Experiences of Discrimination and Incident Type 2 Diabetes Mellitus

The Multi-Ethnic Study of Atherosclerosis (MESA)

Kara M. Whitaker; Susan A. Everson-Rose; James S. Pankow; Carlos J. Rodriguez; Tené T. Lewis; Kiarri N. Kershaw; Ana V. Diez Roux; Pamela L. Lutsey


Am J Epidemiol. 2017;186(4):445-455. 

In This Article


Major experiences of discrimination were associated with greater risk of incident type 2 diabetes mellitus over a median follow-up of 9.4 years in this multiethnic, population-based cohort study. These associations were largely independent of age, sex, race/ethnicity, marital status, income, education, depressive symptoms, other psychosocial and behavioral factors, and obesity. Furthermore, the association between major experiences of discrimination and diabetes was present regardless of whether the discrimination was attributed to race/ethnicity or to other causes. The risk of incident diabetes was greatest in those who reported 2 or more major experiences of discrimination, indicating a potential threshold effect, with no greater risk observed in those reporting only 1 experience of discrimination. Our findings add to the growing body of literature demonstrating an association between discrimination and physical health outcomes,[1,3] and are novel: To our knowledge this is the first study to identify a relationship between discrimination and diabetes incidence.

While few studies have looked at the associations of stressful experiences and type 2 diabetes, Mooy et al.[34] found that the number of stressful life events was associated with prevalent diabetes in a cross-sectional sample of Caucasian adults aged 50–74 years. Our findings complement the existing literature on stress and diabetes by showing that major experiences of discrimination, a form of stress, are associated with incident diabetes. Somewhat surprisingly, we observed minimal attenuation in associations after controlling for psychosocial and behavioral factors as well as obesity, indicating that these potential mediators did not explain the relationship between discrimination and incident diabetes. A theory by Bjorntorp[35,36] may provide an alternate explanation for the biological plausibility of the stress-diabetes association: This theory states that a defeatist reaction to stress leads to activation of the hypothalamic-pituitary-adrenal axis, resulting in endocrine abnormalities, including high cortisol and low sex-steroid levels, which antagonize the actions of insulin. Furthermore, this hormonal imbalance is associated with visceral adiposity, which plays an important role in the development of diabetes. Further research is needed to clarify the mechanisms through which discrimination may lead to diabetes risk.

Major experiences of discrimination that were attributed to race/ethnicity conferred similar risk of incident diabetes as did discrimination attributed to other causes. Relatively few prior studies have considered the source of discrimination (racial vs. other) when examining the association between discrimination and physical health outcomes.[6,24,37] Sims et al.[24] found that the associations between major experiences of discrimination and prevalent hypertension were similar, regardless of whether the discrimination was attributed to racial or nonracial factors. Everson-Rose et al.[6] observed a greater risk of incident cardiovascular events in those who experienced both racial and nonracial discrimination, compared with those exposed to 1 type of discrimination only. Roberts et al.[37] found that nonracial discrimination was associated with risk of hypertension among African American women in the Pitt County Study, while racial discrimination was not. Our findings provide evidence that the association between racial bias and physical health outcomes is not distinct from unfair treatment occurring due to other causes, indicating that similar mechanisms may underlie the association of various types of discrimination with health outcomes.

Our data suggest that experiences of discrimination in major life situations, or more conspicuous sources of discrimination, have lasting consequences for diabetes risk whereas experiences of more subtle forms of day-to-day discrimination do not. Given that chronic, everyday stressors are generally more strongly associated with disease onset as compared with acute stressors,[4,12,16] we were surprised by the lack of association between everyday discrimination and incident diabetes. However, these more subtle types of unfair treatment may not have the same influence as major experiences of discrimination. Several other studies have found weak or no associations between everyday discrimination and various physical health outcomes. For example, investigators have reported associations between major experiences of discrimination, but not everyday discrimination, with incident breast cancer[8] and prevalent hypertension.[24] A prior MESA publication found an association between both major experiences of discrimination and everyday discrimination with incident cardiovascular events; in that study, everyday discrimination showed a weaker association and was significant only in men.[6] While the mechanisms explaining these findings are unclear, it appears that major experiences of discrimination may have differential associations with specific health outcomes as compared with everyday experiences of discrimination. It is also possible that recall of major forms of discrimination is more accurate than for everyday discrimination.

We found no significant differences in associations by race/ethnicity, age, or sex. This suggests that the risk of diabetes as a result of discrimination does not appear to differ across demographic characteristics. However, our power to detect differences by race/ethnicity was limited. Consistent with the existing literature,[6,23] the prevalence of self-reported discrimination differed substantially by race/ethnicity, with African Americans reporting higher levels than all other racial/ethnic groups.

This study has multiple limitations that should be noted. Experiences of discrimination were assessed via self-report, and therefore subject to recall and social desirability bias. Discrimination was assessed at baseline only, and it is possible that a participant's experience of discrimination may have changed during the follow-up period, leading to misclassification bias. Studies have found that some disadvantaged groups may cope with discrimination by denying or minimizing its occurrence,[2] thus leading to underreporting of experiences of discrimination. This may explain why, in our sample, people with lower socioeconomic status were less likely to report having experienced discrimination. Nonetheless, if underreporting occurred, the discrimination-diabetes association would have been biased toward the null. Detailed data were not collected on coping methods. Active coping strategies, such as seeking social support, may be effective in preventing the adverse health consequences associated with discrimination.[38] Future studies should include assessment of coping methods, because coping may potentially mediate the association between discrimination and health outcomes. Finally, while the study is relatively large for a multiracial/multiethnic cohort, the numbers in some subgroups were somewhat small. This was especially true for the Chinese and Hispanic groups, and may have hindered our ability to detect racial/ethnic differences in associations between discrimination and incident diabetes.

Strengths of this study included the use of data from a large, prospective multiracial/multiethnic cohort. Discrimination was assessed using 2 scales, which allowed for a more in-depth examination of the role of discriminatory events on diabetes incidence. Furthermore, the Major Experiences of Discrimination Scale used in this study assesses unfair treatment generally and then asks for specific attributions. Evidence suggests that this method of assessing discrimination reduces bias in comparison with studies that ask about specific types of discrimination first.[4] We used an objective assessment of diabetes and were also able to control for many potential confounders and mediators in our analyses.

In conclusion, we found that major experiences of discrimination were associated with incident diabetes in a diverse sample of middle-aged to older adults, even after controlling for potential confounders and mediators. This association remained when examining discrimination attributed to race/ethnicity or other causes. Given that this is the first study to examine the relationship between discrimination and incident diabetes, it is important for future studies to confirm these findings and to further explore the mechanisms linking discrimination and diabetes.