(FORTUNE Magazine) – For three decades now, thanks to insurance and Medicare, consumers have paid relatively little out of their own pockets for medical services. Lacking compelling reasons to trade off costs against benefits, they naturally have demanded ever-increasing quantities of care. Just as naturally, the suppliers of health care -- doctors, hospitals, laboratories, and so forth -- have expanded their services, both to compete for business and because payment was a sure thing. Thus our health care system has encouraged ''overservicing,'' a main cause of the upward spiral in health care costs. Health maintenance organizations -- HMOs -- and other prepayment plans were supposed to solve the problem. Since HMOs receive a flat fee in advance, they have an incentive to control costs. But prepayment plans do nothing to constrain demand for care. Once a consumer enrolls in an HMO, he is free, in theory at least, to use as many services as he wants. To solve that part of the equation, HMOs have turned to ''gatekeeping.'' The idea is deceptively simple: Gatekeepers propose to reduce costs by making sure patients use the least expensive types of services. The gatekeeper, the first person to examine a prospective patient, has a dual function: to keep those who don't need special treatment from wasting the time of specialists, and to guide those who do need such treatment to the appropriate specialist.

Proponents of gatekeeping argue that it controls runaway demand without harming the quality of care. I believe their stand is based on several false assumptions or myths. In fact, gatekeeping may increase the costs of health care, and it poses a serious threat to patients. Let's examine the myths first, then their consequences. -- Myth 1: Gatekeeping ensures efficient medical care. Gatekeepers, usually general practitoners or internists, are not efficient when they become middlemen, referring the patient to another physician. Referrals increase costs directly, by requiring another visit to a doctor, and indirectly, by delaying diagnosis of conditions that become more expensive to treat the longer they go untreated. Nor are gatekeepers necessarily efficient when they themselves treat patients. The fans of gatekeeping assume that a generalist's fees will be lower than a specialist's, but that's not always true. Cardiologists and neurosurgeons often charge more for an office visit than generalists do; pediatricians, dermatologists, and orthopedic surgeons often charge less. Fees for office visits aren't the only costs of treatment. Consider a 1983 Emory University study that compared how dermatologists and family practitioners would manage treatment of ten different skin diseases. Compared with the generalists, the specialists ordered tests that cost only half as much, and they would have required patients to return less often for treatment. While the specialists wrote more prescriptions, the total cost of medication wasn't much higher. Taking everything into account, the dermatologists would have provided care for 10% less cost than the family practitioners. -- Myth 2: Gatekeeping ensures effective diagnosis and treatment. In the Emory University study, much of the additional expense incurred by family practitioners was due to misdiagnosis. All the physicians were presented with the same symptoms and conditions. In diagnosing these conditions, the generalists averaged just 60% accuracy, compared to 98% for the specialists. Yet less than a third of the generalists said they would refer the patients to another, more specialized physician. Why didn't they refer more often? The doctors in many HMOs are penalized for making referrals and rewarded, usually by bonuses or profit sharing, for treating patients directly. While designed to prevent overservicing, these incentives can produce a real disservice to the patient. -- Myth 3: Consumers can't judge their own need for a specialist. Maybe not, but neither can many gatekeepers. Often the patient is a fine judge. Anybody can recognize most eye problems, broken bones, or skin rashes and seek out the appropriate specialist. Consumers with chronic medical problems become quite sophisticated about their own needs. -- Myth 4: Specialists can't provide primary health care. Because of increasing competition for patients between generalists and specialists, distinctions are blurring. Pediatricians, internists, and obstetrician- gynecologists, all specialists, have long supplied primary health care. Many other specialists also serve as primary physicians for patients with severe, complex diseases. The nephrologist, for example, frequently becomes the primary physician for patients with end-stage renal disease. Now let's look at some consequences of these myths. -- Consequence 1: Less freedom of choice. Limiting a patient's freedom to see a specialist undermines the benefits of competition among doctors. Proponents of gatekeeping, and of HMOs generally, argue that competition isn't eliminated because unhappy subscribers can always choose another HMO. Thus they say HMOs have strong incentives to compete by producing high-quality medical care. I doubt consumers find this kind of competition as effective as competition between one doctor and another. Doctors are more sensitive than HMOs to the possibility that patients who feel mistreated will persuade their friends to go elsewhere. Moreover, consumers find it easier to choose among competing doctors than competing, anonymous health care corporations. -- Consequence 2: Delayed diagnosis and improper treatment. We can expect more of both when gatekeepers have incentives to limit services and referrals. The costs to physicians, patients, and insurers can be huge. The patient may suffer needlessly, even die, or he may spread infectious diseases. The physician may lose his reputation and face malpractice suits. The insurer -- indeed, the whole health care system -- may suffer the higher costs of mistreatment or late treatment. -- Consequence 3: Breaking the law. According to the Professional Liability Guide of the Massachusetts Medical Society, a general practitioner who treats a patient who ought to be referred to a specialist ''may be held to the standards of a specialist in the field.'' That means, of course, larger and more serious malpractice judgments. Moreover, some opponents of gatekeeping claim it's an illegal restraint of trade, since gatekeepers allocate demand among physicians. I leave the details to legal experts, but it's clear to me that limiting a patient's access to care inhibits competition between generalists and specialists. -- Consequence 4: Slower medical progress. If patients use specialists less, fewer medical students will become specialists; with fewer students turning toward specialties, fewer specialists will be available to teach and do research in specialty fields; with less research in the specialties, the well of innovation that brought us antibiotics, open heart surgery, and a wealth of new diagnostic tools will dry up. If gatekeeping is a bad way to solve the very real problem of overservicing, of patients demanding too much care, are there better ways? Probably. A number of HMOs, including ChoiceCare of Cincinnati and Physicians Health Care Plan of Greensboro, North Carolina, seem to have found solutions that avoid gatekeepers and allow patients direct access to specialists. Such creative organizations generally use one of three methods to control overservicing. -- Method 1: Capitation of specialists. Some HMOs have agreed to pay specialists a certain amount per month, based on the number of patients in the plan, regardless of how many patients the specialist actually sees. This puts the burden of control squarely on the shoulders of the specialist. -- Method 2: Co-payment for specialty consultation. Some HMOs allow subscribers direct access to specialists but require them to pay part of the extra cost -- either a flat fee or a percentage of incurred cost -- for each visit. Though co-payment may inhibit some patients from seeking special care they need, it motivates all of them to avoid needless service. -- Method 3: Statistical reviews of utilization. Some HMOs, particularly those with sophisticated computer systems, scrutinize the performance of each specialist along all sorts of dimensions -- number of patients seen, cost of treatment per patient, average visits per patient, number and average cost of hospital admissions, and so on. They compare the physician's performance to the national average for his specialty and scrutinize doctors whose costs exceed the average. When a physician clearly is abusing the system, the HMO asks him to accept a reduced level of reimbursement -- or to resign.

None of these methods is free of negative consequences for the physician, the patient, and the system as a whole. But all of them are preferable to a gatekeeper whose advice and counsel are inefficient, ineffective, and, from the patient's point of view, undesirable. And the options I've outlined preserve three unique, commendable features of American medicine: professionalism and competition among physicians, and freedom of choice for health care consumers.