VOLUME 8: NO. 6, A142 NOVEMBER 2011
ORIGINAL RESEARCH
Racial and Ethnic Disparities in the 
Quality of Diabetes Care in a Nationally 
Representative Sample
Patrick Richard, PhD, MA; Pierre K?breau Alexandre, PhD, MS, MPH; Anthony Lara, MHSA; Adaeze B. 
Akamigbo, PhD, MPP
Suggested citation for this article: Richard P, Alexandre with their white counterparts. Hispanic patients with diabe-
PK, Lara A, Akamigbo AB. Racial and ethnic disparities in tes were also less likely to have received a foot examination 
the quality of diabetes care in a nationally representative during the past year compared with white patients with 
sample. Prev Chronic Dis 2011;8(6):A142. http://www.cdc. diabetes. Conversely, black patients with diabetes were 
gov/pcd/issues/2011/nov/10_0174.htm. Accessed [date]. more likely to have received a foot examination during the 
past year compared with white patients with diabetes. The 
PEER REVIEWED differences in the quality of diabetes care remained signifi-
cant even after controlling for socioeconomic status (SES), 
health insurance status, self-rated health status, comorbid 
Abstract conditions, and lifestyle behavior variables.
Introduction Conclusions
Previous studies have consistently documented that racial/ Although the link between racial/ethnic minority status 
ethnic minority patients with diabetes receive lower qual- and the quality of care for patients with diabetes is not 
ity of care, based on various measures of quality of care completely understood, our results suggest that factors 
and care settings. However, 2 recent studies that used such as SES, health insurance status, self-rated health 
data from Medicare or Veterans Administration beneficia- status, and other health conditions are potential anteced-
ries have shown improvements in racial/ethnic disparities ents of quality of diabetes care.
in the quality of diabetes care. These inconsistencies sug-
gest that additional investigation is needed to provide new 
information about the relationship between racial/ethnic Introduction
minority patients and the quality of diabetes care.
Although diabetes is a prevalent, debilitating, and costly 
Methods chronic condition that affects the general population, evi-
We analyzed 3 years of data (2005-2007) from the Medical dence suggests that racial/ethnic minority groups bear a 
Expenditure Panel Survey and used multivariate models disproportionate burden of the condition (1-5). Racial/eth-
that adjusted for sociodemographic characteristics, region- nic minority groups have a higher prevalence, worse dia-
al location, insurance status, health behaviors, health sta- betes outcomes, and higher rates of diabetes-related com-
tus, and comorbidity to examine racial/ethnic disparities plications than their white counterparts (1-3,6,7). Previous 
in the quality of diabetes care. studies have consistently documented that racial/ethnic 
minority patients with diabetes receive lower quality of 
Results care, based on various measures of quality of care and care 
We found that Asian patients with diabetes were less likely settings (8-12). For instance, a report published in 2006 by 
to have received 2 or more glycated hemoglobin (HbA1c) the Agency for Healthcare Research and Quality showed 
tests or a foot examination during the past year compared that racial/ethnic minority groups, including patients who 
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the 
Centers for Disease Control and Prevention.
 www.cdc.gov/pcd/issues/2011/nov/10_0174.htm ? Centers for Disease Control and Prevention 1
VOLUME 8: NO. 6
NOVEMBER 2011
have diabetes, received poorer quality of care in 22 critical use, health insurance coverage, medical expenditures, 
measures of quality care compared with whites (13). and sources of payment for the US civilian noninstitu-
tionalized population that is cosponsored by the Agency 
However, a recent study that used data from 1997 to 2003 for Healthcare Research and Quality and the National 
from Medicare beneficiaries in managed care plans has Center for Health Statistics. For this analysis, we used the 
shown that improvements have been made in racial/ethnic household component (HC) file of MEPS, which is the core 
disparities in the quality of diabetes care (14). Clinical per- component of the survey that collects data on demographic 
formance for patients with diabetes improved, and the gap characteristics, health conditions, self-rated health status, 
in the quality of diabetes care between whites and blacks medical services use, access to care, satisfaction with care, 
narrowed for 7 of the Health Plan Employer Data and health insurance coverage status, and income for each 
Information Set (HEDIS) measures, including glycated person surveyed.
