n 1999, shortly before I left private practice, I received a phone call from the wife of a patient in his early seventies. She feared that her husband was having some form of seizure or “epileptic event” during his sleep and wanted me to refer him to a neurologist or the sleep center at a nearby university. I had the couple come in right away. During the appointment, the man’s wife explained that her husband had dramatic episodes of “moaning and groaning and kicking his feet” during his sleep at night. From several years of being their family doctor, I knew something of the man’s habits, and I asked him how much he was drinking. “Three or four” martinis each night, was the answer, to which his wife added, “Heavy martinis.” After doing a neurological examination (watching his gait and balance, checking strength and reflexes, and examining his eyes) and finding nothing wrong, I advised the man to have only one martini for each of the next five nights, then to quit all alcohol and see me again in about ten days. A couple of weeks later, the pair returned for a visit. This time, the wife reported that her husband was “sleeping like a baby.” After a discussion of how to deal with the husband’s possible alcohol problem, I arranged to see them a third time in three weeks to see if further counseling would be appropriate.

The cost to Medicare of the three visits to my office was slightly over $100. If the man had gone to a neurologist, he might have had an electroencephalogram or an MRI that would have cost at least $1,500. Consultation and overnight evaluation in the sleep center would have cost just as much. Instead, his problem was accurately diagnosed and addressed in the setting of an ongoing relationship where progress or regression could be monitored.

Episodes like this occur almost every day in the offices of properly trained and motivated internists and family practitioners. Areas of the United States where the most care is delivered by primary care physicians have lower overall costs, higher patient satisfaction, and, as a rule, better outcomes. A primary care doctor can be a trusted, friendly advisor who sees a patient over many years. When serious health problems strike, the primary care doctor can become the patient’s medical shepherd, helping to guide him through a complicated system of specialists and hospitals.

In any rational health care system, primary care doctors are central to keeping quality of care high and costs low. Unfortunately, the system in the United States is far from rational, and the number of primary care doctors is plummeting. In 1949, 59 percent of doctors worked in primary care, but by 1995 that number was down to 37 percent. Over the past ten years, as many as one in five primary care providers have left the profession. There’s a broad expert consensus that we face a critical shortage of general practitioners and that the problem is only getting worse.

Health care has become one of the major issues addressed by the current crop of presidential candidates, and each of the leading Democratic contenders has released a plan. None, however, has addressed the need to put primary care at the center of the American health care system. It’s one thing to make sure everybody has health coverage. It’s another to make sure that the coverage is good and, for the nation, affordable. Unless we wean ourselves off of an increasing reliance on specialists at the expense of primary care, we’ll soon be a lot unhealthier and, with all of our spending, a lot poorer, too.

went into medicine because I was interested both in the science of medicine and in the art of interacting with people. In Norfolk, Virginia, where I was raised, doctors were widely considered to be the most important people in the community, and I’d grown up admiring them. In 1954, I entered medical school at the University of Virginia, and I never doubted that I wanted to practice general internal medicine. At the time, that was considered to be the most intellectual part of medicine, a field in which one could be most like the great diagnosticians of the past.

In the 1960s, I did my national service in Washington, D.C., at the National Institutes of Health. I also met a woman (at Clyde’s in Georgetown) whom I eventually married. Both of us liked D.C., and a very good practice opportunity for me turned up. So it was that I went into practice in 1968, joining a group of three older internists in an office on Eye Street.

This was a happy time. Our office had one of the most interesting groups of patients in the country, from waitresses to Supreme Court justices. The “annual physical” was a ritual, perhaps the major part of our income, and some of our patient charts extended back to the 1930s. A typical visit might take nearly an hour (in the case of the annual physical), and with the older patients I’d offer a careful update on their symptoms and my findings, as well as advice on lifestyle changes and medications. For any visit, we always allotted at least thirty minutes.

These were still prosperous years in the postwar boom, and most Americans had what was called “fee-for-service” or “indemnity” insurance coverage, meaning that their employers covered most of the premium and insurers picked up 80 percent of the cost. Patients could go to any health care provider or specialist they wanted and be reimbursed nearly without question. Many received good care, but the system depended on economic factors that weren’t sustainable.

All of us felt the stagflation of the late 1970s, but the era also coincided with an especially sharp upsurge in health care costs. As journalist Shannon Brownlee explains in her book Overtreated, this led to a change in the way insurance companies did business. Experts noticed that certain entitiestraditional health maintenance organizations, such as Kaiser Permanentewere managing to keep costs under control. Among the things HMOs did differently was to assign primary care providers a central role, setting yearly budgets for their clinics and hospitals and leaving it up to doctors to decide how best to prioritize their spending. Although HMOs rarely offered the latest in technology and kept visits to specialists to a minimum, researchers found that they provided care that was often superior to what other health facilities offered.

Insurers took note and tried to import the HMO model into their plans. But they wound up doing it in a shortsighted way. Instead of paying doctors fixed salaries, as Kaiser Permanente did, insurance companies paid them according to how many patients they saw. At the same time, they bargained with doctors and forced them to lower rates of insurance reimbursement. To compensate for the drop in revenue, doctors started to see more patients per hour to make up for the loss in income. Soon, patients began to notice the drop in personal attention, and many became unhappy. This poisoned relationships with primary care providers and helped push patients to seek out specialists.

