Computerized diagnosis: implications for clinical education.

Area of Science:

• Medical Informatics
• Clinical Decision Support

Background:

• Small computers are becoming essential clinical tools, impacting medical practice and training.
• Probability theory has driven the creation of diagnostic programs for small computers.
• Expert systems, acting as 'electronic consultants,' are now available for various clinical scenarios.

Purpose of the Study:

• To review the current state and future prospects of expert systems in medicine.
• To assess the suitability of different computer programs for clinical environments.
• To highlight the importance of clinician input and diagnostic skills for computerized diagnosis.

Main Methods:

• Review of existing literature on expert systems in medicine (e.g., INTERNIST/CADUCEUS, MYCIN).
• Analysis of the role of probability theory in developing diagnostic programs.
• Evaluation of less complex programs, including clinical prediction rules and expert system shells.

Main Results:

• Expert systems like INTERNIST/CADUCEUS and MYCIN offer valuable insights into medical diagnosis.
• Simpler programs and expert system shells are well-suited for clinical settings.
• The effectiveness of computerized diagnosis hinges on the quality of clinician-provided data.

Conclusions:

• Computerized medical diagnosis, aided by expert systems, holds significant promise.
• Enhanced understanding of data-gathering strategies is necessary for effective bedside diagnosis.
• Emphasis on fundamental diagnostic skills in clinical training is vital for leveraging diagnostic computer programs.