In the news

Bogardus et al. [(1999) JAMA 281, 1037-1041 (Abstract)] present a succinct and lucid discussion of the factors involved in discussing medical risks with patients. Risk has five basic dimensions: identity (what is the unwanted outcome?), permanence (is it permanent or temporary?), timing (when is the risk likely to occur?), probability, and value (or "subjective badness"). The last of these is likely to be of the most importance to an individual, i.e. how bad does the patient perceive the unwanted outcome to be. Bogardus et al discuss the problems involved in presenting risks in an unbiased fashion, and in presenting quantitative information in an understandable way. A very interesting section of the article deals with common errors in risk estimation, which are categorised as anchoring bias (the patient estimates the risk based on the risk of some other event that is already familiar); availability bias (overestimation of a risk that gets a lot of publicity); compression (overestimation of small risks and underestimation of large ones); and miscalibration (a tendency to overestimate the accuracy of the person's own knowledge).

Comment: The issue of risk presentation and perception is of great importance in the field of genetics. It will become increasingly important, but also much more complex, as susceptibility testing for common late-onset disorders becomes possible. The value of genetic testing in this situation depends crucially on the perception of risk at many levels: the population risk of the disease; the individual's risk due to his or her age, family history and lifestyle, and the way this risk estimate might change after a test result. Whether a risk is perceived as high or low can depend on whether it is presented as a relative risk or an age-related absolute risk. In addition, the person's motivation to undergo testing at all, and the likelihood that they will modify their behaviour in the light of a test result, both depend on the "subjective badness" of the disease in question. All the common errors in risk estimation outlined by Bogardus et al. must be considered. As they stress, there is a need to improve health professionals' understanding of risk and the way in which they discuss risk with patients. A useful article entitled "Communicating about Risks to Public Health", by Peter Bennett, has  recently been published on the Department of Health web site.