Absolute risk, relative risk and number needed to treat
Translate a relative effect into actual event counts, showing the baseline event risk and the follow-up period.
Start with event counts and time
Risk is the proportion of a defined group experiencing an event over a specified period. State the population, outcome and timeframe. For each group, divide its event count by its own total using the same follow-up window.
Using Incidence rates via person-time and hazard ratios are different measurements. Because of unequal follow-up times and/or loss to follow-up, a simple Risk calculation could lead to inaccurate results.
Compare in two ways
Relative risk is the intervention group’s risk divided by the comparison risk. An RR of 0.6 means 60% of the comparison risk, or a 40% relative reduction for an undesirable event.
The absolute Risk reduction (arr) is the difference between the Risk of the comparison group and the intervention group. This should always be expressed as either percentage points or events per number of participants in a common denominator. For example, although two interventions may have the exact same relative Risk reductions, they can still demonstrate vastly different absolute Risk reductions based upon baseline risks.
Translate benefit carefully
NNT, or number needed to treat, is 1/ARR when ARR is written as a fraction. It gives the average number treated per additional undesirable event prevented over that follow-up under the comparison’s assumptions. Round a positive whole-person NNT upward.
A zero risk difference has no finite NNT. An increase in risk requires a harm interpretation. Odds ratios compare odds and may differ considerably from risk ratios for common events. Report uncertainty. An observational risk ratio is an association that may be confounded; a causal interpretation requires a defensible study design.
Worked example: the same result in three forms
Out of a total of 500 comparison participants (control group) in a study over a period of one year, 40 experience an undesireable event. Of the 500 participants of the intervention group (treatment group), 25 also experience an undesireable event.
Risks = 8% and 5%. RR = 0.05/0.08 = 0.625. Relative reduction = 37.5%. Absolute reduction = 3 percentage points. NNT = 1/0.03 ≈ 33.33, rounded up to 34.
The absolute summary is three fewer events per 100 people over one year. The causal interpretation still depends on the study design.
Provide the baseline risk, absolute difference and timeframe alongside any relative reduction.