Time preference in medical decision making and cost-effectiveness analysis.

Cost-effectiveness analyses usually quantify peoples' attitudes towards delayed outcomes using the exponential discount model. The authors examined three assumptions of this model by assessing the time preferences of individuals towards hypothetical health states and calculating implicit annual discount rates. Of a random sample of medical students, house officers, and attending physicians, 121 participated, reflecting a response rate of 81%. The participants considered three temporary events (colostomy, blindness, depression) that were destined to occur at five sequentially distant times in the future (one day, six months, one year, five years, and ten years). The utility of each prospect was measured using two elicitation techniques (standard gamble and categorical scaling), and 1,394 implicit discount rates were calculated. Of all the discount rates, 62.1% equalled zero, 10.0% were less than 0.00, and 15.7% were greater than 0.10. Mean discount rates for relatively proximal time intervals tended to be larger than those for relatively more distant intervals (0.041 vs. 0.025, p < 0.01). Mean discount rates for blindness tended to be smaller than those for colostomy or depression (0.023 vs. 0.039 vs 0.037, respectively, p < 0.005). Hence, peoples' implicit discount rates are not always small positive numbers that are constant over time and the same for all settings. The authors suggest that the conventional exponential discount model may not fully characterize the time preferences held by individuals.