Cover Story
A Measured Approach
By Kate Ledger
The goal is quality. The incentive, money. And the yardsticks payers use to measure physicians’ performance are many.
When the nation’s quality improvement organizations stressed the need, only a handful of years ago, for physicians to improve health care by focusing on better outcomes, no one quibbled with the goal. But it turned out to be health plans that quickly took charge of achieving it.
Using a basic business model aimed at improving care and reducing costs, health plans in pockets across the country began introducing financial incentives to the reimbursement formula in 2002, following the Institute of Medicine’s Crossing the Quality Chasm report that showed a gap between what is known to be good health care and the care that patients actually receive. In Minnesota, many of the major payers began rolling out programs that same year, and as of last year, nine different incentive programs were well underway.
The swift and resolute appearance of pay-for-performance programs may have taken physicians by surprise, but now both private and public payers are offering some form of financial incentive for clinics to improve their treatment of patients with certain diseases and increase the frequency of specific disease screenings. In July 2007, Medicare announced it would offer doctors a
1.5 percent bonus for reporting quality measures established by the Physician Quality Reporting Initiative, a voluntary program that provides the public with information about physicians’ outcomes. “Pay for performance is already a real phenomenon that’s occurring everywhere now,” says Phillip Kibort, M.D., vice president of medical affairs and chief medical officer at Children’s Hospitals and Clinics of Minnesota.
Although physicians voice support for the intent to improve care, they’ve been less enamored with the pay-for-performance programs themselves. Some find themselves navigating a morass of paperwork. Others encounter frustrations with the health plans’ disparate expectations and the reimbursement process. Some struggle with social complexities, expressing concern that pay-for-performance programs favor wealthier clinics or that incentives will encourage doctors to “cherry pick” healthier patients. Published studies remain divided about whether health outcomes are improving as a result of these initiatives. And how well pay for performance is working on a day-to-day basis seems to depend on whom you ask.
One thing is certain, however, says University of Minnesota family physician Brian Sick, M.D., medical director for chronic disease management in the department of internal medicine: “It’s a different world than it was three years ago.”
Complex, Cumbersome
In November of last year, the Minnesota Medical Association published a report evaluating how well various pay-for-performance programs operating in the state aligned with the organization’s pay-for-performance principles. After more than six months of meetings of the MMA quality committee, which included primary care physicians and specialists from practices around the state as well as health plan representatives, it concluded that “there is an opportunity to refine the current programs so that they better serve the community and more effectively link performance measurement with physician payment incentives.”
One of the main criticisms of the state’s pay-for-performance programs is the sheer variety of their measures. Across Minnesota, the report points out, nine programs offer physicians some kind of incentive for meeting quality standards. Some of those programs offer incentives to clinics for developing infrastructure such as implementing an electronic medical record; others focus only on outcomes. Among the nine programs are 10 methods for collecting the necessary data. Some of the programs, such as those created by Medica, Blue Cross and Blue Shield, and HealthPartners are voluntary. Publicly funded UCare enlists clinics automatically if they serve its enrollees. Depending on the program, clinics either self-report their outcomes, have auditors come in to review charts, have their numbers culled by the health plan through claims, or rely on a combination of these methods. Collectively, the nine programs address 63 disease states; but each program might focus on only a few of those conditions. Moreover, the programs together have 117 measures, the specifics of which do not always overlap.
What’s at stake can vary from a bonus of a few thousand dollars for meeting a particular benchmark such as improving the percentage of generic drugs prescribed, to hundreds of thousands for meeting a whole array of measures. In academic settings, the money may go to the academic department or clinical service unit; but in private practices, the money can turn into a bonus for individual physicians. As one physician noted, the programs can be “part carrot, part stick.” HealthPartners, for one, invites clinics at the beginning of the year to set their goals. If a clinic fails to meet its anticipated numbers, the health plan will withhold reimbursement.
