Preference assessment method affects decision-analytic recommendations: a prostate cancer treatment example.

Purpose: To evaluate the effect of preference assessment method on treatment recommended by an individualized decision-analytic model for early prostate cancer.
Methods: Health state preferences were elicited by time tradeoff, rating scale, and a power transformation of the rating scale from 63 men ages 55 to 75. The authors used these values in a Markov model to determine whether radical prostatectomy or watchful waiting yielded the greater quality-adjusted life expectancy.
Results: Time tradeoff and transformed rating scale recommendations differed widely. Time tradeoff and transformed rating scale utilities differed in their treatment recommendation for 21% to 52% of men, and the mean difference in quality-adjusted life years varied from less than 0.5 to greater than 1.0.
Conclusions: Treatment recommendations from the prostate cancer decision model were sensitive to the method of preference assessment. If decision analysis is used to counsel individual patients, careful consideration must be given to the method of preference elicitation.