., Moskosky, S., & Gavin, L. (2017). Does Contraceptive Use in the United States Meet Global Goals? Perspectives on Sexual and Reproductive Health, 49(4), 197?205. https://doi.org/10.1363/psrh.12042 260 French, V., & Darney, P. D. (2015). Implantable Contraception. In D. Shoupe & D. R. Mishell Jr (Eds.), The Handbook of Contraception: A Guide for Practical Management. Humana Press. Freundl, G., Sivin, I., & Bat?r, I. (2010). State-of-the-art of non-hormonal methods of contraception: IV. Natural family planning. European Journal of Contraception & Reproductive Health Care, 15(2), 113?123. https://doi.org/10.3109/13625180903545302 Frost, J. J., & Darroch, J. E. (2008). Factors Associated with Contraceptive Choice and Inconsistent Method Use, United States, 2004. Perspectives on Sexual and Reproductive Health, 40(2), 94?104. https://doi.org/10.1363/4009408 Frost, J. J., Finer, L. B., & Tapales, A. (2008). The Impact of Publicly Funded Family Planning Clinic Services on Unintended Pregnancies and Government Cost Savings. Journal of Health Care for the Poor and Underserved, 19(3), 778?796. https://doi.org/10.1353/hpu.0.0060 Frost, J. J., Singh, S., & Finer, L. B. (2007a). Factors Associated with Contraceptive Use and Nonuse, United States, 2004. Perspectives on Sexual and Reproductive Health, 39(2), 90?99. https://doi.org/10.1363/3909007 Frost, J. J., Singh, S., & Finer, L. B. (2007b). U.S. Women?s One-Year Contraceptive Use Patterns, 2004. Perspectives on Sexual and Reproductive Health, 39(1), 48?55. https://doi.org/10.1363/3904807 Gambera, A., Corda, F., Papa, R., Bastianelli, C., Bucciantini, S., Dessole, S., Scagliola, P., Bernardini, N., de Feo, D., & Beligotti, F. (2015). Observational, prospective, multicentre study to evaluate the effects of counselling on the choice of combined hormonal contraceptives in Italy?The ECOS (Educational COunselling effectS) study. BMC Women?s Health, 15(1), 69. https://doi.org/10.1186/s12905-015-0226-x Garbers, S., Chiasson, M. A., Baum, R., Tobier, N., Ventura, A., & Hirshfield, S. (2015). ?Get It and Forget It:? online evaluation of a theory-based IUD educational video in English and Spanish. Contraception, 91(1), 76?79. https://doi.org/10.1016/j.contraception.2014.09.002 Geronimus, A. T. (1992). The weathering hypothesis and the health of African-American women and infants: Evidence and speculations. Ethnicity & Disease, 2(3), 207?221. 261 Geronimus, A. T. (1996). Black/white differences in the relationship of maternal age to birthweight: A population- based test of the weathering hypothesis. Social Science & Medicine, 42(4), 589?597. https://doi.org/10.1016/0277-9536(95)00159-X Giho, Y., Jones, K. A., Dick, R. N., Gold, M. A., Talis, J. M., Gmelin, T. A., Laird, H. J., Vanek, M. S., & Miller, E. (2020). Feasibility and acceptability of using a web-based contraceptive support tool in a university health clinic. Journal of American College Health, 68(4), 336?340. https://doi.org/10.1080/07448481.2019.1577859 Gilliam, M. L., Martins, S. L., Bartlett, E., Mistretta, S. Q., & Holl, J. L. (2014). Development and testing of an iOS waiting room ?app? for contraceptive counseling in a Title X family planning clinic. American Journal of Obstetrics and Gynecology, 211(5), 481.e1-481.e8. https://doi.org/10.1016/j.ajog.2014.05.034 Gilliam, M. L., Warden, M., Goldstein, C., & Tapia, B. (2004). Concerns about contraceptive side effects among young Latinas: A focus-group approach. Contraception, 70(4), 299?305. https://doi.org/10.1016/j.contraception.2004.04.013 Gipson, J. D., Koenig, M. A., & Hindin, M. J. (2008a). The Effects of Unintended Pregnancy on Infant, Child, and Parental Health: A Review of the Literature. Studies in Family Planning, 39(1), 18?38. https://doi.org/10.1111/j.1728-4465.2008.00148.x Gipson, J. D., Koenig, M. A., & Hindin, M. J. (2008b). The Effects of Unintended Pregnancy on Infant, Child, and Parental Health: A Review of the Literature. Studies in Family Planning, 39(1), 18?38. https://doi.org/10.1111/j.1728-4465.2008.00148.x Glasier, A., Scorer, J., & Bigrigg, A. (2008). Attitudes of women in Scotland to contraception: A qualitative study to explore the acceptability of long-acting methods. BMJ Sexual & Reproductive Health, 34(4), 213?217. https://doi.org/10.1783/147118908786000497 Goffman, D., Madden, R. C., Harrison, E. A., Merkatz, I. R., & Chazotte, C. (2007). Predictors of maternal mortality and near-miss maternal morbidity. Journal of Perinatology, 27(10), 597?601. https://doi.org/10.1038/sj.jp.7211810 262 Gold, R. B. (2014). Guarding Against Coercion While Ensuring Access: A Delicate Balance (No. 17; 3; Guttmacher Policy Review). Guttmacher Institute. https://www.guttmacher.org/gpr/2014/09/guarding-against-coercion- while-ensuring-access-delicate-balance Gold, R. B., Sonfield, A., Frost, J. J., & Richards, C. L. (2009). Next Steps for America?s Family Planning Program: Leveraging the Potential of Medicaid and Title X in an Evolving Health Care System. Guttmacher Institute. https://www.guttmacher.org/report/next-steps-americas-family-planning-program-leveraging-potential- medicaid-and-title-x Gomez, A. M., & Clark, J. B. (2014). The Relationship Between Contraceptive Features Preferred by Young Women and Interest in IUDs: An Exploratory Analysis. Perspectives on Sexual and Reproductive Health, 46(3), 157?163. https://doi.org/10.1363/46e2014 Gomez, A. M., Fuentes, L., & Allina, A. (2014). Women or LARC First? Reproductive Autonomy and the Promotion of Long-Acting Reversible Contraceptive Methods. Perspectives on Sexual and Reproductive Health, 46(3), 171?175. https://doi.org/10.1363/46e1614 Gomez, A. M., & Wapman, M. (2017). Under (implicit) pressure: Young Black and Latina women?s perceptions of contraceptive care. Contraception, 96(4), 221?226. https://doi.org/10.1016/j.contraception.2017.07.007 Grady, W. R., Billy, J. O. G., & Klepinger, D. H. (2002). Contraceptive Method Switching in the United States. Perspectives on Sexual and Reproductive Health, 34(3), 135?145. https://doi.org/10.2307/3097712 Grady, W. R., Klepinger, D. H., & Nelson-Wally, A. (1999). Contraceptive Characteristics: The Perceptions and Priorities of Men and Women. Family Planning Perspectives, 31(4), 168?175. https://doi.org/10.2307/2991589 Greenberg, K. B., Makino, K. K., & Coles, M. S. (2013). Factors Associated With Provision of Long-Acting Reversible Contraception Among Adolescent Health Care Providers. Journal of Adolescent Health, 52(3), 372?374. https://doi.org/10.1016/j.jadohealth.2012.11.003 Grindlay, K., Burns, B., & Grossman, D. (2013). Prescription requirements and over-the-counter access to oral contraceptives: A global review. Contraception, 88(1), 91?96. https://doi.org/10.1016/j.contraception.2012.11.021 263 Grossman, D. A., Grindlay, K., Buchacker, T., Potter, J. E., & Schmertmann, C. P. (2012). Changes in Service Delivery Patterns After Introduction of Telemedicine Provision of Medical Abortion in Iowa. American Journal of Public Health, 103(1), 73?78. https://doi.org/10.2105/AJPH.2012.301097 Grossman, D., & Fuentes, L. (2013). Over-the-counter access to oral contraceptives as a reproductive healthcare strategy. Current Opinion in Obstetrics & Gynecology, 25(6), 500?505. https://doi.org/10.1097/GCO.0000000000000019 