Background: Widening of social inequality in health is often discussed. Data regarding trends of the association between diabetes prevalence and social status are lacking.
Methods: Using two German health surveys (age 25-69 years), we estimated diabetes prevalences in 1998 compared to 1990-1992. Interaction of secular time with educational level, adjusted for age and BMI, were estimated in men and women using multiple regression models, considering the sample design.
Results: Diabetes prevalences in 1990-1992 and 1998 were 5.1% (95% CI 4.1-6.0) and 4.3% (3.5-5.1) in men, and 4.7% (4.0-5.4) and 3.8% (3.0-4.6) in women. It was significantly higher in older subjects and in obese subjects, and tended to be higher in lower educated subjects. Overall, prevalence tended to be lower in 1998 compared to 1990-1992, however, not statistically significant after adjustment for education and BMI (odds ratio, 95% CI: men 0.73; 0.39-1.37; women 0.41; 0.17-1.03). On a descriptive level, in the lowest education group, the diabetes prevalence was higher in 1998 compared to 1990-1992, whereas, it has decreased in higher educated subjects. However, confidence intervals were large, and we found no statistically significant interaction between calendar year and the educational level.
Conclusions: Diabetes prevalence tended to decrease in Germany during the 1990s. A widening of social disparity in diabetes risk might be present, but a significant increase could not be confirmed.