Abstract
This paper discusses the evaluation of innovations in health care technology using an illustrative example drawn from computerized electrocardiography. It focuses on the increasing role of such soft technologies that involve computer- assisted diagnostic problem-solving. The purpose is not only to present one evaluative methodology, but also to detect any unanticipated side effects. Since knowledge of the patient's true state is usually not possible in these situations, an alternative evaluative approach has been used to assess the automated system. It is argued that if computer-physician agreement is as high or higher than that between different physicians or the same physician over time, then the computer is interpreting electrocardiograms on an equal par with physicians even though at times both could be wrong. By this criterion, computer interpretation was judged as inadequate for one of five diagnostic classifications. In addition, it was found that the computer systematically biased the judgment of those physicians previously unfamiliar with computerized electrocardiography.
Get full access to this article
View all access options for this article.