In many cases, clinics are finding pay for performance has introduced an administrative headache. “Every [plan] seems to have a different [data-collection] tool,” notes Barbara Graham, R.N., manager of quality services for Olmsted Medical Group, an independent multispecialty practice in southeastern Minnesota, with nearly 100 physicians serving in 11 offices. “Some you need to do on paper, some you enter directly into their system. Some … send people to the facility, and they do the chart review. But we still need to pull the medical records and spend time with them, orienting them to our electronic [system],” she says. In addition to committing staff for chart reviews and data collection, the group has had to enlist its information technology staff to smooth the hook-up between the reviewers and the clinic’s electronic medical records. Medicare, for instance, uses a computer program that has a tool for clinics to enter their data, and that tool requires frequent upgrading. “It doesn’t take long,” Graham says, “but it’s another step.”
Finding the programs cumbersome, Olmsted decided to participate in only those that dovetail with its own internal quality review efforts and those that yield enough money to make the additional work worthwhile. Medicare’s most recent addition, the voluntary reporting system instituted last summer, does not tell physicians how much they will be reimbursed until after they’ve submitted results; moreover, it can require a major computer readjustment. “We did not participate in the Medicare PQRI program initiated this year because it was way too much effort with very minimal returns compared to the expenditures that we would have to go through to comply with all of the requirements,” acknowledges Roy Yawn, M.D., Olmsted’s president and CEO.
To keep up with the administrative demands of the various programs, some clinics have added staff or committed part-time nurses for data collection. Some clinics have even pulled physicians away from patient care. At the University of Minnesota’s Smiley’s Clinic in Minneapolis, family physician Peter Harper, M.D., now spends two half-days a week working on what he calls “quality infrastructure,” doing such things as coordinating data collection, building patient registries, helping structure clinic processes, and interacting with the health plans—sometimes arguing numbers when the clinic’s internal reviews and the health plans’ tabulations come up with differing results. “You need an M.D. involved,” he asserts.
Moreover, he points out that the data-collection process is seldom straightforward. “If [the program is] using claims, you’re not sure they’re always correct.” And there are other aggravations. For instance, while striving to improve the rates of child immunizations, the clinic was not allowed by one program to include patients who received shots at free sites. “If there’s no billing for it, we don’t get credit, even though [the children are] up to date,” Harper says, adding “data collection … feels out of control at times.”
For some clinics, the record combing becomes protracted when they have to ask for a re-review of records. In one instance, the insurance company’s data “did not jibe with our own numbers,” according to Burnell Mellema, M.D., care improvement medical director of Affiliated Community Medical Centers in Willmar, a multispecialty service with 97 physicians. When the charts of less than 1 percent of the clinic’s diabetes patients were audited, the clinic was found not to meet its goals for getting the disease under control. “That small sample size affected our numbers, and we knew it wasn’t right because it wasn’t close to our own audits,” Mellema says. The discrepancy was resolved in favor of the clinic.
But, as Olmsted’s Yawn points out, medical practices have found such disputes are not always just a matter of dollars. Now that outcomes and relative rankings of practices are being reported online by many of the health plans and by advocacy groups such as MN Community Measurement, a pay-for-performance review also can affect a group’s reputation. Although clinics are able to recoup their bonus when auditors come back to reinvestigate their records, the Olmsted group was disappointed that its revised standing in the report issued by MN Community Measurement came after the erroneous listing, in which it had scored low with one outcome, had been widely distributed to physicians.
University of Minnesota family physician David Satin, M.D., emphasizes that the scrutiny of the pay-for-performance era means clinics these days have to collect their own data, which will protect them when payers make mistakes. Many doctors agree that keeping a closer eye on populations of patients, and how well physicians are attending to the details of their disease, may be an effective way to make a dent in some of the nation’s most debilitating and widespread health problems. Such record-keeping may simply add up to “a new kind of professional responsibility,” Satin says.
