Objective: To estimate the overall and age-specific incidence of known diabetes and its total duration through prevalence data and to assess the consistency of the results by mortality analysis of the same cohort.
Research design and methods: Two different sources were used. The first was a representative sample of 2,274 prevalent known-diabetic subjects. These data provided overall and age-specific incidence estimates by fitting a logistic model to the partial incidence rates for different diagnosis cohorts and to the disease duration. The age at diagnosis structure was built from the age-specific estimates. Prevalence data also provided total duration estimates by converting the prevalent duration-to-date structure into an incident total duration structure. The second source was 145 deceased subjects who were taken from the 6-year follow-up sample of 1,132 prevalent subjects. The age at diagnosis and estimates of total disease duration were provided for these subjects, who paralleled the characteristics of the incident cohort.
Results: The two independent estimates of total disease duration were similar (prevalent subjects, 15.7 years; deceased subjects, 14.1 years): the average duration was 14.9 years. The ratio between prevalence and total duration yielded an independent yearly incidence estimate of 2.2 per 1,000 person-years (men, 2.0; women, 2.4), which was close to the value given by the model of 2.1 per 1,000 person-years (men, 1.9; women, 2.3). Also, the independently determined age structures overlapped, and their average was used to calculate the age-specific incidence. Incidence was negligible for individuals < 30 years of age, and it was about 6.0 per 1,000 person-years for individuals > 50 years of age.
Conclusions: This study provided reliable estimates of NIDDM age-specific incidence rates and total disease duration, data that are seldom investigated in this type of disease.