Grossman, D., Grindlay, K., Li, R., Potter, J. E., Trussell, J., & Blanchard, K. (2013). Interest in over-the-counter access to oral contraceptives among women in the United States. Contraception, 88(4), 544?552. https://doi.org/10.1016/j.contraception.2013.04.005 Grunloh, D. S., Casner, T., Secura, G. M., Peipert, J. F., & Madden, T. (2013). Characteristics associated with discontinuation of long-acting reversible contraception within the first 6 months of use. Obstetrics and Gynecology, 122(6), 1214?1221. https://doi.org/10.1097/01.AOG.0000435452.86108.59 Guendelman, S., Denny, C., Mauldon, J., & Chetkovich, C. (2000). Perceptions of Hormonal Contraceptive Safety and Side Effects Among Low-Income Latina and Non-Latina Women. Maternal and Child Health Journal, 4(4), 233?239. https://doi.org/10.1023/A:1026643621387 Guttmacher Institute. (2019). Unintended Pregnancy in the United States. https://www.guttmacher.org/fact- sheet/unintended-pregnancy-united-states Hall, J. A., Benton, L., Copas, A., & Stephenson, J. (2017). Pregnancy Intention and Pregnancy Outcome: Systematic Review and Meta-Analysis. Maternal and Child Health Journal, 21(3), 670?704. https://doi.org/10.1007/s10995-016-2237-0 Hall, K. S. (2012). The Health Belief Model Can Guide Modern Contraceptive Behavior Research and Practice. Journal of Midwifery & Women?s Health, 57(1), 74?81. https://doi.org/10.1111/j.1542-2011.2011.00110.x Halpern, V., Brache, V., Taylor, D., Lendvay, A., Coch?n, L., Jensen, J. T., & Dorflinger, L. J. (2021). Clinical trial to evaluate pharmacokinetics and pharmacodynamics of medroxyprogesterone acetate after subcutaneous administration of Depo-Provera. Fertility and Sterility, 115(4), 1035?1043. https://doi.org/10.1016/j.fertnstert.2020.11.002 264 Harper, C. C., Brown, B. A., Foster-Rosales, A., & Raine, T. R. (2010). Hormonal contraceptive method choice among young, low-income women: How important is the provider? Patient Education and Counseling, 81(3), 349?354. https://doi.org/10.1016/j.pec.2010.08.010 Harper, C. C., Rocca, C. H., Thompson, K. M., Morfesis, J., Goodman, S., Darney, P. D., Westhoff, C. L., & Speidel, J. J. (2015). Reductions in pregnancy rates in the USA with long-acting reversible contraception: A cluster randomised trial. The Lancet, 386(9993), 562?568. https://doi.org/10.1016/S0140- 6736(14)62460-0 Harrison, D. D., & Cooke, C. W. (1988). An elucidation of factors influencing physicians? willingness to perform elective female sterilization. Obstetrics and Gynecology, 72(4), 565?570. Hartnett, C. S., & Margolis, R. (2019). Births that are Later-than-Desired: Correlates and Consequences. Population Research and Policy Review, 38(4), 483?505. https://doi.org/10.1007/s11113-019-09513-6 Harvey, S. M., Beckman, L. J., & Murray, J. (1991). Perceived Contraceptive Attributes and Method Choice. Journal of Applied Social Psychology, 21(9), 774?790. https://doi.org/10.1111/j.1559-1816.1991.tb00548.x Haynes, M. C., Ryan, N., Saleh, M., Winkel, A. F., & Ades, V. (2017). Contraceptive Knowledge Assessment: Validity and reliability of a novel contraceptive research tool. Contraception, 95(2), 190?197. https://doi.org/10.1016/j.contraception.2016.09.002 Hebert, L. E., Hill, B. J., Quinn, M., Holl, J. L., Whitaker, A. K., & Gilliam, M. L. (2018). Mobile contraceptive application use in a clinical setting in addition to standard contraceptive counseling: A randomized controlled trial. Contraception, 98(4), 281?287. https://doi.org/10.1016/j.contraception.2018.07.001 Hellerstedt, W. L., Pirie, P. L., Lando, H. A., Curry, S. J., McBride, C. M., Grothaus, L. C., & Nelson, J. C. (1998). Differences in preconceptional and prenatal behaviors in women with intended and unintended pregnancies. American Journal of Public Health, 88(4), 663?666. https://doi.org/10.2105/AJPH.88.4.663 Henshaw, S. K. (1998). Unintended Pregnancy in the United States. Family Planning Perspectives, 30(1), 24?46. https://doi.org/10.2307/2991522 265 Herd, P., Higgins, J., Sicinski, K., & Merkurieva, I. (2016). The Implications of Unintended Pregnancies for Mental Health in Later Life. American Journal of Public Health, 106(3), 421?429. https://doi.org/10.2105/AJPH.2015.302973 Higgins, J. A. (2014). Celebration meets caution: LARC?s boons, potential busts, and the benefits of a reproductive justice approach. Contraception, 89(4), 237?241. https://doi.org/10.1016/j.contraception.2014.01.027 Higgins, J. A. (2017). Pregnancy Ambivalence and Long-Acting Reversible Contraceptive (LARC) Use Among Young Adult Women: A Qualitative Study. Perspectives on Sexual and Reproductive Health, 49(3), 149? 156. https://doi.org/10.1363/psrh.12025 Higgins, J. A., Wright, K. Q., Turok, D. K., & Sanders, J. N. (2020). Beyond safety and efficacy: Sexuality-related priorities and their associations with contraceptive method selection. Contraception: X, 2, 100038. https://doi.org/10.1016/j.conx.2020.100038 Hirshberg, A., & Srinivas, S. K. (2017). Epidemiology of maternal morbidity and mortality. Seminars in Perinatology, 41(6), 332?337. https://doi.org/10.1053/j.semperi.2017.07.007 Hoopes, A. J., Teal, S. B., Akers, A. Y., & Sheeder, J. (2018). Low Acceptability of Certain Contraceptive Methods among Young Women. Journal of Pediatric and Adolescent Gynecology, 31(3), 274?280. https://doi.org/10.1016/j.jpag.2017.11.008 Hoopes, A., Timko, C. A., & Akers, A. Y. (2020). What?s Known and What?s Next: Contraceptive Counseling and Support for Adolescents and Young Adult Women. Journal of Pediatric and Adolescent Gynecology. https://doi.org/10.1016/j.jpag.2020.12.008 Hubacher, D., Spector, H., Monteith, C., & Chen, P.-L. (2018). Not seeking yet trying long-acting reversible contraception: A 24-month randomized trial on continuation, unintended pregnancy and satisfaction. Contraception, 97(6), 524?532. https://doi.org/10.1016/j.contraception.2018.02.001 Huynh, L., McCoy, M., Law, A., Tran, K. N., Knuth, S., Lefebvre, P., Sullivan, S., & Duh, M. S. (2013). Systematic literature review of the costs of pregnancy in the US. PharmacoEconomics, 31(11), 1005?1030. https://doi.org/10.1007/s40273-013-0096-8 266 Inter-agency Working Group on Reproductive Health in Crises. (2010). Inter-Agency Field Manual on Reproductive Health in Humanitarian Settings: 2010 Revision for Field Review. In Inter-Agency Field Manual on Reproductive Health in Humanitarian Settings: 2010 Revision for Field Review. Inter-agency Working Group on Reproductive Health in Crises. https://www.ncbi.nlm.nih.gov/books/NBK305152/ Jaccard, J., Dodge, T., & Dittus, P. (2002). Parent-adolescent communication about sex and birth control: A conceptual framework. New Directions for Child and Adolescent Development, 2002(97), 9?42. https://doi.org/10.1002/cd.48 Jaccard, J., Helbig, D. W., Wan, C. K., Gutman, M. A., & Kritz-Silverstein, D. C. (1996). The Prediction of Accurate Contraceptive Use From Attitudes and Knowledge. Health Education Quarterly, 23(1), 17?33. https://doi.org/10.1177/109019819602300102 Jain, J., Dutton, C., Nicosia, A., Wajszczuk, C., Bode, F. R., & Mishell, D. R. (2004). Pharmacokinetics, ovulation suppression and return to ovulation following a lower dose subcutaneous formulation of Depo-Provera?. Contraception, 70(1), 11?18. https://doi.org/10.1016/j.contraception.2004.01.011 Jamin, C. G., H?usler, G., Abascal, P. L., Fiala, C., Lasa, L. I. L., Nappi, R. E., Micheletti, M.