Concept Architecture
Concept
Theoretically, Value of Information (VOI) is a decision-theoretic framework that quantifies the expected benefit of reducing uncertainty before making a healthcare decision. It is founded on Bayesian decision theory and measures the economic value of obtaining additional information that may improve decision-making. In health economics, value of information analysis is used to determine whether further research is worthwhile, prioritise research investments and optimise evidence generation.
Mathematically, value of information compares the expected net benefit achievable after reducing uncertainty with the expected net benefit obtainable using current evidence. The difference represents the expected value of additional information. The framework encompasses several related measures, including Expected Value of Perfect Information, Expected Value of Partial Perfect Information, Expected Value of Sample Information and Expected Net Benefit of Sampling, each addressing different aspects of uncertainty and research design.
In practice, value of information is estimated using probabilistic sensitivity analysis combined with Bayesian decision models and simulation methods. Health economists use value of information analysis to quantify decision uncertainty, evaluate the potential return from future research and determine whether additional evidence is likely to improve reimbursement, pricing and resource allocation decisions.
Purpose
Used to quantify the economic value of reducing uncertainty, prioritise future research and support efficient healthcare decision-making under uncertainty.
Mathematical Formulae
Primary Formula
VOI = Expected Value with Additional Information ? Expected Value with Current Information
Supporting Formulae
Expected Value of Perfect Information:
EVPI = E[max(NMB)] ? max(E[NMB])
Expected Value of Sample Information:
EVSI = E[max(E(NMB|Data))] ? max(E[NMB])
Expected Net Benefit of Sampling:
ENBS = EVSI ? C
where:
NMB = Net Monetary Benefit
C = expected research cost
Related Mathematical Methods
- Expected Value of Perfect Information
- Expected Value of Partial Perfect Information
- Expected Value of Sample Information
- Expected Net Benefit of Sampling
- Population EVPI
- Bayesian Decision Theory
- Probabilistic Sensitivity Analysis
- Net Monetary Benefit
Example
A probabilistic sensitivity analysis identifies substantial uncertainty regarding the cost-effectiveness of a new oncology treatment. The per-person EVPI is estimated at �380, indicating that eliminating all uncertainty could improve expected decision-making by this amount. Further analyses estimate EVPPI for treatment effectiveness, EVSI for a proposed clinical trial and ENBS after accounting for research costs. Together, these measures determine whether additional research is justified and identify the most efficient study design.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| MAX | =MAX(B2:D2) | Identify the intervention with the greatest net monetary benefit in each simulation. |
| AVERAGE | =AVERAGE(ResultRange) | Estimate expected net monetary benefit across simulations. |
| SUMPRODUCT | =SUMPRODUCT(EVPI,Population) | Calculate population value of information measures. |
| IF | =IF(ENBS>0,"Research justified","Research not justified") | Determine whether additional research is economically worthwhile. |
VBA (Optional)
VBA can automate value of information analyses by calculating EVPI, EVPPI, EVSI and ENBS across probabilistic sensitivity analyses and alternative research designs.
Sources
- Claxton K. The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies. Journal of Health Economics. 1999;18(3):341?364.
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press.
- Ades AE, Lu G, Claxton K. Expected value of sample information calculations in medical decision modelling. Medical Decision Making. 2004;24(2):207?227.
- ISPOR Value of Information Good Practice Reports.
- NICE. NICE Health Technology Evaluations: The Manual.
Related Concepts (2)
Institutional Perspectives (2)
- ZIN
EVPI and EVPPI Mandatory in the Reference Case
The 2024 Dutch guideline makes value-of-information analysis part of the reference case: reporting the expected value of perfect information (EVPI) and expected value of partial perfect information (EVPPI) is mandatory, estimated from the probabilistic analysis and presented as patient- and population-level curves over a 5-year horizon; EVSI and ENBS are optional additions.