hemoglobin (HbA1c) and eye examination. Similarly, a 
more recent study that used nationally representative We pooled 3 years of data to increase the sample size of the 
data from the Medical Expenditures Panel Survey (MEPS) study and used a study design that attempted to address 
found no significant differences in the quality of diabetes previous shortcomings and inconsistencies in the lit-
care between racial/ethnic minority groups and white erature. The overlapping design of MEPS allows repeated 
patients (15). observations of the same people several times during the 
year. We combined data from the HC files with the pooled 
Additional studies that examined the Medicare and estimation linkage file of MEPS to restrict the analytic 
Veterans Administration (VA) populations have suggested sample to unique individuals. By restricting the sample in 
that recent investment in public resources to address this way, we were able to compute appropriate standard 
racial/ethnic inequalities in health and health care may errors. To construct the analytic sample we used data 
have resulted in the reduction or elimination of racial/eth- from the MEPS Diabetes Care Survey, a self-administered 
nic disparities in the quality of diabetes care (14,16,17). questionnaire to adult respondents aged 18 years or older 
The results of these 2 studies are encouraging but cannot who reported that they had been diagnosed with diabetes 
be generalized to the US population because of systematic by a health care professional. This survey contains a series 
differences between the general population and Medicare of questions about diabetes management for 2005, 2006, 
or VA beneficiaries. These inconsistencies also suggest and 2007, including the number of times respondents 
that additional investigation is needed to provide new reported having had an HbA1c test, the number of times 
information about the relationship between racial/ethnic they reported having had their feet checked for sores or 
minority groups and the quality of diabetes care. irritation, and the last time they reported having had an 
eye examination during the same period.
Racial/ethnic differences in the quality of diabetes care 
may arise from multiple factors and complex interactions The resulting sample was 2,671 people who reported 
between patients, their providers, and the health care sys- that they had been diagnosed with diabetes by a health 
tems in which they operate (18). Therefore, we investigated care professional during the 3 years combined. We sub-
factors that are amenable to policy changes in our models, sequently excluded 182 people, either because they did 
including socioeconomic status (SES) and health insurance not respond to the self-administered questionnaire them-
coverage, to determine racial/ethnic differences in the qual- selves or did not have any office visit during the time 
ity of diabetes care. Our objective was to examine racial/ of the study. We limited the analytic sample to people 
ethnic disparities in the quality of care provided to patients who responded to the self-administered questionnaire by 
with diabetes by using nationally representative data sets. themselves, not by their spouse or another proxy, to limit 
reporting bias and included only patients with diabetes 
who had at least 1 visit to a health care professional dur-
Methods ing the past 12 months to capture patient-provider inter-
actions in measuring the quality of care for patients with 
Sample diabetes. Finally, we excluded an additional 37 people 
who had missing observations on the different variables 
We analyzed 3 years of data (2005-2007) from MEPS, used in the analysis. The final analytic sample was 2,452 
a nationally representative survey of health services patients who reported that they had been diagnosed with 
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position  
of the Centers for Disease Control and Prevention.
2 Centers for Disease Control and Prevention ? www.cdc.gov/pcd/issues/2011/nov/10_0174.htm
VOLUME 8: NO. 6
NOVEMBER 2011
diabetes by a health care professional and were aged 18 Statistical analysis
years or older.
We used logistic regression models to determine the odds 
Variables of receiving at least 2 HbA1c tests, a foot examination, or 
an eye examination in the past year. We conducted ?2 tests 
Consistent with American Diabetes Association guidelines to determine differences in outcomes among the different 
for patients with diabetes, we used 3 binary indicators to racial/ethnic minority groups. Significance was set at P < 
measure quality of care for patients with diabetes, which .10. Because of the complex survey design of the MEPS 
were reporting receipt of the following during the past HC file, we used special diabetes weights from MEPS to 
year: 1) 2 or more HbA1c tests, 2) 1 foot examination, compute robust standard errors of the estimates. Because 
and 3) 1 eye examination. For the HbA1c tests, MEPS we pooled data over several years for a subsample of 
asked, ?During [survey year], how many times did a doc- patients with diabetes, we used the balanced repeated rep-
tor, nurse, or other health professional check your blood lication method of variance estimation to account for the 
for glycosylated hemoglobin or ?hemoglobin A-one-C??? For full set of survey stratum and primary sampling units, as 
the foot examination, MEPS asked, ?How many times did recommended by MEPS. Weighted proportions, adjusted 
a health professional check your feet for any sores or irrita- odds ratios (AORs), and 95% confidence intervals (CIs) 
tions?? For the eye examination, MEPS asked, ?In which were used to present the results. We used Stata version 
year did you have an eye examination in which your pupils 11 (StataCorp LP, College Station, Texas) to conduct the 
were dilated?? analysis.
