Absolute and Relative Effects

#Two ways to describe a change

An absolute effect describes the difference between outcome frequencies in two groups. A relative effect compares one frequency with the other as a ratio or proportional change. Both can be correct, but they answer different questions and can create very different impressions of the same result.

Consider a hypothetical study in which an event occurs in 4 out of 100 people in a comparison group and 2 out of 100 in an intervention group over one year. The absolute risk reduction is 2 percentage points, or 2 fewer events per 100 people.

#Why baseline risk matters

In that example, the risk is halved, giving a relative risk reduction of 50%. Saying only '50% lower risk' leaves out how common the event was to begin with. The starting frequency, often called baseline risk, is essential for understanding the practical size of the difference.

If baseline risk differs, the same relative reduction can correspond to a different absolute change. This does not mean a study's relative effect necessarily applies everywhere. Population, setting and follow-up may change both measures, so transferring a reported effect to another group requires evidence rather than assumption.

#Read numbers with their context

The number needed to treat expresses an absolute reduction as the number of people who would need the intervention, on average, for one additional person to benefit over a stated period. In the hypothetical example, it is 50 over one year, assuming the difference reflects an intervention effect.

This number is not a promise that every group of 50 will contain exactly one beneficiary. Effect estimates have uncertainty and should be considered alongside harms. Always check the outcome, time period and measure used: an odds ratio or hazard ratio is not automatically the same as a risk ratio.

#Common misunderstandings

A relative reduction of 50% does not mean that 50 out of every 100 people benefit. It means the measured risk is half the risk in the comparison group. In a hypothetical example, a fall from 2 in 100 to 1 in 100 is a 50% relative reduction, but an absolute reduction of 1 percentage point.

Percentages and percentage points are not interchangeable. Saying a risk “fell by 1%” can be ambiguous unless the report explains which measure it uses.

Another misunderstanding is that a large relative effect guarantees a large practical benefit. The absolute difference may still be small, and its importance depends on the outcome, the time involved, and possible harms.

Neither measure tells us exactly who will benefit or experience harm. Both describe results across groups, and both can be uncertain. Reporting an effect in two ways does not make the underlying evidence more reliable.

#Questions worth asking a clinician

• What is my baseline risk without treatment, and how does the reported relative effect translate into an absolute risk change for someone like me?
• Does the reported percentage describe a relative change or a percentage-point difference, and what are the risks in both groups?
• How might the absolute benefit differ for people with my risk factors and over a shorter or longer follow-up period?
• What are the confidence intervals for the absolute and relative effects, and do they include no effect or possible harm?
• What are the absolute and relative increases in each important harm, measured over the same period as the benefits?