The accuracy of a death certificate checkbox for diabetes: early results from New Jersey.

OBJECTIVES: The rapid growth in diabetes prevalence has increased interest in measuring the burden of this disease. One response has been to add a checkbox for diabetes status to the death certificate, which New Jersey did in 2004. This study assessed the accuracy of the diabetes checkbox and its effect on cause-of-death coding. We analyzed whether a diabetes checkbox is a useful addition to the death certificate.
METHODS: We examined the trend in cause-of-death coding for diabetes as an underlying and contributing cause of death by analyzing New Jersey mortality data between 1990 and 2005. We assessed the accuracy of the checkbox by examining inconsistencies between cause-of-death coding and checkbox status, and assessed sensitivity by analyzing linked hospital and death data for a cohort of decedents with diabetes.
RESULTS: Between 2003 and 2005, there was approximately a 15% increase in the number of deaths listing diabetes as a contributing cause. The number of deaths where diabetes was listed as an underlying cause changed little. Approximately 10% of death certificates had an inconsistency between cause of death and checkbox status. The sensitivity analysis showed that approximately 40% of diabetic decedents had the appropriate checkbox status.
CONCLUSION: The addition of the checkbox was accompanied by a change in the reporting of diabetes as a contributing cause of death. Results from the sensitivity analysis raise questions about the accuracy of the checkbox as a measure of the diabetic status of decedents.