On the basis of previous research, we controlled for a set of 
patient characteristics known to be associated with differ- Results
ences in quality of care including age, race/ethnicity, sex, 
SES, health insurance status, smoking status, obesity sta- More than 68% of respondents were aged 55 years or older, 
tus, general health status, comorbid cardiovascular condi- 34% of respondents resided in families with incomes high-
tions, and regional location (19-21). To assess patients? er than 400% of the FPL, and 31% had some form of public 
race/ethnicity, respondents were asked, ?Which of these insurance such as Medicaid or Medicare. Approximately 
would you say is your main racial or ethnic group?? 78% of the sample had other comorbid cardiovascular con-
Response options were non-Hispanic white, non-Hispanic ditions (Table 1).
African American, Hispanic, American Indian or Alaska 
Native, Asian or Pacific Islander, mixed race, or some On average, about 83% of patients reported receiving at 
other single race. From these responses, we constructed 4 least 2 HbA1c tests; 70%, a foot examination; and 61%, an 
categories: non-Hispanic white, black, Hispanic, and Asian. eye examination during the past 12 months. Chi-square 
We used the MEPS body mass index (BMI) measure, cal- tests indicated significant differences between whites and 
culated from respondents? self-reported height and weight, Asians receiving at least 2 HbA1c tests (P = .007) and a 
to create an indicator variable for obesity (BMI ?30 kg/m2). foot examination (P = .002) (Table 2). Hispanic patients 
Different categories of education and income were used to were less likely to receive an eye examination during the 
account for the nonlinearity of the relationship between past year than were white patients (P = .005). Conversely, 
these 2 variables and the quality of care for patients black patients were more likely to receive a foot examina-
with diabetes. Education levels were defined as receiving tion than were white patients (P = .009) and less likely to 
less than a high school degree, high school degree, col- receive an eye examination than were white patients (P = 
lege degree, or postgraduate degree. Income levels were .03) during the past year.
defined as incomes below 100% of the federal poverty 
level (FPL), between 100% and 199% of the FPL, between Multivariate logistic regression analyses indicated that 
200% and 400% of the FPL, and above 400% of the FPL. Asian patients with diabetes were less likely to receive at 
Comorbid cardiovascular conditions included patients who least 2 HbA1c tests or a foot examination in the past year 
reported being diagnosed with at least 1 of the following than were their white counterparts (Table 3). Likewise, 
conditions: hypertension, angina, mild or coronary heart Hispanic patients with diabetes were less likely to receive 
attack, stroke, or other form of heart disease. a foot examination in the past 12 months than were white 
patients with diabetes. Conversely, black patients with 
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position 
of the Centers for Disease Control and Prevention.
 www.cdc.gov/pcd/issues/2011/nov/10_0174.htm ? Centers for Disease Control and Prevention 
VOLUME 8: NO. 6
NOVEMBER 2011
diabetes were more likely to receive a foot examination patients with diabetes were more likely to have received 
than were white patients with diabetes. a foot examination in the past 12 months compared 
with white patients with diabetes. This finding may be 
High school graduates were less likely to receive at least 2 explained by the fact that black patients with diabetes 
HbA1c tests or a foot examination compared with partici- tend to have higher rates of diabetes complications and 
pants who did not graduate from high school. Similarly, amputations, and recent guidelines have highlighted the 
patients with diabetes who resided in the Midwest, South, need to carefully monitor these patients as their conditions 
or West were less likely to receive a foot examination than progress (22). These differences remained significant even 
were those who lived in the northeastern part of the coun- after controlling for SES, insurance status, health status, 
try. We also found negative associations between receipt of comorbid conditions, and lifestyle behavior variables.
