LAUSR

ABSTRACT Objective Recent research identified that workplace factors play a role in the development of diabetes mellitus (DM). This study examines the longitudinal association of work-related overqualification with the incidence of DM over a 14-year follow-up period. Methods We used data from the 2003 Canadian Community Health Survey linked to the Ontario Health Insurance Plan and the Canadian Institute for Health Information Discharge Abstract databases. Cox proportional hazards regression models were performed to evaluate the relationship between overqualification and the incidence of DM. Results Over the study period, there were 91,835 person-years of follow-up (median follow-up = 13.7 years). The final sample included 7026 respondents (mean [standard deviation] age at baseline = 47.1 [8.2]; 47% female). An elevated risk of DM was associated with substantial overqualification (hazard ratio = 1.58, 95% confidence interval = 1.01–2.49) after adjustment for sociodemographic, health, and work variables. Additional adjustment for body mass index and health behaviors attenuated this risk (hazard ratio = 1.30, 95% confidence interval = 0.81–2.08). Underqualification was not associated with the incidence of DM in adjusted regression models. We did not observe any statistical difference in the effects of overqualification on DM risk across sex or education groups. Conclusions This study adds to the growing body of research literature uncovering the relationships between work exposures and DM risk. The results from the study suggest that higher body mass index and, to a lesser extent, health behaviors may be mediating factors in the association between overqualification and incident DM. Further research on the association of overqualification with DM is warranted.