Projection of prediabetes and diabetes population size in Singapore using a dynamic Markov model

Abstract

The aim of the present study was to forecast the prevalence and number of adult Singapore residents with prediabetes and diabetes in 2035. A dynamic Markov model with nine mutually exclusive states was developed based on the clinical course of diabetes using time-dependent rates and probabilities. A 1-year cycle over a 25-year time horizon from 2010 to 2035 was used in the model. With publicly available data and a chronic disease register, the model forecast annual disease burden by simulating transition of cohorts across different health states using prevalence rates, incidence rates, mortality rates, disease transition, disease detection, and complication rates. An aging index was used in the model in anticipation of population aging to minimize risks of underestimating disease burden. From 2010 to 2035, the number of Singapore residents with prediabetes and diabetes is projected to more than double, from 434 685 to 903 596 and from 373 104 to 823 802, respectively. The prevalence of prediabetes and diabetes will rise steadily from 15.5% to 24.9% and from 13.3% to 22.7%, respectively. By 2035, a further estimate of 733 174 and 100 250 patients with prediabetes and uncomplicated diabetes, respectively, will remain undiagnosed. The prevalence of detected and undetected complications is forecast to rise from 60.0% in 2010 to 70.2% by 2035. By 2035, the prevalence of prediabetes and diabetes among Singapore residents aged 21+ years is expected to be one in four and one in five, respectively. There is an impetus to adopt more aggressive interventions to contain disease progression.