eye examination and patients who were uninsured or who 
smoked compared with those who were privately insured However, our results differ from those found by Lee and 
or did not smoke, respectively. Patients who resided in colleagues (19), who found no differences in receipt of these 
families with incomes more than 400% of the FPL, were measures among racial/ethnic minorities. Their analysis 
in fair or poor health, and suffered from comorbid cardio- of 2000 MEPS data found no differences among racial/
vascular conditions were more likely to report receiving ethnic groups for most of the outcomes in diabetes care 
HbA1c tests compared with those who lived in families management, including respondents who had received an 
with incomes below 100% of the FPL, were in excellent or HbA1c test, had their feet checked for sores or irritation, 
good health, and did not have a comorbid cardiovascular or received an eye examination in the past year. A possible 
condition. For example, patients with a comorbid cardio- explanation for the different findings may be differences 
vascular condition were 34% more likely to have received in study design. Contrary to the study conducted by Lee 
2 or more HbA1c tests than were those who did not. et al (19), we restricted our sample to unique individuals 
to compute appropriate standard errors in pooled estima-
Patients who resided in families with an income above tions. Additional differences were the use of more recent 
400% of the FPL, were publicly insured with either data sets, the use of special diabetes weights from MEPS, 
Medicaid or Medicare, reported fair/poor health, or had and the use of the balanced repeated replication method 
a comorbid cardiovascular condition were more likely to variance estimation to account for the full set of survey 
have a foot examination compared with those who lived stratum and primary sampling units, as recommended 
in families with incomes below 100% of the FPL, were by MEPS (23). Our results also differ from findings of a 
privately insured, were in excellent/good health, or did not study by Trivedi et al that found narrowing of the gap in 
have a cardiovascular comorbid condition. Patients who the quality of diabetes care between whites and blacks 
had incomes greater than 400% of the FPL were more like- (14). However, this study was limited to Medicare benefi-
ly to receive a foot examination than patients who lived in ciaries in managed care, and the authors did not stratify 
families with incomes below 100% of the FPL. by other racial/ethnic minority groups such as Hispanics 
and Asians. The findings by Trivedi et al may not be gen-
eralizable to other health systems or to other racial/ethnic 
Discussion groups that may experience greater racial/ethnic dispari-
ties in the quality of diabetes care. Conversely, our find-
Our study advances the literature on racial/ethnic dis- ings are consistent with those of other studies that used 
parities in quality of care for patients with diabetes. We both clinical and community-based data (24-29).
assessed racial/ethnic disparities in the quality of diabetes 
care on the basis of receipt of recommended HbA1c tests Our study has limitations. First, the data we used were 
and foot and eye examinations in the previous year. We cross-sectional, so causal relationships cannot be estab-
hypothesized that racial/ethnic minority patients with lished. Second, the dependent variables were self-reported 
diabetes would receive lower quality of care than their measures of process outcomes of diabetes care. Although 
white counterparts. Compared with white patients with we controlled for patients who reported poor or fair health 
diabetes, Asian patients with diabetes were less likely to or comorbid cardiovascular conditions, these patients may 
have received at least 2 HbA1c tests and both Asian and have visited their providers more often and thus were 
Hispanic patients were less likely to have received a foot more likely to receive diabetes tests compared with those 
examination in the past 12 months. Conversely, black who reported excellent or good health and no comorbid 
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position  
of the Centers for Disease Control and Prevention.
4 Centers for Disease Control and Prevention ? www.cdc.gov/pcd/issues/2011/nov/10_0174.htm
VOLUME 8: NO. 6
NOVEMBER 2011
cardiovascular conditions. Furthermore, no information on References
glycemic control among patients with diabetes was avail-
able. Asians may have better glycemic control and may  1. Lustman PJ, Anderson RJ, Freedland KE, de Groot 
have received HbA1c tests and foot and eye examinations M, Carney RM, Clouse RE. Depression and poor gly-
less frequently than their white counterparts. cemic control: a meta-analytic review of the literature. 
Diabetes Care 2000;23(7):934-42
Although the link between racial/ethnic minority status  2. Sundquist J, Winkleby MA, Pudaric S. Cardiovascular 
and the quality of care for patients with diabetes is not disease risk factors among older black, Mexican-
completely understood, our study suggests that factors American, and white women and men: an analysis 
such as health insurance status, SES, and self-rated of NHANES III, 1988-1994. Third National Health 
health status are potential antecedents of quality of dia- and Nutrition Examination Survey. J Am Geriatr Soc 
betes care. Therefore, assessing the association between 2001;49(2):109-16.
racial/ethnic disparities in the quality of diabetes care  3. Mokdad AH, Bowman BA, Ford ES, Vinicor F, 
and factors such as SES, insurance status, and health Marks JS, Koplan JP. The continuing epidemics of 
behaviors is warranted because these factors are modifi- obesity and diabetes in the United States. JAMA 
able and can serve as the focus of interventions to reduce 2001;286(10):1195-200.
