The absolute difference in event probabilities between a control (or untreated) group and an intervention group over a specified period: ARR = risk_control − risk_treatment; reported as a proportion or percentage and forming the basis for NNT calculation.

Principle

ARR quantifies the concrete change in outcome probability attributable to the intervention in that population and timeframe; it depends on baseline risk and the relative effect size, and is directly interpretable as an absolute percentage‑point change.

Demonstration

Illustrative scenario: In a trial with one‑year follow‑up, the control group event rate is 10% and the treatment group rate is 6%; ARR = 0.10 − 0.06 = 0.04 (4 percentage points), which corresponds to NNT = 25 for that year.

Misapplication

Quoting relative risk reduction (RRR) instead of ARR when communicating absolute benefit, or omitting the time frame and baseline risk. The error is presenting proportional reductions as absolute patient‑level benefit, which can exaggerate perceived effect size when baseline risk is small.

Consequence

ARR provides a direct, population‑relevant measure for clinical communication, health economic evaluation, and derivation of NNT; misreporting or ignoring heterogeneity of baseline risk can mislead clinical interpretation and policy decisions.

Reversal

When baseline risk varies meaningfully across subgroups, ARR will change even if relative effects remain constant; in recurrent event settings or with competing risks, simple ARR on cumulative incidence may misrepresent net clinical benefit and require alternative estimands.

Boundary

Clearly within: between‑group difference in cumulative incidence of a binary outcome over a defined follow‑up. Boundary case: hazard ratios from time‑to‑event analyses require conversion to cumulative incidence to compute ARR. Clearly outside: relative measures (RR, OR) and continuous outcome differences are not ARR.

Semantic Tension

Absolute (ARR) versus relative (RRR) effect measures: ARR gives concrete change in risk but can be small when baseline risk is low; RRR can appear large but obscure absolute patient impact.

Synthesis

ARR is the straightforward, context‑dependent measure of absolute benefit that underpins NNT and patient counselling; always report its time frame, baseline risk, and uncertainty and interpret it alongside relative measures to present a complete effect picture.