Academic journal article Memory & Cognition

How Bad Is a 10% Chance of Losing a Toe? Judgments of Probabilistic Conditions by Doctors and Laypeople

Academic journal article Memory & Cognition

How Bad Is a 10% Chance of Losing a Toe? Judgments of Probabilistic Conditions by Doctors and Laypeople

Article excerpt

We presented a Web questionnaire to 139 physicians and medical researchers and 109 laypeople. The subjects made judgments of badness and importance of prevention for eight medical conditions at each of seven different probability levels. By assuming that the response to each of the 56 risks was monotonically related to transformations of the probability and of the disutility of the condition, we could assess tine relative effect of probability and disutility on each subject's judgments. Physicians' judgments were more sensitive than laypeople's judgments to changes in probability. Older and female laypeople were less sensitive to probability (and correspondingly, more responsive to differences in severity among medical conditions). Laypeople varied more than physicians in their responsiveness to probability. These results point to general individual differences in the effect of probability on evaluations of medical risks. They may also provide insight into causes and noncauses of physician-patient miscommunication.

Medical decisions often involve evaluations of risks, such as a 5% chance of death or a 50% chance of sexual impotence, as a result of some treatment. Health professionals frequently communicate these risks to patients in roughly this form, a probability and an outcome, so that patients can make a decision themselves, understand why their health care provider made a recommendation, or give informed consent. Arguably, physicians and patients should evaluate such risks in terms of their expected utility (EU). If they evaluate risks in this way, they can easily translate their evaluations into decisions that best achieve their goals. If other things are equal, they will choose the best option for risk reduction if they consider reduction in terms of EU (Baron, 2004). They should, roughly, multiply the probability by the disutility of the outcome in order to arrive at an overall assessment of the seriousness of the risk. Such an approach, and many similar approaches, require an understanding of communicated probabilities and their trade-offs with the disutility (i.e., severity) of possible outcomes.

People may differ in how they respond to probability, relative to the disutility of the outcome, and health professionals may differ systematically from their patients. For example, patients may think of a very low probability (e.g., 0.1%) as quite high, thus failing to distinguish high and low probabilities, whereas physicians may make greater distinctions because they are more familiar with probabilistic reasoning (Nisbett, Fong, Lehman, & Cheng, 1987). Physician-patient differences in judgments of the probability of some outcome may account for a host of problems in the physician-patient relationship, such as miscommunication, patients' misunderstandings of their situation and options, and patient dissatisfaction with physician communications (Collins, Clark, Petersen, & Kressin, 2002; Gurmankin, Domchek, Stopfer, Pels, & Armstrong, 2005).

In addition, departures from EU may help explain some suboptimal patient decisions, such as nonadherence to medications or to recommended screening tests, on the one hand, or excessive health care utilization, on the other. For instance, if a physician tells a 35-year-old woman at average breast cancer risk that her risk of developing breast cancer in the next 5 years is 0.3%, the physician will regard this as a very low risk, whereas the patient may perceive it to be high and may, therefore, conduct daily breast self-exams, seek frequent mammograms, and undergo biopsies of any masses or lumps, despite the extremely low probability of their being malignant.

In the present study-the first of its type, to our knowledge-we presented laypeople and physicians with simple risks (e.g., "a 10% chance of amputation of a big toe") and asked for judgments of badness and of importance of prevention for each risk. We assessed how these judgments depended on the outcome and the probability. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.