Pay-for-Performance: At Last or Alas?

What a lovely phrase … pay-for-performance. For those who toil harder, work smarter, go the extra mile, satisfy customers, follow the rules, comply with regulations—the surest incentive of all—more money on payday.

The underlying principle of pay-for-performance compensation, now the rage in health care, is the creation of financial incentives that reward quality-improving, cost-saving, and more efficient behavior by medical professionals.

For many in health care, the question is not why now, but rather what took so long with this pay-for-performance? Isn't it about time that someone recognized the quality work of physicians and hospitals? After all, for years everyone has known that a fee schedule that rewards volume will result in … more volume. Therefore, isn't it a no-brainer that a payment mechanism that compensates clinical excellence will lead to better quality? What's not to like?

Barely five years old, a myriad of different pay-for-performance strategies have already spread from New England to California, and the notion has been embraced by hospitals, physicians, health plans, and employers, particularly large corporations. Medical journals, management literature, the lay press, government leaders, and CEOs all tout the concept as a solution to the ills of health care.

But does pay for performance, long the mantra of corporate America, offer a realistic means for improving quality and efficiency in health care? Or is this another ill-conceived strategy that falls flat on the way from bench to bedside? Will pay for performance be the answer to national concerns regarding rising health care costs and uneven quality in medicine? Or will this payment mechanism founder in a sea of complexity, imperfect data and provider pessimism?

Pay-for-performance made it to Main Street commerce in the 1990s when it emerged as the model for executive compensation, especially in publicly traded, for-profit companies. The idea was logical enough: senior management would have their pay (in the form of salary, equities, or bonuses) tied to quantitative outcomes such as earnings or stock price—presumably a win-win strategy for shareholders and company officials.

Though mentioned in health care management circles as early as 1985, the pay-for-performance movement did not really get going until the creation of the Leapfrog Group in 2000. Prompted by the 1999 Institute of Medicine report¹ about the parlous state of quality in American medicine, companies such as General Electric, IBM, General Motors, and Boeing launched Leapfrog with an original mission of disseminating information about quality and fashioning a payment mechanism that rewarded value and efficiency.

Leapfrog quickly settled on three standards for judging hospitals: computerized physician order entry, full-time intensivist staffing of ICUs, and referral to hospitals with high-volume surgical practices. Hospital compliance with these voluntary standards is published annually in the group's Hospital Quality and Safety Survey.

While tangible benefits for complying with these guidelines have been limited, some hospitals in New York have been given bonuses for meeting the standards, and in Seattle employees have had copayments waived at cooperating institutions.

A second major project began in 2003 when Premier, Inc, a medical center purchasing alliance, partnered with Medicare in a pilot project to improve quality in more than 300 member hospitals. The trial involves following patients with myocardial infarction, knee and hip replacement, congestive heart failure, community acquired pneumonia, and coronary artery bypass surgery.

As an example, in orthopedic surgery patients, outcomes such as antibiotic usage in the perioperative period, post-operative bleeding, and readmissions within 30 days will be measured. In coronary artery bypass patients, rates of internal mammary grafts and inpatient mortality will be assessed.

For the first time, there was an explicit financial incentive for participation. Hospitals in the top 10% will receive an additional 2.0% in payments, the second 10% will earn an extra 1.0%, while the lowest 10% can be docked as much as 2.0%.

Bridges to Excellence, originated by General Electric in 2003, goes one step further than Leapfrog by creating a financial bonus system for physicians, at least as pertains to caring for patients with diabetes and heart disease. By adhering to National Committee for Quality Assurance guidelines, a physician can earn $80 for diabetic and $160 for heart patients per year. The guidelines are straightforward—for example, monitoring lipids, blood pressure and renal functions in patients with diabetes; smoking cessation; and antithrombotic use in cardiac patients.

From its origins in New England, Bridges has now spread across the country as the product has been licensed nationally to insurers including BlueCross BlueShield, Cigna, and United Healthcare.

While all these initiatives provided a boost for the pay-for-performance movement, the whole landscape for physicians was jolted recently when the 2000-pound gorilla, namely Medicare, got into the game. On February 1, 2005, Mark McClellan, MD, an economist/internist and the Director of the Center for Medicare and Medicaid Services (CMS), announced that ten physician groups would be enrolled in a pay-for-performance trial, dubbed the Medicare Physician Group Practice Demonstration.²

These practices included the Geisinger Health System in Pennsylvania, Dartmouth-Hitchcock in New Hampshire, Deaconess Billings in Montana, Forsyth in Winston-Salem, the University of Michigan Faculty Group Practice, and others. Unlike the Premier experiment, which involved only technical or hospital fees, this new venture would put physician revenue at risk.

Each group will have a different area of concentration. For instance, Geisinger will emphasize the use of its electronic medical record to improve access to health care information among Medicare beneficiaries in rural Pennsylvania. Other participants will optimize treatment of patients with diabetes, congestive heart failure, hypertension and COPD utilizing means such as home care, preventive health, and disease management programs.

Other programs have arisen that share a similar philosophy. In California, the Integrated Healthcare Association, formed in 1994 by six health plans encompassing over seven million members, is unique in its incorporation of patient satisfaction into the bonus formula for its pay-for-performance initiative. In 2004, more than $50 million was distributed on the basis of clinical quality (50%), patient satisfaction (40%) and computer investments (10%). (Kaiser Permanente initially played an advisory role only, because the medical group incentive payments did not fit with its integrated health plan-medical group structure. However, the two California Permanente Medical Groups began reporting data on clinical and satisfaction measures to the IHA initiative in 2005 and 2006.)

