EFFECTS OF PRESENTING BASELINE RISK INFORMATION ON PEOPLE'S REACTIONS TO THE RELATIVE RISKS AND BENEFITS OF TREATMENTS

Purpose. A critical contribution to better informed decisions by patients is the effective communication of risk. However, this is not straightforward and there is growing evidence which shows that people can be confused by risk information. Some numerical presentations, such as relative risks, are particularly prone to confusion and, compared to absolute risk or number needed to treat formats, can lead people to over-state the benefits of treatments (see review by Covey, 2007). However, most studies have not provided information about the baseline risk level in the population – information that is crucial for people to be able to put the size of relative risks into perspective (Leung, 2001). The experiments presented here examined the effects of presenting baseline risk information on people’s reactions to the relative risks and benefits of treatments.

 Method. Two experiments were conducted which compared three formats (relative, relative-with-baseline and absolute) for presenting risk reducing or risk increasing information about different healthcare treatments. Experiment 1 (N=180) tested how participants’ ratings of the effectiveness or dangerousness of the interventions was affected by the three formats. Experiment 2 (N=100) tested how the different types of risk information were understood by examining participants’ choices between treatments with different levels of absolute benefits.

 Results. Experiment 1 found no significant differences between the relative-with-baseline and relative formats with effect sizes ranging from d=0.07-0.28. Moreover, both the relative and relative-with-baseline formats produced significantly higher ratings than the absolute format (d=0.41-1.45). Experiment 2 showed that, compared to the relative format, participants’ preferences for a treatment with a higher relative risk reduction but lower absolute risk reduction were weaker in the relative-with-baseline (d=1.6) and absolute (d=2.95) formats – although ratings in the relative-with-baseline format were not significantly in favour of the better treatment.

 Conclusions. When treatments were evaluated in isolation baseline risk information was not very helpful in reducing the tendency for people to over-state the benefits or risks. However, when treatments were presented side-by-side providing the baseline risk did weaken the tendency for people to choose a treatment with a smaller absolute risk reduction. Hence, sole reliance on relative risk statistics should be avoided in risk communications.