(FORTUNE Magazine) – I started having vague concerns about managed care a few years ago when my new doctor, selected from a thick roster approved by a managed-care company with which my employer has a contract, began giving puzzling responses to my minor complaints.

When I had abdominal pain I thought might be a hernia, his advice was succinct: "Don't lift anything heavy." When he heard I often awoke at night, he prescribed a child-sized dose of an ancient but very inexpensive antidepressant. All it did was make me groggy and hungry all day. A pinched nerve in my neck? Wear a brace and eat 12 Advil a day, he said, and he told me to rearrange my computer, desk, telephone, and life so that I could see things without bending my neck. Well, okay, I said to all this. With medical costs running amok, we all need to chip in and accept managed care's austere approach to medicine.

But when I asked my doctor an off-hand question in the middle of a checkup about a year ago, I nearly had a heart attack. He had sent me off for routine blood tests every year or two but never mentioned what they revealed. What, I asked out of idle curiosity, was my cholesterol level? He shuffled some papers, paused a moment, and said sheepishly, "318." Holy Hippocrates! Three hundred eighteen? That's dreadful! One-eighty is great, 200 is okay, and 240 means it's time to take serious action. But 318? I was a walking coronary! Why didn't he tell me?

Paranoia struck. I had picked some cheesy managed-care company that wouldn't pay my doctor for much of anything and offered him perverse incentives whereby he could get rich only by withholding care from me! Every nickel he saved by prescribing cheap medicine or avoiding neck therapy went into his pocket! "Don't tell O'Reilly about his cholesterol because those medicines are expensive and he'll need a lot of follow-up visits," he must have calculated. "Besides, O'Reilly probably won't have his heart attack until 20 years from now, when he's with a different doctor and a different plan, and I won't have to pay for his bypass!" If HMO is supposed to stand for health-maintenance organization, that was some irony, I said to myself, as I joined the millions of Americans frustrated and alarmed by the vast, rapid shift to managed care.

But simple-minded fantasies about greedy doctors and grotesque economic incentives rarely explain, I discovered after months of poking around, how managed care really works and why the shift of 50 million Americans into it in the past decade has been so disconcerting.

The problem, I think, isn't that managed care delivers inferior routine health care or routinely withholds big-ticket treatments like cancer care or heart surgery--there is no evidence that it actually does. What looks like a problem with managed care is actually a problem with medicine: Doctors don't know the best way to treat most illnesses. As a result, the quality and cost of a huge number of medical procedures vary enormously from doctor to doctor and from region to region. And a big, clumsy, mysterious bureaucracy that seems accountable to nobody has come along to highlight and exploit that unnerving variation in how doctors work.

Beginning about two decades ago, elaborate research by Dr. John Wennberg, now at Dartmouth, revealed fantastic differences around the U.S. in how doctors perform tonsillectomies and breast cancer surgery, and treat colds, stomach ulcers, heart disease, and just about everything else but hip fractures. Medical students, it turns out, don't learn what to do by consulting some central national repository of the best medical practices but by following a senior doctor around a hospital and seeing what he does. Yes, older doctors read bulletins and attend conferences, but their practices don't change much as a result. Researchers at Rand Corp. looked at this unevenness in the late 1980s and made an important discovery: The docs who used the most expensive and invasive techniques didn't cure patients quicker or faster. Frequently, in fact, aggressive use of medicines, surgeries, and hospitalizations produced worse outcomes at higher cost.

Bingo! A new branch of the managed-care industry took off: managed cost. If cheap, minimalist medicine was not demonstrably worse than the expensive stuff, why not steer doctors toward more economical methods? (Doctors screeched, naturally, viewing this as a sacrilegious intrusion on the art of medicine and their considerable autonomy.)

