Refining an Automatic EDSS Scoring Expert System for Routine Clinical Use in Multiple Sclerosis

The Expanded Disability Status Scale (EDSS) has been the most widely used measure of disability in multiple sclerosis clinical trials. Although EDSS has the advantage of familiarity with respect to recent proposals, and remains the de facto standard, it is difficult to use consistently between evaluators. AEDSS (Automatic EDSS) is an expert system designed to overcome this problem. It constrains the neurologist to follow precise reasoning steps, enhancing EDSS reliability. In this paper we show how a deep analysis of the neurological knowledge involved has been essential for adopting AEDSS in routine clinical use. We present an ontology for the EDSS domain and we highlight the enhancements to AEDSS due to this additional knowledge. A validation experiment in four MS centers in Italy showed that AEDSS reduces inter-rater variability and in many cases is able to correct errors of neurologists.