Skip to main content
HEOR Africa

Interactive tool

ICER and cost-effectiveness plane explorer

See how incremental cost, incremental effect and a willingness-to-pay threshold combine into an ICER, a net monetary benefit and a position on the cost-effectiveness plane.

Example values for demonstration. Calculations run in your browser; nothing you enter is stored or sent.

Inputs

Currency
Outcome measure

Intervention cost minus comparator cost. Negative means savings.

USD

Health gained by the intervention versus the comparator. Negative means health lost.

Maximum acceptable cost per DALY averted. Choose a threshold appropriate to the decision context.

USD/ DALY

Fixed exchange rate: 1 USD = 1,475 RWF, as of 28 September 2026 (National Bank of Rwanda reference rate).

Results

ICER (ΔC ÷ ΔE)
USD 600per DALY averted
Net monetary benefit (ΔE × λ − ΔC)
USD 8,000

Cost-effective at this threshold: net monetary benefit is zero or positive.

Intervention vs comparatorThreshold λ = USD 1,000 per DALY averted; points below the line are cost-effective

How it works

The incremental cost-effectiveness ratio (ICER) is the difference in cost between an intervention and its comparator divided by the difference in health effect: ICER = ΔC ÷ ΔE.

The net monetary benefit (NMB) converts health gains into money at the willingness-to-pay threshold λ: NMB = ΔE × λ − ΔC. A positive NMB means the intervention is cost-effective at that threshold. Unlike the ICER, NMB is well defined in every quadrant, which makes it easier to interpret and to use in probabilistic analysis.

On the cost-effectiveness plane, the threshold is a line through the origin with slope λ. Results below the line are cost-effective. Results in the south-east quadrant (more effective, less costly) dominate the comparator; results in the north-west quadrant are dominated.

This tool shows a single point estimate. A full analysis would characterise uncertainty, for example with a cloud of probabilistic simulations and a cost-effectiveness acceptability curve, and would justify the threshold for the decision context.

Need decision-grade evidence?

We build locally adapted, independently checked models with every input traced to a source.