DECISION-MAKING UNDER UNCERTAINTY: THE EFFECTS OF EXPLICIT PROBABILITY INFORMATION AND RISK PREFERENCES OF DOCTORS

Purpose: Decision theory suggests that providing probability information improves the quality of decision-making and that when faced with an uncertain outcome, people more often risk a greater loss rather than incur a certain lesser cost (Prospect Theory). Clinicians frequently make decisions in situations where the outcome is uncertain, explicit probability information is unavailable and risk aversion may influence choices. The extent to which decision theories developed in non-clinical domains with non-clinical participants explain clinical decision making is unknown.

Methods: 462 participants (n=117 non-medical undergraduates, n=113 medical students, n=117 resident trainees, and n=115 medical/surgical faculty) completed a 3-part online task. First, participants completed a non-clinical simulation in which they decided whether to salt roads to prevent icing or risk freezing, based on temperature forecasts with or without explicit probability information. Participants received a monetary award based on performance (better performance described by less negative estimated decision value). Second, participants chose between less or more risk-averse(“defensive medicine”) clinical decisions using standardized scenarios. Last, participants chose between recommending therapy with certain outcomes or risking additional years gained or years lost. Outcome was reported as frequency of prospect-theory-concordant decisions(choosing to gamble only to avoid years lost). We compared results of the tasks between participants and across levels of medical training.

Results: In the salting task, the mean expected decision value of clinicians(-$1,022) was lower than for non-clinicians(-$1,061; p<0.001)(Figure 1). Probability information improved decision making for all participants, but non-clinicians improved more (mean improvement $64 versus $33; p=0.027). Mean defensive decisions decreased across training level (undergraduates 2.7±0.9, medical students 2.1±0.9, residents 1.6±0.8, faculty1.6±1.1; p-trend<0.001). Defensive medicine decisions were not associated with expected values in the salting task(p=0.07). The proportion of medical participants making prospect-theory-concordant decisions increased with medical education level (25.4% of medical students, 33.9% of residents, and 40.7% of faculty, p-trend=0.016).

Conclusions: Both clinicians and non-clinicians made more economically rational decisions when given explicit probability information in a non-medical domain. However, choices in the non-medical domain were not related to prospect-theory-concordant decision making and risk aversion tendencies in the medical domain. Recognizing this discordance may be important when applying decision theories to interventions aimed at improving clinical care.

 

Figure 1: Expected Decision Value of Non-medical and Medical* participants, stratified by presentation of explicit probability information.