Expected Value of Perfect Information

Concept

---

Theoretically, Expected Value of Perfect Information (EVPI) is a value of information measure that quantifies the maximum expected benefit of completely eliminating uncertainty before making a healthcare decision. It is founded on Bayesian decision theory and represents the economic value of having perfect knowledge of all uncertain model parameters. In health economics, EVPI is used to determine whether current decision uncertainty is sufficiently important to justify further research.

Mathematically, EVPI is calculated as the difference between the expected value obtained when decisions are made with perfect information and the expected value obtained when decisions are made using current information. It therefore represents the expected opportunity loss associated with making decisions under uncertainty and provides an upper bound on the value of any additional research.

In practice, EVPI is estimated using probabilistic sensitivity analysis by comparing the expected maximum net benefit across all possible parameter values with the maximum expected net benefit based on current evidence. Population EVPI is obtained by multiplying the per-person EVPI by the expected number of patients affected during the decision relevance period and is widely used to prioritise future research in health technology assessment.

Purpose

---

Used to quantify the maximum economic value of eliminating all decision uncertainty and determine whether additional research may be worthwhile before making healthcare decisions.

Mathematical Formulae

Primary Formula

EVPI = E[max(NMB)] ? max(E[NMB])

where:

NMB = Net Monetary Benefit

E = expectation over parameter uncertainty

Supporting Formulae

Net Monetary Benefit:

NMB = ?E ? C

Population EVPI:

Population EVPI = EVPI ? N

where:

N = affected population over the decision relevance period

Related Mathematical Methods

• Value of Information Analysis
• Expected Value of Partial Perfect Information
• Expected Value of Sample Information
• Expected Net Benefit of Sampling
• Probabilistic Sensitivity Analysis
• Bayesian Decision Theory
• Net Monetary Benefit

Example

---

A probabilistic sensitivity analysis estimates an expected maximum net monetary benefit of �18,250 per patient if perfect information were available. Under current evidence, the maximum expected net monetary benefit is �17,900 per patient.

EVPI = �18,250 ? �17,900

EVPI = �350 per patient

If 25,000 patients are expected to be affected during the decision relevance period, the population EVPI is:

Population EVPI = �350 ? 25,000 = �8.75 million

This indicates that eliminating all uncertainty could generate a maximum societal value of �8.75 million.

Excel Implementation

Function	Example Formula	Health Economics Application
MAX	=MAX(B2:D2)	Identify the maximum net monetary benefit for each simulation.
AVERAGE	=AVERAGE(E2:E10001)	Calculate the expected value of the maximum net monetary benefit.
MAX	=MAX(AVERAGE(B2:B10001),AVERAGE(C2:C10001),AVERAGE(D2:D10001))	Calculate the maximum expected net monetary benefit under current information.
SUMPRODUCT	=EVPI_per_patient*Population	Calculate the population EVPI.

VBA (Optional)

---

VBA can automate EVPI estimation from probabilistic sensitivity analysis outputs and calculate population EVPI across alternative decision scenarios.

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.
• Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness acceptability curves. Health Economics. 2001;10(8):779?787.
• ISPOR Value of Information Good Practice Reports.
• NICE. NICE Health Technology Evaluations: The Manual.