-C., Fern?ndez- Dorado, A., Pintiaux, A., & Chabbert-Buffet, N. (2017). Development and conceptual validation of a questionnaire to help contraceptive choice: CHLOE (Contraception: HeLping for wOmen?s choicE). The European Journal of Contraception & Reproductive Health Care, 22(4), 286?290. https://doi.org/10.1080/13625187.2017.1364719 Johnson, S., Pion, C., & Jennings, V. (2013). Current methods and attitudes of women towards contraception in Europe and America. Reproductive Health, 10(1), 7. https://doi.org/10.1186/1742-4755-10-7 Jones, J., Mosher, W. D., & Daniels, K. (2012). Current Contraceptive Use in the United States, 2006?2010, and Changes in Patterns of Use Since 1995 (p. 26) [National Health Statistics Reports]. U.S. Department of Health and Human Services. Jones, R. K., & Jerman, J. (2017). Population Group Abortion Rates and Lifetime Incidence of Abortion: United States, 2008?2014. American Journal of Public Health, 107(12), 1904?1909. https://doi.org/10.2105/AJPH.2017.304042 267 Jones, R. K., Singh, S., Finer, L. B., & Frohwirth, L. F. (2006). Repeat abortion in the United States (Occasional Report No. 9). Guttmacher Institute. https://www.policyarchive.org/handle/10207/5886 Jones, R. K., Tapales, A., Lindberg, L. D., & Frost, J. (2015). Using Longitudinal Data to Understand Changes In Consistent Contraceptive Use. Perspectives on Sexual and Reproductive Health, 47(3), 131?139. https://doi.org/10.1363/47e4615 Joseph, K. S., Lisonkova, S., Muraca, G. M., Razaz, N., Sabr, Y., Mehrabadi, A., & Schisterman, E. F. (2017). Factors Underlying the Temporal Increase in Maternal Mortality in the United States. Obstetrics and Gynecology, 129(1), 91?100. https://doi.org/10.1097/AOG.0000000000001810 Joyce, T. J., Kaestner, R., & Korenman, S. (2000a). The effect of pregnancy intention on child development. Demography, 37(1), 83?94. https://doi.org/10.2307/2648098 Joyce, T. J., Kaestner, R., & Korenman, S. (2000b). The Stability of Pregnancy Intentions and Pregnancy-Related Maternal Behaviors. Maternal and Child Health Journal, 4(3), 171?178. https://doi.org/10.1023/A:1009571313297 Kaiser Family Foundation. (2019, October 1). Contraceptive Implants. KFF. https://www.kff.org/womens-health- policy/fact-sheet/contraceptive-implants/ Kaiser Family Foundation. (2020, September 9). Intrauterine Devices (IUDs): Access for Women in the U.S. KFF. https://www.kff.org/womens-health-policy/fact-sheet/intrauterine-devices-iuds-access-for-women-in-the-u- s/ Kaneshiro, B., Kon, Z., Tschann, M., Williams, A., Kajiwara, K., & Soon, R. (2020). Meeting Women?s Requests for Intrauterine Device and Contraceptive Implant Discontinuation: An Exploratory Survey of Physicians. Hawai?i Journal of Health & Social Welfare, 79(10), 296?301. Kaunitz, A., Hughes, J., Grimes, D., Smith, J., Rochat, R., & Kafrissen, M. (1985). Causes of maternal mortality in the United States. Obstetrics and Gynecology, 65(5), 605?612. Kavanaugh, M. L., Frohwirth, L., Jerman, J., Popkin, R., & Ethier, K. (2013). Long-acting Reversible Contraception for Adolescents and Young Adults: Patient and Provider Perspectives. Journal of Pediatric and Adolescent Gynecology, 26(2), 86?95. https://doi.org/10.1016/j.jpag.2012.10.006 268 Kaye, K., Suellentrop, K., & Sloup, C. (2009). The Fog Zone: How Misperceptions, Magical Thinking, and Ambivalence Put Young Adults at Risk for Unplanned Pregnancy. Power to Decide (formerly The National Campaign to Prevent Teen and Unplanned Pregnancy). https://powertodecide.org/what-we- do/information/resource-library/fog-zone Kelly, A., Lindo, J. M., & Packham, A. (2020). The power of the IUD: Effects of expanding access to contraception through Title X clinics. Journal of Public Economics, 192, 104288. https://doi.org/10.1016/j.jpubeco.2020.104288 Kemet, S., Lundsberg, L. S., & Gariepy, A. M. (2018). Race and ethnicity may not be associated with risk of unintended pregnancy. Contraception, 97(4), 313?318. https://doi.org/10.1016/j.contraception.2017.12.014 Kilbourne, A. M., Switzer, G., Hyman, K., Crowley-Matoka, M., & Fine, M. J. (2006). Advancing Health Disparities Research Within the Health Care System: A Conceptual Framework. American Journal of Public Health, 96(12), 2113?2121. https://doi.org/10.2105/AJPH.2005.077628 Kim, S., Im, E.-O., Liu, J., & Ulrich, C. (2020). Maternal Age Patterns of Preterm Birth: Exploring the Moderating Roles of Chronic Stress and Race/Ethnicity. Annals of Behavioral Medicine, 54(9), 653?664. https://doi.org/10.1093/abm/kaaa008 Klein, D. A., Arnold, J. J., & Reese, E. S. (2015). Provision of Contraception: Key Recommendations from the CDC. American Family Physician, 91(9), 625?633. Klerman, L. V. (2000). The Intendedness of Pregnancy: A Concept in Transition. Maternal and Child Health Journal, 4(3), 155?162. https://doi.org/10.1023/A:1009534612388 Kofinas, J. D., Varrey, A., Sapra, K. J., Kanj, R. V., Chervenak, F. A., & Asfaw, T. (2014). Adjunctive Social Media for More Effective Contraceptive Counseling: A Randomized Controlled Trial. Obstetrics & Gynecology, 123(4), 763?770. https://doi.org/10.1097/AOG.0000000000000172 Koren, A., & Mawn, B. (2010). The context of unintended pregnancy among married women in the USA. BMJ Sexual & Reproductive Health, 36(3), 150?158. https://doi.org/10.1783/147118910791749380 269 Korenman, S., Kaestner, R., & Joyce, T. J. (2002). Consequences for Infants of Parental Disagreement in Pregnancy Intention. Perspectives on Sexual and Reproductive Health, 34(4), 198?205. https://doi.org/10.2307/3097730 Kost, K. (2015). Unintended pregnancy rates at the state level: Estimates for 2010 and trends since 2002 (p. 19). Guttmacher Institute. https://www.guttmacher.org/report/unintended-pregnancy-rates-state-level-estimates- 2010-and-trends-2002 Kost, K., Landry, D. J., & Darroch, J. E. (1998). Predicting Maternal Behaviors During Pregnancy: Does Intention Status Matter? Family Planning Perspectives, 30(2), 79?88. https://doi.org/10.2307/2991664 Kost, K., & Lindberg, L. (2015). Pregnancy intentions, maternal behaviors, and infant health: Investigating relationships with new measures and propensity score analysis. Demography, 52(1), 83?111. https://doi.org/10.1007/s13524-014-0359-9 Kost, K., Maddow-Zimet, I., & Arpaia, A. (2017). Pregnancies, Births and Abortions Among Adolescents and Young Women In the United States, 2013: National and State Trends by Age, Race and Ethnicity (p. 71). Guttmacher Institute. Kost, K., Maddow-Zimet, I., & Kochhar, S. (2018). Pregnancy Desires and Pregnancies at the State Level: Estimates for 2014. Guttmacher Institute. https://doi.org/10.1363/2018.30238 Lamvu, G., Steiner, M. J., Condon, S., & Hartmann, K. (2006). Consistency between most important reasons for using contraception and current method used: The influence of health care providers. Contraception, 73(4), 399?403. https://doi.org/10.1016/j.contraception.2005.10.007 Landau, S. C., Tapias, M. P., & McGhee, B. T. (2006). Birth control within reach: A national survey on women?s attitudes toward and interest in pharmacy access to hormonal contraception. Contraception, 74(6), 463? 