We propose to test the feasibility of constructing medical decision support systems using fuzzy logic by hospital personnel previously untrained in decision systems, believing that such systems can help lower health care costs without sacrifice in quality. We will develop and test two major pieces of software and their tutorial documentation. First is a general programming system for writing decision support programs in a medical setting using fuzzy logic, written by Southern Dynamic Systems personnel; second is a prototype decision support program for screening hemodynamic data for validity in an intensive care unit, to be written by hospital personnel using the first general system. The general system will be based on existing software packages. The ICU system will be developed by a Clinical Nurse Specialist with support of an advisory committee of physicians, nurses and a biomedical engineer, using a tutorial manual for domain experts. The code for the ICU system will be written by a University Hospital programmer. Learning curves for both domain expert and programmer will be established. Support for hospital personnel will be similar to the that expected from medical equipment manufacturers; 800- line phone support plus necessary on-site inservice training. The ICU system will be calibrated and validated in the University Hospital Heart/Lung Transplant ICU, and its sensitivity and selectivity determined. Development and validation of the ICU system will test the feasibility of using the general decision support system in a hospital setting.
Potential commercial market includes major hospitals possessing programming staff who must reduce direct health care costs, and developers of expert systems.
