Until the early years of the twentieth century, there were plenty of physicians to go around. Many earned their stripes as apprentices, just as a plumber or butcher might train, while others studied at for-profit medical colleges owned by doctors. Still others got their degrees from a relative handful of university medical schools modeled on European institutions. Then, as the twentieth century dawned, scientific advances led to calls for a more standardized, rigorous medical education, and in 1908 the recently formed Council on Medical Education asked educator Abraham Flexner of the Carnegie Foundation for the Advancement of Teaching to evaluate the American way of training doctors. During the following 18 months, Flexner visited every one of North America’s 155 medical schools, and in 1910 he delivered his report, a landmark study that decried the poor quality and broad variation characterizing much of the continent’s medical training. In the wake of the Flexner Report, dozens of medical schools, mostly diploma mills that required no university training, were forced to shut down.

Fewer medical schools meant fewer doctors, and in the years that followed the national ratio of physicians to population declined to about 125 per 100,000 people, compared with almost 240 per 100,000 today. But with supply constrained, demand soared, and so did doctors’ incomes. Then, in the late 1950s, policy experts projected a shortfall of nearly 40,000 physicians by 1975. Congress increased funding for medical schools and helped finance the establishment of new ones, and the number of graduates gradually rose to today’s level of nearly 16,000 annually.

This push to increase the supply of physicians was such a success that, before long, experts began to wonder whether it had worked too well. In 1980 the Graduate Medical Education National Advisory Committee (GMENAC), a group of health care policy analysts, academics and physicians, produced a study predicting the United States would have 145,000 more physicians than it needed by the turn of the twenty-first century. Subsequent studies predicted an even greater excess, and in 1997 Congress limited the number of residencies Medicare would fund to about 80,000 each year. Although another 20,000 residencies are financed by the Veterans Administration and Medicaid, and teaching hospitals pay for a small number without government assistance, the Medicare spending freeze effectively capped the physician workforce.

The year 2000 came and went, and a physician surplus never materialized. “Physicians were still working as doctors, not driving cabs,” says internist David Blumenthal, director of the Institute of Health Policy at the MGH and a professor at Harvard Medical School. “Everyone started wondering what was wrong with the models.”

One thing that was wrong was that the GMENAC analysts had failed to anticipate changes in medical practice that would affect which specialties doctors chose and how many physicians the system could keep gainfully employed. Two lucrative branches, cosmetic surgery and sports medicine, began attracting large numbers of physicians. An explosion in new imaging technologies has meant many more jobs for radiologists. But the popularity of particular specialties can turn on a dime, confounding predictions of surplus or shortage. Twenty years ago, residencies in cardiac surgery were among the most coveted, but with reduced compensation and a migration toward less-invasive heart procedures done by cardiologists, these days there are more slots available than there are applicants.

Richard Cooper, professor of medicine and senior fellow at the Leonard David Institute of Health Economics at the University of Pennsylvania, had always been skeptical of studies projecting a physician surplus. “They reflected a political objective-—that to contain health spending you had to constrain the supply of health care,” Cooper says. “But physicians don’t drive medical costs; demand does.”

On the demand side of the equation, wealth plays a role, according to Cooper’s research. As per capita income increases, so does a country’s use of health care services—and at a faster rate than the growth in affluence, Cooper says. During the past 70 years, the demand for medical services in the United States has grown half again as fast as the gross domestic product, a trend Cooper expects to continue.

“Baby boomers and their offspring will tap health care services more than prior generations have because they’ve grown up in an era of unprecedented wealth and medical advances,” he says. Based on these and other assumptions, Cooper calculates the physician shortage could rise to 200,000 by 2020. “We’ve lost 10 to 15 years of training physicians because of surplus projections, and now we’re in a real predicament,” he says. “The medical profession has been seriously damaged.”