System for computerized interactive protocols using H.E.L.P

The expanded use of physicians' assistants could reduce the burden on ambulatory care physicians and increase the availability of health care to rural underserved areas. The acceptable use of assistants depends in part on the availability of appropriate support methods t insure the quality of care provided by these nonphysicians. The utilization of paper protocols--which define decision-making logic and provide data collection guidance for patient complaints--has been plagued with usage errors and problems of user alienation. The computerized interactive protocol system described in this dissertation represents a unqieu and workable solution to the problems of protocol implementation. Four prominent features enabled the system to provide effective support to physicians' assistants in tow clinical settings: 1. The HELP system of the Department of Medical Biophysics and Computing allowed flexible definitiona and easy modification of patient care protocols. 2. A thorough and specific evlauation of patient problems was insured by interactive computer prompting when data collection was inadequate. 3. Dependabel decisions were provided by computer execution of protocol logic and through data entry routines which resulted in about one error per 100 data entries. 4. The system provided a method of auditing care which was acceptable to physicians and assistants. Protocols for urinary tract infections, upper respiratory-ear, nose, and throat complaints, chronic diabetes, adn chronic hypertension were implemented on the system and utilized in the two clinical settings. Evaluation data from use of the upper respiratory0ear, nose, and throat protocol--involving 108 patient encounters--and the hypertension protocol--involving 145 retrospective chart audits--is presented. The upper respiratory-ear, nose, and throat protocol effectively managed 89 percent of patients whil the yypertension protocol managed 72 percent of patients without physician referrals. Both protocols demonstrated low error rates for false-positive and false-negative decisions. The automated self-histories provided with the computerized protocol system were demonstrated to be effective in collecting thorough histories relating to presenting complaints. As part of the conclusion section, a proposal is presented for the implementation of the protocol system on a small in-house minicomputer for use in rural clinics with physicians' assistant. Using a computerized interactive protocol system, protocols can be effectively implemented and provide adequate support to physicians' assistants. By insuring accurate decision-making and data collection for patient problems, assistants cn manage a large fraction of complaints normally directed to physicians and provide health care in many underserved areas.