Diagnostic decision support in ophthalmology

Abstract

Isabel is a Web-based, diagnostic decision support tool designed to provide a differential diagnosis of a patient's condition for interpretation by a qualified health-care professional. We investigated the accuracy of the Isabel system in ophthalmic primary care. A total of 100 case histories were prospectively collected from ophthalmic primary care clinic records. The patient demographics and clinical features of each case were then entered into the Isabel system, and the results generated by the decision support tool for each case were compared with the diagnosis reached by the ophthalmic team. Of the 100 cases in the dataset, there was no matching diagnosis in the first 2 pages of Isabel results in 40 cases. Of the 60 cases in which there was a matching diagnosis on the first 2 pages of results, 31 had a >50% match between the terms of the query and the Isabel diagnosis reminder system's database. It remains to be established whether this is high enough to be clinically useful in a practice setting. Inclusion of specific ophthalmic knowledge would probably improve the accuracy of the Isabel clinical diagnostic decision support system.