Gender differences in the socioeconomic gradient in self-reported diabetes: does health service access play a role?

To investigate the hypothesis that gender difference in the socioeconomic gradient of self-reported diabetes might be partially determined by a detection bias: among low SEP men, but not women, a less frequent use of medical care would result in lower diabetes detection and awareness.

We conducted a cross-sectional analysis of 2387 (57.1% women) Brazilian university non-faculty civil servants. We quantified the association between educational attainment (a marker of socioeconomic position) and self-reported diabetes through gender-specific logistic regression models adjusting for age. Health insurance coverage (a marker of potential health care access) and never having had serum cholesterol tested (a marker of actual care access) were analyzed to investigate the role of detection bias.

Compared to participants with college education or higher, the adjusted odds ratio (OR) and 95% confidence interval for diabetes for those with less than high school was 2.5 (1.0-6.5) in men and 5.0 (2.1-11.7) in women. Only among men we observed an increment in this OR after an additional adjustment for markers of care access [men 3.5 (1.3-9.1); women 4.9 (1.9-12.1)].

Our findings suggest that health service access may explain some of the gender difference in the socioeconomic gradient of self-reported diabetes.