Patient Preferences and Treatment Thresholds under Diagnostic Risk: An Economic Laboratory Experiment

We study risk-aversion and prudence in medical treatment decisions. In a laboratory experiment, we investigate the frequency and intensity of second- and third-order risk preferences, as well as the effect of the medical decision context. Risk preferences are assessed through treatment thresholds (the indifference point between not treating and treating). Under diagnostic risk, medical decision theory predicts lower treatment thresholds for risk-averse than for risk-neutral decision makers. Given a comorbidity risk in the sick state, prudent individuals have an even lower threshold. Our results demonstrate risk-averse and prudent behavior in medical decisions, which reduce the (average) treatment threshold by 41% relative to risk-neutrality (from 50.0% to 29.3%). Risk aversion accounts for 3/4 of this effect, prudence for 1/4. Medical decision framing does not affect risk aversion, but is associated with more and stronger prudent behavior. These findings can have consequences for diagnostic technologies and QALYs, and thus for clinical guidelines.