Doctored (16 page)

I spoke with a senior health care quality consultant about this problem at a hospital quality meeting. “We're in a difficult situation,” he said. “We're introducing these things without thinking, without looking at the consequences. Doctors who wrote care guidelines never expected them to become performance measures.” In other words, he explained, recommended care in certain situations had become mandated care in all situations.

The guidelines could have a chilling effect, he said. “What about hospitals that stray from the guidelines in an effort to do even better? Should they be punished for trying to innovate? Will they have to take a hit financially until performance measures catch up with current research?” Moreover, how do you correct for patient mix? Hospitals with larger catchment areas will have longer delays in treating heart attacks because of increased patient transit times. How do you adjust for these demographic factors?

To better understand the potential problems, one simply has to look at another quality-improvement program: surgical report cards. In the early 1990s, New York and Pennsylvania started publishing mortality statistics on hospitals and surgeons performing coronary bypass surgery. The purpose of these report cards was to improve the quality of cardiac surgery by pointing out deficiencies in hospitals and surgeons. The idea was that surgeons who did not measure up would be forced to improve.

Of course, surgical deaths are affected by myriad factors, so models were created to predict surgical risks and avoid penalizing surgeons who took on the most difficult cases. For example, a fifty-year-old man, otherwise healthy, who underwent coronary bypass surgery was judged to have a risk of death of about 1 percent. For a seventy-year-old on intravenous nitroglycerin with a history of heart surgery, congestive heart failure, emphysema, and other medical problems, the risk was about 20 percent or higher. Many surgeons, however, thought that the models underestimated surgical risk, particularly for the sickest patients. They criticized the models for oversimplifying heart surgery. Surgery, they argued, is a team sport, involving referring physicians, technicians, nurses, anesthesiologists, and surgeons. Many variables can affect patient outcomes that are beyond a surgeon's control. Among other things, they said, the models did not account for simple bad luck.

Surgeons began to fastidiously report—some would say overreport—medical conditions that could affect the outcome of surgery. In some New York hospitals the prevalence in surgical patients of emphysema, a condition known to increase surgical risk, increased from a few percent to more than 50 percent after report cards came into use. Some surgeons even made a habit of routinely putting patients on intravenous nitroglycerin because being on the drug conferred added risk, thus covering for a possibly poor outcome. Others tried to “hide” surgical deaths by transferring patients to hospice programs right before they died.

“It's all about the numbers,” a surgeon at NYU, where I first learned about report cards during my fellowship, told me. “We have to start coding for everything, and you guys have to help us out.” There would be no high-risk surgeries, he added, unless the risk was documented in detail. “If I don't operate again until next year, that's okay with me.”

Despite these excesses, in the beginning there were high hopes for this quality-improvement program. In the first few years there were major gains in surgical outcomes. The most striking results were in New York State, where mortality rates for coronary bypass surgery declined a whopping 41 percent, and outcomes improved for all hospitals at all levels. In a 1994 article in
Annals of Internal Medicine,
the cardiologists Eric Topol and Robert Califf wrote that “appropriate implementation of score cards could ultimately lead to a substantial improvement in the quality of U.S. cardiovascular medicine.”

But not everyone believed that report cards were causing real improvements in care. Some entertained a more disturbing possibility. Were surgeons' numbers improving because of better performance or because sicker patients were not getting the operations they needed?

In 2003, researchers at Northwestern and Stanford tried to answer this question. Using Medicare data, they studied all elderly patients in the United States who had had heart attacks or coronary bypass surgery in 1987 (before report cards were used) and 1994 (after they had taken effect). They compared New York and Pennsylvania, states with mandatory surgical report cards, with the rest of the country. They discovered a significant amount of “cherry-picking” in the states with mandatory report cards, and learned that patients generally were worse-off for it. They wrote: “Mandatory reporting mechanisms inevitably give providers the incentive to decline to treat more difficult and complicated patients,” adding that “observed mortality declined as a result of a shift in incidence of surgeries toward healthier patients, not because report cards improved the outcomes of care for individuals with heart disease.” Doctors agreed with these conclusions. In a survey in New York State, 63 percent of cardiac surgeons acknowledged that because of report cards, they were accepting only relatively healthy patients for heart bypass surgery. And 59 percent of cardiologists said it had become harder to find a surgeon to operate on their most severely ill patients.

(Of course, it isn't only heart surgeons who are feeling this pressure. Similar pressures are being brought to bear in other areas of medicine. For example, there is evidence that report cards on interventional cardiologists have resulted in a drop in the number of angioplasty procedures performed on very sick patients. “I said, ‘Just treat her with medicine,'” an interventional cardiologist at NYU told me about a critically ill patient in shock on whom he had refused to do angioplasty. “I didn't tell them it was because I didn't want a death on the table.”)

Whenever you try to dictate professional behavior, there are bound to be unintended consequences. With surgical report cards, surgeons' numbers improved not only because of better performance but also because dying patients were not getting the operations they needed. Pay for performance is likely to have similar repercussions.

For example, doctors today are being encouraged to voluntarily report to Medicare on sixteen quality indicators, including prescribing aspirin and beta-blocker drugs to patients who have suffered heart attacks and strict cholesterol and blood pressure control for diabetics. Those who perform well receive cash bonuses.

But what to do about complex patients with multiple medical problems? Half of Medicare beneficiaries over sixty-five have at least three chronic conditions. Twenty-one percent have five or more. P4P quality measures are focused on acute illness. It isn't at all clear they should be applied to elderly patients with multiple disorders who may have trouble keeping track of their medications. With P4P doling out bonuses, many doctors worry that they will feel pressured to prescribe “mandated” drugs, even to elderly patients who may not benefit, and to cherry-pick patients who can comply with the measures.

