The dark side of P4P.

In 2003, gastroenterologist Jim Petersen, MD, of Chesterfield, Missouri, had 279 encounters with patients covered by a UnitedHealth Group PPO medical plan. The cost of their care, according to the Minneapolis-based insurer, totaled $375,000.

Unfortunately, under the "efficiency" calculus of a new "UnitedHealth Performance[SM]" pilot program introduced by the insurer in 13 states in early 2005, Petersen's costs--based not only on his own charges but also on all the hospital, diagnostic testing and associated medical expenses incurred under those patients' treatment regimens--were $1,600 too high.

To be considered an efficient doctor, ruled UnitedHealth, Petersen would have had to have kept the bills to $373,400. He'd exceeded the efficiency cutoff by 4/10 of 1 percent.

As a consequence, Petersen did not qualify for the laudatory designation "Performance Physician." He did not receive a star after his name in the UnitedHealth Group's directory of local PPO network doctors.

To add injury to insult, that meant that anyone insured by the company--which covers some 900,000 people in the St. Louis area, notably hourly employees of General Motors, Daimler Chrysler and UPS--would have to pay out of their own pocket 100 percent of the cost of any office visit or procedure provided by Petersen.

On UnitedHealth's terms, Petersen, the respected president of his county medical society, was a second-rate doctor.

On one loyal patient's terms, "UnitedHealth is like I have insurance but I really don't."

Welcome to the world of physician pay for performance.

Millions and millions

In 2004, William McGuire, MD, earned a salary of about $2.2 million. He received "other" compensation worth $2.5 million, according to Forbes Magazine. For his success in meeting a set of internally established corporate performance targets, he collected a bonus of some $5.6 million. And on top of that, he was awarded stock options worth $114.5 million.

Altogether, according to Forbes, which publishes an annual comparison of CEO earnings and corporate performance, McGuire banked $124.8 million for his year's work. That made him the third-highest-paid chief executive in the Fortune 500--the heaviest of the heavyweights among U.S. corporations. (McGuire was out-earned only by the chief executive of Yahoo!, Terry Semel, and the celebrated head of IAC/Interactive Corp., Barry Diller.)

McGuire has been CEO of UnitedHealth Group for the past 14 years. His 2004 salary was 133 percent higher than the median for CEOs in the health care equipment and services sector; his "other compensation" was 562 percent higher than the median among his peers; his performance bonus was 472 percent higher than the median.

And yet, according to Forbes, McGuire was surpassed in leadership efficiency by 32 of the 189 chief executives who'd steered their companies for at least six years. Number 33 in outcomes as measured by UnitedHealth Group's sales, profits, assets and overall market value during the period, McGuire ranked No. 3 in heft of paycheck.

Welcome to the world of executive pay for performance.

[ILLUSTRATION OMITTED]

It's hard to find anyone who will badmouth the concept of selective monetary rewards for excellent performers, whether doctors or chief executives.

To be fair, McGuire can boast that Fortune Magazine's analysis of his efficacy as a corporate leader places him in the 82nd percentile of Fortune 500 CEOs.

Petersen says he takes particular pride in "the way I do colonoscopies." After a moment's consideration and a generous adjustment out of modesty, he pegs himself at, "well, I'd say I'm certainly within the top 25 percent in the area."

So it rankles, he admits, that there is no premium in reimbursement for him, a specialist who performs the procedure with practiced skill and rare complications, over "some surgeon who might do one or two colonoscopies a month" and whose inadequacies may lead to pain, misdiagnosis or injury.

"In my mind I know some of these doctors are not very good," he says. "And yet, they get paid the same as I do. There's no satisfaction in that. And you're not being rewarded on volume."

So, yes, he says, "We all support pay for performance if it means the best doctors are getting something extra. But the way United went about this was just a total mess."

One physician who has voiced unequivocal disdain for pay for performance is Santa Monica, Calif., thoracic and cardiovascular surgeon William G. Plested III, MD.

As chairman of the AMA Board of Trustees in 2003, Plested wrote a vitriolic editorial in American Medical News assailing pay for performance as "the newest scam dreamed up by the multimillionaire CEOs of health insurance companies and HMOs to reduce payments to the vast majority of physicians."

(In fact, one effect of the application of UnitedHealth's performance template in St. Louis was that 88 percent of the 932 physicians in the prestigious Washington University School of Medicine faculty practice group found themselves labeled inefficient, reports Petersen.)

