O1-5.6 Forecasting Diabetes Prevalence: validation of a simple model with few data requirements

Abstract

<h3>Introduction</h3> Current projections of diabetes prevalence are mostly based on demographic change. Explicitly including trends in obesity and other risk factors could improve the accuracy of the projections and assist in evaluating policy options for prevention. <h3>Methods</h3> The model integrates population, obesity and smoking trends to estimate future diabetes prevalence using a Markov approach. Model parameters were derived from the literature, except for diabetes incidence which was estimated using DISMOD from the baseline estimation of diabetes prevalence. We developed a model for the US population (2000–2006), and validated the model outputs (NHANES prevalence and projections using a different model). <h3>Results</h3> US Diabetes mellitus prevalence estimated by the model (aged 25+) was 9.7% in 2000–2002 (7.8%–11.6%), increasing to 10.7% (8.6.3%–12.7%) by 2003–2006. Comparisons of the model results with the observed prevalence in the NHANES survey showed a close fit to the observed estimates (NHANES prevance 2003–2006 10.3%, 9.3%–11.3). The forecasts for 2030 was 19 19.3% (15.3%–23.0%). A different model (Narayan <i>et al</i>) for the same period and age group were 20.2%, 18.8%–21.6%. We tested the model for the England and Wales population obtaining a similar performance. <h3>Conclusions</h3> This model provides a reasonably close estimate of diabetes prevalence for the USA over the 2000–2006 period, compared with contemporary independent prevalence surveys in the same population and with a different model. Because of its few data requirements, the approach is now being tested in different middle income countries as a potential global diabetes prevalence forecast tool.