Carrots for Doctors
By Bill Keller
The New York Times, January 27th, 2013
With its ambitious proposal to pay doctors in public hospitals based on the quality of their work — not the number of tests they order, pills they prescribe or procedures they perform — New York City has hopped aboard the biggest bandwagon in health care. Pay for performance, or P4P in the jargon, is embraced by right and left. It has long been the favorite egghead prescription for our absurdly overpriced, underperforming health care system. The logic seems unassailable: Reward quality, and you will get quality. Stop rewarding waste, and you will get less waste. QED! P4P!
If only it worked.
For if you spend a little time with the P4P skeptics — a data-bearing minority among physicians and health economists — you will come away full of doubts. In practice, pay for performance does little to improve outcomes or to control costs. But if you look hard enough at why this common-sense approach doesn’t deliver, you find some clues to what might.
The New York plan would give hospitals bonuses to distribute to their doctors based on such indicators as patient satisfaction, speeding the flow of cases through the system and administering specific therapies. It is an attempt to get ahead of the federal law we all know as Obamacare, which includes rewards for hospitals that peg pay to a list of quality metrics. I’ve said before that I consider Obamacare an important leap forward, mainly for extending the basic safety net to millions of Americans. And there are aspects of the law, notably the Medicare Independent Payment Advisory Board, that may help bring down costs, if Congress can resist the temptation to interfere. But the pay-for-performance provisions are a triumph of theory over experience.
The first problem with P4P is that it does not address the biggest problem. Americans spend more than twice as much per capita as other developed countries on health care — a crippling 18 percent of the country’s economic output, and growing. Before studying the statistics, I assumed the root of the problem was doctors who, paid piecemeal for the services they provide, load up patients with marginal tests and treatments. In fact, America’s health care system is not much different from other developed countries in the volume of service. Our doctors prescribe more or less the same number of pills and X-rays, perform similar numbers of blood tests and surgeries, as doctors in the best European countries. While there are undoubtedly savings to be had by cutting unnecessary services (shortening hospital stays, for example), the main problem is that our system charges far more for each service — each office visit, each hip replacement, each day in a hospital bed, each dose of antibiotic. “The facile explanation is that doctors do too much,” said Peter Bach, a doctor at Memorial Sloan-Kettering who studies quality of cancer care. “But if you compare us to other countries on volume, we’re not leading in any category. The flip side is, we pay double for a lot of stuff.” (Actually, we lead in tonsillectomies and knee replacements, but his point is generally right.)
Doctors cite a number of reasons our medical treatments cost more — the high price of malpractice insurance being a favorite, and genuine, culprit. But the main reason everything costs less in other countries is that other countries tend to have one big payer — usually the government — with the clout to bargain down prices. A single-payer system has, so far, proven politically unpalatable in this country. And even Medicare, which has the power of scale and uses it to drive down prices, wields its power sparingly, because doctors threaten to stop serving Medicare patients if the reimbursements fall too low. As hospitals merge into mightier megachains, they may be able to bargain down the payments to doctors and drug companies and device-makers, and create economies of scale by standardizing treatments. (The physician and New Yorker writer Atul Gawande proposed in a provocative August article that hospitals could drastically improve productivity by studying the example of restaurant chains like the Cheesecake Factory.) But that’s not what P4P is about.
Instead of leverage, P4P employs incentives. Reduce the length of stay for acute-care patients, cut the rate of readmission for pneumonia cases, make sure heart-attack victims get a talk about diet before they are discharged, and you stand to find a little windfall in your paycheck.
Critics, who have evidence from a host of pilot programs, say that the bonuses are typically too small to change behavior; New York’s would be a maximum of 2.5 percent of a doctor’s salary, and most P4P programs pay less than that. Often the performance indicators measure things that are not within the doctor’s control. Luis Marcos, a former president of the agency that runs New York’s public hospitals, plucked several examples from the list of 13 metrics proposed for New York. Reducing waits in the emergency room (No. 4) is a worthy goal, he said, but it depends on a whole cast of people, from the clerk who greets the patient to the orderly who makes the bed ready. Shortening hospital stays (No. 13) requires having someone in the community — family, a social worker — to receive the patient; that’s an especially hard requirement for hospitals that treat a lot of low-income or homeless people.
Another problem with P4P is that providers learn how to manipulate the results. “Once you define performance, people manage toward those metrics and neglect other things that don’t get counted,” said the Princeton health economist Uwe Reinhardt, who writes for The Times’s Economix blog. New York doctors have a favorite, possibly apocryphal, story of medical providers gaming the system: Miami hospitals that use patient feedback as a performance measure wait until spring to do their surveys, because that’s when the cranky, hard-to-please New Yorkers go home from their winter refuge.
Ashish Jha, a doctor at the Harvard School of Public Health, argues that P4P might get results if the incentive payments to hospitals were substantially bigger, the formulas were simplified and the performance indicators were kept to a few, clear measures that doctors and patients agree matter: mortality rates, infection rates, recurrence of heart attacks. “If hospitals achieve great outcomes, it matters little how they did so,” Jha wrote in the latest Journal of the American Medical Association. And death rates are harder to game than administrative procedures.
Jha also proposes a little experiment in behavioral psychology: give hospitals 100 percent of the bonus at the beginning of the year, but require them to send a refund at the end of the year to pay for any shortcomings. Givebacks really focus the mind.
Peter Provonost, a physician and professor at Johns Hopkins, says that rather than bribe doctors to adopt better practices, we should play to doctors’ professionalism. Pronovost is famous for a scheme that drastically reduced infections associated with catheters by disseminating to doctors and nurses a simple, five-point operating-room checklist of reminders: Wash your hands, wear a sterile mask, etc. No pay incentives were involved, just an appeal to professional pride — and, later, public reporting of the results. “I wouldn’t rule out economic incentives,” Pronovost told me. “But you can do a lot just working with professional norms.”
Pronovost has a more radical piece of common sense to offer. He proposes that the United States create an agency that would do for medicine what the Securities and Exchange Commission does for securities markets: compile and audit information about the performance of hospitals, and make it public. Rankings of hospitals now come either from news organizations or consumer groups working with untested and inconsistent data, or they are propaganda produced by medical providers themselves. “There are stronger regulations about what a company can say about toothpaste than what a hospital can claim about quality,” Pronovost said.
For image-conscious doctors and hospitals, pride and shame might be the most effective kind of carrots and sticks.