Monday, Aug. 10, 1959

Dr. Automation

The patient sits alone in a sterile-looking cubicle, electrodes taped to his chest and extremities, and hunches over a series of buttons on a metal console. He presses a button. On a viewing screen, up pops a question, such as "Do you suffer from shortness of breath?" The patient thinks he does, so he presses another button marked "Yes." The machine records this, and his yes or no answers to a hundred other questions. From the electrodes, a polygraph ("lie detector") notes which questions pack a heavy emotional charge for him. The machine produces a printed and punched, easy-to-read case history.

Feeling more and more like an automaton himself, the patient goes down the line. In one cubicle, a technician takes a blood sample, feeds it into a machine that spins out and counts the cells, measures the concentration of certain key chemicals. In another, the patient gives a urine specimen. Again, a machine reduces it to neat chemical symbols and figures.

At last the patient gets to see a physician. The man in white has the case history and lab reports before him. At the plaintive, immemorial question, "What do you think the trouble is, Doc?", the physician simply presses more buttons. The recorded data are fed into an electronic computer. The cybernetic brain compares the patient's symptoms with those of diseases it has previously learned, discards all but three, offers these to the doctor by code number. A couple of questions enable the doctor to rule out two, and he has his diagnosis. But there are several ways of treating this disease, none 100% effective and all with some risk. The physician punches more buttons, lets the computer decide mathematically which treatment has the best chance of success with the least risk. Resorting at last to the historically illegible penmanship of his profession, he writes a prescription.

Nightmare Feeling. Nowhere in the world is medicine yet practiced in this manner. And automation experts insist that it never will be--quite. But at half a dozen U.S. medical and cybernetic research centers, scores of human computers are at work trying to bring the card-shuffling business machines and the electronic computer into more areas of medicine. At System,Development Corp. in Santa Monica, Calif., an eleven-man team under Engineer Charles J. Roach, 38, has figured after a half-year study that no fewer than six areas invite automation. Of greatest direct interest to the patient: taking and "retrieval" of case histories; diagnosis and treatment; automated control of a medical procedure, e.g., anesthesia during an operation.

System Development engineers have already teamed with the Veterans Administration to compare the effects of various treatments on patients with a form of heart disease. Physicians at Manhattan's Mount Sinai Hospital are using IBM computers to sift mountains of data on blood diseases, and getting answers that may suggest changes in treatment.

Patients' histories are automation's meat. The University of Southern California's Dr. Frederick Moore says that medical workers get a "nightmare feeling of frustration," knowing that the data they are collecting have already been collected somewhere else and filed and forgotten--as their own efforts will be. Automation's ultimate answer: a universal dog tag with key numbers that a doctor would phone into a central records service, soon have the patient's history projected for him on a closed-circuit TV screen.

Is Art Good? That automation will keep moving in on medicine is certain. Tulane University has a computer course for its-medical faculty, plans to extend it to med students. A.M.A. President Louis M. Orr says that automation can improve both the quality and quantity of service, adds reassuringly: "I feel there is no danger of practice being taken over by machines. Some patients and some diseases just do not respond to impersonal treatment from either a doctor or a machine."

But the automation wizards may drive right on past Dr. Orr's stop sign. Says Engineer Roach: "Doctors are inclined to insist that diagnosis is an art. Perhaps it is--now. But must it be? And is that good?"

This file is automatically generated by a robot program, so reader's discretion is required.