Through the years my own practice felt these trends. My patient load increasingly comprised many elderly people, and Medicare, in an attempt to cut costs, was also lowering the fees it was willing to pay. This made things difficult, because while medical advances were helping everyone to live longer, my clients were also developing more complicated sets of problems. I was spending more and more time on these patients, even as Medicare was paying less and less for my services (certainly far less than for the annual physicals of old). While I’d grown very close to my existing patients, whom I’d gotten to know over decades as individual human beings, I also found that I was trying to avoid taking on new patients who weren’t young and healthy. Medicare paid so little that I couldn’t afford to do otherwise. Like fellow primary care providers, I also had to take on more patients in order to cover costs.

By the late 1990s, the pressure to see so many clients became too heavy. I was working about twelve hours a day, and, as I got into my sixties, I could recognize the physical signs of stress in myself. In 2000, I retired.

wasn’t the only one leaving the profession. I began to hear more and more stories of doctors who’d had enough and decided to get out of the field. They were finding the workload too high and the compensation too low. This problem will only increase. About a third of physicians are over the age of fifty-five, and, without a sudden easing of their burdens, many, if not most, will be tempted to retire sooner rather than later.

That will leave us with a severe shortage of physicians, because young medical professionals aren’t replacing those who are leaving. Today, students from medical school typically graduate with about $130,000 in debt (far more than students in my generation owed, even adjusting for inflation) and know that choosing a high-paying specialty will allow them to pay it off many years sooner. There’s also a prestige gap between primary care, once considered the creme de la creme of medicine, and what is now viewed as the more attractive position of specialist. Professor Thomas Bodenheimer of the University of California in San Francisco has noted that between 1997 and 2005, the number of U.S. graduates entering family practice residencies dropped by 50 percent.

At the same time, those who do choose to remain at their posts are finding, as I did, that their work is getting even harder and less rewarding. Doctors are rushed, leaving them unable to develop much of a relationship with their patients over time. Also, as a legacy of HMO efforts in the 1990s to reward doctors for denying clients access to specialty care, patients trust their physicians less and less, viewing their doctor mainly as a gatekeeper to the care they “really” need. As a result, doctors today tend to give in to patients who want to see specialists, whatever the costs to the system. It’s much easier to send a patient to an orthopedist than to take a few extra minutes to talk to a patient with a frozen shoulder about the use of heat, simple exercises, and an inexpensive anti-inflammatory drug.

As primary care providers exit the scene, patients make bad choices. Some pay unnecessary visits to emergency rooms, which is far and away the most expensive way to deliver medical care. (This also means that other patients with true emergencies, such as acute coronaries, are often not promptly or effectively handled.) Others delay seeking necessary medical care, simply because they don’t know how to find the right specialist for an acute problem such as a painful abdomen. And still others wind up visiting an array of specialists for mundane problems. (It is often under-fifty patients with high-end insurance who use this option, thinking they’ll get the “best care available.”) But relying on expensive specialists, while preferable to forgoing all medical care, is much less effective over the long run than going to a single doctor whom a patient knows and trusts. In studies undertaken at the Dartmouth Medical School, researchers have found that the best health care has little to do with high expenditure and much to do with placing primary care at the center of patient treatment.

It is part of my self-image to be a doctor and see patients. When I retired, I started to practice medicine in a way I liked, by volunteering at the Free Clinic in Arlington, Virginia. In the clinic, I can work at a pace that I set, and the patients, who are without health coverage, are grateful. I think what my colleagues and I do at the clinic is very successful in helping our patients manage their health problems and in keeping them out of the hospital emergency room. We spend the time we think is necessary and give people the care they need. This is what primary care used to be, and it’s what it ought to be again.

But that won’t happen if we expect all doctors to join free clinics and work without pay. To repair the damage to the profession that has occurred, the federal government needs to take steps to make sure that doctors are properly paid. Medicare, which sets the course for many private insurers, could lead the way. Since 1992, the Centers for Medicare and Medicaid Services have been setting fees for reimbursement based on advice from an entity called the Relative Value Update Committee, or RUC. This committee, composed of twenty-nine members, twenty-three of whom are appointed by various specialist medical societies, operates under the auspices of the American Medical Association. The problem with the RUC is that most of its members are specialists, and the result is that the reimbursement levels it recommends are clearly weighted in favor of surgical or procedure-oriented specialties.

Instead of relying on the RUC, Medicare administrators should consider taking a fresh look at the decline in the number of primary care physicians and come up with some conclusions of their own about proper reimbursement fees. The cognitive services rendered by primary care doctors are perhaps the most valuable medical commodity on the market, and they merit higher fees than, say, the removal of a skin lesion. Medicare won’t make such changes without backing from Congress and the White House. Lobbying by medical professionals who appreciate the current status quo will be fierce, and Congress must be prepared to resist it. There are other ways to increase the number of primary care physicians, certainly, but we need agreement on the ends before we can come up with proper suggestions for the means.

Once reforms are in place, primary care doctors will have to meet new obligations too. They’ll need to make their offices more efficient, to handle what are sure to be increased demands as health care becomes more complex. But no matter how many improvements in efficiency they might manage to introduce on their own, time is a fixed commodity. Spending fifteen minutes one on one with a patient either happens or it doesn’t; no innovation can allow doctors to spend fifteen minutes one on one with five patients an hour.

There’s still time to do something about this problem, but we can’t keep letting it get worse. Otherwise, by the time Washington works up the political will to implement a sensible health system with primary care at its core, it will be a solution with no doctors left to make it work.