"What amazes me," Plested added, "is how many physicians actually believe this claptrap."

As it happens, in March 2005 the AMA formally sided with the believers.

"We support the idea that in order to keep physicians up to speed and on the cutting edge of medicine, some financial incentives might be included," acknowledges AMA senior public information officer Robert Mills.

"Physician pay-for-performance programs that are designed primarily to improve the effectiveness and safety of patient care may serve as a positive force in our health care system," the new AMA policy statement allows. To earn its favor, though, a program must conform to five basic principles, the AMA cautions.

Plested has since climbed a rung on the AMA ladder. He's now its president-elect. And his organization has a pay-for-performance policy. That means any personal reservations he may still harbor have to be muzzled.

"I'd be happy to bend your ear," he told a reporter recently, asking only that the interview be scheduled through the AMA's media relations office. But when contacted, Mills made it clear Plested would only--could only--echo the AMA party line.

The right way and the wrong way

By far the most ambitious experiment in physician pay for performance is under way in California. More than 35,000 physicians in some 225 medical groups statewide--every one with at least five doctors--is eligible to receive incentive payments from the state's seven major health plans (covering 6.2 million HMO enrollees) keyed to their scores on a set of quality benchmarks.

In 2004, its first year, the program distributed bonuses totaling $50 million.

Ronald Bangasser, MD, a wound specialist in Redlands, Calif., and a former president of the California Medical Association, is a member of the board of directors of the Integrated Healthcare Association, which designed and runs this pioneering pay-for-performance initiative. He also chairs the IHA's technical committee. (The program was described in the May-June 2004 issue of The Physician Executive.)

Bangasser admits that he started out in the "claptrap" school.

"At first I was convinced that pay for performance wasn't the right thing to do," he says. And although finally persuaded that a framework could be constructed that would be "fair, honest and transparent," he remains "very concerned that of the 120-odd programs that are now in operation in the United States, many certainly have problems."

"There are ways to do it right," he declares, "and there are ways to do it wrong."

Speaking from experience--with two years under its belt, the IHA has documented significant gains across 14 clinical, information technology and patient satisfaction quality parameters--Bangasser rattles off a series of critical questions to pose about any pay-for-performance initiative.

1. Have physicians been involved in the process of picking the performance measures?

"If there's no physician buy-in, there'll be a pushback," he explains. "A little money will change behavior very quickly if doctors believe it's a good measure; if not, a million dollars won't make me change my practice pattern."

2. How many measurements are being used to gauge physician performance? And who's filling out the scorecard?

"Often too many measures have been set up," says Bangasser. Moreover, if physicians lack the sophisticated wherewithal to track every one of them, insurer claims data usually fills the void. But this can be seriously misleading--"particularly when one health plan has control of all the data," he warns.

"Take the example of mammograms," he suggests. For various reasons, patients sometimes go outside a practice to get them, he says. So claims data for that practice will not reflect the true percentage of patients who've complied with the breast cancer prevention protocol--a common clinical quality indicator. Although the patient has been appropriately treated, and the physician knows it, no associated claim is on record with the insurer. Points out Bangasser, "I'd get dinged."

3. How is patient satisfaction--a common measurement component of physician performance quality--being gauged?

"Some patient satisfaction surveys are not very robust," Bangasser observes.

It's all very well and good to poll patients on "How long was the wait to see the doctor?" and "Was the office staff kind to me?" he concedes. "Those are nice tokens, but in terms of quality, what's really important is 'Did my physician help me quit smoking? Did my doctor make sure I got a flu shot?'

If the answer is 'no,' then you have a problem."

4. Are physicians being assisted financially to acquire the expensive information technology that underpins pay for performance?

"There's a cart and horse," asserts Bangasser. "Pay for performance without an electronic health record or registry means you won't know who's falling through the cracks." But if the financial onus of gearing up for pay-for-performance data demands falls entirely on the doctors being judged, "I can bet," says Bangasser, "that almost any place in the U.S., they'll push back."

5. Who's marking out the field and making the calls?

"It's absolutely essential to have an independent, non-profit convener," says Bangasser.

The IHA, a consortium of health plans, physician groups, health systems, plus large academic, purchaser, pharmaceutical industry and consumer representatives, meets that definition. A single insurer doesn't.

"It's also absolutely essential to have an independent data aggregator," he stresses, "to audit the results, make sure they're accurate and come up with the scores."

In California, the National Committee for Quality Assurance fills that role. In Missouri, UnitedHealth called its own shots.

