Diabetes Mellitus and Marital Status: Evidence from the National Longitudinal Mortality Study on the Effect of Marital Dissolution and the Death of a Spouse

Abstract

Purpose

This study evaluates the full impact of marital status on diabetes mellitus by stratifying the analysis by gender, including socioeconomic covariates and, unlike most studies, extending marital status by separating out previously conflated status categories.

Methods

Release 5 of the National Longitudinal Mortality Study (NLMS) was used for the data. Logistic regression was applied to the data from 1990 to 2011. The effective sample size consists of 1,384,507 individuals age 18 and above recruited into the study (via the Current Population Surveys), 3,955 of whom had died of diabetes mellitus by 2011.

Results

For minority men and non-Hispanic white men, divorced/separated status was significantly related to diabetes mortality, respectively (OR=1.318, CI=1.010, 1.719; and OR=1.283, CI=1.054, 1.562). For minority women and non-Hispanic white women, widowed status was related to diabetes mortality, respectively (OR=1.349, CI=1.107, 1.643; and OR=1.262, CI=1.113, 1.431).

Conclusion

Contrary to recent epidemiological studies in which divorced/separated and widowed status were combined into one covariate, this United States study finds that divorced/separated men and widowed women are at increased risk for diabetes mellitus mortality, and that among these populations at risk, minorities are at higher risk than whites. The study highlights the importance of marital status and gender differences in the risk of death from diabetes.

Keywords:

• diabetes
• gender
• divorced
• widowed
• socioeconomic disparities

Ethical Approval and Informed Consent

Research for this study did not require an institutional review board or ethics committee. NLMS data are publicly available to researchers on the World Wide Web and all the data are anonymized prior to release. Without identifying information consent was not needed.

Acknowledgments

The US Bureau of the Census and the National Center for Health Statistics are gratefully acknowledged for providing the data for this research. The views and opinions presented in this paper do not necessarily reflect those of these institutions. We also thank the anonymous reviewers for their generous and constructive comments.

Disclosure

The authors report no conflicts of interest in this study.