Alas, the reason nobody likes managed care is that managed-care companies--and the corporations that buy medical insurance for their employees--have found it far easier to measure the cost of care than the quality of care. They have become far too obsessed with cutting costs and have used clumsy, unsophisticated, inept, even brutal methods to extract care from doctors and hospitals on the cheap. Just as unforgivable, managed-care companies have blown a golden opportunity to install computerized data systems into their vast networks of patients and doctors to help determine just what medical techniques deliver the best, quickest cures at the lowest cost.

Partly as a result, the original and comforting model of managed care has become increasingly rare. This is the Kaiser Permanente approach, in which all the doctors, including specialists, work exclusively for Kaiser--the so-called group model. The salaried doctors have no incentive to overspend or underspend on your care. But because their small annual bonus is tied to the economic health of Kaiser, the docs are motivated to spot and cure disease early and efficiently. Unfortunately, group-model HMOs are relatively expensive and have been underbid by other types of managed-care company, and their share of the 150 million people covered by managed care is shrinking. Kaiser, a not-for-profit, lost $270 million on revenues of $14.5 billion last year.

The far more prevalent managed-care model, enveloping nearly half the population, is the very different "independent-practice association," a loose-knit confederation of doctors pulled together by an insurance company. How does it get established? Typically the Econo-Care managed-care company seeks the health insurance business of Amalgamated Widget, a giant corporation with 50,000 employees in 30 states. Before it can land the contract, Econo-Care has to line up doctors, labs, and hospitals for Amalgamated's workers all around the country. Since there is usually no preexisting network of doctors to contract with, Econo-Care has to send out recruiters, who try to get practitioners--doctors working solo or in groups of a dozen or two--to sign up. But because Amalgamated Widget is frantic about rising health-care costs, Econo-Care also has to get the doctors to charge less than before.

The recruiting methods have been mostly crude. If Amalgamated runs a giant smelter in a small town in Missouri, and half of every local doctor's patients are Amalgamated employees and their families, the task is simple. Tell the doctor to do the work at a deep discount off the regular price--the so-called discounted fee for service. Arrange it so that Amalgamated workers can't go to a specialist without permission from a primary-care doc beholden to Econo-Care, and make the doc ask permission to send patients to specialists or to the hospital. Otherwise, Econo-Care will stop paying his bills and send hundreds of patients off to a more cooperative doctor, whose work will be transformed into a high-volume, low-margin business. Econo-Care works similar voodoo on specialists and hospitals, lining up gastroenterologists and dermatologists who will work for less and reserving the right to decide when a patient is ready to leave the hospital.

One common, highly controversial technique of managed-care companies is the use of bonuses or incentive pools to reward doctors for not wasting medical resources. If bonuses are structured properly, they are not bad. They can force doctors of the old school, who were accustomed to running every conceivable test and parking patients in hospitals for weeks, to become heedful of cost. Unfortunately, how pools are set up, and how powerful an incentive they create for doctors to deny appropriate care, is often kept secret by the managed-care companies, creating the kind of paranoia evident in the opening paragraphs of this story. A rough rule of thumb is that no more than 10% of a doctor's income should be at risk with pools and incentives; 20% is way too high, suggests an official of the American Medical Association. As doctors in Fort Worth discovered, you will see later, a managed-care company's poorly conceived bonus arrangement can be a disaster for everyone.

Doctors, naturally, often resist getting caught up in managed care. Joseph Martingale, a managed-care consultant at Towers Perrin in New York, says that when managed care controls 10% to 15% of the patients in a region, a stampede often develops. Unconverted doctors see patients disappear and revenues start to dry up. (A decade ago about 70% of employees with company-sponsored insurance had traditional fee-for-service plans. Now 65% are in an HMO or a network of discounted or "capitated" doctors, and only 1% are in an unmanaged indemnity plan, according to KPMG Peat Marwick. Wide variations in penetration remain, however: In 1996 only 41% of patients in the mid-South were in managed care, according to an October report in Medical Economics magazine, vs. 55% in the Northeast and 66% on the West Coast.)

