Mathematical models in decision analysis

Abstract

Decision analysis offers powerful techniques to understand and evaluate uncertain clinical situations better. Decision analytic models are appearing with increasing frequency in health policy planning, clinical information and decision-support computer systems, evaluations of clinical pathways, development of clinical practice or utilization review guidelines, and epidemiologic research. This article describes the structure, application, and limitations of the more popular decision analytic methods, including decision trees, Markov models, Monte Carlo simulation, survival and hazard functions, fuzzy logic, and sensitivity analysis. Understanding the nature of these methods will help readers to assess better the appropriateness of their use in published reports.