Eliciting uncertainty for complex parameters in model-based economic evaluations: quantifying a temporal change in...

Area of Science:

• Health Economics
• Gerontology
• Clinical Trial Analysis

Background:

• Model-based economic evaluations often rely on extrapolated treatment effects from short-term clinical trials.
• Extrapolation using limited follow-up data may inadequately represent uncertainty.
• Quantifying uncertainty in treatment effect extrapolation is crucial for accurate economic modeling.

Purpose of the Study:

• To utilize structured expert elicitation to quantify uncertainty in the extrapolation of treatment effects observed in clinical trials.
• To compare expert-derived uncertainty with extrapolation solely from trial data.

Main Methods:

• Conducted a structured expert elicitation exercise with 38 healthcare professionals and academics.
• Used a web application to gather experts' probabilistic beliefs on falls and fracture rates under intervention versus usual care.
• Derived temporal changes in treatment effects from expert priors to extrapolate trial outcomes.

Main Results:

• Thirty-eight experts participated, predominantly believing treatment effects diminish over time.
• Experts expressed greater uncertainty in extrapolation compared to methods using only trial outcomes.
• Observed relatively small variation in predicted outcomes among experts.

Conclusions:

• Structured expert elicitation is a viable method for informing uncertainty in extrapolation for economic evaluations.
• Incorporating expert judgment requires careful consideration of assumptions and simplifications.
• This approach enhances the robustness of long-term treatment effect predictions.