Once a managed-care company has a large network of doctors in a region, it is in a better position to win new business from large corporations. With more new patients, it can squeeze even more concessions from the doctors. Not surprisingly, managed-care companies like to merge with one another, since size gives them negotiating power. There are close to 1,000 managed-care companies in the U.S. now, but about 35 firms, such as Blue Cross, Aetna/U.S. Healthcare, United Healthcare, Prudential, and Pacificare, handle some 80% of the business. Many managed-care operations grew out of traditional indemnity insurance companies that previously handled medical claims under the fee-for-service model.

One great frustration for patients and doctors is the confusion that results when numerous managed-care companies swarm into an area, each with different and rapidly changing rules, specialists, profit goals, ethics, and managerial competence.

Witness, for instance, the scene that unfolds in Freehold, an attractive small town in central New Jersey, where the workday is well under way at the Monmouth Family Medicine Group by 9 a.m. on a recent Wednesday.

Between 8:30 and 5, the three doctors, eight receptionists, two practical nurses, and five medical assistants on duty today will see 110 patients. The 23 phone lines will handle several million phone calls, or so it seems. A dozen patients are sitting calmly in the waiting room, but the back office is already the scene of controlled chaos that barely lets up for a moment all day. It will be a crazier day than usual. The 20 doctors in the Family Medicine Group actually operate out of a half-dozen offices in the region, and Dr. John Gumina, who formed the group in 1977, will be in Freehold today. He has a large and devoted following of older patients who won't see anyone else, confides his office manager, Roberta Barry, and so they pile in to the office on Wednesdays when he is there.

Gumina, 50, a relaxed and friendly man with a smile that often broadens across his entire face, is tolerant of managed care, he admits. He was one of the first to try it when U.S. Healthcare, then a small new company (now part of Aetna), appeared in New Jersey in the early 1980s. Now his group handles one of Aetna's biggest patient loads in the state--6,000. In all, the group grosses nearly $5 million per year, contracts with 30 managed-care companies, and provides primary care for 70,000 people.

On the surface, the rules are fairly simple. Most of the managed-care companies have negotiated "discounted fee-for-service" rates with Dr. Gumina, in which they agree to pay for any work he does but at a price lower than what he charges patients not in managed care. Some companies, though, including Aetna/U.S. Healthcare, make very different arrangements, called "capitation." Under capitation, Dr. Gumina's group gets about $15 per month for each patient that has selected Monmouth Family Medicine to handle his primary care. Whether all the capitated patients come five times a month or none of them show up for a year, the group gets the same fixed amount. Like doctors everywhere who accept capitation contracts, Gumina can only hope that the $15 per month (and the $5 to $15 co-payments that all patients make) will cover their cost of care. With his other patients, the more they show up, the more money Gumina's group will get. But if capitated patients show up in droves, his hours and expenses will rise while his income will barely budge.

It's when a patient needs to see a specialist or get an outside test that the chaos and lack of coordination of managed care is manifest. Each of the 30 companies with which the group contracts has struck its own economic arrangements with nearby hospitals, specialists, labs, and x-ray centers, and all have maddeningly different rules about how and when Gumina et al. can use them. Send a patient to the wrong one, and the insurance company won't pay or will bounce the bill to the patient. On this day, a sore-looking kid with a baseball injury to his arm shows up and needs a bone scan. But he can't go to the hospital across the street. His plan got a discount on bone scans at a hospital a half-hour away, and he must go there.

But first, one of Gumina's office assistants must get approval from the managed-care company. One receptionist spends nearly all her time calling managed-care companies and asking for approvals. A few plans are a breeze--they have installed devices similar to what retailers use to swipe credit cards, and the process takes seconds. But a big new HMO (Gumina's group prefers not to identify it) is an organizational nightmare: The receptionist routinely spends 30 to 60 minutes waiting on the phone, listening to elevator music and "pressing three for more options" before a nurse finally comes on the line. Other plans aren't much better. (Gumina's staffers go silently crazy when patients show up and announce they have a long-standing appointment with a specialist starting in 20 minutes, and they need an authorization immediately.) The mother of the kid with the damaged arm appears somewhere between numb and furious over the delay that managed care is causing.

