OBJECTIVES: Despite the growing burden of type 2 diabetes (T2D) in Japan and the corresponding need to optimize healthcare resource use, little has been published on T2D models that could support economic evaluation of new diabetes interventions in this setting. The aim of the present study was to perform a validation analysis using the PRIME T2D Model to understand its potential validity for use in Japanese cost-effectiveness analyses.

METHODS: A literature review was performed to identify recently published, long-term cohort studies (≥3 years of follow up) in patients with T2D in Japan. Studies identified by literature search were screened to select those that reported long-term cumulative incidence of diabetes-related complications, with information on baseline characteristics and treatment regimens during follow up. The PRIME T2D Model was used to run simulations designed to reproduce the outcomes of the published studies, by matching cohort characteristics, treatment effects/risk factor progression and running simulations over an equivalent time horizon.

RESULTS: Four key studies were identified with endpoint data suitable for validation analysis from the broader literature review. Validation analyses were performed for ischemic heart disease (IHD, 2 comparisons), myocardial infarction (MI, 3 comparisons), stroke (4 comparisons) and end-stage renal disease (ESRD, 1 comparison). Where endpoint definitions were well matched, the model produced results that were comparable with real world data for MI and stroke (root means squared difference 0.45%). However, mismatched endpoint definitions meant that other outcomes could not be directly compared (IHD and ESRD).

CONCLUSIONS: The PRIME T2D Model, a novel, product independent analysis tool, is one of the few available diabetes models that utilize risk data specific to Asian populations to evaluate long-term outcomes in patients with T2D. This analysis provides evidence that the PRIME T2D Model is capable of reproducing real-life cohort outcomes in the Japanese setting.