Maximum net monetary benefit decision rule

Selects, at a stated threshold, the mutually exclusive option with the highest net monetary benefit. The rule reaches the same choice as a correctly conducted fully incremental analysis without ordering options or removing dominated ones.

Signature

j* = argmax_(j in J) (lambda * E_j - C_j)
Inputs
InputsDefinitionUnit
lambdaMonetary value placed on one unit of health effectcurrency per unit of health effect
E_jExpected health effect of option jhealth-outcome unit per defined population or person
C_jExpected total relevant cost of option jcurrency per defined population or person
Output
j*Option with the highest net monetary benefit at threshold lambdaoption index
  • J Set of mutually exclusive options in the decision problem, indexed 1 to n (set of options)

Function

Net monetary benefit function

Values an option's expected health effect in money at a stated cost-effectiveness threshold and subtracts its expected cost, placing health and cost on one monetary scale.

Implementations

  • Excel

    Return the highest-NMB option

    Excel returns the option name with the highest NMB and flags a tie when more than one option shares the maximum.

    =IF(COUNTIF(NMBValues,MAX(NMBValues))>1,"Equal highest NMB",INDEX(OptionNames,MATCH(MAX(NMBValues),NMBValues,0)))

Assumptions

  • Complete set of mutually exclusive options

    J contains every relevant option, and each is evaluated on the same basis. Leaving out an option can change which option has the highest NMB.

  • Expected net benefit under uncertainty

    Under probabilistic analysis the rule uses expected NMB, the mean across simulations, and not the probability that each option has the highest NMB.

Worked examples

  • Four options at £20,000 per QALY

    Options A to D have expected costs of £10,000, £13,000, £18,000 and £22,000 and 4.0, 4.3, 4.5 and 4.8 expected QALYs. At £20,000 per QALY their NMBs are £70,000, £73,000, £72,000 and £74,000, so option D, the fourth, is selected. Option C, which is extendedly dominated, is not selected at this or any other threshold. The figures are illustrative.

    J = [1,2,3,4]; lambda = 20000; E_j = [4.0,4.3,4.5,4.8]; C_j = [10000,13000,18000,22000]; j* = 4

Common errors

  • Choosing the option most likely to be cost-effective

    The option with the highest probability of having the highest NMB across simulations can differ from the option with the highest expected NMB. The decision rule uses expected NMB.

Sources

  • Decisions based on mean net benefit

    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.

    View source →

Canonical Identity

Stable URI · Machine-readable · Resolvable · CC BY 4.0