Roberta's phone rings. A patient has been sent to a radiology lab approved by her plan. But the lab's x-ray machine is down, so the lab sent her to a nearby hospital. But the patient doesn't have any paperwork authorizing her to get the x-ray at the hospital, so the hospital wants Dr. Gumina's office to be responsible for the cost. Roberta is incredulous and annoyed. "It's the radiology lab that has the problem," she says sharply. "It's their machine that's broken." To accommodate the patient, she finally agrees to eat the cost and hangs up. More phone calls. The patient, it turned out, had gone to the wrong section of the hospital. "She went to outpatient," Roberta fumes, rolling her eyes. "We specifically told her to go to inpatient." The patient is finally mollified, at least for today. Told that her job looks exhausting, Roberta laughs. She seems energized by the chaos. "If it wasn't like this, I'd be bored."

In fact, Gumina has had to hire a huge staff to manage managed care--four for every doctor in the group. The receptionists have to scrutinize everybody's insurance card, checking big books to make sure the employer hasn't changed plans and the patient is still covered. Otherwise, the insurance companies won't pay. Some insurers have multiple plans with nearly identical names but a completely different set of hospitals and specialists that patients must use. If the receptionist codes the wrong plan and sends Mrs. Brown to the wrong gastroenterologist, the patient will get the bill. Some plans change specialists and labs but don't tell Gumina's office right away, then refuse to pay the bills. More phone calls required. Some plans pay for any MRI, an expensive form of x-ray; others pay for head and chest MRIs but rarely allow them for backs and legs. Nearly every time a doctor refers a patient to an outsider, the nurses must lug out directories the size of telephone books, see what is permitted, see which specialist or lab must be used, and seek approval by phone, fax, or computer. (Sometimes the rules are absurd. A doctor in Texas told me about a plan that requires patients to go to a certain radiology lab, but the only radiologist working there hasn't been approved by the plan; it will pay for the x-ray but won't pay the radiologist to interpret it.)

There is a mild, if unintended, benefit brought about by the conflicting rules of the 30 managed-care companies Monmouth Family Medicine contracted with: The group's doctors don't care whether you're with Prudential or Oxford or United. They can't keep track of which plans allow which treatments, or how profitable or expensive a particular patient will be. If one plan is a bigger headache than another to deal with, doctors generally leave it to their staffs to iron out. Dr. Gumina and several doctors interviewed at other groups insist convincingly that they ignore the often conflicting guidelines with which the companies bombard them and practice medicine pretty much as they always have. Gumina's staff agrees. "I handle the billings, and I know what tests the doctors order," says administrator Debbie Welsh. "They spend the same time and do the same tests and treatments for everybody. They don't know if a plan pays $35 per visit or $49, and don't concern themselves about it."

But a badly put-together managed-care plan will indeed insert itself into doctors' day-to-day treatment decisions, even if they don't want it to. In Fort Worth two years ago, for example, Harris Methodist Medical Center, one of the oldest and biggest managed-care organizations in town, realized it was in trouble. Twenty-five competitors had entered the Fort Worth market in 1994, setting off a price war in an effort to win business from big local employers. The premiums Harris could collect for each of its 225,000 members dropped from an average of $140 per month in 1992 to $126 in 1995. Harris was losing $20 million a year on revenues of $1 billion, according to Neil Godbey, a Harris vice president. The company moved swiftly, sending registered letters to 2,000 doctors around Fort Worth, canceling contracts.

At the Fort Worth Clinic, the fourth-largest group practice in town, the news was stunning. "They terminated 20% of our business overnight. We didn't know what the new agreement would be," says Dr. Kendra Belfi, an internist and a member of the group's board. Harris, they found out, had allocated the premiums they got from employers into precise pools. In some areas Harris seemed reasonable: If a doctor group went over budget on specialists, Harris would eat the cost, and if a group stayed under budget, Harris would refund the difference. But if a group spent too much on prescription drugs, it had to pay 35% of the excess.

