geb@cadre.UUCP (02/10/84)
We have become aware of a program being sold by N-Squared Computing of Silverton, Oregon called INTERNIST or The INTERNIST (their advertising uses both names). As you may be aware, our laboratory is involved in development of artificial intelligence programs for AI diagnosis. Our program, called INTERNIST in our publications as early as 1975 has attained some recognition in the field of expert systems. The program being sold by N-Squared purports to be a program to assist with medical diagnosis. It is available for Apple and IBM-PC computers, and is written in BASIC. Their program bears no relation whatsoever to our program. No one associated with our project had anything to do with N-Squared Computing, and no permission was asked or given to use the name. In addition, we have evaluated their program in its Apple version and in our opinion it is severely lacking even as a teaching tool. The first two cases we ran through their program were the following: (the terminology used to describe symptoms here is the same terminology used by their program). 1. A patient with numbness-feeling of in the arms, pain-chest, feeling of fluttering, and irregular pulse. Any physician should worry about myocardial infarction with ventricular ectopic beats when presented with such a patient. The program's "Exclusive Analysis" failed to find any diagnosis, while its "Pooled Analysis" provided a huge list of possible diagnoses. Heart attack was #12 on this list with such things as diabetes and scleroderma ahead of it. 2. A patient with clouding of consciousness, headache, chills, fever, dizziness, stiff neck, and pain in the neck. Again, a clear case in which life-threatening meningitis must be ruled out. Again, exclusive analysis failed to find any disease to explain the symptoms, and pooled analysis had meningitis #14 on the list, and said that it could explain only 3 of the findings. Diagnoses ahead of meningitis included trichinosis, anemia, and poliomyelitis. Hopefully, we as physicians are not yet in such bad shape that we require this kind of help! And also, I hope that patients (the language of the program is more that of lay people than physicians) are not going to try to use this program to (mis)diagnose themselves rather than consulting their physicians, because from the two examples above, it may be hazardous to their health.