Evaluation of a population-based screening for type 2 diabetes: a community-based screening project in Puli, Taiwan.

Abstract

A Markov method incorporating the relationships between prevalence, incidence, and mortality with respect to type 2 diabetes was used to assess a population-based screening for this disease. Data from a population-based screening project for residents of Puli, Taiwan, over 30 years of age (n= 1,219) were used to estimate the annual incidence of asymptomatic type 2 diabetes, the prevalence to incidence (P/I) ratio, and the hazard rate of death due to type 2 diabetes. These parameters were employed to develop a Markov process to evaluate the effects of early detection of type 2 diabetes on the risk of death from this disease in a simulated population (n= 10,000) receiving biennial, 5-year interval, or no screening. The estimated annual incidence, average duration from asymptomatic to symptomatic type 2 diabetes (P/I ratio), and hazard rate for death from this disease were 0.86% (95% CI 0.50-1.48), 10 years (95% CI 7.69-14.01), and 1.1% per year, respectively. This yields an optimal screening interval of 5 years. Simulation of a 5-year interval screening regimen versus no screen ing yielded a relative risk reduction of 31% (95% CI 12-46%). A similar value was found for a biennial screening regime. The results suggest that early detection of type 2 diabetes via a community-based screening project in developing countries with high prevalence is worthwhile.