Crediting Treatments for Good Outcomes That Would Have Happened Anyway

Abstract

Background. In some patient populations, many would experience good outcomes even if untreated (described as a small baseline risk). It appears that treatments for populations with small baseline risks are perceived as more effective than those for populations with large baseline risks: in essence, treatments are credited for good outcomes that would have happened anyway. Previous research failed to control for differences in treatment effects. Objective. To evaluate if the baseline risk of developing an acute condition influences the perceived effectiveness of preventive treatments besides the genuine treatment effect. Methods. In study 1 (n = 1100) and study 2 (n = 336), general population samples were shown information that systematically differed in size of treatment effect (absolute and relative risk reduction) and the size of the baseline risk of developing a condition. In study 3, medical students (n = 110) were shown treatments that systematically differed in the size of the baseline risk of developing a condition and the type of condition to be prevented (i.e., migraines, heart disease, and pancreatic cancer). Measures included the perceived effectiveness of treatments, intentions to use these treatments, and numeracy. Results. The baseline risk of developing an acute condition influenced the perceived effectiveness of treatments (study 1: F _1,1043 = 66.17, P < 0.001; study 2: F _1,329 = 16.574, P < 0.001; study 3: F _1,98 =14.31, P < 0.001) and, in turn, intention to use the treatment (study 1: F _1,1045 = 7.12, P = 0.008). It caused treatments to be judged as unduly ineffective when the baseline risk was large and vice versa. The influence was independent of treatment effect and was not moderated by numeracy. Conclusions. Treatments for populations with a small baseline risk of developing the condition are perceived as more effective than those for populations with a large baseline risk.