Why physician incentives don't work
So you've reviewed your financial performance and are appalled by the results. Faced with business trouble, more physicians, and even payers, are turning to incentives or pay-for-performance plans.
Not meeting productivity benchmarks? Just rearrange the compensation plan and pay more to physicians who meet or exceed the goal. Staff reluctant to collect copays? Give them 25 cents for each one they collect.
Payers are giving pay-for-performance plans an especially warm embrace recently. While tracking mechanisms don't exist yet, there certainly seems to be a growing payer movement to give more to physicians who meet quality goals set by the payer. Atlantic Information System, a major healthcare publisher, just announced the launch of a new newsletter devoted to the topic - "Case Studies in Health Plan Pay-for-Performance Programs" - and counts 78 such programs running nationwide.
It makes sense that pay-for-performance programs are popular. Money motivates. And it seems fair to pay more to the people who work the hardest. Isn't that the American way? If you are a physician who provides better quality care, shouldn't you get more than your lackluster competitor who wouldn't know an evidence-based guideline from an old wives' tale? It's simply intuitive that incentive programs make sense.
But that doesn't mean they actually work.
Even as pay-for-performance plans grow in popularity, critics are pointing out their limitations. The debate is raging. When Managed Care Information Center surveyed healthcare leaders asking whether physician incentive plans are working for managed care plans, 46 percent said yes. But 51 percent said no.
Marc Halley, President and CEO of Ambulatory Management Services in Westerville, Ohio, has overseen change initiatives at many practices. He still says pay-for-performance is an idea with good promise, but he is very aware of its problems.
"There has been some tremendous work done by hospitals and health systems as well as by some individual physicians and medical groups, but by and large, I think the challenges ... still remain," he says.
One major challenge: how to define the goal, especially when the aim is better medical quality. What makes one sort of treatment better than the other? What makes one physician better than another?
On some level it makes sense to measure outcomes. How many patients did well in the hands of Doctor A versus Doctor B? But what if Dr. A sees more poor, minority, or elderly patients whose health status is less stable to begin with? His results will be worse than those of Dr. B -- who only sees well-insured, well-fed, young people -- no matter how good his actual performance. Surely no one would suggest docking physicians for agreeing to treat non-normative patients. But might that be the result?
Debate can also erupt over whose desired outcomes to measure. Imagine a patient about to undergo back surgery. The payer's desired outcome may be to get that patient out of the hospital as fast as possible. The physician sees a perfect repair as the desired outcome. And the patient just wants to be able to get back out on the golf course. Whose measure do you use?
And what about the vast aspects of medicine that are nonprocedural? "The outcomes of cognitive service are extremely difficult [to assess]," says Halley. "As many of our providers tell us, 'Many of the patients [who] come to see us would have gotten better anyway. We're treating symptoms'. ... Those things are really difficult to measure from an outcomes perspective."
Others base pay-for-performance on following clinical guidelines, but that approach might encourage physicians to offer the same treatments over and over instead of pursuing possibly better options. What if all physicians were rewarded for treating every patient with the flu with a bloodletting? Clinical guidelines only change as doctors try new things and find that they work better. Indeed, some physicians maintain that clinical guidelines provide no more than a historical record of what the best practices were a few years ago.
Moreover, it can be hard to agree on what the guidelines should be, and doctors find it annoying when they must demonstrate the good clinical reasons why they sometimes aren't followed, or to have to prove to a payer that they actually were followed. Physicians are notoriously bad at documenting, let alone filing claims for the sorts of service payers measure.
Even setting standards for things like physician or staff productivity can be surprisingly difficult. For example, many practices rely on relative value units (RVUs) to measure physician productivity, but does everyone in your practice agree that this is a suitable gauge? Does anyone in your practice even understand where RVUs come from? Not to mention that pushing productivity in lieu of other measures can easily lead to a practice style that means pushing patients through as fast as possible. Is that really what you want?
