For years we have focused on morbidity and mortality data of all kinds, usually derived from hospital settings and data. While useful, this required significant acuity adjustments to be comparable amongst providers, and most individual physicians do not have enough patients in a given category during a single year to derive statistically significant results. Medicare and other payers have published hospital and physician-specific mortality data in the past, but in almost all cases it has not been useful in creating change. We have also collected vast amounts of Health Plan Employer Data Information Set (HEDIS) data over the years, but again, this has not been truly useful in creating clinical improvements. Given the many criticisms of the quality of American health care by the Institute of Medicine and others, the rapidly rising costs of our system, and the new health care consumerism, innumerable new proposals for change are now being discussed.

The California P4P Initiatives promoted by the Pacific Business Group on Health (PBGH) and the Integrated Healthcare Association (IHA) have begun a significant transformation in the way patients with HMO coverage receive care. In the diverse areas of childhood immunizations, diabetes care, cardiac disease, asthma, breast cancer, cervical cancer, and recently Chlamydia infections, significant clinical improvement can now be measured. This has required the development of patient "disease registries," aggregation of laboratory results and filled prescriptions, physician support reporting, and innovative strategies to improve patient compliance. The experience within the Mills-Peninsula Medical Group (MPMG) demonstrates that 97 percent of pediatric patients receive recommended MMR immunizations, 96 percent of asthma patients ages 5 to 9 receive appropriate medications, and 95 percent of patients with diabetes are getting regular cholesterol measurements. The care provided by the MPMG community of physicians is excellent on both absolute and relative performance scales. And there is more improvement ahead.

This effort comes at a cost, however. There are significant expenses related to data collection and reporting as well as the communication of the information to physicians and patients. For the California HMO P4P program, groups may participate by relying either on aggregated health plan reported performance data or producing their own data. Because the health plan data is invariably less complete, plan reported performance is uniformly poorer than actual, often by 10 percent or more. To avoid this "under-performance," a group must develop substantial data sophistication—hardware, software, and personnel—and undergo an annual external audit to verify compliant methodology in reporting. And then there are the costs associated with taking action to improve clinical performance.

To the credit of the employer and health plan communities, they have put their money where their mouth is and funded P4P with a little more than $50 million in new payments to P4P participating groups. Although a good start, given that there are in excess of 17 million Californians covered by HMO health insurance (10 million excluding Kaiser Permanente), this is still inadequate funding by the standards of any other industry. And, as health insurance premium increases are slowing, many plans are now reducing their financial commitment to P4P.

As with all things health care, give enough "experts" enough time, and they can seriously impair most good intentions. The California P4P program is no different. Each health plan wants to use its money in a manner that distinguishes itself from the other plans. Although this may be good marketing, it dilutes efforts and causes confusion. Until this latest P4P year (2005 results reported in 2006), Pacificare did not participate in the statewide effort and now HealthNet has recently announced that this year’s award payments will be based on year-over-year improvement, not absolute results. IHA has contributed to the confusion as well with changes in the technical specifications of the measures (most important, the definition of the subject population, e.g. the age range) from one year to another, making longitudinal comparisons impossible. Then there is the public reporting, also known as transparency, by both the regulatory Department of Managed Healthcare (DMHC) and the employer/PBGH sponsored consumer Web site, healthscope.org , each with a different rating system, to "help" consumers. Does this mean P4P is of limited use? Perhaps for patients; but for those of us focusing on the patient care and process improvement, it is still very important. It is the only way we can ensure the entire population of patients gets the care they deserve and expect.

We can and should learn from the experience of others as well. Many other countries are using P4P-type payment incentives in health care. One very interesting example recently reported in the New England Journal of Medicine (July 27, 2006) is from the British National Health Service (NHS). In an enormous effort to document and improve the clinical care in the general practitioner (PCP) offices, 146 quality indicators were identified, each with associated points contributing to a payment award. The funds for this were largely derived from new or incremental funding to the NHS, so this was not generally viewed as punitive. Because the potential payments were substantial, the physicians responded aggressively, hired additional staff, added computers to track patients, and created new patient notification systems. In the end, the first year of the program saw the PCP incomes increase by an average of almost $40,000, targets for the P4P metrics were met for 83 percent of patients, and

97 percent of the available funds were awarded. This was a huge success for all, even if the NHS budget was substantially exceeded. The system can be changed if there is commitment and funding.

Now the Medicare program is planning to implement its own P4P system. It faces enormous challenges, including how to find appropriate performance measures for all specialties. Some 100 physician measurements are under consideration with the current concept of "paying" physicians to self-report using a new coding system and data submission with Medicare claims. While some Medicare hospital-based P4P payments are already implemented, the physician system is far more complex. In the setting of inadequate funding of the current fee-for-service system—after several years of trivial or no physician fee increases and the annual threat of fee decreases—there is no new money. Current proposals call for as little as 1 percent payment increases for physicians complying with a Medicare P4P program. This is how other efforts have failed.

Although P4P can be a useful process for health care improvement, I would add a few points of caution. In the absence of an electronic medical record (EMR), all of our efforts will be dependent on administrative data (primarily claims information) or extremely intensive chart reviews. Those physicians using an EMR will be significantly advantaged. Most current P4P measures are patient-population-based and work well where physician responsibility can easily be identified as in most HMO plans that assign patients to a group or PCP. For PPO or Medicare patients, typically no such assignment is made and physicians cannot identify "their" patients. Again, large multispecialty groups will be advantaged as likely to identify patients by virtue of multiple points of contact. Most P4P programs, including stated intent by Medicare, are designed to produce physician-specific results. Given the way medicine is practiced today, with large on-call groups and many part-time practices, it is often very difficult to accurately identify physician-specific performance. As a result of this, MPMG does much of its performance measurement at the office or practice level.

Most important, we must avoid de-professionalizing our work. These measurements are valuable, but the health care we provide is much more than the percentage of patients receiving beta-blockers. We must avoid "playing to the test" and losing sight of our goals for better health care. This must not become just a competition.

Performance measurement and payment strategies are here and will grow. With some care, these can be constructive for better patient care. It is imperative that physicians remain active participants in guiding this effort.