470. https://doi.org/10.1016/j.contraception.2006.07.006 Lang, C. T., & King, J. C. (2008). Maternal mortality in the United States. Best Practice & Research Clinical Obstetrics & Gynaecology, 22(3), 517?531. https://doi.org/10.1016/j.bpobgyn.2007.10.004 270 Larsson, M., Eurenius, K., Westerling, R., & Tyd?n, T. (2004a). Emergency contraceptive pills in Sweden: Evaluation of an information campaign. BJOG: An International Journal of Obstetrics & Gynaecology, 111(8), 820?827. https://doi.org/10.1111/j.1471-0528.2004.00206.x Larsson, M., Eurenius, K., Westerling, R., & Tyd?n, T. (2004b). Emergency contraceptive pills over-the-counter: A population-based survey of young Swedish women. Contraception, 69(4), 309?315. https://doi.org/10.1016/j.contraception.2003.11.013 LaVeist, T. A. (2005). Disentangling race and socioeconomic status: A key to understanding health inequalities. Journal of Urban Health, 82(3), iii26?iii34. https://doi.org/10.1093/jurban/jti061 Lee, J. K., Papic, M., Baldauf, E., Updike, G., & Schwarz, E. B. (2015). A checklist approach to caring for women seeking pregnancy testing: Effects on contraceptive knowledge and use. Contraception, 91(2), 143?149. https://doi.org/10.1016/j.contraception.2014.11.003 Lee, J. K., Parisi, S. M., & Schwarz, E. B. (2013). Contraceptive Counseling and Use among Women with Poorer Health. Journal of Women?s Health, Issues & Care, 2(1), 103. https://doi.org/10.4172/2325-9795.1000103 Lessard, L. N., Karasek, D., Ma, S., Darney, P., Deardorff, J., Lahiff, M., Grossman, D., & Foster, D. G. (2012). Contraceptive Features Preferred by Women At High Risk of Unintended Pregnancy. Perspectives on Sexual and Reproductive Health, 44(3), 194?200. https://doi.org/10.1363/4419412 Lete, I., Doval, J. L., P?rez-Campos, E., S?nchez-Borrego, R., Correa, M., de la Viuda, E., G?mez, M. A., Gonz?lez, J. V., Lertxundi, R., Mart?nez, M. T., Mendoza, N., & Robledo, J. (2007). Factors affecting women?s selection of a combined hormonal contraceptive method: The TEAM-06 Spanish cross-sectional study. Contraception, 76(2), 77?83. https://doi.org/10.1016/j.contraception.2007.04.014 Lewis, C., Darney, P., & Thiel de Bocanegra, H. (2013). Intrauterine contraception: Impact of provider training on participant knowledge and provision. Contraception, 88(2), 226?231. https://doi.org/10.1016/j.contraception.2013.06.004 Lindo, J. M., & Packham, A. (2017). How Much Can Expanding Access to Long-Acting Reversible Contraceptives Reduce Teen Birth Rates? American Economic Journal: Economic Policy, 9(3), 348?376. https://doi.org/10.1257/pol.20160039 271 Little, P., Griffin, S., Kelly, J., Dickson, N., & Sadler, C. (1998). Effect of educational leaflets and questions on knowledge of contraception in women taking the combined contraceptive pill: Randomised controlled trial. BMJ, 316(7149), 1948?1952. https://doi.org/10.1136/bmj.316.7149.1948 Littlejohn, K. E. (2013). ?It?s those Pills that are Ruining Me?: Gender and the Social Meanings of Hormonal Contraceptive Side Effects. Gender & Society, 27(6), 843?863. https://doi.org/10.1177/0891243213504033 Lopez, L. M., Grey, T. W., Chen, M., Tolley, E. E., & Stockton, L. L. (2016). Theory?based interventions for contraception. Cochrane Database of Systematic Reviews, 11. https://doi.org/10.1002/14651858.CD007249.pub5 Luna, Z., & Luker, K. (2013). Reproductive Justice. Annual Review of Law and Social Science, 9(1), 327?352. https://doi.org/10.1146/annurev-lawsocsci-102612-134037 Madden, T., Mullersman, J. L., Omvig, K. J., Secura, G. M., & Peipert, J. F. (2013). Structured contraceptive counseling provided by the Contraceptive CHOICE Project. Contraception, 88(2), 243?249. https://doi.org/10.1016/j.contraception.2012.07.015 Madden, T., Secura, G. M., Nease, R. F., Politi, M. C., & Peipert, J. F. (2015). The role of contraceptive attributes in women?s contraceptive decision making. American Journal of Obstetrics and Gynecology, 213(1), 46.e1- 46.e6. https://doi.org/10.1016/j.ajog.2015.01.051 Mahony, H., Spinner, C., Vamos, C. A., & Daley, E. M. (2021). Social Network Influences on Young Women?s Choice to Use Long-Acting Reversible Contraception: A Systematic Review. Journal of Midwifery & Women?s Health, 66(6), 758?771. https://doi.org/10.1111/jmwh.13280 Mann, E. S. (2022). The power of persuasion: Normative accountability and clinicians? practices of contraceptive counseling. SSM - Qualitative Research in Health, 2, 100049. https://doi.org/10.1016/j.ssmqr.2022.100049 Mansour, D. (2014). International survey to assess women?s attitudes regarding choice of daily versus nondaily female hormonal contraception. International Journal of Women?s Health, 6, 367?375. https://doi.org/10.2147/IJWH.S59059 272 Manzer, J. L., & Bell, A. V. (2021). ?We?re a Little Biased?: Medicine and the Management of Bias through the Case of Contraception. Journal of Health and Social Behavior, 62(2), 120?135. https://doi.org/10.1177/00221465211003232 Manzer, J. L., & Bell, A. V. (2022a). ?Did I Choose a Birth Control Method Yet??: Health Care and Women?s Contraceptive Decision-Making. Qualitative Health Research, 32(1), 80?94. https://doi.org/10.1177/10497323211004081 Manzer, J. L., & Bell, A. V. (2022b). The limitations of patient-centered care: The case of early long-acting reversible contraception (LARC) removal. Social Science & Medicine, 292, 114632. https://doi.org/10.1016/j.socscimed.2021.114632 Marcinkowski, A., Gauf, A., Goedken, P., Sales, J., Brown, J., & Kottke, M. (2021). 7. Using the Theory of Planned Behavior to Identify Predictors of Contraceptive Use Intentions and Behaviors in Adolescents. Journal of Pediatric and Adolescent Gynecology, 34(2), 242. https://doi.org/10.1016/j.jpag.2021.02.011 Marshall, C., Guendelman, S., Mauldon, J., & Nuru?Jeter, A. (2016). Young Women?s Contraceptive Decision Making: Do Preferences for Contraceptive Attributes Align with Method Choice? Perspectives on Sexual and Reproductive Health, 48(3), 119?127. https://doi.org/10.1363/48e10116 Marshall, C., Kandahari, N., & Raine-Bennett, T. (2018). Exploring young women?s decisional needs for contraceptive method choice: A qualitative study. Contraception, 97(3), 243?248. https://doi.org/10.1016/j.contraception.2017.10.004 Marsiglio, W., & Mott, F. L. (1988). Does Wanting to Become Pregnant with a First Child Affect Subsequent Maternal Behaviors and Infant Birth Weight? Journal of Marriage and the Family, 50(4), 1023. https://doi.org/10.2307/352112 Marthey, D., Rashid, H., Xie, L., & Boudreaux, M. (2021). An evaluation of the Be Your Own Baby public awareness campaign. Health Services Research, 56(5), 766?776. https://doi.org/10.1111/1475-6773.13698 Martinez, G. M., & Abma, J. C. (2020). Sexual Activity and Contraceptive Use Among Teenagers Aged 15?19 in the United States, 2015?2017. NCHS Data Brief, 366, 8. 