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

Dina Jankovic,Mona Kanaan,Katherine Payne,Laura Bojke

doi:10.1017/s0266462322000022

Abstract

In model-based economic evaluations, the effectiveness parameter is often informed by studies with a limited duration of follow-up, requiring extrapolation of the treatment effect over a longer time horizon. Extrapolation from short-term data alone may not adequately capture uncertainty in that extrapolation. This study aimed to use structured expert elicitation to quantify uncertainty associated with extrapolation of the treatment effect observed in a clinical trial. A structured expert elicitation exercise was conducted for an applied study of a podiatry intervention designed to reduce the rate of falls and fractures in the elderly. A bespoke web application was used to elicit experts' beliefs about two outcomes (rate of falls and odds of fracture) as probability distributions (priors), for two treatment options (intervention and treatment as usual) at multiple time points. These priors were used to derive the temporal change in the treatment effect of the intervention, to extrapolate outcomes observed in a trial. The results were compared with extrapolation without experts' priors. The study recruited thirty-eight experts (geriatricians, general practitioners, physiotherapists, nurses, and academics) from England and Wales. The majority of experts (32/38) believed that the treatment effect would depreciate over time and expressed greater uncertainty than that extrapolated from a trial-based outcome alone. The between-expert variation in predicted outcomes was relatively small. This study suggests that uncertainty in extrapolation can be informed using structured expert elicitation methods. Using structured elicitation to attach values to complex parameters requires key assumptions and simplifications to be considered.

Highlights

Decision-analytic model-based economic evaluation (MBEE) is intended to generate relevant costs and treatment effects for competing alternatives, over an appropriate time horizon [1–3]
Temporal uncertainty arises when the time horizon for the specified MBEE exceeds the observed end point for available evidence
A multifaceted podiatry intervention involving education, exercise, foot orthoses, and footwear was designed to reduce the rate of falls and fractures in the elderly, with an associated impact on costs and health-related quality of life (HRQoL)

Summary

Introduction

Decision-analytic model-based economic evaluation (MBEE) is intended to generate relevant costs and treatment effects for competing alternatives, over an appropriate time horizon [1–3]. Outcomes can be extrapolated directly from the trial or instead, through scenarios representing assumptions or statistical models of the relative effect of treatment after the evidence end point [7;8]. Neither of these approaches appropriately quantify uncertainty around the generated predictions for use in MBEE. A bespoke web application was used to elicit experts’ beliefs about two outcomes (rate of falls and odds of fracture) as probability distributions (priors), for two treatment options (intervention and treatment as usual) at multiple time points These priors were used to derive the temporal change in the treatment effect of the intervention, to extrapolate outcomes observed in a trial. Using structured elicitation to attach values to complex parameters requires key assumptions and simplifications to be considered

Objectives

Methods

Results