A New Type 2 Diabetes Microsimulation Model to Estimate Long-Term Health Outcomes, Costs, and Cost-Effectiveness

Area of Science:

• Health economics
• Epidemiology
• Biostatistics

Background:

• Type 2 diabetes poses a significant public health burden in the US, necessitating effective interventions.
• Accurate estimation of health effects, costs, and cost-effectiveness is crucial for evaluating interventions.
• Existing models may have limitations in predicting outcomes for US populations.

Purpose of the Study:

• To develop and validate a microsimulation model for estimating health effects, costs, and cost-effectiveness of type 2 diabetes interventions.
• To utilize US-based data for enhanced prediction accuracy in the US context.
• To demonstrate the model's utility in assessing intervention cost-effectiveness.

Main Methods:

• Developed a microsimulation model incorporating newly derived equations for complications, mortality, risk factor progression, patient utility, and costs.
• Utilized US-based studies for all model equations.
• Performed internal and external validation, and simulated outcomes for a cohort of 10,000 US adults with type 2 diabetes.

Main Results:

• The model demonstrated good prediction accuracy, with an average absolute difference < 8% for 17 complications in internal validation.
• External validation showed better performance in clinical trials than observational studies.
• Simulated outcomes for a US cohort included 19.95 life-years, $187,729 medical costs, and 8.79 quality-adjusted life-years (QALYs).
• Reducing hemoglobin A1c from 9% to 7% yielded an incremental cost-effectiveness ratio of $9103 per QALY.

Conclusions:

• The developed microsimulation model, based exclusively on US data, achieves good prediction accuracy for US populations.
• The model is a valuable tool for estimating the long-term health impact, costs, and cost-effectiveness of type 2 diabetes interventions in the United States.