273 Mason, V., McEwan, A., Walker, D., Barrett, S., & James, D. (2003). The use of video information in obtaining consent for female sterilisation: A randomised study. BJOG: An International Journal of Obstetrics & Gynaecology, 110(12), 1062?1071. https://doi.org/10.1111/j.1471-0528.2003.03041.x Mayer, J. P. (1997). Unintended Childbearing, Maternal Beliefs, and Delay of Prenatal Care. Birth, 24(4), 247?252. https://doi.org/10.1111/j.1523-536X.1997.00247.pp.x Mazza, D., Watson, C. J., Taft, A., Lucke, J., McGeechan, K., Haas, M., McNamee, K., Peipert, J. F., & Black, K. I. (2020). Increasing long-acting reversible contraceptives: The Australian Contraceptive ChOice pRoject (ACCORd) cluster randomized trial. American Journal of Obstetrics and Gynecology, 222(4, Supplement), S921.e1-S921.e13. https://doi.org/10.1016/j.ajog.2019.11.1267 McCauley, H. L., Silverman, J. G., Jones, K. A., Tancredi, D. J., Decker, M. R., McCormick, M. C., Austin, S. B., Anderson, H. A., & Miller, E. (2017). Psychometric properties and refinement of the Reproductive Coercion Scale. Contraception, 95(3), 292?298. https://doi.org/10.1016/j.contraception.2016.09.010 McCrory, C., & McNally, S. (2013). The effect of pregnancy intention on maternal prenatal behaviours and parent and child health: Results of an irish cohort study. Paediatric and Perinatal Epidemiology, 27(2), 208?215. https://doi.org/10.1111/ppe.12027 McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An Ecological Perspective on Health Promotion Programs. Health Education Quarterly, 15(4), 351?377. https://doi.org/10.1177/109019818801500401 McNicholas, C., Tessa, M., Secura, G., & Peipert, J. F. (2014). The Contraceptive CHOICE Project Round Up: What we did and what we learned. Clinical Obstetrics and Gynecology, 57(4), 635?643. https://doi.org/10.1097/GRF.0000000000000070 Melo, J., Peters, M., Teal, S., & Guiahi, M. (2015). Adolescent and Young Women?s Contraceptive Decision- Making Processes: Choosing ?The Best Method for Her.? Journal of Pediatric and Adolescent Gynecology, 28(4), 224?228. https://doi.org/10.1016/j.jpag.2014.08.001 Menon, S., COMMITTEE ON ADOLESCENCE, Alderman, E. M., Chung, R. J., Grubb, L. K., Lee, J., Powers, M. E., Upadhya, K. K., & Wallace, S. B. (2020). Long-Acting Reversible Contraception: Specific Issues for Adolescents. Pediatrics, 146(2), e2020007252. https://doi.org/10.1542/peds.2020-007252 274 Mestad, R., Secura, G., Allsworth, J. E., Madden, T., Zhao, Q., & Peipert, J. F. (2011). Acceptance of long-acting reversible contraceptive methods by adolescent participants in the Contraceptive CHOICE Project. Contraception, 84(5), 493?498. https://doi.org/10.1016/j.contraception.2011.03.001 Michie, L., Cameron, S. T., Glasier, A., & Johnstone, A. (2016). Giving information about the contraceptive implant using a DVD: Is it acceptable and informative? A pilot randomised study. Journal of Family Planning and Reproductive Health Care, 42(3), 194?200. https://doi.org/10.1136/jfprhc-2015-101186 Mogos, M. F., Liese, K. L., Thornton, P. D., Manuck, T. A., O?Brien, W. D. J., & McFarlin, B. L. (2020). Inpatient Maternal Mortality in the United States, 2002?2014. Nursing Research, 69(1), 42?50. https://doi.org/10.1097/NNR.0000000000000397 Mohllajee, A. P., Curtis, K. M., Morrow, B., & Marchbanks, P. A. (2007). Pregnancy Intention and Its Relationship to Birth and Maternal Outcomes. Obstetrics & Gynecology, 109(3), 678?686. https://doi.org/10.1097/01.AOG.0000255666.78427.c5 Monea, E., & Thomas, A. (2011). Unintended Pregnancy and Taxpayer Spending. Perspectives on Sexual and Reproductive Health, 43(2), 88?93. https://doi.org/10.1363/4308811 Monta?o, D., & Kasprzyk, D. (2008). Theory of reasoned action, theory of planned behavior, and the integrated behavioral model. In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education: Theory, research and practice (4th ed., Vol. 70, p. 231). Monta?o, D., & Kasprzyk, D. (2015). Theory of reasoned action, theory of planned behavior, and the integrated behavioral model. In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior: Theory, research and practice (5th ed.). John Wiley & Sons. Montgomery, T. M., Stephens-Shields, A. J., Schapira, M. M., & Akers, A. Y. (2020). Dual-Method Contraception Use Among Young Women Pre- and Post-ACA Implementation. Policy, Politics, & Nursing Practice, 21(3), 140?150. https://doi.org/10.1177/1527154420923747 Moreau, C., Cleland, K., & Trussell, J. (2007). Contraceptive discontinuation attributed to method dissatisfaction in the United States. Contraception, 76(4), 267?272. https://doi.org/10.1016/j.contraception.2007.06.008 275 Moreau, C., Hall, K. S., Trussell, J., & Barber, J. (2013). Effect of prospectively measured pregnancy intentions on the consistency of contraceptive use among young women in Michigan. Human Reproduction, 28(3), 642? 650. https://doi.org/10.1093/humrep/des421 Morse, J. E., Ramesh, S., & Jackson, A. (2017). Reassessing Unintended Pregnancy: Toward a Patient-centered Approach to Family Planning. Obstetrics and Gynecology Clinics, 44(1), 27?40. https://doi.org/10.1016/j.ogc.2016.10.003 Mosher, W. D., & Jones, J. (2010). Use of contraception in the United States: 1982-2008 (No. 29; Vital and Health Statistics. Series 23). U.S. Dept. of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. Mosher, W. D., Jones, J., & Abma, J. C. (2012). Intended and unintended births in the United States: 1982-2010. National Health Statistics Reports, 55, 1?28. Mumford, S. L., Sapra, K. J., King, R. B., Louis, J. F., & Buck Louis, G. M. (2016). Pregnancy intentions?A complex construct and call for new measures. Fertility and Sterility, 106(6), 1453?1462. https://doi.org/10.1016/j.fertnstert.2016.07.1067 Murphy, M. K., Burke, P. J., & Haider, S. (2017). A Qualitative Application of Diffusion of Innovations to Adolescents? Perceptions of Long-Acting Reversible Contraception?s Attributes. Journal of Pediatric and Adolescent Gynecology, 30(4), 484?490. https://doi.org/10.1016/j.jpag.2016.11.005 Musick, K. (2002). Planned and Unplanned Childbearing Among Unmarried Women. Journal of Marriage and Family, 64(4), 915?929. https://doi.org/10.1111/j.1741-3737.2002.00915.x Nappi, R. E., Vermuyten, N., & Bannemerschult, R. (2022). Missed opportunities in contraceptive counselling: Findings from a European survey-based study with simulated patient consultation. The European Journal of Contraception & Reproductive Health Care, 27(2), 85?94. https://doi.org/10.1080/13625187.2021.2010040 National Center for Health Statistics. (2019a). Key Statistics from the National Survey of Family Growth?C Listing. https://www.cdc.gov/nchs/nsfg/key_statistics/c.htm 276 National Center for Health Statistics. (2019b). Key Statistics from the National Survey of Family Growth?NSFG - Listing C. https://www.cdc.gov/nchs/nsfg/key_statistics/c_2015-2017.htm Nobili, M. P., Piergrossi, S., Brusati, V., & Moja, E. A. (2007). The effect of patient-centered contraceptive counseling in women who undergo a voluntary termination of pregnancy. Patient Education and Counseling, 65(3), 361?368. https://doi.org/10.1016/j.pec.2006.09.004 O?Donnell, L., Doval, A. S., Duran, R., & O?Donnell, C. R. (1995). The Effectiveness of Video-Based Interventions in Promoting Condom Acquisition Among STD Clinic Patients. Sexually Transmitted Diseases, 22(2), 97? 