Not a good deal, as it turned out. The pharmacy budget was way too small. Long-established urban group practices like Fort Worth Clinic tend to have older and sicker patients than a two-doctor office located in a new suburban subdivision. Belfi and several of her colleagues say they always tried to ignore what medical plan a patient belonged to, but found themselves cringing when Harris patients came in, especially those who needed lots of expensive drugs. "You'd find yourself saying, 'How do I get rid of this person?' " admitted one embarrassed-looking doctor. Belfi wrestled with medical decisions that would cost her money: "Do you screen a patient for osteoporosis and put her on an expensive medicine like Fosamax to prevent possible fractures in the future?"

The Fort Worth Clinic decided it would not stint on medications. After the first year the clinic learned that Harris was deducting $90,000 from its payments for over-budget prescriptions. Nine months later Harris was demanding another $90,000. The doctors refused to pay, went to court, and got an injunction against Harris. (The state insurance department proposed an $800,000 fine against Harris, which it is appealing.) Harris turned a $20 million profit the year after it rewrote the rules, but a spokeswoman says the plan is losing money again.

Being hammered financially by managed-care companies over the past decade has made doctors' incomes highly unpredictable. The average income for all doctors was a respectable $164,000 in 1996, the last year for which data are available, virtually unchanged from ten years earlier. But the average dipped in 1991 and again in 1993, according to surveys by Medical Economics magazine, and some specialty groups have been especially hard hit (see table). Psychiatrists, for instance, made $113,000 in 1996, 14% less than in 1987, mainly because managed-care companies generally refuse to pay for long-term navel-gazing anymore.

What is unnerving about watching managed care up close is not that the care is worse than before, but that the exercise reveals what a highly variable, almost random approach doctors and managed-care companies take toward medicine. When I tell other doctors about the failure of my own physician to pay attention to my cholesterol, some are astonished, others shrug. Virtually all agree the lapse is a reflection not of managed care but of what my doctor chooses to worry about. He had probably been ignoring cholesterol levels even before managed care came along, several suggested. Does Aetna's approach, giving Dr. Gumina $15 a month to handle each patient, produce better medicine than Prudential's paying him for each patient visit? Why do 60% of the Medicaid patients in Kentucky still get antibiotics for a cold, even though that is completely inappropriate? Why were 16% of the hysterectomies approved by seven managed-care companies later deemed to be inappropriate, and another 25% of dubious value? How come the Blue Cross HMO in New Jersey thinks an MRI for an aching back is okay, while other plans balk? Did employers in Fort Worth have any reason to think that letting Harris punish doctors for writing too many prescriptions would improve the health of their employees?

It would be comforting to think that the vast and rapid shift to managed care was producing a consensus on the best way to practice medicine. But the inconsistent and often conflicting messages and methods companies impose on doctors is anything but reassuring. Doctors' love of autonomy and their suspicion that advice from managed-care companies is rooted in avarice have not advanced the cause either. Dr. John E. Wennberg, whose father at Dartmouth chronicled the enormous variations in the practice of medicine, is doing similar research at the Maine Medical Center in Portland and says there is little evidence that managed care has nudged doctors toward producing the best outcomes at the best cost. "Most of the savings managed care has achieved has been from forcing doctors to do less of what they did before and to accept less for doing it. Managed care hasn't changed behaviors, just squeezed margins," says the younger Dr. Wennberg.

