Methods Inf Med 1994; 33(02): 205-213
DOI: 10.1055/s-0038-1635005
Original Article
Schattauer GmbH

Case-Based Explanation for Medical Diagnostic Programs, with an Example from Gynaecology

R. Stamper
1   The Programming Research Group, Oxford University Computing Laboratory, Oxford, UK
,
B. S. Todd
1   The Programming Research Group, Oxford University Computing Laboratory, Oxford, UK
,
P. Macpherson
2   The Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford, UK
› Author Affiliations

We are grateful to Tony Hoare for providing valuable suggestions and advice. We would like to thank the Consultants in the Nuffield Department of Obstetrics and Gynaecology, John Rad-cliffe Hospital, Oxford, for permission to collect and use clinical data from patients under their care. Richard Stamper is supported by a SERC Research grant (SERC GR/F47077). Bryan Todd is supported by a SERC Advanced Fellowship in Information Technology (SERC B/ITF/269).
Further Information

Publication History

Publication Date:
08 February 2018 (online)

Preview

Abstract:

One of the most accountable methods of providing machine assistance in medical diagnosis is to retrieve and display similar previously diagnosed cases from a database. In practice, however, classifying cases according to the diagnoses of their nearest neighbours is often significantly less accurate than other statistical classifiers. In this paper the transparency of the nearest neighbours method is combined with the accuracy of another statistical method. This is achieved by using the other statistical method to define a measure of similarity between the presentations of two cases. The diagnosis of abdominal pain of suspected gynaecological origin is used as a case study to evaluate this method. Bayes’ theorem, with the usual assumption of conditional independence, is used to define a metric on cases. This new metric was found to correspond as well as Hamming distance to the clinical notion of “similarity” between cases, while significantly increasing accuracy to that of the Bayes’ method itself.