103. Office of Disease Prevention and Health Promotion. (2020). Reduce the proportion of unintended pregnancies? FP?01. Healthy People 2030. https://health.gov/healthypeople/objectives-and-data/browse- objectives/family-planning/reduce-proportion-unintended-pregnancies-fp-01 Office of Women?s Health. (2017). Birth control methods. Womenshealth.Gov. https://www.womenshealth.gov/a-z- topics/birth-control-methods Orr, S. T., Miller, C. A., James, S. A., & Babones, S. (2000). Unintended pregnancy and preterm birth. Paediatric and Perinatal Epidemiology, 14(4), 309?313. https://doi.org/10.1046/j.1365-3016.2000.00289.x Ott, M. A., Sucato, G. S., & Committee on Adolescence et al. (2014). Contraception for Adolescents. Pediatrics, 134(4), e1257?e1281. https://doi.org/10.1542/peds.2014-2300 Owen, C. M., Goldstein, E. H., Clayton, J. A., & Segars, J. H. (2013). Racial and Ethnic Health Disparities in Reproductive Medicine: An Evidence-Based Overview. Seminars in Reproductive Medicine, 31(5), 317? 324. https://doi.org/10.1055/s-0033-1348889 Pardthaisong, T., Gray, RonaldH., & Mcdaniel, EdwinB. (1980). Return of fertility after dicontinuation of depot medroxyprogesterone acetate and intra-uterine devices in Northern Thailand. The Lancet, 315(8167), 509? 512. https://doi.org/10.1016/S0140-6736(80)92765-8 Patel, P. R., Laz, T. H., & Berenson, A. B. (2014). Patient Characteristics Associated with Pregnancy Ambivalence. Journal of Women?s Health, 24(1), 37?41. https://doi.org/10.1089/jwh.2014.4924 277 Paulen, M. E., & Curtis, K. M. (2009). When can a woman have repeat progestogen-only injectables?depot medroxyprogesterone acetate or norethisterone enantate? Contraception, 80(4), 391?408. https://doi.org/10.1016/j.contraception.2009.03.023 Pazol, K., Zapata, L. B., Dehlendorf, C., Malcolm, N. M., Rosmarin, R. B., & Frederiksen, B. N. (2018). Impact of Contraceptive Education on Knowledge and Decision Making: An Updated Systematic Review. American Journal of Preventive Medicine, 55(5), 703?715. https://doi.org/10.1016/j.amepre.2018.07.012 Pedrazzini, A., McGowan, H., Lucking, L., & Johanson, R. (2000). ?The trouble with sex ? it always gets in the way?: An evaluation of a peer-produced teenage pregnancy video. BMJ Sexual & Reproductive Health, 26(3), 131?134. https://doi.org/10.1783/147118900101194544 Peipert, J. F., Madden, T., Allsworth, J. E., & Secura, G. M. (2012). Preventing Unintended Pregnancies by Providing No-Cost Contraception. Obstetrics and Gynecology, 120(6), 1291?1297. Philliber Research Associates, B. C. for G. R. H. (2012). Reducing Unintended Pregnancies in Iowa by Investing in Title X Clinics. https://www.astho.org/Maternal-and-Child-Health/Long-Acting-Reversible- Contraception/Iowa-Initiative-Title-X-Issue-Brief/ Ragland, D., Payakachat, N., Ounpraseuth, S., Pate, A., Harrod, S. E., & Ott, R. E. (2011). Emergency contraception counseling: An opportunity for pharmacists. Journal of the American Pharmacists Association, 51(6), 756? 761. https://doi.org/10.1331/JAPhA.2011.10157 Ragland, D., Payakachat, N., & Stafford, R. A. (2015). Emergency Contraception Counseling in a Retail Pharmacy Setting: A Pilot Study. Journal of Pharmacy Practice, 28(3), 261?265. https://doi.org/10.1177/0897190013516507 Raine, T. R., Foster-Rosales, A., Upadhyay, U. D., Boyer, C. B., Brown, B. A., Sokoloff, A., & Harper, C. C. (2011). One-Year Contraceptive Continuation and Pregnancy in Adolescent Girls and Women Initiating Hormonal Contraceptives. Obstetrics and Gynecology, 117(2 Pt 1), 363?371. https://doi.org/10.1097/AOG.0b013e31820563d3 Ranji, U., Gomez, I., Salganicoff, A., Rosenzweig, C., Kellenberg, R., & Gifford, K. (2022, February 17). Medicaid Coverage of Family Planning Benefits: Findings from a 2021 State Survey. KFF. 278 https://www.kff.org/womens-health-policy/report/medicaid-coverage-of-family-planning-benefits-findings- from-a-2021-state-survey/ Raymond, E. G., & Grimes, D. A. (2012). The Comparative Safety of Legal Induced Abortion and Childbirth in the United States. Obstetrics & Gynecology, 119(2 Part 1), 215?219. https://doi.org/10.1097/AOG.0b013e31823fe923 Ricketts, S., Klingler, G., & Schwalberg, R. (2014). Game Change in Colorado: Widespread Use Of Long-Acting Reversible Contraceptives and Rapid Decline in Births Among Young, Low-Income Women. Perspectives on Sexual and Reproductive Health, 46(3), 125?132. https://doi.org/10.1363/46e1714 Robbins, C. L., Gavin, L., Carter, M. W., & Moskosky, S. B. (2017). The Link Between Reproductive Life Plan Assessment And Provision of Preconception Care At Publicly Funded Health Centers. Perspectives on Sexual and Reproductive Health, 49(3), 167?172. https://doi.org/10.1363/psrh.12030 Roberts, D. (2000). Black Women and the Pill. Family Planning Perspectives, 32(2), 92. https://doi.org/10.2307/2648220 Roberts, D. (2015). Reproductive Justice, Not Just Rights. Dissent, 62(4), 79?82. https://doi.org/10.1353/dss.2015.0073 Rocca, C. H., & Harper, C. C. (2012). Do Racial and Ethnic Differences in Contraceptive Attitudes and Knowledge Explain Disparities In Method Use? Perspectives on Sexual and Reproductive Health, 44(3), 150?158. https://doi.org/10.1363/4415012 Romero, D., & Ag?nor, M. (2009). US Fertility Prevention as Poverty Prevention: An Empirical Question and Social Justice Issue. Women?s Health Issues, 19(6), 355?364. https://doi.org/10.1016/j.whi.2009.08.004 Rosenstock, J. R., Peipert, J. F., Madden, T., Zhao, Q., & Secura, G. M. (2012). Continuation of Reversible Contraception in Teenagers and Young Women. Obstetrics and Gynecology, 120(6), 1298?1305. Ross, L., & Solinger, R. (2017). Reproductive Justice: An Introduction. Univ of California Press. Rozin, P., & Royzman, E. B. (2001). Negativity Bias, Negativity Dominance, and Contagion. Personality and Social Psychology Review, 5(4), 296?320. https://doi.org/10.1207/S15327957PSPR0504_2 279 Rubin, V., & East, P. L. (1999). Adolescents? pregnancy intentions: Relations to life situations and caretaking behaviors prenatally and 2 years postpartum. Journal of Adolescent Health, 24(5), 313?320. https://doi.org/10.1016/S1054-139X(98)00082-2 Russo, J. A., Miller, E., & Gold, M. A. (2013). Myths and Misconceptions About Long-Acting Reversible Contraception (LARC). Journal of Adolescent Health, 52(4, Supplement), S14?S21. https://doi.org/10.1016/j.jadohealth.2013.02.003 Sanders, J. N., Myers, K., Gawron, L. M., Simmons, R. G., & Turok, D. K. (2018). Contraceptive Method Use During the Community-Wide HER Salt Lake Contraceptive Initiative. American Journal of Public Health, 108(4), 550?556. https://doi.org/10.2105/AJPH.2017.304299 Sanders, J. N., Turok, D. K., Gawron, L. M., Law, A., Wen, L., & Lynen, R. (2017). Two-year continuation of intrauterine devices and contraceptive implants in a mixed-payer setting: A retrospective review. American Journal of Obstetrics and Gynecology, 216(6), 590.e1-590.e8. https://doi.org/10.1016/j.ajog.2017.02.003 Sawhill, I. V., & Guyot, K. (2019). Preventing Unplanned Pregnancy: Lessons from the States (p. 36). Brookings Institution. Schapiro, N. A. (2020). Title X Regulatory Changes and Their Impact on Adolescent Health. Journal of Pediatric Health Care, 34(2), 171?176. https://doi.org/10.1016/j.pedhc.2019.12.001 Schwallie, P. C., & Assenzo, J. R. (1974). The effect of depo-medroxyprogesterone acetate on pituitary and ovarian function, and the return of fertility following its discontinuation: A review. Contraception, 10(2), 181?202. https://doi.org/10.1016/0010-7824(74)90073-0 Schwarz, E. B., Burch, E. J., Parisi, S. M., Tebb, K. P., Grossman, D., Mehrotra, A., & Gonzales, R. (2013). Computer-assisted provision of hormonal contraception in acute care settings. Contraception, 87(2), 242? 