The news on improving the quality of managed care is not entirely bleak. Large employers like Xerox, GTE, and AT&T have demanded more accountability from the managed-care companies they use. In response, an organization in Washington called the National Committee for Quality Assurance has begun performing increasingly elaborate analyses of hundreds of HMOs. It requires information on a plan's preventive-medicine programs, its record keeping, denials of treatment requests by physicians, and so forth. (Managed care, the NCQA found, is better at getting docs to do preventive-medicine screenings than fee-for-service is.) Critics say it's too easy for plans to score well by focusing on just what the NCQA measures, and that the center does not measure which plans produce the healthiest or sickest members. But the group is generally viewed as fair and objective. Margaret E. O'Kane, head of the NCQA, notes that about 12% of plans applying for the first time flunk the review, and another 60% get only temporary or provisional approvals (NCQA posts accreditation status on the Internet at www.ncqa.org/accred/asr/asrlist.htm). Although many employers require plans to undergo the review, too often they still select the plan with the lowest price.

Gradually doctors have been organizing to push back against managed-care companies, and that could lead, circuitously, toward a better practice of medicine. Particularly on the West Coast, where managed care has been around far longer than in most places, doctors realized they can negotiate much better deals with managed-care companies if they band into large groups of primary-care doctors and specialists. In New Jersey, Dr. Gumina sold his group practice last year to a new and even larger organization of doctors and hospitals, the Meridian Health System. Big multispecialty groups can take on riskier--and potentially more profitable--contracts than a tiny practice by agreeing to handle all of a patient's needs, from checkups to heart surgery, for a fixed monthly fee, says Burton Blumberg, a Meridian vice president. The groups become, in effect, miniature managed-care companies, but owned and run by doctors and hospitals.

Doctor-owned HMOs may be a mixed blessing. Some, such as the Palo Alto Medical Foundation in California, are highly motivated to develop the best medical outcomes and economic treatments, says Dr. David Mirkin, a consultant on actuarial risk and medical practice at Milliman & Robertson in New York. Paul Ginsburg, head of Health System Change in Washington, D.C., says the rise of the large, semiautonomous group practices is "one of the most exciting developments in medicine. They have a physician culture that says, 'What's the best way to treat diabetes?' It's far better than calling 1-800-NURSE and asking a managed-care company what they can and cannot do." But O'Kane at NCQA says some of the worst managed-care abuses come from big doctor groups that get a wad of money from an HMO that then ceases to monitor the quality of care the group delivers. "They say, 'Here's the money. See you around.' Some of the groups overwithhold care."

It may be quite a while before a big group of high-minded doctors arrives in a community near you. One of the most popular developments in managed care is even looser networks of doctors, in which patients have considerable freedom to bypass their "gatekeeper" and can pay extra to visit specialists at will. Called a point-of-service plan, it's like tacking the old fee-for-service approach onto an HMO. About half the new patients enrolling in managed care in the past two years have joined point-of-service plans. The trend disturbs Dr. Wennberg, who says shifting back toward fee-for-service will make it harder to coax doctors toward best practices. "It's not a recipe for reducing the variability," he says.

What, then, should you do? There isn't any rule of thumb to steer you to the best plan. A big survey of patient satisfaction by J.D. Power & Associates and the Medstat Group released in June found no clear pattern for which model produced the happiest customers. Says Dennis Becker, a Medstat executive: "It's not so much the type of plan you're in as how well it's implemented." The simplest advice: Find a great doctor, and he or she will fight to make managed care work for you. That's easier said than done.

I have decided I don't trust my doctor after the cholesterol mishap, and I'm getting a new one. A neighbor in the pharmaceuticals business recommended his doc. I have to wait a mere ten weeks before I can see him. Will he be better than the last? Who knows? I called my managed-care company to ask whether the new guy is any good: lots of complaints, long waits for appointments, etc.? "No way," responded the agent on the phone, apparently surprised that I would ask. "We don't keep records on that." The plan investigates any serious complaints, she explained, but those are confidential. Great. Now what?

Maybe my old doctor wasn't so bad. I'm sleeping okay, my neck is better, the hernia never materialized, and he put me on pills for my cholesterol as soon as I panicked. Good thing he decided not to operate. Or did the managed-care company decide? I'm not sure I want to know.

REPORTER ASSOCIATE Cora Daniels