’ 94 Third Conference on CLIPS

The prescription of wheelchairs for the Multiple Sclerosis (MS) patients involves the examination of a number of complicated factors including ambulation status, length of diagnosis, funding sources, to name a few. Consequently, only a few experts exist in this area. To aid medical therapists with the wheelchair selection decision, a prototype medical expert system (ES) was developed. This paper describes and discusses the steps of designing and developing the system, the experiences of the authors, and the lessons learned from working on this project. Wheelchair_Advisor, programmed in CLIPS, serves as a diagnosis, classification, prescription, and training tool in the MS field. Interviews, insurance letters, forms, and prototyping were used to gain knowledge regarding the wheelchair selection problem. Among the lessons learned are that evolutionary prototyping is superior to the conventional system development life-cycle (SDLC), the wheelchair selection is a good candidate for ES applications, and that ES can be applied to other similar medical subdomains. INTRODUCTION The medical field was one of the first testing grounds for Expert System (ES) technology; the now classic expert system, MYCIN, has often been cited as one of the great breakthroughs in Expert Systems. MYCIN, however, is only one of a large number of expert system applications introduced over the last two decades in the medical field alone [17]. Other examples include NURSExpert [1], CENTAUR, DIAGNOSER, MEDI and GUIDON [17], MEDICS [3], and DiagFH [10] to mention only a few. However, no expert system, to our knowledge, has been developed for the wheelchair selection problem. In this paper, we report on a new application of ES in the medical field; the paper discusses the experiences of the authors with a prototype system developed, using CLIPS, to delineate a wheelchair selection for multiple sclerosis (MS) patients. Our work, therefore, contributes to the existing applications of medical expert/support systems by expanding the domain of applications to the wheelchair selection problem and demonstrating the utility of CLIPS on this problem domain. We will demonstrate that the complexity of the wheelchair selection decision makes it a prime target for an expert system application. To prescribe a wheelchair for a patient with MS involves more than knowing the patient’s disease condition and available choices of potential wheelchairs. A complex web of factors has to be untangled to reach an appropriate choice of a wheelchair. The decision is complicated by such factors as physical body measurements, age and life style, degree of neuralgic impairment, and environmental factors.