HTAtlas
← Back to explore

Moment matching with nonlinear regression for EVSI across sample sizes

Methodpeer-reviewed✓ Source-grounded

A fast and accurate method to estimate the economic value of future research studies of different sizes, helping to choose the most cost-effective study design by predicting how much uncertainty can be reduced.

At a glance

Use when

Designing clinical trials where economic value of information is a key criterion, comparing multiple study designs efficiently

Avoid when

Using extremely complex models where even moment matching is computationally prohibitive, or when prior information is highly unstable

Inputs

Health economic model outputs, prior distributions, sample size range, simulation results from the model

Outputs

Estimated EVSI values across specified sample sizes, predicted optimal sample size for maximum net benefit

How it works

This method extends the moment-matching approach for estimating Expected Value of Sample Information (EVSI) across varying sample sizes. It computes posterior variances of net monetary benefit for different sample sizes and applies Bayesian nonlinear regression to estimate EVSI efficiently, enabling optimization over trial designs with minimal computational overhead compared to single EVSI calculations.

HTA domains
Clinical Effectiveness, Costs & Economic Evaluation
Assumptions
The relationship between sample size and posterior variance can be approximated; the health economic model is stable across simulations; Bayesian nonlinear regression fits the data well
Strengths
Fast computation relative to repeated EVSI estimation, enables comparison of many trial designs, retains accuracy for realistic models
Limitations
Requires rerunning the health economic model, which may be costly for complex models; performance depends on model stability and regression fit
Also known as
Moment matching for EVSI optimization, Nonlinear regression for EVSI

Questions this answers

References & sources

Similar by meaning

Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.