Objective: To develop alternative estimates of the excess mortality attributable to diabetes in Mississippi.
Methods: For each age/sex group, the total (all cause) mortality rate in the population was partitioned into the mortality rates for diabetics (M1) and non-diabetics (M0), using age- and sex-specific mortality rate ratios for diabetics from the Verona Diabetes Study. M1 and M0 were used, respectively, to calculate the predicted number of deaths and the 'usual' number of deaths (the number expected if diabetics had the same mortality rates as non-diabetics) in diabetics. The difference between predicted and 'usual' deaths is the number of excess deaths in diabetics--that is, deaths attributable specifically to diabetes. This was compared to the number of deaths in which diabetes was registered on death certificates as an underlying and/or contributory cause.
Results: In Mississippi in 1998-2000, 2,322 male and 3,436 female deaths in diabetics were calculated to be excess. These numbers are approximately three times the numbers of deaths in which diabetes was registered on the death certificate as the underlying cause of death. The excess numbers of deaths are much closer to the total numbers of deaths due to diabetes (that is, where diabetes is listed as either an underlying or a contributory cause of death).
Conclusions: When reporting diabetes mortality numbers, public health agencies should consider reporting total registered diabetes deaths, not just registered underlying deaths. Better education of providers and others to increase the awareness of the importance of listing diabetes on the death certificate would help to increase concern and make diabetes a priority for public health action. Persons with diabetes have higher mortality rates and excess premature mortality, and diabetes ranks consistently among the leading causes of death in Mississippi and the U.S. Yet it has long been known that mortality attributable to diabetes is underestimated. Cause-of-death statistics are unreliable due to underreporting of diabetes on death certificates, and estimates based on attributable risk calculations do not take into account competing risks for mortality. In view of the public health importance of diabetes, it would be useful to develop alternative estimates of the burden of excess mortality attributable to it. In this paper I estimate the excess mortality attributable to diabetes for the diabetic population of Mississippi using epidemiologic data on the prevalence of diabetes and the relative risk of mortality among persons with diabetes.