Article Text
Abstract
Background That type 2 diabetes mellitus (T2DM) varies significantly by race and ethnicity is a widely accepted fact. It is often invoked as a base rate (a priori probability) during the process of clinical reasoning. Epidemiological studies repeatedly show undiagnosed T2DM varies more by socioeconomic status (SES), than by race/ethnicity. This study seeks to understand the discrepancy between the true prevalence of undiagnosed T2DM by SES and its continually reported prevalence by race/ethnicity.
Methods Data from two different but complementary studies are employed: a) a large Boston Area Community Health (BACH) survey; and b) a factorial experiment conducted with primary care doctors to examine variations in clinical decision making. The BACH epidemiologic survey (n=5502) employed a stratified, multi-stage cluster sample design and used multivariable techniques including logistic regression. The factorial experiment concerning decision making employed clinically authentic videotaped scenarios presented to primary care doctors (n=192), and used ANCOVA analyses.
Results Results from the epidemiologic survey show that both undiagnosed signs and symptoms and diagnosed T2DM vary similarly by socio-economic status (SES). This finding is independently corroborated by National Health and Nutrition Examination Survey (NHANES) data for diagnosed T2DM. Complementary data from the clinical decision making experiment show the diagnosis of T2DM varies significantly by a patients’ race/ethnicity, controlling for SES, age and gender in the design. While undiagnosed signs and symptoms of T2DM in the community vary significantly by SES, rather than race/ethnicity, following diagnosis by primary care doctors they vary more by race/ethnicity, rather than by SES.
Conclusion Race/ethnicity and SES in the US are almost totally confounded, such that measuring one is essentially also measuring the other. Consequently, doctors generally get the social patterning of T2DM right, but for entirely the wrong reason. Continued patterning of T2DM by race/ethnicity motivates the search for genetic and biophysiologic explanations and distracts attention from the more important and potentially modifiable contribution of SES circumstances to the prevalence of T2DM.