racial/ethnic disparities in the quality of diabetes care.  4. Fox CS, Coady S, Sorlie PD, Levy D, Meigs JB, 
Findings from this study may have clinical, public health, D?Agostino RB Sr, et al. Trends in cardiovascular com-
public policy, and research implications. Specifically, these plications of diabetes. JAMA 2004;292(20):2495-9.
results may underscore the importance of providing diver-  5. McBean AM, Li S, Gilbertson DT, Collins AJ. 
sity training to providers to improve the quality of care to Differences in diabetes prevalence, incidence, and mor-
patients with diabetes. Furthermore, evidence from this tality among the elderly of four racial/ethnic groups: 
study may play a key role in informing policy makers in whites, blacks, Hispanics, and Asians. Diabetes Care 
their continuous efforts to translate effective research into 2004;27(10):2317-24.
nationwide practices to eliminate racial/ethnic differences  6. Lanting LC, Joung IM, Mackenbach JP, Lamberts 
in quality of care, which is relevant in the context of the SW, Bootsma AH. Ethnic differences in mortality, end-
current health care reform law that seeks to eliminate stage complications, and quality of care among diabetic 
racial/ethnic disparities. Additional research is needed to patients: a review. Diabetes Care 2005;28(9):2280-8.
fully evaluate the mechanisms and sources of racial/ethnic  7. Chin MH, Walters AE, Cook SC, Huang ES. 
disparities in the quality of diabetes care. Interventions to reduce racial and ethnic disparities 
in health care. Med Care Res Rev 2007;64(5 Suppl):7S-
28S.
Author Information  8. Peek ME, Cargill A, Huang ES. Diabetes health dis-
parities: a systematic review of health care interven-
Corresponding Author: Patrick Richard, PhD, MA, tions. Med Care Res Rev 2007;64(5 Suppl):101S-56S.
Assistant Research Professor, Department of Health  9. McGlynn EA, McGlynn EA, Asch SM, Adams J, 
Policy, The George Washington University School of Keesey J, Hicks J, et al. The quality of health care 
Public Health and Health Services, 2021 K St, NW Ste delivered to adults in the United States. New Engl J 
800, Washington, DC 20006. Telephone: 202-994-4176. E- Med 2003;348(26):2635-45.
mail: patrick.richard@gwumc.edu. 10. Thackeray R, Merrill RM, Neiger BL. Disparities in 
diabetes management practice between racial and 
Author Affiliations: Pierre K?breau Alexandre, Johns ethnic groups in the United States. Diabetes Educ 
Hopkins Bloomberg School of Public Health, Baltimore, 2004;30(4):665-75.
Maryland; Anthony Lara, Department of Health Policy, 11. Jha AK, Li Z, Orav EJ, Epstein AM. Care in US hospi-
The George Washington University School of Public tals ? the Hospital Quality Alliance program. N Engl 
Health and Health Services, Washington, DC; Adaeze J Med 2005;353(3):265-74.
B. Akamigbo, Health Research & Educational Trust, 12. Vaccarino V, Rathore SS, Wenger NK, Frederick 
Chicago, Illinois. PD, Abramson JL, Barron HV, et al. Sex and racial 
differences in the management of acute myocar-
dial infarction, 1994 through 2002. N Engl J Med 
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position 
of the Centers for Disease Control and Prevention.
 www.cdc.gov/pcd/issues/2011/nov/10_0174.htm ? Centers for Disease Control and Prevention 
VOLUME 8: NO. 6
NOVEMBER 2011
2005;353(7):671-82. 25. McBean AM, Huang Z, Virnig BA, Lurie N, Musgrave 
13. 2004 National healthcare disparities report. Rockville D. Racial variation in the control of diabetes among 
(MD): Agency for Healthcare Research and Quality; elderly Medicare managed care beneficiaries. Diabetes 
2004. Care 2003;26(12):3250-6.
14. Trivedi AN, Zaslavsky AM, Schneider EC, Ayanian 26. Kirk JK, D?Agostino RB, Bell RA, Passmore LV, Bonds 
JZ. Trends in the quality of care and racial dis- DE, Karter AJ, Narayan KM. Disparities in HbA1c 
parities in Medicare managed care. N Engl J Med levels between African-American and non-Hispanic 
2005;353(7):692-700. white adults with diabetes: a meta-analysis. Diabetes 
15. Cohen JW, Monheit AC, Beauregard KM, Cohen Care 2006;29(9):2130-6.
SB, Lefkowitz DC, Potter DE, et al. The Medical 27. Brown AF, Gregg EW, Stevens MR, Karter AJ, 
Expenditure Panel Survey: a national health informa- Weinberger M, Safford MM, et al. Race, ethnicity, socio-
tion resource. Inquiry 1996;33(4):373-89. economic position, and quality of care for adults with 
16. Kerr EA, Gerzoff RB, Krein SL, Selby JV, Piette diabetes enrolled in managed care: the Translating 
JD, Curb JD, et al. Diabetes care quality in the Research Into Action for Diabetes (TRIAD) study. 
