Decision analysis is a useful clinical tool, but the calculational burden placed on the physician is a major impediment to its application. We have developed a computer program, implemented on a personal computer, that guides the physician in constructing and analyzing decision trees and simplifies the performance of sensitivity analysis, the process that assesses the probability and utility data used in an analysis. Senior physicians, subspecialty fellows, house staff, and students have used this program to analyze complex patient treatment problems. Because the calculational burden is shifted from physician to computer, trade-offs between one clinical variable (eg, the chance that a given disease is present) and another (eg, the risk of treatment) can be rapidly conveniently assessed. This capability provides clinical insights that can be obtained only with great difficulty by previously available methods.