Background: Clinicians must communicate risks and benefits of treatments effectively so that patients can appropriately weigh their options and make decisions consistent with their wishes. Relative risk comparisons (e.g. relative risk and odds ratios) can be difficult to interpret, partly because patients need to know both the relative comparison and the baseline risk (BR), plus make the necessary multiplicative calculation to combine the two numbers. Consequently, most recommend reporting risk difference (RD) and/or number needed to treat (NNT), with the rationale being that these absolute measures incorporate BR. The implication is that once the RD (or NNT) is known, BR adds no further information.

Objectives: To show why baseline risk (BR) may affect the interpretation of RD/NNT without introducing cognitive biases, and is therefore essential to make an informed clinical decision.

Methods: Critical overview of the literature.

Results: Five rational reasons why BR may alter the interpretation of RD/NNT are: (1) added value when risk is eliminated; (2) BR changes the value one associates with benefits and risks; (3) decisions made in the context of uncertainty are different than decisions made in the context of certainty; (4) values associated with losses are different than gains; and (5) the value associated with the cost of unavoidable undesirable aspects of treatment is dependent on BR.

Conclusions: BR is essential to interpret all summary effect measures (e.g. RR, OR, RD or NNT). Therefore, the minimum information required for decision making includes BR and risk under the treatment condition. All summary effect measure estimates should be considered as only adjunct information.