Trouble with Medical Computers

TROUBLE WITH MEDICAL COMPUTERS ????. BUSH* Trouble with Medical Computers Behind every computer L· a man . . .far behind. [1] The potential uses of computers in medicine have been vigorously proclaimed for over 2 decades. In research and some specialized areas of medical care much of the early promise has been achieved [2-8], and the best of medical computing compares favorably with the best in other fields. Despite considerable expenditures of both public and private funds, however, there remain serious problems with the class of larger systems known as "medical information systems" (MIS). Although these problems have been obvious to specialists for at least a decade [9, 1 0] and more widely publicized recently [2, 1 1-15], most physicians and medical administrators seem unaware either of their magnitude or of their serious implications for the near future. This article attempts to review the present status of medical computers in terms understandable by medical professionals lacking special expertise in computing. MIS are computer systems designed to improve the informational transactions which make up such a large part of everyday medical care [15]. As with other computer systems, the goals of their designers fall into three main classes: (1) to replace conventional procedures with much less expensive ones, that is, doing the same thing at reduced cost; (2) to increase the speed or number of procedures without increasing cost, that is, doing more of the same thing at the same cost; and (3) to»Departments of Psychiatry and Neurology, Dartmouth Medical School, Hanover, New Hampshire 03755. I am grateful to the National Institute of General Medical Sciences, the National Institutes of Health, and the National Science Foundation for early support of my work. I must also thank numerous workers in this difficult field for their personal experiences , which were obtained under conditions of confidence and hence reported anonymously . I am particularly grateful to Roger Ball, Octo Barnet, William Hagamen, Tom Kurtz, G. H. Mandell, Charles Musselman, Warner Slack, Raymond Sobel, George Stibbitz, Eric von Hippel, and Homer Warner for valuable suggestions and/or information supplementing published material. Address reprint requests to author, V. A. Center, White River Junction, Vermont 05001.© 1979 by The University of Chicago. 0031-5982/79/2204-0091$01.00 600 I Ian E. Bush ¦ Trouble with Medical Computers improve effectiveness by doing things which would be impossible or impractical without a computer (e.g., eliminating human errors, making highly sophisticated calculations or other functions available on a routine basis), that is, doing novel things at a reasonable cost. The first two are to some extent complementary and are the basis for the claim that computers ought to liberate human minds from the drudgery of "inhuman" tasks. The third is the basis for considering computers as tools extending the reach and capabilities of the human mind. In high technology and research, these goals have often been achieved with spectacular results. These successes have been skillfully used by the marketers of computers to conceal the facts—well known to leading professionals for years—that the majority of computer systems are almost unbelievably primitive in their capabilities, that the application of computers to any complex human activity is still extremely difficult, and that these goals are rarely achieved to more than a marginal extent. But despite some published work and much unpublished experience confirming this, the majority of policymakers and potential users still seem to remain victims of the myth of the inevitably successful computer . Largely unobserved by nonspecialists, the development of MIS has passed through two crises of confidence in the past 2 decades. The first, peaking around 1968 [9, 10], was muted because of the continued availability of research funds. The second peaked approximately 4 years ago and was more visible because of the painful shrinkage of research funds [2, 11] and a series of disasters in the British program to "computerize " hospital services [12-14]. These crises exhibit three strange features. First, the frustrations and problems were remarkably similar in the United States [2] and the United Kingdom [12-14] despite different social and funding arrangements. Second, the problems of the second decade were surprisingly similar to those of the first. Finally, very few systems came within sight...