250. https://doi.org/10.1016/j.contraception.2012.07.003 Schwarz, E. B., Papic, M., Parisi, S. M., Baldauf, E., Rapkin, R., & Updike, G. (2014). Routine counseling about intrauterine contraception for women seeking emergency contraception. Contraception, 90(1), 66?71. https://doi.org/10.1016/j.contraception.2014.02.007 280 Secura, G. M., Allsworth, J. E., Madden, T., Mullersman, J. L., & Peipert, J. F. (2010). The Contraceptive CHOICE Project: Reducing barriers to long-acting reversible contraception. American Journal of Obstetrics and Gynecology, 203(2), 115.e1-115.e7. https://doi.org/10.1016/j.ajog.2010.04.017 Sedgh, G., Singh, S., & Hussain, R. (2014). Intended and Unintended Pregnancies Worldwide in 2012 and Recent Trends. Studies in Family Planning, 45(3), 301?314. https://doi.org/10.1111/j.1728-4465.2014.00393.x Senderowicz, L. (2020). Contraceptive Autonomy: Conceptions and Measurement of a Novel Family Planning Indicator. Studies in Family Planning, 51(2), 161?176. https://doi.org/10.1111/sifp.12114 Shah, P. S., Balkhair, T., Ohlsson, A., Beyene, J., Scott, F., & Frick, C. (2011). Intention to become pregnant and low birth weight and preterm birth: A systematic review. Maternal and Child Health Journal, 15(2), 205? 216. https://doi.org/10.1007/s10995-009-0546-2 Shay, L. A., & Lafata, J. E. (2015). Where Is the Evidence? A Systematic Review of Shared Decision Making and Patient Outcomes. Medical Decision Making, 35(1), 114?131. https://doi.org/10.1177/0272989X14551638 Shoupe, D. (Ed.). (2020). The Handbook of Contraception: Evidence Based Practice Recommendations and Rationales. Springer Nature. Simmons, R. G., Sanders, J. N., Geist, C., Gawron, L., Myers, K., & Turok, D. K. (2019). Predictors of contraceptive switching and discontinuation within the first 6 months of use among Highly Effective Reversible Contraceptive Initiative Salt Lake study participants. American Journal of Obstetrics and Gynecology, 220(4), 376.e1-376.e12. https://doi.org/10.1016/j.ajog.2018.12.022 SisterSong Women of Color Reproductive Justice Collective. (2021). Reproductive Justice. https://www.sistersong.net/reproductive-justice Skra?i?, I., Lewin, A. B., & Roy, K. M. (2021). Evaluation of the Delaware Contraceptive Access Now (DelCAN) initiative: A qualitative analysis of site leaders? implementation recommendations. Contraception, 104(2), 211?215. https://doi.org/10.1016/j.contraception.2021.03.015 Snyder, A. H., Weisman, C. S., Liu, G., Leslie, D., & Chuang, C. H. (2018). The Impact of the Affordable Care Act on Contraceptive Use and Costs among Privately Insured Women. Women?s Health Issues, 28(3), 219?223. https://doi.org/10.1016/j.whi.2018.01.005 281 Sonfield, A. (2021, January 27). A fragmented system: Ensuring comprehensive contraceptive coverage in all U.S. health insurance plans. Guttmacher Institute. https://www.guttmacher.org/gpr/2021/02/fragmented-system- ensuring-comprehensive-contraceptive-coverage-all-us-health-insurance Sonfield, A., Hasstedt, K., & Gold, R. B. (2014). Moving Forward: Family Planning in the Era of Health Reform. Guttmacher Institute. https://www.guttmacher.org/report/moving-forward-family-planning-era-health- reform# Sonfield, A., Hasstedt, K., Kavanaugh, M. L., & Anderson, R. (2013). The Social and Economic Benefits of Women?s Ability To Determine Whether and When to Have Children (p. 48). Guttmacher Institute. Sonfield, A., Kost, K., Gold, R. B., & Finer, L. B. (2011). The Public Costs of Births Resulting from Unintended Pregnancies: National and State-Level Estimates. Perspectives on Sexual and Reproductive Health, 43(2), 94?102. https://doi.org/10.1363/4309411 Spies, E. L., Askelson, N. M., Gelman, E., & Losch, M. (2010). Young Women?s Knowledge, Attitudes, and Behaviors Related to Long-Acting Reversible Contraceptives. Women?s Health Issues, 20(6), 394?399. https://doi.org/10.1016/j.whi.2010.07.005 Sridhar, A., Chen, A., Forbes, E. R., & Glik, D. (2015). Mobile application for information on reversible contraception: A randomized controlled trial. American Journal of Obstetrics and Gynecology, 212(6), 774.e1-774.e7. https://doi.org/10.1016/j.ajog.2015.01.011 Sridhar, A., & Roher, R. (2017). Birth Control Tales. http://www.birthcontroltales.com/index.html Stanback, J., Steiner, M., Dorflinger, L., Solo, J., & Cates, W. (2015). WHO Tiered-Effectiveness Counseling Is Rights-Based Family Planning. Global Health: Science and Practice, 3(3), 352?357. https://doi.org/10.9745/GHSP-D-15-00096 Stanwood, N. L., & Bradley, K. A. (2006). Young Pregnant Women?s Knowledge of Modern Intrauterine Devices. Obstetrics & Gynecology, 108(6), 1417?1422. https://doi.org/10.1097/01.AOG.0000245447.56585.a0 Steinberg, J. R., Marthey, D., Xie, L., & Boudreaux, M. (2021). Contraceptive method type and satisfaction, confidence in use, and switching intentions. Contraception, 104(2), 176?182. https://doi.org/10.1016/j.contraception.2021.02.010 282 Stern, A. M. (2005). Sterilized in the Name of Public Health. American Journal of Public Health, 95(7), 1128?1138. https://doi.org/10.2105/AJPH.2004.041608 Stern, A. M. (2016, January 8). That Time The United States Sterilized 60,000 Of Its Citizens. HuffPost. https://www.huffpost.com/entry/sterilization-united-states_n_568f35f2e4b0c8beacf68713 Stevens, J., & Berlan, E. D. (2014). Applying Principles from Behavioral Economics To Promote Long-Acting Reversible Contraceptive (LARC) Methods. Perspectives on Sexual and Reproductive Health, 46(3), 165? 170. https://doi.org/10.1363/46e0614 Sundaram, A., Vaughan, B., Kost, K., Bankole, A., Finer, L., Singh, S., & Trussell, J. (2017). Contraceptive Failure in the United States: Estimates from the 2006?2010 National Survey of Family Growth. Perspectives on Sexual and Reproductive Health, 49(1), 7?16. https://doi.org/10.1363/psrh.12017 Sundstrom, B., Ferrara, M., DeMaria, A. L., Baker-Whitcomb, A., & Payne, J. B. (2017). Integrating Pregnancy Ambivalence and Effectiveness in Contraceptive Choice. Health Communication, 32(7), 820?827. https://doi.org/10.1080/10410236.2016.1172294 Sundstrom, B., Szabo, C., & Dempsey, A. (2018). ?My Body. My Choice?: A Qualitative Study of the Influence of Trust and Locus of Control on Postpartum Contraceptive Choice. Journal of Health Communication, 23(2), 162?169. https://doi.org/10.1080/10810730.2017.1421728 Sznajder, K. K., Tomaszewski, K. S., Burke, A. E., & Trent, M. (2017). Incidence of Discontinuation of Long- Acting Reversible Contraception among Adolescent and Young Adult Women Served by an Urban Primary Care Clinic. Journal of Pediatric and Adolescent Gynecology, 30(1), 53?57. https://doi.org/10.1016/j.jpag.2016.06.012 Taylor, D. J., Halpern, V., Brache, V., Bahamondes, L., Jensen, J. T., & Dorflinger, L. J. (2022). Ovulation Suppression following Subcutaneous Administration of Depot Medroxyprogesterone Acetate. Contraception: X, 100073. https://doi.org/10.1016/j.conx.2022.100073 Taylor, J. S., & Cabral, H. J. (2002). Are women with an unintended pregnancy less likely to breastfeed? The Journal of Family Practice, 51(5), 431?436. 