PubMed1995
Studies of health-provider decision-making, and of their practice patterns, play a central role in efforts to improve the quality and effectiveness of care and in decreasing costs of healthcare delivery systems. Researchers from a variety of disciplines have studied a broad range of clinical conditions, using a number of methodological approaches and measurement tools, including self-report, written clinical vignettes, simulated clinical encounters using actors as patients and analysis of medical records and administrative data. Although these provide information about the outcomes of clinical decisions, they provide little or no information about the process of the decision. Most clinicians agree that the decision process is as important as the outcome, and indeed it is not unusual to have an exemplary process but a poor outcome. Process information is therefore a crucial dimension of care evaluation. In this paper, we describe a new software product that was originally used to measure diagnostic reasoning in the basic medical science of immunology; subsequently adapted to measure key steps in the clinical decision-making process. This Windows-based software is user-friendly, inexpensive, and requires only commonly available hardware for its operation. It is very flexible, permitting the creation of unlimited numbers and types of clinical scenarios, with diagnostic and/or management approaches. Being clinically "real-world," the scenarios are familiar to the user, who is therefore likely to respond in a "real-world" fashion, with the consequent improved accuracy of data. In addition, a wide range of users may be accommodated. The clinical activities of physicians, nurses, pharmacists, and any other clinical providers may be measured and analyzed by the system. Non-clinical providers, such as managers and administrators, could also be assessed. The system has three major modes. In the Authoring Mode, the author creates a menu, which is common to a number of linked scenarios. For example, the menu for physicians might include the History, Physical examination, Laboratory tests, Radiology, Consultations, etc. The actual details of each related clinical case may then be varied. There is virtually unlimited flexibility in the design of the menu and the clinical details, depending on the needs of the author, and the type of information desired. Both diagnostic and management scenarios are easily constructed. The cost for each individual step may be assigned, using any scoring scale desired. Actual dollar costs, or a suitable point score, are equally possible. Once the menu and associated scenarios are generated, the candidate is asked to solve the clinical problem in the User Mode. The candidate obtains information by "mouse-clicking," so it is not necessary to be a computer expert to use the system. Eventually, the candidate is presented with a short vignette outlining the desired solution, which may include the authors comments, sources for further information, etc. In the Data Collection and Analysis Mode, the candidate proceeds to solve the scenario, the software captures and stores each individual information request i.e., each step in the candidates reasoning process. Thus, the reasoning process can be examined, including timing and order and types of information used; this may be done both for individual candidates, and also for groups. A "gold standard" reasoning may be predetermined by the author for comparison purposes. The software has already been used to teach immunology to medical students, and is currently being expanded to train gynecology surgeons in the use of Clinical Practice Guidelines. The software has potential applications in many aspects of the healthcare field. For educators, it could serve in traditional exit examinations for the clinical disciplines, both undergraduate and postgraduate. (abstract truncated)
Computer-Assisted Instruction↗