A User-oriented Approach to Decision Making in Differential Diagnosis and Therapeutic Advice

Abstract

SummaryA search for a user-oriented, computer-assisted expert system resulted in a simple method of simulation of a human consultant’s thinking and action. Making a quick decision in differential diagnosis and selection of the proper line of treatment when faced with a problematic patient has always been a challenging task for clinicians, especially those working in emergency wards, intensive care wards and acute wards. Consulting a vast memory of a number of past cases of a similar nature is near to the natural behaviour of a medical expert and well acceptable to most of the users and patients alike. In this method, the case details of an undiagnosed fresh patient are compared with the details of similar cases diagnosed and treated earlier by experienced consultants and stored in the computer memory. These old cases with confirmed final diagnoses, proven methods of treatment and correctly recorded outcomes enable the system to arrive at the most likely diagnosis, the next likely diagnosis and the most suitable treatment for the new case.A check facility to ensure the accuracy of the decisions made by MEMCONSULT through comparison with a textbook description and any particular old case in the memory has been added in view of the education and training of newcomers in the field of medicine. A self revising capability depending on the changed clinical condition of the patient is also implemented to suit the basic requirements of an ideal expert system in medicine.