Estimating Lifetime Duration of Diabetes by Age and Gender in the Korean Population Using a Markov Model

Abstract

Background: Duration of type 2 diabetes is clinically important. Duration of morbidity is an independent and critical predictor of developing its complications. This study aims to explore an applicability of a Markov model to estimate the duration of diabetes in the Korean population. Methods: We constructed the Markov model with two Markov states, diabetes and death, for estimation of duration of diabetes. The cycle of the Markov model was 1 year. Each diabetes onset by 5 years was considered from 30 to 85 years old or above. The endpoint of the Markov was 100 years old. Type 2 diabetes was operationally defined using the 10th revision of International Statistical Classification of Diseases and prescriptions of anti-diabetic drugs from the National Health Insurance Services-National Sample cohort. In each incident and existing prevalence cases, survival probabilities were obtained. Durations of diabetes from the Markov model were compared with those from the DisMod II program. Reductions of life expectancy due to diabetes were defined as differences of life expectancies between diabetic patients and the general public. Sensitivity analyses were also conducted using a cure rate and 95% confidence interval of survival probability. Results: The duration of diabetes gradually decreased with incident age in both genders. In the early 30s, the duration was the largest at 48.9 and 41.9 years in women and men, respectively. In the average incident age group of type 2 diabetes, the late 50s, the reduction of life expectancy due to diabetes was estimated to be about two years in both genders. As annual cure probabilities increased, the durations of diabetes were reduced. Conclusion: This study estimated the duration of diabetes using a Markov model. The model seems to work well and diabetes could reduce life expectancy by about 2 years on average. This approach could be useful to estimate the duration of illness, calculate disability-adjusted life years, and conduct economic evaluation studies on interventions for diabetic patients.