UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Validation of a diabetes model : a process to select and assess the validity of an existing model for economic evaluation of type 2 diabetes in British Columbia Smith, Kirby Lee


The growing burden of chronic disease like Type 2 Diabetes Mellitus (T2DM) raises concern about the sustainability of Canada’s health system. Consequently, greater emphasis is being placed on economic evaluation using computer simulation models to enhance decision making in health care. This research was initiated to assess if model output for a cost-effectiveness analysis (CEA) in T2DM would be generalizable to a diverse BC population. The thesis outline is as follows: • Chapter 1 overviews the epidemiology, risk factors and complications of T2DM, the role of model based economic evaluation and knowledge gaps used to guide the thesis objectives. • Chapter 2 assesses the methods of existing T2DM models and utilizes criteria to select a model for validation. • Chapter 3 describes an internal validation of a T2DM model by comparing the expected versus actual impact of changes to input parameters on output. • Chapter 4 describes an external validation comparing the predicted rate of vascular events to the observed rate by gender, age and ethnic sub-cohorts from a BC population based on the coefficient of determination (R²) and a 95% confidence interval (CI). • Chapter 5 summarizes the results, discusses limitations, and highlights future research to enhance the credibility of T2DM models. Results: The Ontario Diabetes Economic Model (ODEM) was selected and an internal validation demonstrated the simulated rate of vascular events responded as expected to changes in baseline variables in a 10 year simulation. The external validation in cohorts with no history of complications had a modest positive correlation (R² = 0.68) and a tendency to over predict vascular events in older adults. Adding individuals with previous events improved the correlation (R² > 0.99) and statistical accuracy of the ODEM. A higher correlation was observed in those of younger age, male gender (R² = 0.71) and SA ethnicity (R² = 0.77). Conclusion: The ODEM was demonstrated to be a functional model with output considered generalizable for the economic evaluation of a diverse BC T2DM population. There were trends in model to overestimate complications in cohorts with no previous vascular events and those of older age that require further research to validate.

Item Media

Item Citations and Data


Attribution-NonCommercial-NoDerivatives 4.0 International