Similarly, I was fascinated by the idea that the computer might provide a powerful diagnostic tool, but diagnostic computer systems have found little acceptance in medicine. Doctors don't trust them and patients don't like them; they would rather give case histories to a paramedic or aide. On the other hand, everyone accepts automated lab tests, which are quick, accurate, and inexpensive. But overall, the effect of automation in medicine has been mixed; for example, even the mundane use of computers for hospital record-keeping has proven unexpectedly problematic, as our society struggles with issues of accuracy and privacy in the era of electronic data.

What was not foreseen by me, or by anyone else in the late 1960s, was that computers would soon become almost unimaginably cheap. A computer that cost ten million dollars in 1970 cost only a few thousand dollars in 1980, and only a few hundred dollars in 1990. Ubiquitous, cheap computer power has made possible a variety of non-invasive imaging procedures-computer assisted tomography, magnetic resonance imaging, and sonography- which have transformed the daily practice of medicine, and which seem, to someone from that era, almost magical in their results.

As medical technology has proliferated we have gained more sophisticated understanding of its limitations. Indeed, one trend in medicine has turned away from technology altogether. The long-term improvement in statistics for heart disease is primarily credited to life-style changes in the population. Diet, exercise, and meditation are now solemnly prescribed where once they were laughed at. And the growing interest in psychoimmu-nology, the interaction of mind and disease, is shared by patients and physicians alike. (When I wrote Five Patients, the most famous doctor in America was probably Michael DeBakey, the Houston cardiac surgeon. Now, he may very well be Deepak Chopra.)