Research review: a computer-based diagnostic model for individual case review

A previous study found that Iliad, a diagnostic expert system, detects diagnostic errors missed by peer review organization (PRO) review. That study used volunteer physicians from an institution as gold standard reviewers, however. The article discusses a second experiment employing Utah PRO (UPRO) review physicians as gold standards. Iliad was compared with the Unified Clinical Data Set used by the UPRO and was found to detect otherwise unsuspected diagnostic errors. The confirmation rate of Iliad flags was much higher in the earlier study, however. No agreement was found between institution and UPRO physicians, but there was agreement between a unique physician (who was both an institution and UPRO physician) and each of the two groups. Because Iliad screens for potential diagnostic errors to be confirmed or denied by gold standard physician review, the different types of physicians in the two experiments might have been the cause.