Relative risk, absolute risk and number needed to treat: basic concepts.

The benefits of therapeutic interventions, whether pharmacologic, surgical, psychological, or others, are usually determined through clinical trials. The design of such trials involves the definition of a control group (placebo or active agent), which will be compared to one or more experimental groups (active agents). Benefits and/or harms are identified using so-called measures of association, including relative risk reduction (RRR), absolute risk reduction (ARR) and number needed to treat (NNT).

This is the first of several papers on basic methodological concepts for analysis of the medical literature. We will present the concepts of RRR, ARR and NNT in a deliberately simple fashion, using published studies. On the basis of this knowledge the reader will be able to analyze the literature and make decisions based on it.