Moreover, which doctor should be held responsible for meeting the quality guidelines? Medicare patients see on average two primary care physicians in any given year and five specialists working in four different practices. Care is widely dispersed, so it is difficult to assign responsibility to one doctor. If a doctor assumes responsibility for only a minority of her patients, then there is little financial incentive to participate in P4P. If she assumes too much responsibility, she may be unfairly blamed for any lapses in quality.

Nor is it even clear that pay for performance actually results in better care, because it may end up benefiting mainly those physicians who already meet the guidelines. A few years ago, researchers at Harvard conducted a study on the impact of P4P at one of the nation's largest health plans, PacifiCare Health Systems. In 2003, PacifiCare began paying bonuses to medical groups in the Pacific Northwest if they met or exceeded ten quality targets. The researchers compared the performance of these groups with a control group on three measures of clinical quality: cervical cancer screening, mammography, and diabetes testing. For all three measures, physician groups with better performance at baseline improved the least but got the most payments (per enrollee, the maximum annual bonus was about $27). If they could collect bonuses by maintaining the status quo, what was the incentive for these doctors to improve? Another study also showed no difference in thirty-day mortality for patients hospitalized with one of four conditions—heart failure, myocardial infarction, coronary bypass surgery, or pneumonia—at 252 hospitals that participated in P4P as compared with more than 3,000 control hospitals that did not.

Several simple reforms could improve P4P. Insurers could use less stringent requirements for antibiotic delivery. For example, to minimize antibiotic misuse they could set the clock running after a diagnosis of pneumonia is made, instead of when a patient is first brought into the ER. They could also use percentile-based rankings or rolling averages over extended time periods so hospitals don't feel pressured to be at 100 percent compliance all the time. The irony is that lowering the benchmark—allowing for a bit of wiggle room—is more likely to result in proper care.

P4P not surprisingly is deeply unpopular among most American physicians. It forces them to follow certain clinical priorities—“cookbook medicine,” “a rule book”—leaving them deeply dissatisfied with the loss of autonomy. Many doctors say they feel like pawns in a game being played by regulators. Instead of being allowed to exercise their professional judgments and deliver “patient-centered” care, physicians believe they are being guided on what to do with a burgeoning menu of incentive payments or strict regulations. One recently wrote online:

We as a profession are partly to blame. We allowed the insurance companies to become the intermediary between us and our patients. There was more money at first but now we suffer the “controls” they have put in place. With regulations reaching the point of insanity, the Department of Health says jump and we don't have the backbone to question or fight back. If we don't begin to take back the controls of our profession, we will become mere technicians working on the government dole.

Another wrote:

What makes this particularly difficult is that [P4P] was not imposed on us against our will. We, through our professional societies, have adopted it voluntarily. If we could simply band together and fight the external enemy who did this to us, I would have high hopes. Since we did it to ourselves, the solution will be orders of magnitude more difficult.

Doctors have seldom been rewarded for excellence, at least not in any tangible way. In medical school there are tests, board exams, and lab practicals, but once you go into clinical practice, these traditional measures fall away. Whether pay for performance can remedy this problem is still unclear. But from what I learned in my first year as an attending, it has the potential to compromise patient care in unexpected ways.

*   *   *

What then is the solution to health care overuse? One option is to hire doctors as employees and put them on a salary, as they do at the Mayo and Cleveland clinics, taking away the financial incentive to overtest. Chronically ill patients in the last two years of life cost Medicare tens of thousands of dollars less when treated at the Cleveland Clinic, where teams of doctors follow established best-practice models, than at many other medical centers. Nevertheless, many self-employed doctors recoil at the idea of institutional employment and intrusion on their decision-making authority. Another option is to use bundled payments. A major driver of overutilization is that doctors are paid piecework. There is less of an incentive to increase volume if payments are packaged (e.g., for an entire hospitalization) rather than discrete for every service. Yet another possibility is “accountable care organizations” advanced by Obamacare, in which teams of doctors would be responsible (and paid accordingly) for their patients' clinical outcomes. Of course, such a scheme would force doctors to work together and to coordinate care. Unfortunately, most doctors, notoriously independent and already smothered in paperwork, have generally performed poorly in this regard.

However, if we want to maintain the current fee-for-service system, reforms will have to focus less on payment models and more on education. Medical specialty societies recently have released lists of tests and procedures that are not beneficial to patients, including MRIs for most lower-back pain and nuclear stress tests when there are no signs of heart disease. These “appropriate-use criteria” (bolstered by “comparative effectiveness research”) are essential for educating physicians and patients alike about medical services that are wasteful and should be avoided. (By employing these criteria, cardiologists have been able—or forced—to decrease their use of imaging by 20 percent.) In fact, better-informed patients might be the most potent restraint on overutilization. A large percentage of health care costs is a consequence of induced demand—that is, physicians persuading patients to consume services they would not have chosen had they been better educated. If patients were more involved in medical decision-making (admittedly not easy to put into practice in the hyperspeed that is contemporary American medicine, and also obviously at odds with regulation-driven P4P care), there would be more constraints on doctors' behavior, thus decreasing the possibility of unnecessary testing. Shared decision-making would be more likely to get patients the treatments they want, consistent with their values. This could serve as a potent check on what the doctor ordered.

Of course, health information is imperfect, and patients, ill and under duress, are often poorly equipped to understand it. (And busy doctors—our name ironically derives from the Latin word for teacher—clearly do a haphazard job of advising and instructing patients.) But even when good information is available, patients too often are passive consumers, still operating on the model of “Doctor knows best” (which many doctors admittedly encourage). For example, studies have shown that patients take little interest in the informed consent process. In a study of cataract surgery, only 4 percent of patients recalled more than two out of five risks disclosed to them by their doctors. Only a third remembered later that blindness was a potential risk.