Since the plans paid only top performers, some groups received little or nothing. Of the 215 organizations (representing 45,000 doctors) that participated, only 74 scored significantly high on most of the clinical measures, which accounted for 50 percent of the points needed for rewards in 2003. (A consumer assessment survey counted for 40 percent of a group's potential reward and use of information technology, the final 10 percent.)

Inspired by the IHA and by a corporate-financed program known as Bridges to Excellence, the Centers for Medicare & Medicaid Services is also getting into pay for performance. In January, CMS plans to start a three-year pay-for-performance pilot in four states—Arkansas, California, Massachusetts, and Utah—that could well turn into a national program for all Medicare providers.

The initial impetus in pay for performance programs has come from HMOs, because these plans make sure that every enrollee has a primary care physician. It's more difficult to rate groups or physicians on their quality of care for PPO members, who can see any physician they like without a referral. But the Blues plan in Hawaii has done it, and Blue Cross of California also has a pilot program in its PPO network. You can expect more of this, since PPOs now cover 55 percent of workers with health insurance.

Why has pay for performance gathered momentum so quickly? Because everything else has failed. Employers are desperate to control healthcare costs, and they now believe that raising quality standards—and rewarding those who meet them—will result in cost savings. While nobody knows yet whether that's true, quality incentives will almost certainly reach your neck of the woods soon, if they haven't already.

How pay for performance measures doctors Pay-for-performance programs aren't just about clinical quality. Some incorporate utilization measures, such as the percentage of generic drugs prescribed or the number of medically unnecessary MRI or CT tests ordered. Many include patient satisfaction and service measures. And a number of programs reward doctors for acquiring various forms of information technology.

The major focus, however, is to determine how well doctors take care of their patients and to reward the ones who do the best. The most common measures are various criteria used by the NCQA to rank and accredit health plans. These include measures of preventive services such as cervical cancer screening, mammography, and immunizations, as well as measures of care for such chronic diseases as asthma, diabetes, and coronary artery disease. While a dozen or so measures are typical, a few plans use many more. Also, the use of "outcomes" measures—such as lab values for blood glucose and LDL cholesterol levels—is becoming widespread.

The choice of these measures is often determined by the accessibility of claims or lab data related to them. In some of the better-organized programs, however, physicians have input into the measures. For instance, in Minneapolis and Rochester, NY, guidelines have been created through community-wide collaborations of plans and physicians.

This process has corrected some obvious problems. For example, primary care doctors in the Rochester IPA to which Robert Ostrander belongs objected to the goal of having diabetics come in four times a year. "The doctors said the patients thought they were trying to rip them off," recalls internist Howard Beckman, medical director of the IPA. "And it turned out that the evidence of better outcomes was poor, so we took it out."

The sample-size issue is problematic If an HMO looks at how physicians in a 300-doctor group do on a measure that applies to many patients, it can probably obtain a sample that's large enough to yield statistically valid results. The same could be true for individual physicians in a plan like Blue Choice, which has 80 percent of the commercial HMO market in Rochester. Beckman says that some of his IPA's doctors have a couple of hundred Blue Choice patients who are eligible for mammograms. But when a plan that has only 10 or 15 percent of a market analyzes the performance of individual doctors, "they're not going to get the numbers we get," he says.

That hasn't stopped some plans from trying to measure a large portion of their physician panels. For instance, Blue Cross of California initially measured 12,000 PPO doctors who treated as few as 10 members who were eligible for a particular service. The East Coast division of Blues giant Anthem has set the cutoff at 25 total patients per doctor.

To put this in perspective, healthcare quality expert and internist Sheldon Greenfield says that you must look at a minimum of 20 to 30 diabetic patients to get valid data on diabetes measures. Greenfield, who's co-director of the Health Policy and Research Unit at the University of California at Irvine, adds that this 20-to-30 number works only under the right conditions. Those conditions including setting quality goals at a level that takes into account, among other things, noncompliance and severity of illness. Moreover, he says, the use of claims data for performance measurement "needs more field testing." He greatly prefers the use of chart reviews.

The other side: the profit potential The upside of pay for performance, don't forget, is extra money in your pocket—and that can be substantial. Primary care physicians in California's IHA program, for example, had the opportunity to make as much as $8,000 extra this year. The program used by Hawaii's Blues plan gives top performers the chance to earn up to $24,000 a year extra. This year, the average award was about $5,500.

The program drawn up by Geisinger Health Plan, based in Danville, PA, which kicks off in 2005, will give bonuses equivalent to an extra $3 per member per month to high-performing physicians, and $1 per member per month to physicians in the middle tier. This translates into an average raise of 25 percent for those in the top tier and 9 percent for medium performers. FP Anthony Aquilina, Geisinger's regional medical director, believes these incentives will grab doctors' attention if they treat enough Geisinger patients.

Other health plans are considering a shift in their strategy from paying cash bonuses to raising fees for top physicians. The reason: Self-insured employers don't like to pay bonuses for services they believe they've already paid for. To get around this, Blue Cross of California is talking about boosting fees 6 to 12 percent for high-performing physicians. Anthem's pilot program in New England is lifting fees only 2 to 6 percent for top physicians.

Physicians in the CMS pilot program have a shot at earning an extra $5,000 to $20,000 a year from Medicare, depending on how many seniors they care for, says the agency's William Rollow. But they'll have to invest in computerizing their practices. CMS will pay doctors a certain amount for acquiring various forms of IT, and then will give them more for achieving quality goals.