Alfie Kohn, who has written numerous books and articles exploring the impact of incentives in the workplace and schools -- where compliant students get gold stars -- suggests that establishing a consensus on measurements can be practically futile:
"[A]s many thoughtful observers ... have realized, it's simply impossible to quantify quality and arrive at a consensus on what constitutes a fair way to reward for it. This was made abundantly clear to me a couple of years ago when I began to research the absurd RVUs that are used to determine incentive pay for radiologists. To privilege one factor is to neglect another. Pretty soon all incentives become perverse in one way or the other."
But Kohn's more scathing criticism of incentive programs is that they don't work -- not for the long haul. Once the reward stops, so does the improvement. No one has learned anything about why they should change; they work just for the reward, creating an incentive economy.
"[I]t rarely dawns on us that while people may seem to respond to the goodies we offer, the very need to keep offering those treats to elicit the same behavior may offer a clue about their long-term effects (or lack of them)," Kohn writes in Punished by Rewards.
Indeed, because pay-for-performance plans set up an unequal relationship between the reward-er and the rewarded, they can actually hurt performance. The person getting the reward suddenly feels judged, and even if she would choose to, say, do foot checks on all her diabetic patients anyway, the fact that some nosy insurance company will be checking up how often she does so makes her suddenly resent practicing good medicine.
"The very idea of pay-for-performance is not merely ineffective but ultimately counterproductive: the effects it has on relationships with colleagues and with patients; the extent to which it feels manipulative and insulting; the way its simplistic focus on behavior fails to get at -- and actively distracts us from -- the deeper explanations for problems with medical care," Kohn explains in an interview.
"There's a huge difference between intrinsic motivation (the desire to do right by one's patients and practice the best possible kind of medicine) and extrinsic motivation (the desire to puff up one's paycheck)," he adds.
"In fact, scores of studies have demonstrated that these two are often inversely related. The more we reward people for doing something, the more they tend to lose interest in whatever they had to do to get the reward. Maybe that's why, for more than a decade now, I've had no takers in the healthcare field -- or anywhere else in the business world -- when I've challenged defenders of incentives to name a single controlled study that has ever demonstrated a long-term increase in the quality of performance as a result of any reward system."
Getting to the core
So what does improve performance and make one physician -- or staff member -- better than another? Physician Atul Gawande decided to find out when he wrote about cystic fibrosis and performance ratings in The New Yorker. The Cystic Fibrosis Foundation has collected data from CF treatment centers for 40 years. Some participants recently agreed to reveal their names, making it obvious which were having better patient outcomes.
It is surprising that there are differences. As Gawande writes in the December 6, 2004, issue, "The 117 CF centers across the country are all ultra-specialized, undergo a rigorous certification process, and have lots of experience in caring for people with CF. They all follow the same detailed guidelines for CF treatment." That's a lot more consistency than you'd find in other areas of medicine, but the same treatments provided by very similar centers still had different outcomes.
When Gawande visited a center with less-than-stellar results, he couldn't immediately see any problem. All the protocols were followed, staff were dedicated, and they even sent patients a questionnaire before each visit so staff and physicians could prepare.
But when he visited a second, high-performing center, he could immediately see the difference -- one creative physician, Warren Warwick. "Warwick's combination of focus, aggressiveness, and inventiveness is what makes him extraordinary. ... He believed that excellence came from seeing, on a daily basis, the difference between being 99.5 percent successful and being 99.95 percent successful," Gawande explains.
His treatment was basically the same treatment as everywhere else, but he is profoundly motivated to stop any progression of the disease dead in its tracks and forces his patients to have the same level of commitment.
A $500 bonus -- even a $500,000 bonus -- can't make a Warren Warwick. Passion for excellence can't be bought. It is intrinsic. And very hard to influence - to the bane of managers everywhere.
Pamela Moore, PhD, CPC, last wrote about successful coding in the March issue of Physicians Practice. She can be reached at firstname.lastname@example.org.
This article originally appeared in the April 2005 issue of Physicians Practice.