6. How is cost efficiency being determined and promoted?

"To do it the right way," says Bangasser, "you set up a team to work together, using and sharing the data to create efficiencies. You attribute the cost of care to all the physicians involved, rather than setting them up to compete one-on-one for dollars."

7. Where is the [pay-for-performance] money coming from?

"If you're taking it from one physician and giving it to another, it won't work," asserts Bangasser. "You can't just put some money aside and then pay it back for quality. You have to have new money, from a separate source. Otherwise, you're going to get into a very negative, confrontational situation."

Getting it wrong

UnitedHealth's ill-fated foray into pay for performance suffered from almost every failing cited by Bangasser.

First, the insurer decided unilaterally to base its scoring on two parameters: "quality" and "efficiency." Physicians being rated were given no input into the process of choosing the indicators.

Second, United relied on its own claims data from 2002 and 2003--outdated to begin with, critics argued--to evaluate physician performance. Doctors for whom data were lacking, meaning as many as 40 percent of those practicing in Missouri, were automatically ineligible to be classified as Performance physicians. And stars were awarded or withheld in some cases based on as few as 10 claims.

For physicians in nine specialties for which UnitedHealth could find no officially agreed-upon clinical quality standards, efficiency--read cost--constituted the sole rating measure.

Moreover, Petersen and his colleagues throughout the region took strenuous exception to the way UnitedHealth assessed its two performance parameters.

Under the quality metric, for example, a physician could be classified a star performer with a complication rate up to three times higher than the regional average for peers, according to Petersen. To merit a star under the efficiency standard, a physician's costs had to be 20 percent lower than the average.

"I don't know about you," he comments, "but if you're going to judge on quality, I'd want my doctor to have at least the average complication rate."

UnitedHealth's calculus was already fatally flawed, Petersen adds. For instance, if a physician on call in a hospital emergency room had treated a patient for an infection following surgery performed elsewhere by some other doctor, claims data from the hospital would pin the wound complication on the physician who happened to be staffing the ER.

Indeed, Petersen and his colleagues complained loudly at the very inclusion of hospital charges in UnitedHealth's pay-for-performance formula. Since each system negotiates its own contract rates with insurers, Petersen points out, costs vary significantly.

"When I do a colonoscopy at a BJC hospital, I'm an inefficient doctor," he scoffs. That's because BJC HealthCare, the St. Louis-area's largest hospital system, wrangled comparatively generous reimbursements from UnitedHealth, he explains. "When I did one at St. Luke, which has a crappy contract, I was an efficient doctor."

Petersen wryly suggested to UnitedHealth that it needed to reveal the details of each of its contracts so he could figure out "where I can go to do a cheap colonoscopy. Well, of course, the hospitals got all hot and bothered at that."

In fact, providers of every stripe had erupted in howls of protest at UnitedHealth's actions. BJC HealthCare threatened to pull out of the insurer's network, as did Tenet Healthcare Corp., operator of two local hospitals. Washington University Medical School doctors announced the same intention.

The American Academy of Family Physicians, among a slew of organizations, fired off an outraged letter to McGuire. Under the guise of pay for performance, charged AAFP president-elect Larry Fields, MD, of Ashland, Kent., UnitedHealth had actually introduced "a tiered network benefit plan that promotes patients going to certain physicians based on cost without regard to the physician/patient relationship."

"They destroyed about 75 percent of patient relationships," nods Petersen. "But they really took a hammering."

Rocked by the intensity of the backlash, UnitedHealth sheepishly denied that its Performance plan had ever been conceived as a pay-for-performance initiative. After some tinkering with the formula and the award of a few more stars, the company simply suspended the program altogether.

"Emerging efforts to use data as a tool to both improve results and enhance health care decision making are gaining momentum," insisted UnitedHealth CEO McGuire. But he conceded that "it is imperative that such programs be as accurate as possible, continuously evolve and improve, and appropriately reflect the realities of clinical practice."

Thus the company has now recruited a "standing advisory committee" of "respected physician leaders, clinicians and other experts drawn from medical specialty societies, state medical societies, academic medical centers, and medical practice groups, as well as health care performance assessment experts."

Their input will be invoked to legitimize UnitedHealth's next excursion into pay-for-performance programs.

Accidents waiting to happen

"I found myself dumfounded that we had to be paying people for doing what they should be doing," says San Francisco health care consultant Lucy Johns of her initial reaction to pay-for-performance programs for doctors.