CMS will follow the Bridges model CMS' program is closely patterned after that of Bridges to Excellence. Launched by corporations including GE, Ford, Procter & Gamble, UPS, and Verizon, the Bridges program seeks to reduce medical errors and waste by encouraging doctors to invest in information technology and improve the quality of care.

Toward that end, the coalition is paying doctors who meet performance criteria in diabetes care, cardiac care, and physician office processes. For each diabetic patient employed by a Bridges member company, qualifying physicians get $80 per year; for each Bridges patient with cardiac disease, they get $160. And physicians who qualify for bonuses based on office re-engineering, including the use of IT, receive $50 per Bridges patient.

Information technology that meets the coalition's goal for incentives includes e-prescribing, online lab orders and results, electronic registries of chronic-disease patients, and a full electronic health records system. Francois deBrantes, chairman of Bridges to Excellence, says his group has even qualified doctors who are using index cards or an Excel spread sheet to track patients. But the program plans to raise the bar every year, he says. "You're going to have to demonstrate you're moving from a box of index cards to a fully functional electronic health record."

The program will also examine how doctors handle chronic diseases and complex patients, promote health, and prevent hospitalizations. And physicians will have to show that they provide an adequate level of patient education. "It's really about systems and processes, rather than 'did you do this for this patient or that for that patient,'" says Richard Sorian of the NCQA.

Pay for performance is a major trend that will probably affect every physician. How long it will last, however, is a matter for conjecture. Health plans are conflicted about whether they'll see a return on investment; and if they don't, pay for performance could vanish as quickly as it has arrived. But the involvement of Medicare might ensure a longer life for this approach—and a greater impact on doctors: "If the Medicare program starts getting involved in something, you can bet the system will respond fairly rapidly," says Sorian.

Reactions of organized medicine The American College of Physicians recently issued a position paper on pay for performance. It recommends that the programs use widely accepted, evidence-based measures that "provide valid and reliable comparative assessment across populations"; that the programs not impose extra administrative burdens on physicians or rate doctors on things they can't control; that incentives be positive, not punitive; and that pay for performance foster quality improvement, not just competition.

FP Michael O. Fleming, board chair of the American Academy of Family Physicians, says that not only must the data be valid, but so must the measures. "We look forward to true pay for performance. However, you have to be able to define quality, and on some things, we just don't have a good definition of what good quality of care is. So we want the payers to work with us to define what those quality measures are." He adds that he hopes the plans will help physicians acquire EHRs and develop an electronic infrastructure.

Pay for performance, California style Two months ago, six of California's largest HMOs—Aetna, Blue Cross of California, Blue Shield of California, Cigna, Health Net, and PacifiCare—paid groups and IPAs about $50 million for their 2003 performance on standardized quality measures. The Integrated Healthcare Association, a consortium of health plans and physician organizations, used data on these measures to rate the groups.

The groups' quality scores have been posted on the Web site of the California Office of the Patient Advocate ( www.opa.ca.gov/report_card) These scorecards include overall rankings in clinical quality and patient service, as well as drill-downs on how well each group did on each of the measures within those categories.

Groups have mixed reactions to the initial payouts. Greater Newport Physicians, a 450-doctor IPA based in Newport Beach, CA, was one of the better-performing groups on the IHA scorecard. Yet it received far less than it had anticipated from some plans because of their payment methods, says Diane Laird, chief executive officer of the IPA. One HMO, she says, gave most of its rewards to the top few groups.

If the plans pay only the top 10 to 20 percent of groups, Laird says, "it makes it difficult for groups to keep investing in improvement." FP Ronald P. Bangasser of the Beaver Medical Group in Redlands, CA, wants to see groups paid for improvement as well as for sheer performance. He says some plans are considering doing that, but others oppose it.

While the IHA program has been praised for using a single yardstick to measure physician groups, two of the six founding plans—Blue Cross of California and PacifiCare—still base payments on their own set of measures (including some of the IHA metrics) and publish their own scorecards.

This rankles internist Andrew P. Siskind, president of the Bristol Park Medical Group in Irvine, CA, who points out that groups may be scored and paid differently by different plans. Noting that Blue Cross rated Bristol Park as one of the top five groups in southern California, Siskind expresses satisfaction with the bonus that the group received from that plan. But he's surprised that the group rated low on PacifiCare's report card, and chagrined that it received so little money from PacifiCare.

FP Sam Ho, the HMO's corporate chief medical officer, responds that if Bristol Park didn't get as much as it expected, that was because it didn't perform well on PacifiCare's measures. (The group's score on the IHA report card was "average" for clinical quality and "good" for patient experience.)

Four of the six plans allowed groups to report data on the IHA measures, and one of the others, PacifiCare, is allowing them to do it this year. Groups that self-reported generally did better on the measures than those that allowed health plans to rely on their own data. Sharon Katz, vice president of medical management for the Mills Peninsula Medical Group, a 350-doctor IPA in the San Francisco Bay area, says the plans' and the groups' data differed by up to 20 percent. Since the self-reporting groups were audited, she adds, much of this difference is related to problems the HMOs have had in managing encounter data from the groups.

Group executives interviewed by Medical Economics said that most of the quality awards would go to the physicians in their groups. After deducting some of the money for IT investments, for instance, the Beaver group is dividing the rest of its award evenly between primary care physicians and specialists. Bangasser says primary care doctors can make up to $8,000 each.

Despite grumbles over the size and methodology of the quality payments, most participants praised the approach. "I'd much rather get paid for quality than for utilization," says Siskind. And Bangasser predicts most groups will do better next year. "Now that they know the money's there, they'll respond more quickly," he says.