The development and evaluation of two computer-based diagnostic aids in the field of inherited skeletal dysplasias and malformation syndromes.

A retrospective review of all skeletal dysplasias referred in a period of one year for a diagnostic opinion, identified that, of all referring clinicians, general radiologists in particular had a low diagnostic accuracy. Two computerised systems were developed. The first was an expert, knowledge-based system in which the knowledge of an experienced paediatric radiologist in the field of skeletal dysplasias was captured and diagnostic reasoning pathways defined. The second comprised a database of radiographic images with their radiological findings, of patients with skeletal dysplasias, combined with a powerful search facility for matching findings and images. Each system has been tested individually in a standardised clinical trial by comparison with standard diagnostic methods used by radiologists, that is, by referring to lists of gamuts, standard reference textbooks and journals. Because the trial protocols and material were identical, a comparison of the relative value of each system could be made. The results of the two trials showed that the use of each computerised system individually achieved significantly improved levels of diagnostic accuracy when compared with standard methods of diagnosis. Of the two systems the expert system showed slightly improved diagnostic accuracy compared to the image database. The development and use of such systems will help to improve diagnostic accuracy and the quality of genetic advice given to patients and their families and patient management and treatment in difficult medical domains