Multiple Measures
What many would like to see aligned, beyond the methods of data collection, are the performance measures. The plans have varying standards when it comes to calculating “quality” health care. For instance, six of the nine programs operating in Minnesota offer incentives to primary care doctors to improve treatment of diabetes using a constellation of measures; patients must have a particular blood pressure, cholesterol, and blood sugar measure, take a daily aspirin, and cease smoking. “If you have all five [in a patient],” explains Satin, “it’s a grand slam,” and the practice gets credit for successfully managing a diabetes case. However, the standard for blood sugar may vary from program to program, with some requiring patients to have a hemoglobin A1c level no higher than 7 percent and others strictly stipulating 6.9 percent and below. The range of goals frustrates physicians because it does not reinforce the message that the standard is evidence-based.
Further, according to Sick, “some people don’t agree with what the goal of treatment should be.” A plan may require patients to have blood pressures of 130/80, with a single point over disqualifying them from counting as health care successes. “But if a patient was previously 160/100 and brings their blood pressure down to 131/80,” Sick says, “that’s a very dramatic difference that will save that patient years of illness.”
Some of the discrepancies between measures may change sooner rather than later. According to Julie Brunner, executive director of the Minnesota Council of Health Plans, since the MMA’s report in November, many plans have expressed interest in aligning their measures more closely with standards like those of MN Community Measurement, which were developed by the Institute for Clinical Systems Improvement.
Ethical Blind Spots
Satin, a post-doctoral fellow in the University’s Center for Bioethics who has studied pay-for-performance programs in Britain and New Zealand, points out that the problems with these programs are more than merely frustrating. Most programs treat clinics with a wealthy patient base the same as those that serve poor communities. “We know [from research] that the greatest determinant for health outcomes is not the doctor or the clinic,” Satin says. “It’s economic status [of the patient].” It may not be reasonable to expect a clinic with poor patients who struggle simply to make the trip to the doctor’s office to achieve the same health results as clinics where most patients have health insurance and are able to follow through with referrals. In Britain, he says, pay-for-performance programs adjust the health expectations based on the postal code of the clinic. Here, however, “we don’t risk adjust.”
The effect, he says, ultimately punishes poorer clinics, where patients may not only be more difficult to treat but also where a smaller fraction of diabetes patients might appear to have their disease under control.
Satin and others at Smiley’s Clinic, which serves an urban, poor, and largely immigrant population that doesn’t speak English, became concerned one month when a health plan pointed out that too many of its asthma patients had been seen in the ER. The doctors investigated and found two of the 12 patients noted by the health plan had never even been to the clinic. Another one had been seen but didn’t have asthma at the time of his visit. In the case of two others who’d been seen in the ER with asthma-like symptoms and received treatment there for asthma, subsequent spirometry tests showed they didn’t have the disease. “Poor patients tend to use the ER for primary care,” he points out. “Should our clinic be penalized?”
Satin calls the economics of such a system a “reverse Robin Hood,” where fee-for-service payments get redistributed as bonuses to those treating healthier patient populations. Without some kind of adjustment of goals, he warns, “clinics serving the poor get poorer and those serving the rich get richer.”
Such tensions also weigh on the esprit de corps of a practice, notes Mellema. “One negative aspect,” he says, “is the all-or-nothing approach the plans take.” Since some incentive programs reward the top 15th percentile of clinics, a clinic that strives hard and invests money in improvement but ends up in the 84th percentile will come up short. His group, whose rural clinics serve many elderly and poor patients, sought to improve generic drug prescription by 4 percent. When their numbers in 2006 showed they’d made strides but only reached 3.9 percent, they lost a potential bonus worth as much as $10,000. Add that to other measures where they came close but didn’t quite make their goals, and they were looking at nearly $100,000 that they could have earned. “It adds up,” he says.
