Accounting for uncertainty in health economic decision models by using model averaging.

Christopher H Jackson,Linda D Sharples,Simon G Thompson

doi:10.1111/j.1467-985x.2008.00573.x

Abstract

Health economic decision models are subject to considerable uncertainty, much of which arises from choices between several plausible model structures, e.g. choices of covariates in a regression model. Such structural uncertainty is rarely accounted for formally in decision models but can be addressed by model averaging. We discuss the most common methods of averaging models and the principles underlying them. We apply them to a comparison of two surgical techniques for repairing abdominal aortic aneurysms. In model averaging, competing models are usually either weighted by using an asymptotically consistent model assessment criterion, such as the Bayesian information criterion, or a measure of predictive ability, such as Akaike's information criterion. We argue that the predictive approach is more suitable when modelling the complex underlying processes of interest in health economics, such as individual disease progression and response to treatment.

Highlights

The parameter uncertainty that is inherent in these quantities is accounted for by probabilistic sensitivity analysis, which provides simulated distributions of incremental costs and benefits. This leads to the probability PCE.λ/ that endo-vascular aneurysm repair (EVAR) is cost effective compared with open repair, defined as the probability of a positive incremental net benefit: PCE.λ/ = P.λΔB − ΔC > 0/:
Substantive conclusions Estimated probabilities that EVAR is cost effective compared with open repair, for a threshold of £20 000 per quality-adjusted life years (QALYs), ranged from 0.01 under the most probable base case, to the alternative of 0.081 where one covariate effect was omitted, to 0.52 under a scenario where all covariate effects were omitted
Using Bayesian model averaging based on AIC, we obtained an estimate of 0.069 for this probability, which forms a statistically principled compromise between the alternative assumptions

Summary

Uncertainty in health economic decision models

The parameter uncertainty that is inherent in these quantities is accounted for by probabilistic sensitivity analysis, which provides simulated distributions of incremental costs and benefits This leads to the probability PCE.λ/ that EVAR is cost effective compared with open repair, defined as the probability of a positive incremental net benefit: PCE.λ/ = P.λΔB − ΔC > 0/:. This was taken to be a hazard ratio of 3.06 (95% confidence interval 1.12–8.36) for EVAR compared with open repair This was calculated from a piecewise exponential survival model on the EVAR1 trial data and assumed only to operate in the second year after the surgery. This technique is applicable to a wide range of model uncertainty problems as well as covariate selection

Model averaging

Akaike’s information criterion

Marginal likelihood and Bayes factors

Findings

Application to surgery for aneurysm repair

Discussion