Computer assisted instruction system for teaching decision-making in clinical medicine

Clinically-oriented computer assisted instruction (CAI) has been receiving greater emphasis in the last few years. Similarly progress has been made in the development of (a) generative CAI techniques and (b) the refinement of a formalized medical decision-making criterion. These developments have led to the implementation of a nonauthored CAI system for training clinicians in medical decision-making using actual patient data. The nonauthored CAI system is composed of several subsystems. The principal components include (a) a statistical analysis system providing many options including the generation of histograms, (b) an automated medical decision-making system, and (c) a system for finding patients with desired characteristics and generating populations interactively. The functions and interrelationships of each component are described in detail. The nonauthored CAI system's primary advantages include: (a) it provides the capability of studying medical decision-making procedures in a formal setting; (b) it is a subject-independent system and therefore is relevant to a broad spectrum of clinicians (it is limited by the extent of the clinical data available); (c) because of its generative qualities no courseware authoring"" is required; and (d) because the system requires little training to use, investigations may be conducted in either a formalized group setting or in an independent study mode. Due to the system's relevance to clinical medicine, several persons affiliated with academic medicine have seen the system demonstrated. In addition to these observers, a demonstration of the system in a more structured setting v;as explored. A broad spectrum of clinicians repeatedly used the CAI system to investigate subjects of interest during these structured sessions. Details of their investigations and comments are included. Finally, a few future possibilities and enhancements for the system are explored.""