Financial hits potentially threaten the doctor/patient relationship, Mellema notes. “Physicians do express concerns and apprehension about it,” he says. “Every doctor has worked with patients on tobacco cessation [in cases where the patient] is not willing to make changes. Now, if there’s a financial penalty, the doctor might say, ‘Our lack of progress is hurting you and hurting me—maybe it’s in both of our best interests if you see someone else,’” Mellema says. “That’s an unhealthy ... unintended [potential] consequence of excessive pay for performance.”
Ticket to Change
Despite the various pitfalls of pay for performance, many physicians acknowledge they’ve seen positive changes since the programs have been in effect. At Smiley’s Clinic, pay for performance has prompted the creation of a diabetes registry to keep tabs on patients and made inspecting immunization records during well-child visits an entrenched practice. As a system, Harper adds, “we have moved toward being able to generate our own data.”
Satin acknowledges having a checklist has provided new structure for interacting with diabetes patients. “I didn’t know before I wasn’t paying close attention to cholesterol. I was good at [attending to] blood pressure and … blood sugar.” Now he makes sure to hit all five points during a visit, even when the patient comes in for a reason other than a diabetes-related complaint. “It’s helped me focus my visit from the most important thing to the least important.”
Sick’s group has made what he considers to be “dramatic improvements” in bringing patients in to monitor their diabetes and in coordinating diabetes care.
Fairview Health Services medical director and family physician Barry Bershow, M.D., credits pay for performance for several dramatic improvements throughout the entire system. In two years, he says, Chlamydia screening in Fairview’s OB clinics has jumped from 29 percent to nearly 90 percent. The system of more than 50 primary care clinics also went from having approximately 4 percent of patients with diabetes having their disease under control to 22 percent two years later, and now several Fairview sites are ranked among the best places in the state to be treated for diabetes. He credits the turnaround to Fairview implementing its own internal pay-for-performance program, which he believes is far more effective than those offered by any of the health plans. Two years ago, a group of physicians approached the board and asked for seed money to spur a “culture change,” steering physicians’ attention toward improving outcomes. The board agreed, putting $3 million of bonus money at stake if all its physicians met quality goals. With the program in place, Fairview’s primary care doctors, who stand to gain about $2,500 each from meeting external goals also may earn as much as $15,000 a year for meeting internal quality goals that are aligned with the strictest health plan standards.
Even with such a substantial incentive, Bershow notes, not all physicians bought in initially. But the money immediately changed the tenor of meetings, turning the focus from budgets and breaking even to plans for obtaining quality outcomes. “What the literature shows nationally,” he states, “is that pay for performance doesn’t work, but that’s because they never paid enough critical-mass dollars to make it work.” Further, because Fairview began tracking data so closely, “we could get back to our physicians on 100 percent of their patients every single month on how they were doing.”
Kibort of Children’s believes paying for performance is just the ticket to spur change in a hospital system, “but only if the incentive is available on a hospital and physician level.” Doctors who work in the hospital don’t have a stake in making sure every single discharged asthma patient has not only a prescription for the appropriate medication for controlling their asthma but also information about how to use it—unless they’re targeted individually with rewards or penalties, he says. He favors programs that establish stretch goals, even if failure would mean withholding salaries. “It’s a good motivator,” he says. “Pay for performance will be one of the key modalities in how to change hospitals’ and physicians’ behaviors. Physicians are always cynical at the beginning,” when a new program is introduced, he says. “They always say it’s not a perfect system. But over time people adapt.”
What’s clear for medicine is that a culture change is taking hold. Some physicians are hopeful that the incentive system will eventually steer toward rewarding the coordination of care for chronically ill patients like diabetics, rather than asking physicians to aim for a bottom-line number of symptoms. Others hope for more substantial rewards for clinics that improve infrastructures such as patient registries and information systems. For most physicians, the long-term goal of improving quality remains paramount, and “the headaches,” as Sick says, “are outweighed by [a desire to do] what’s better for the patient.” Exactly how the rough spots will be refined remains to be seen. As Olmsted’s Yawn says, “We’re in the midst of a big experiment.” MM
Kate Ledger is a freelance writer in St. Louis Park.