Zorginstituut Nederland, Guideline for Economic Evaluations in Healthcare (2024)View source → - NICE
Value-of-Information Reasoning Informs Research Recommendations (Not Mandated)
NICE does not mandate value-of-information analysis in every submission, but uses value-of-information reasoning — supported by DSU methods — to identify where decision uncertainty is high and to frame research recommendations (including managed access / data-collection arrangements) rather than requiring EVPI/EVPPI as a standing reference-case output.
NICE DSU value-of-information methods; NICE manual (PMG36)View source →
Library
Publications
1
Value of Information Analysis for Research Decisions — An Introduction: Report 1 of the ISPOR Value of Information Analysis Emerging Good Practices Task Force — Fenwick, Steuten, Knies, Ghabri, Basu, Murray, Koffijberg, Strong, Sanders Schmidler & Rothery, Vol. 23, No. 2 ed., 2020 (Value in Health)
The introductory ISPOR good-practice report on value-of-information (VOI) analysis, explaining how VOI quantifies the value of reducing decision uncertainty through further research and where it fits in resource-allocation decisions.
Journal ArticleView source →
Frequently Asked Questions (6)
What is value of information?
A framework quantifying the expected benefit of reducing decision uncertainty, comparing expected outcomes under current uncertainty against outcomes if it were resolved.
Source: Raiffa & Schlaifer 1961
Why does value of information treat uncertainty as costly?
Uncertainty is costly because, when the true values are unknown, the option chosen may turn out not to be the best, and the health or money lost through such wrong decisions is a real cost. Value of information measures this by asking how much better decisions would be if the uncertainty were reduced or removed, expressing the gain as the value of the information that would achieve it. Framing uncertainty as an expected loss that evidence can reduce is what gives the approach its power. It prices knowledge by its effect on decisions. Claxton and Sculpher (2006) describe this.
Source: Claxton & Sculpher 2006
How does value of information work?
Value of information works by comparing decisions and their expected outcomes with and without additional information: under current uncertainty, the best option is chosen on expected value, but it may not be truly best, so resolving uncertainty allows better choices for each possible truth. The expected gain from making better-informed decisions, averaged over the uncertainty, is the value of information. Perfect information resolves all uncertainty, partial information a subset of parameters, and sample information the finite data from a study, each giving a corresponding value.
Source: Claxton & Posnett 1996
Why is value of information useful?
Value of information is useful because it links the uncertainty in an analysis to the decisions it informs, quantifying the benefit of reducing uncertainty in the same terms as the decision, such as health or money. This shows whether further research is worthwhile, which parameters are most worth investigating, and which study designs offer the best value. By valuing information through better decisions, it provides a principled basis for prioritising and designing research, directing evidence collection to where it most improves outcomes rather than relying on convention or intuition.
Source: Claxton & Posnett 1996
What are the main value of information measures?
The main value-of-information measures are the expected value of perfect information, the value of resolving all uncertainty, giving an upper bound; the expected value of partial perfect information, the value of resolving a specific subset of parameters, identifying which matter most; and the expected value of sample information, the value of the finite information from a specific study. Scaled to the affected population and combined with research costs through the expected net benefit of sampling, these measures indicate whether, on what, and how to conduct research.
Source: Raiffa & Schlaifer 1961
What are the limitations of value of information?
Value of information depends on the decision model and its assumed parameter distributions, so its estimates inherit any errors or uncertainty in these, and on the population size and decision relevance period used to scale it, which require forecasting the future. The calculations, especially for partial and sample information, can be computationally demanding. It values information only through the modelled decision, potentially missing wider considerations. These limitations mean value-of-information results are treated as informative but approximate, computed with appropriate methods and their sensitivity to key assumptions examined.
Source: Claxton & Posnett 1996
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 3 Nov 2025
Content version: 1.0.0
Canonical Identity
- Term code
- HE-EM-VI-036
Stable URI · Machine-readable · Resolvable · CC BY 4.0