Veterans Affairs health care system and commercial Diabetes Care 2005;28(12):2864-70.
managed care: the TRIAD study. Ann Intern Med 28. Sequist TD, Fitzmaurice GM, Marshall R, Shaykevich 
2004;141(4):272-81. S, Safran DG, Ayanian JZ. Physician performance 
17. Gordon HS, Johnson ML, Ashton CM. Process of care and racial disparities in diabetes mellitus care. Arch 
in Hispanic, black, and white VA beneficiaries. Med Intern Med 2008;168(11):1145-51.
Care 2003;40(9):824-33. 29. McWilliams JM, Meara E, Zaslavsky AM, Ayanian JZ. 
18. Heisler M, Vijan S, Anderson RM, Ubel PA, Bernstein Differences in control of cardiovascular disease and 
SJ, Hofer TP. When do patients and their physicians diabetes by race, ethnicity, and education: US trends 
agree on diabetes treatment goals and strategies, from 1999 to 2006 and effects of Medicare coverage. 
and what difference does it make? J Gen Intern Med Ann Intern Med 2009;150(8):505-15.
2003;18(11):893-902.
19. Lee J-A, Liu C-F, Sales AE. Racial and ethnic differ-
ences in diabetes care and health care use and costs. 
Prev Chronic Dis 2006;3:1-12. http://www.cdc.gov/pcd/
issues/2006/jul/05_0196.htm. Accessed January 22, 
2011.
20. Safran DG, Taira DA, Rogers WH, Kosinski M, Ware 
JE, Tarlov AR. Linking primary care performance to 
outcomes of care. J Fam Pract 1998;47(3):213-20.
21. Mainous AG 3rd, Griffith CH 3rd, Love MM. Patient 
satisfaction with care in programs for low income indi-
viduals. J Community Health 1999;24(5):381-91.
22. Centers for Disease Control and Prevention. Racial/
ethnic and socioeconomic disparities in multiple risk 
factors for heart disease and stroke ? United States, 
2003. MMWR Morb Mortal Wkly Rep 2005;54(5):113-
7.
23. Agency for Healthcare Research and Quality. MEPS 
HC-036BRR: MEPS 1996-2008 Replicates for calcu-
lating variances file. http://www.meps.ahrq.gov/mep-
sweb/data_stats/download_data/pufs/h113/h113doc.
pdf. Accessed January 22, 2011.
24. Harris MI, Eastman RC, Cowie CC, Flegal KM, 
Eberhardt MS. Racial and ethnic differences in gly-
cemic control of adults with type 2 diabetes. Diabetes 
Care 1999;22(3):403-8.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position  
of the Centers for Disease Control and Prevention.
6 Centers for Disease Control and Prevention ? www.cdc.gov/pcd/issues/2011/nov/10_0174.htm
VOLUME 8: NO. 6
NOVEMBER 2011
Tables
Table 1. Dependent and Independent Variables for Racial/Ethnic Disparities in the Quality of Care for Patients Aged 18 to 64 Years 
With Diabetes (n = 2,452), 2005-2007 Medical Expenditure Panel Surveya
Characteristics Weighted %b Characteristics Weighted %b
Dependent Variables Independent Variables (continued)
Clinical testing Income as % of FPLc
Received ?2 HbA1c tests in past year 8.2 <100 14.1
Received foot examination in past year 70.9 100-199 21.2
Received eye examination in past year 61.1 200-400 1.0
Independent Variables >400 .7
Female sex 2.1 Insurance status
Age, y Private insurance 62.7
18-24 0.6 Public insurance 1.2
2-4 2.7 Uninsured 6.1
-44 9. Health status/conditions
4-4 18. Fair/poor health 9.7
-64 0. Obesed 4.7
6-74 19.9 Comorbid cardiovascular conditionse 77.8
?75 18. Current smoker 1.
Race/ethnicity Region
Non-Hispanic white 68.2 Northeast 19.1
Black 1.0 Midwest 21.
