Heath-Manolopoulou-Baio moment-matching EVSI method
This method estimates how much value a future research study could add by using existing simulation results and statistical techniques to avoid lengthy computations.
At a glance
Use when
Estimating EVSI in health economic models where computational efficiency is important and probabilistic sensitivity analysis has already been performed.
Avoid when
The health economic model is extremely computationally intensive, making even the additional simulations prohibitive, or when the preposterior distribution is highly non-normal and moment matching is inadequate.
Inputs
Probabilistic sensitivity analysis samples, health economic model structure, prior distributions, potential future study design and sample size
Outputs
Expected Value of Sample Information (EVSI) estimate, distribution of the preposterior mean of incremental net benefit
How it works
A Bayesian approximation method for calculating the Expected Value of Sample Information (EVSI) by estimating the distribution of the preposterior mean of the incremental net benefit via moment matching, leveraging samples from probabilistic sensitivity analysis to improve computational efficiency.
- HTA domains
- Costs & Economic Evaluation
- Assumptions
- The distribution of the preposterior mean can be adequately approximated using moment matching; the future study design is known or can be simulated; the underlying model is suitable for Bayesian updating.
- Strengths
- Computationally efficient by reusing existing PSA samples; accurate compared to other EVSI methods; avoids nested Monte Carlo simulations; applicable to practical health economic models.
- Limitations
- Requires additional simulations which may be costly in complex models; accuracy depends on the quality of moment matching and underlying distributional assumptions.
- Also known as
- moment-matching EVSI method, Heath et al. EVSI method
Questions this answers
- › What is the expected economic value of conducting a future research study?
- › How much would additional evidence reduce decision uncertainty in a health economic model?
- › Can we estimate EVSI without running computationally expensive nested simulations?
- › How can probabilistic sensitivity analysis samples be reused to approximate EVSI?
- › What is the distribution of the preposterior mean of the incremental net benefit under different future data scenarios?
- › Is there a computationally efficient alternative to standard nested Monte Carlo methods for EVSI estimation?
References & sources
Similar by meaning
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.