283 Thorburn, S., & Bogart, L. M. (2005). Conspiracy Beliefs About Birth Control: Barriers to Pregnancy Prevention Among African Americans of Reproductive Age. Health Education & Behavior, 32(4), 474?487. https://doi.org/10.1177/1090198105276220 Tobias, E., & Enriquez, M. (2018). Increasing Long-Acting Reversible Contraceptive Method Use Among Alaska Native Women. The Journal for Nurse Practitioners, 14(5), e105?e108. https://doi.org/10.1016/j.nurpra.2017.12.024 Trussell, J., Aiken, A. R. A., Micks, E., & Guthrie, K. A. (2018). Efficacy, Safety, and Personal Considerations. In R. A. Hatcher, A. L. Nelson, J. Trussell, C. Cwiak, P. Cason, M. S. Policar, A. Edelman, A. R. A. Aiken, J. Marrazzo, & D. Kowal (Eds.), Contraceptive Technology (21st ed.). Ayer Company Publishers, Inc. Trussell, J., & Vaughan, B. (1999). Contraceptive Failure, Method-Related Discontinuation and Resumption of Use: Results from the 1995 National Survey of Family Growth. Family Planning Perspectives, 31(2), 64?93. https://doi.org/10.2307/2991641 Trussell, Lalla, A. M., Doan, Q. V., Reyes, E., Pinto, L., & Gricar, J. (2009). Cost Effectiveness of Contraceptives in the United States. Contraception, 79(1), 5?14. https://doi.org/10.1016/j.contraception.2008.08.003 Tsui, A. O., McDonald-Mosley, R., & Burke, A. E. (2010). Family Planning and the Burden of Unintended Pregnancies. Epidemiologic Reviews, 32(1), 152?174. https://doi.org/10.1093/epirev/mxq012 Turok, D. K., Gero, A., Simmons, R. G., Kaiser, J. E., Stoddard, G. J., Sexsmith, C. D., Gawron, L. M., & Sanders, J. N. (2021). Levonorgestrel vs. Copper Intrauterine Devices for Emergency Contraception. New England Journal of Medicine, 384(4), 335?344. https://doi.org/10.1056/NEJMoa2022141 Upadhyay, U. D., Brown, B. A., Sokoloff, A., & Raine, T. R. (2012). Contraceptive discontinuation and repeat unintended pregnancy within 1 year after an abortion. Contraception, 85(1), 56?62. https://doi.org/10.1016/j.contraception.2011.05.009 Upstream.org. (2021). About Us. Upstream USA. https://upstream.org/about/ U.S. Food and Drug Administration. (2021, June 18). Birth Control. FDA. https://www.fda.gov/consumers/free- publications-women/birth-control 284 U.S. Government Accountability Office. (2001, October 12). Welfare Reform: More Research Needed on TANF Family Caps and Other Policies for Reducing Out-Of-Wedlock Births. https://www.gao.gov/products/gao- 01-924 Usinger, K. M., Gola, S. B., Weis, M., & Smaldone, A. (2016). Intrauterine Contraception Continuation in Adolescents and Young Women: A Systematic Review. Journal of Pediatric and Adolescent Gynecology, 29(6), 659?667. https://doi.org/10.1016/j.jpag.2016.06.007 Vogt, C., & Schaefer, M. (2012). Knowledge matters ? Impact of two types of information brochure on contraceptive knowledge, attitudes and intentions. The European Journal of Contraception & Reproductive Health Care, 17(2), 135?143. https://doi.org/10.3109/13625187.2011.643837 Weisberg, E., Bateson, D., Knox, S., Haas, M., Viney, R., Street, D., & Fiebig, D. (2013). Do women and providers value the same features of contraceptive products? Results of a best-worst stated preference experiment. The European Journal of Contraception & Reproductive Health Care, 18(3), 181?190. https://doi.org/10.3109/13625187.2013.777830 Welti, K., & Manlove, J. (2018). Estimated reductions in unintended pregnancy among Delaware Title X family planning clients after a contraceptive access intervention. Child Trends. https://www.childtrends.org/blog/estimated-reductions-in-unintended-pregnancy-among-delaware-title-x- family-planning-clients-after-a-contraceptive-access-intervention Whitaker, A. K., Terplan, M., Gold, M. A., Johnson, L. M., Creinin, M. D., & Harwood, B. (2010). Effect of a Brief Educational Intervention on the Attitudes of Young Women Toward the Intrauterine Device. Journal of Pediatric and Adolescent Gynecology, 23(2), 116?120. https://doi.org/10.1016/j.jpag.2009.09.012 Wildsmith, E., Guzzo, K. B., & Hayford, S. R. (2010). Repeat Unintended, Unwanted and Seriously Mistimed Childbearing in the United States. Perspectives on Sexual and Reproductive Health, 42(1), 14?22. https://doi.org/10.1363/4201410 Williams, D. R., Priest, N., & Anderson, N. (2016). Understanding Associations between Race, Socioeconomic Status and Health: Patterns and Prospects. Health Psychology?: Official Journal of the Division of Health Psychology, American Psychological Association, 35(4), 407?411. https://doi.org/10.1037/hea0000242 285 Wiltz, T. (2019). Family Welfare Caps Lose Favor in More States. Pew Charitable Trusts. https://pew.org/2JdcJxx Winter, L., & Breckenmaker, L. C. (1991). Tailoring Family Planning Services to the Special Needs of Adolescents. Family Planning Perspectives, 23(1), 24?30. https://doi.org/10.2307/2135397 Wollum, A., Trussell, J., Grossman, D., & Grindlay, K. (2020). Modeling the Impacts of Price of an Over-the- Counter Progestin-Only Pill on Use and Unintended Pregnancy among U.S. Women. Women?s Health Issues, 30(3), 153?160. https://doi.org/10.1016/j.whi.2020.01.003 World Health Organization. (2016). Selected practice recommendations for contraceptive use. World Health Organization. http://www.who.int/reproductivehealth/publications/family_planning/SPR-3/en/ World Health Organization. (2018). Family Planning: A global handbook for providers (3rd ed.). World Health Organization Department of Reproductive Health and Research (WHO/RHR) and Johns Hopkins Bloomberg School of Public Health/Center for Communication Programs (CCP), Knowledge for Health Project. https://www.who.int/publications/i/item/9780999203705 Wyatt, K. D., Anderson, R. T., Creedon, D., Montori, V. M., Bachman, J., Erwin, P., & LeBlanc, A. (2014). Women?s values in contraceptive choice: A systematic review of relevant attributes included in decision aids. BMC Women?s Health, 14(1), 28. https://doi.org/10.1186/1472-6874-14-28 Yazdkhasti, M., Pourreza, A., Pirak, A., & Abdi, F. (2015). Unintended Pregnancy and Its Adverse Social and Economic Consequences on Health System: A Narrative Review Article. Iranian Journal of Public Health, 44(1), 12?21. Yee, L. M., & Simon, M. A. (2010). The Role of the Social Network in Contraceptive Decision-making Among Young, African American and Latina Women. Journal of Adolescent Health, 47(4), 374?380. https://doi.org/10.1016/j.jadohealth.2010.03.014 Yee, L. M., & Simon, M. A. (2011). Perceptions of Coercion, Discrimination and Other Negative Experiences in Postpartum Contraceptive Counseling for Low-income Minority Women. Journal of Health Care for the Poor and Underserved, 22(4), 1387?1400. https://doi.org/10.1353/hpu.2011.0144 286 Yland, J. J., Bresnick, K. A., Hatch, E. E., Wesselink, A. K., Mikkelsen, E. M., Rothman, K. J., S?rensen, H. T., Huybrechts, K. F., & Wise, L. A. (2020). Pregravid contraceptive use and fecundability: Prospective cohort study. BMJ, 371, m3966. https://doi.org/10.1136/bmj.m3966 Yoo, S. H., Guzzo, K. B., & Hayford, S. R. (2014). Understanding the Complexity of Ambivalence Toward Pregnancy: Does It Predict Inconsistent Use of Contraception? Biodemography and Social Biology, 60(1), 49?66. https://doi.org/10.1080/19485565.2014.905193 Zapata, L. B., Pazol, K., Dehlendorf, C., Curtis, K. M., Malcolm, N. M., Rosmarin, R. B., & Frederiksen, B. N. (2018). Contraceptive Counseling in Clinical Settings: An Updated Systematic Review. American Journal of Preventive Medicine, 55(5), 677?690. https://doi.org/10.1016/j.amepre.2018.07.006 287