Hispanic 1.2 South 9.4
Asian .6 West 21.2
Education  Abbreviations: HbA1c, glycated hemoglobin; FPL, federal poverty level. 
Less than a high school degree .0 a Data are pooled for years 200, 2006, and 2007 of the Household 
Component of the Medical Expenditure Panel Survey (MEPS). The sample 
High school degree 1.0 was restricted to unique individuals in each year of the pooled data. There 
College degree 10.6 are no repeated observations for the same individual across the different years. 
Postgraduate degree .4 b Percentages weighted to yield a nationally representative sample of US 
households. 
c FPL is the set minimum amount of income that a family needs for food, 
clothing, transportation, shelter and other necessities and is used to deter-
mine eligibility income limits for public assistance programs as some per-
centage of FPL. FPL varies according to family size and is determined by the 
US Department of Health and Human Services. 
d Reported body mass index of ?30 kg/m2. 
e Respondents with any of the following conditions: hypertension, angina, 
mild or coronary heart attack, stroke, or other form of heart disease.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position 
of the Centers for Disease Control and Prevention.
 www.cdc.gov/pcd/issues/2011/nov/10_0174.htm ? Centers for Disease Control and Prevention 7
VOLUME 8: NO. 6
NOVEMBER 2011
Table 2. Weighted Proportions Receiving 2 HbA1c Tests, Foot Examination, and Eye Examination in the Past Year, by Racial/Ethnic 
Minority Group, Adult Patients Aged 18 to 64 Years With Diabetes (n = 2,452), 2005-2007 Medical Expenditure Panel Surveya
Weighted %b (95% CI)
Race/ethnicity Received ?2 HbA1c Tests in Past Year Received Foot Examination in Past Year Received Eye Examination in Past Year
White 8.0 (80.-8.8) 71.2 (67.6-74.9) 6. (9.2-67.6)
Black 84. (79.1-89.9) 76.9 (71.4-82.) .1 (4.1-61.2)
Hispanic 77.8 (71.4-9.6) 6.1 (6.1-70.2) 2.6 (4.7-9.)
Asian 8. (71.1-9.6) 70.2 (.-86.9) 7.9 (8.8-92.9)
 
Abbreviations: CI, confidence interval; HbA1c, glycated hemoglobin. 
a Data are pooled for years 200, 2006, and 2007 of the Household Component of the Medical Expenditure Panel Survey (MEPS). The sample was restricted 
to unique individuals in each year of the pooled data. There are no repeated observations for the same individual across the different years. 
b Percentage is weighted to yield a nationally representative sample of US households.
Table 3. Logistic Regression Results for Racial/Ethnic Disparities in the Quality of Care for Patients with Diabetes (n = 2,452), by 
Quality Indicators, 2005-2007 Medical Expenditure Panel Surveya
Received ?2 HbA1C Tests in the Past Received Foot Examination in the Received Eye Examination in the Past 
Year Past Year Year
Independent Variable AOR (95% CI) P Value AOR ( 95% CI) P Value AOR (95% CI) P Value
Female sex 1.14 (0.96-1.) .1 1.10 (0.90-1.) .4 1.14 (0.9-1.40) .20
Age, y
18-24 1 [Reference]
2-4 0.1 (0.06-2.6) .84 0.41 (0.0-2.2) .82 0.44 (0.06-2.44) .81
-44 0.81 (0.10-2.6) .9 0.66 (0.08-2.6) .92 0.4 (0.09-2.72) .82
4-4 0.90 (0.14-.6) .97 0.64 (0.1-.41) .91 0.8 (0.1-.4) .87
-64 0.97 (0.16-4.08) .99 0.74 (0.1-4.09) .94 0.6 (0.16-4.1) .90
6-74 0.7 (0.12-.41) .9 0.79 (0.10-.4) .9 0.68 (0.1-.4) .91
?75 0.82 (0.12-.4) .9 0.8 (0.11-.) .97 0.8 (0.11-.40) .96
Race/ethnicity
Non-Hispanic white 1 [Reference]
Black 1.10 (0.8-1.42) .48 1.1 (1.00-1.72) .0 0.86 (0.6-1.17) .