"But the idea that people respond to financial incentives is widely believed. And whether or not it's philosophically congenial with the health services philosophy, [paying physicians for performance] probably doesn't harm anything. At least not as long as the key measures are things that are good for patients."

Still, like many observers, Johns is concerned that pay for performance "creates an incentive to fudge the numbers and to manipulate the database."

In the United Kingdom, where the latest National Health Service contract with primary care physicians pegs as much as a third of their income to meeting performance goals, Muttucumarasamy Mahendran, MD notes that his practice in Milton Keynes now has to "work out when we will measure blood pressure to coincide with the peak effect of the medications." Although "good 24-hour control should be our aim," he complained in an April 2004 letter to the New England Journal of Medicine, the British pay-for-performance approach means doctors have to "work like managers--'milking figures' for targets."

Which is precisely what has happened when chief executives are rewarded for performance, according to monitors of American big business. They point to recent scandals at Delphi, Fannie Mae, OfficeMax, Qwest and WorldCom--"companies that heaped big performance-based bonuses on their bosses but subsequently had to restate earnings lower after accounting shenanigans were discovered."

Writing in a May 9, 2005, Forbes.com examination of CEO compensation, Michael Ozanian and Elizabeth MacDonald note, "It's easy to play with how you meet targets.... Give a smart boss the incentive to do it, and he can push the earnings envelope to the limit--or beyond." (UnitedHealth Group is among companies whose accounting methods, in Forbes's estimation, are "aggressive.")

Even proponents acknowledge that there are plenty of bloody accidents waiting to happen when pay for performance hits the road.

David B. Nash MD, MBA, FACPE, chairman of the department of health policy at Jefferson Medical College of Thomas Jefferson University, in Philadelphia, thinks the "risk is high" that pay for performance could "create a class-system of care. Physicians may become wary of treating those for whom a guaranteed positive outcome doesn't exist. A practice that sees a lot of the poor, the sick and the old will have a lower batting average. And that puts you at risk if you're paid more to be a slugger."

There is also, worries Bangasser, the possibility that programs constructed along UnitedHealth lines will produce only an unfair tiering of physicians, with those on the lower levels penalized by discouragingly higher patient co-pays. That would be especially problematic, says Bangasser, if differentiating tiers turns on measures--say 75 percent patient satisfaction at the top and 72 percent at the bottom--that are statistically insignificant.

What's more, a viable pay-for-performance plan has to reward physicians for incremental improvements, he stresses; it can't simply draw pass-fail lines.

At that, asks Nash, "What do you do with the consistently poor-performing physician? [Pay for performance] creates an obligation to pay far greater attention to the poor performers and offer training opportunities for improvement."

It could happen

There's little question that incentive pay for physicians is the wave of the immediate future--even though, as Nash points out, "the research to support [its effectiveness] is modest--very modest. So from a research perspective, we're in uncharted territory."

Be that as it may, the Center for Medicare and Medicaid Services has launched a five-year demon-stration project in which participating physicians and integrated delivery systems that "achieve safety, effectiveness, efficiency, patient-centeredness, timeliness and equity" goals will qualify for higher reimbursements.

Congress is tackling legislation that will link all Medicare physician payments--slated for a 4.3-percent reduction in 2006, which is the principle topic of the debate--to some form of pay-for-performance framework.

Many doctors like Robert Doherty, senior vice president for governmental affairs and public policy at the American College of Physicians, fear that Congress could prescribe "measures that are not scientifically valid and/or will penalize them for circumstances--such as patient non-compliance--beyond their direct control. They fear that pay for performance will create yet another regulatory hassle that will require more work to produce data of questionable value with little or no improvement in compensation."

As UnitedHealth Group's unhappy experience in St. Louis demonstrates, doctors are not powerless. They are not condemned to helpless buffeting on the seas of change. Poorly conceived and implemented on so massive a basis, a Medicare/Medicaid pay-for-performance mandate could spark a rebellion led by physicians who won't or can't jump through hoops they scorn as illegitimate.

That's Bangasser's nightmare. "If it happens," he says, "patient access is in the crapper. And it could happen. Look how many doctors already refuse to accept Medicaid. Twenty percent of physicians can't take care of all the Medicare patients in the U.S.

"I've tried to have some influence," he declares. "If a pay-for-performance program is done right, there are no downsides. But I don't know. It could happen."

David Ollier Weber is a freelance health writer and frequent contributor to this journal. He can be reached by e-mail in Mendocino, Calif., at [e-mail]