Continuous-risk utility assessment in medical decision-making.

摘要

This dissertation discusses the utility assessment of risky medical treatment decisions. We begin by developing the argument that for such decisions, conventional utility assessment techniques are inadequate due to their emphasis on unrealistic risk magnitudes and sure consequences. This leads to assessment questions that are unfamiliar to most patients, have little educational value, and do not reliably extend to the application at hand. We contend that medical utility assessments should be performed in a continuous-risk domain with risk levels matching those of the actual decision problem.;In support of this position, we describe an empirical study involving utilities over length of life that compares the responses of subjects under a conventional assessment protocol with those under a continuous-risk utility assessment protocol. These results show that conventional assessment protocols result in significantly lower estimates of the degree of risk aversion compared to a more realistic continuous-risk protocol.;Next, we extend the realm of application of continuous-risk utility assessment to more complex and realistic medical situations. We develop, for a new utility function over uncertain health consequences, a continuous-risk assessment protocol.;We then demonstrate the suitability of continuous-risk utility assessment to a realistic clinical decision-making situation. We develop a comprehensive model of the treatment decision faced by a patient with advanced osteoarthritis of the hip who is considering total hip replacement. We use this model, along with a computerized assessment protocol we created to aid in the utility assessment, to test the performance of continuous-risk assessment as a tool in a clinical setting.;We demonstrate that continuous-risk utility analysis can be effectively administered to subjects despite the complexity of the questions. We report empirical results which indicate that the continuous-risk assessment protocol was well received by members of the medical community, and would be welcome as a decision aid in clinical settings.;We also provide insight into the nature of the total hip replacement decision and its aspects of importance. In particular, the surgical treatment recommendation is robust over a large range of risk preferences.;Finally, in our empirical studies we observed a number of different behaviors and attitudes about risk, health and length of life. The disparate nature of the observed preferences and attitudes reinforces our contention that the preferences of individuals are important and should be sought and considered in clinical decisions.