Hispanic 1.1 (0.87-1.46) . 0.76 (0.7-1.0) .08 0.80 (0.60-1.06) .11
Asian 0.44 (0.20-0.98) .04 0. (0.1-0.81) .02 0.67 (0.-1.8) .27
 
Abbreviations: HbA1c, glycated hemoglobin; AOR, adjusted odds ratio; CI, confidence interval; FPL, federal poverty level. 
a Data are pooled for the 200, 2006, and 2007 waves of the household component of the Medical Expenditure Panel Survey. The sample was restricted to 
unique individuals in each of the rounds of the pooled data. There are no repeated observations for the same individual across the different rounds of the 
year. 
b The set minimum amount of income that a family needs for food, clothing, transportation, shelter and other necessities and is used to determine eligibil-
ity income limits for public assistance programs as some percentage of FPL. FPL varies according to family size and is determined by the US Department of 
Health and Human Services.
(Continued on next page)
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position  
of the Centers for Disease Control and Prevention.
8 Centers for Disease Control and Prevention ? www.cdc.gov/pcd/issues/2011/nov/10_0174.htm
VOLUME 8: NO. 6
NOVEMBER 2011
Table 3. (continued) Logistic Regression Results for Racial/Ethnic Disparities in the Quality of Care for Patients with Diabetes (n = 
2,452), by Quality Indicators, 2005-2007 Medical Expenditure Panel Surveya
Received ?2 HbA1C Tests in the Past Received Foot Examination in the Received Eye Examination in the Past 
Year Past Year Year
Independent Variable AOR (95% CI) P Value AOR ( 95% CI) P Value AOR (95% CI) P Value
Education 
No high school degree 1 [Reference]
High school degree 0.81 (0.66-1.00) .0 0.80 (0.6-0.99) .04 1.16 (0.91-1.47) .22
College degree 1.11 (0.77-1.60) .6 0.97 (0.62-1.0) .87 1. (1.0-2.22) .0
Graduate degree 0.98 (0.62-1.4) .9 1.11 (0.61-2.00) .7 0.92 (0.1-1.64) .77
Income as % of FPLb
<100 1 [Reference]
100-199 1.1 (0.84-1.7) .9 1.04 (0.79-1.8) .77 1.21 (0.89-1.6) .22
200-400 1.29 (0.91-1.84) .1 1.00 (0.72-1.40) .99 1.4 (1.0-2.04) .04
>400 1.2 (0.99-1.76) .06 1.6 (1.00-1.8) .0 1.81 (1.18-2.77) .01
Insurance status
Private insurance 1 [Reference]
Public insurance 1.02 (0.8-1.24) .86 1. (1.0-1.68) .02 0.94 (0.72-1.21) .61
Uninsured 0.91 (0.6-1.26) . 0.78 (0.-1.1) .21 0.48 (0.2-0.72) .001
Health status/conditions
Fair/poor health 1.22 (0.98-1.) .08 1.7 (1.10-1.71) .01 0.96 (0.77-1.19) .70
Obese 1.0 (0.86-1.24) .74 1.0 (0.87-1.28) .60 1.0 (0.80-1.) .80
Cardiovascular comorbidity 1.4 (1.04-1.71) .02 1.9 (1.0-1.87) .0 1.24 (0.88-1.74) .21
Current smoker 1.01 (0.78-1.1) .92 1.08 (0.81-1.4) .61 0.72 (0.-0.9) .01
Region 
Northeast 1 [Reference]
Midwest 1.01 (0.69-1.48) .97 0.61 (0.4-0.87) .01 0.96 (0.69-1.) .81
South 0.91 (0.67-1.24) .6 0.69 (0.4-0.89) .004 0.79 (0.6-1.1) .19
West 0.82 (0.6-1.20) .0 0.78 (0.60-1.01) .06 0.84 (0.62-1.14) .26
 
Abbreviations: HbA1c, glycated hemoglobin; AOR, adjusted odds ratio; CI, confidence interval; FPL, federal poverty level. 
a Data are pooled for the 200, 2006, and 2007 waves of the household component of the Medical Expenditure Panel Survey. The sample was restricted to 
unique individuals in each of the rounds of the pooled data. There are no repeated observations for the same individual across the different rounds of the 
year. 
b The set minimum amount of income that a family needs for food, clothing, transportation, shelter and other necessities and is used to determine eligibil-
ity income limits for public assistance programs as some percentage of FPL. FPL varies according to family size and is determined by the US Department of 
Health and Human Services.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position 
of the Centers for Disease Control and Prevention.
 www.cdc.gov/pcd/issues/2011/nov/10_0174.htm ? Centers for Disease Control and Prevention 9