Hospital Compare, a cooperative effort of CMS, the Department of Health and Human Services, and the Hospital Quality Alliance, was introduced in April 2005 to provide the public with quality metrics from every American hospital on their outcomes in patients with myocardial infarction, congestive heart failure, and pneumonia.

So as all these projects show, the pay-for-performance movement is not only alive and well, it's growing. Nonetheless, legitimate concerns have arisen, generally centered on the programs' effectiveness, durability and fairness. Such issues include:

Stifling Innovation

In nonmedical industries, with their vastly larger experience, one of the major worries about pay-for-performance programs has been about overvaluing the status quo and underinvesting in new initiatives.³ A company might find it easy to measure and reward established, easily quantifiable outputs such as sales, but what about strategic planning or new product development? A medical practice can easily tabulate the RVUs for routine CTs of the abdomen, but who pays for the time and effort to develop virtual colonoscopy? Are the short-term gains of increased activity in these conventional areas coming at the expense of a company or medical practice's future?

Undervaluing Teamwork

The new Medicare Physician Group Practice Demonstration proposes paying physicians more for better results in treating patients with congestive heart failure, asthma, diabetes, depression, and other conditions. And in the descriptions of the individual project goals, much emphasis is placed on collaborative care. Yet no mention is made about compensating other members of the health care team—not nurses, technologists, therapists, pharmacists—no one but physicians. In a profession where virtually no task is performed alone, how will this be justified? What will be the impact on the morale and professionalism of valued colleagues?

Exploiting the System

In the Hospital Compare database, facilities are compared on the basis of the time between the diagnosis of pneumonia and the initiation of antibiotic therapy. Who makes the call on the diagnosis of pneumonia? The paramedic? Senior resident? Attending? Does someone in the emergency room trigger a stop watch? What about patients with other infections? Will hospitals shortchange other patients with urinary tract infections, meningitis, or bronchitis in their race to beat the clock?

Although this may seem like a frivolous scenario, teasing out a subgroup of patients for analysis does raise questions about extrapolation of the data to an entire hospital population.

Selection Bias

Benchmarking, which is at the core of pay-for-performance, is not without its limitations. As discussed recently by Denrell,⁴ in the absence of a carefully monitored setting such as a controlled experiment, corporate data is frequently not collected in a rigorous manner. Quite the contrary. In commerce, successful businesses are happy to answer surveys about their triumphs, but the unsuccessful companies—those whose ideas failed—are out of business and no longer around to respond. So, like Lake Wobegon, only above-average results are tabulated.

In medicine, this phenomenon corresponds with the well-accepted fact that investigators frequently do not publish negative results.

Fragmentation

Most community physicians practice at more than one hospital, and nearly all participate in multiple insurance plans. As noted by Epstein, et al,⁵ if only 1% of an internist's patients were in the Bridges to Excellence program, the annual bonus would be $1265, hardly worth the paperwork necessary to enroll. If each pay-for-performance program necessitates an incompatible information system, this could pose an insurmountable burden, particularly to small practices.

Winners and Losers

Behind the headlines about pay-for-performance, some morning-after realities are not so pleasant, specifically zero-sum accounting, better known as winners and losers.⁶ While everyone's first take may be more money for better results, with budget neutrality that also means less money for those on the wrong side of the bell curve.

For example, in the Premier trial, hospital reimbursements could vary from plus 2.0% to minus 2.0%. If Medicare patients were a third of hospital admissions, as is frequently the case, being on the wrong side of that swing could be disastrous, especially in an industry where 4% margins are the stuff of dreams.

Further, will the very hospitals struggling to keep up with information system investments and human resource needs be the ones receiving less compensation? Very likely—thereby raising a multitude of questions about equity and access over the long term.

Provider Acceptance

The Leapfrog Group was the first out of the blocks in the pay-for-performance movement, so their standards would seem likely to be the most accepted. Perhaps, but a query into hospitals within 100 miles of downtown Washington, DC reveals only two institutions, Johns Hopkins in Baltimore and Christiana in Delaware, that had fully responded to their surveys.

Why so limited? In a recent analysis,⁷ Galvin, et al, identified a number of factors including the voluntary nature of hospital surveys and the unrealistically high expectations by Leapfrog's founders. Further, computerized physician order entry and intensivist staffing are expensive and, without tangible returns, hospital executives were reluctant to invest in these programs. And is it realistic to expect that low-volume surgical hospitals are going to rush to answer a survey that recommends diverting their patients to a higher-volume facility?

Impact on the Disadvantaged

Socioeconomic status has been shown to have a strong correlation with HEDIS scores.⁸ That is, the poor are much more likely to have lower baseline scores on measures such as breast, cervical, and colorectal cancer screening, hypertension control, and immunization rates. From so far back in the pack, how likely is it that these populations will meet the lofty targets of most pay-for-performance programs? No doubt, very improbable. And where is the fairness in financially punishing the physicians and hospitals who care for these patients?

The arguments for pay-for-performance are persuasive, and few believe that trials of the methodology are out of order. But many outputs of the health care industry are difficult to define, much less measure. The everyday bazaar of a hospital represents a delicate equilibrium between business and benevolence, empiricism and instinct, complexity and simplicity. The current metrics of pay-for-performance are, by any standard, rudimentary—so elementary as to raise doubts about their real impact or the long-term buy-in by physicians and hospitals.