Decision analysis in nuclear medicine.

This review focuses primarily on the methodology involved in properly reviewing the literature for performing a meta-analysis and on methods for performing a formal decision analysis using decision trees. Issues related to performing a detailed metaanalysis with consideration of particular issues, including publication bias, verification bias and patient spectrum, are addressed. The importance of collecting conventional measures of test performance (e.g., sensitivity and specificity) and of changes in patient management to model the cost-effectiveness of a management algorithm is detailed. With greater utilization of the techniques discussed in this review, nuclear medicine researchers should be well prepared to compete for the limited resources available in the current health care environment. Furthermore, nuclear medicine physicians will be better prepared to best serve their patients by using only those studies with a proven role in improving patient management.