Decision making in surgical practice.

Abstract

Judgment has been defined as the ability to make correct decisions with uncertain, incomplete, or inconsistent information. Decision analysis is a scientific method of choosing between trade-offs. This review provides evidence that the model of decision analysis adequately represents clinical problems and that the use of decision analysis can improve clinical judgment. The technique of analyzing decisions using a decision tree model is reviewed. Using the decision tree model, a decision consists of specifying the options, specifying the outcomes associated with each option, assigning probabilities to those outcomes when they are not known with certainty, and assigning values or utilities to those outcomes. Surgical decisions can be described by a limited number of decision trees despite a much larger number of clinical problems. A model of the surgical decision making process, consistent with the analytical decision tree model, details steps useful in clinical decision making. The surgeon determines whether the characteristics of the patient match those of the typical patient with the condition. If so, the standard solution is recommended. If not, the surgeon concentrates on assessing or revising the specific probabilities or utilities that are different. If the best option is suboptimal, consideration is given to more aggressive diagnosis or to a therapy that offers the potential for a better outcome, but at a risk. The recommendations of surgeons for 6 patient management problems was compared to the analytical solutions using their own subjective estimates of probabilities and utilities. Decision analysis provided a net gain in accuracy in all 6 problems. Decision analysis was also significantly more accurate than the conventional second opinion program.(ABSTRACT TRUNCATED AT 250 WORDS)