ORIGINAL REPORTS:DIABETES DIABETES RISK,DIAGNOSIS, AND CONTROL:DO PSYCHOSOCIAL FACTORS PREDICT HEMOGLOBIN A1C DEFINED OUTCOMES OR ACCURACY OF SELF-REPORTS? Kellee White, PhD; Favel L. Mondesir, BA; Lisa M. Bates, ScD; M. Maria Glymour, ScD Objective: To evaluate the accuracy of self- reported diabetes among multi-ethnic older adults by psychosocial factors and assess pre- dictors of diabetes risk, diagnosis, and control. Design and Methods: The 2006 Health and Retirement Study (N55,594) was used to determine agreement between self-reported diabetes and measured diabetes (HbA1c$ 6.5%) by age, sex, race/ethnicity, nativity, education, health insurance coverage, body mass index, depressive symptoms, and prior report of racial discrimination. We also examined associ- ations between these factors and pre-diabetes (HbA1c $6.0–,6.5%) among individuals with- out diabetes, and those with undiagnosed and poorly controlled (HbA1c $8.0%) diabetes. Results: Accuracy of self-reported diabetes was good (ie, sensitivity $ 80% and specificity $ 95%) among all demographic subgroups and across most social strata. Among those who reported racial discrimination, sensitivity of self-reported diabetes was lower among Blacks who reported racial discrimination in compar- ison to Blacks who did not report racial discrimination (82.7% vs 89.0%) an association that was marginally statistically significant (P5.05). Blacks and Hispanics had higher odds of pre-diabetes, undiagnosed diabetes, and poor glycemic control. Conclusions: Self-reported diabetes corre- sponded well with HbA1c assessed disease for all social strata examined in this sample of multi-ethnic older adults. Blacks with a history of racial discrimination may be less likely to know diabetes status. (Ethn Dis. 2014; 24[1]:19–27) Key Words: Diabetes, Self-report, Accuracy, Race/ethnicity, Racial Discrimination INTRODUCTION Several large-scale chronic disease surveillance and population-based stud- ies rely upon self-report of outcomes such as diabetes to describe and monitor population health status. 1–4 Although the accuracy of self-reports may be compromised by several factors includ- ing undiagnosed disease, ability to recall diagnoses, willingness to report, and accessibility to health services, 5 valida- tion studies of self-reported diabetes conducted in the United States and Europe have suggested moderate to good accuracy. 5–17 These studies have primarily focused on the accuracy of self-reported diabetes by demographic characteristics such as age, sex, and education. To our knowledge, one European study examined the accuracy of self-reported diabetes by ethnicity (ie, Turkish, Surinamese and Dutch). 11 Results from this study suggested that self-reports of diabetes among an ethnic minority (Surinamese) group were less accurate in comparison to the ethnic majority group (Dutch). However, studies assessing the accuracy of self- reported diabetes by race, ethnicity, or nativity in the United States are not as well-described as other demographic characteristics. Moreover, studies exam- ining validity of self-reports do not typically assess the influence of psycho- social factors such as perceived racial discrimination and depression. These factors, which have been associated with barriers to seeking medical care, less trust in providers, and lower usage of preventive care, 18–21 may cause under- diagnosis and lead to underreporting. Adequate management of diabetes requires adherence to a complex array of self-care management behaviors such as monitoring dietary intake, engaging in physical activity, taking prescribed med- ications, and frequent medical check-ups. Psychosocial factors such as depression and perceived discrimination may present obstacles to preventing diabetes and effective management among Blacks and Latinos. For example, depression, which is a significant predictor of pre-diabetes, decreased adherence to self-care behaviors and prescribed medications, and poorer glycemic control, 22–24 has been demon- strated to be more strongly associated with inadequate diabetes control among Blacks in comparison to Whites. 25 Re- search examining perceived racial discrim- ination and diabetes self-management behaviors found mixed results; some studies have shown a positive association with select self-management behaviors, 26 while others have not demonstrated an association. 1 Further, report of discrimi- nation in health care was associated with higher HbA1c levels. 27 However, per- ceived racial discrimination as a risk factor for pre-diabetes, particularly among older adults, has not been addressed in previous studies. Identifying psychosocial factors that reduce effective diabetes management can help improve delivery of medical care and design of patient-centered interven- tions to improve outcomes. In addition to its use as an indicator of diabetes control, HbA1c has now been adopted by the American Diabetes Association as an appropriate measure to screen for and diagnose type 2 From Department of Epidemiology and Biostatistics, University of South Carolina, Columbia (KW, FLM); and Department of Epidemiology, Columbia University Mail- man School of Public Health, New York, New York (LB); and Department of Society, Human Development, and Health, Harvard School of Public Health, Boston (MMG). Address correspondence to Kellee White; 800 Sumter Street, Suite 205; Columbia, SC 29209; 803.777.5057; 803.777.2524 (fax); kwhite@mailbox.sc.edu Ethnicity & Disease, Volume 24, Winter 2014 19