Although pay for performance (P4P) facesnumerous challenges, it may be too soon torelegate it to the list of failed healthcare reformstrategies. In practice, many payers understandthe limitations of performance incentives andhave made efforts to avoid the worst pitfallsthrough targeting provider organizationsrather than individual physicians and directlyaddressing the needs of potentially vulnerablepatient populations. Thoughtful design andimplementation of P4P may therefore yieldimproved value with minimal negative consequences.Because of the critical role paymentincentives play in guiding practice patterns,it is hard to imagine substantial reform of thedelivery system without coordinated changesin payment.
(Am J Manag Care. 2007;13:238-239)
Although unfettered enthusiasm for pay for performance (P4P) is unfounded,quality-based payment should not be dismissed.
Although it has become decidedly mainstream, pay for performance(P4P) still evokes strong objections from some quarters.McMahon et al1 muster the well-known points of contentionagainst what they view as the current slide towards physicianlevelapplication of P4P. Both incentive theory and empirical evidenceshould indeed give pause to any payer or policy maker hoping to improvethe effectiveness and efficiency of care delivery with P4P alone. There areall too many ways that P4P, like quality measurement and reporting, canencounter design and implementation problems and, even in the best ofcircumstances, yield unintended consequences for patients and payers.2
In light of what can go wrong with P4P, it is not without cause thatMcMahon et al raise the alarm against the call for "market-based" incentives.Nonetheless, it may be premature to pronounce, or even hope for,the death of P4P just yet. Without disagreeing with any of the basic argumentsmade by McMahon et al, it is possible to arrive at a somewhat differentconclusion than the authors do for several reasons. First, thepractice of P4P is not as simplistic as some of the statements made by itshigh-level proponents.3 In particular, payers have anticipated some of theadverse consequences of poorly designed incentive programs and havenot, by and large, implemented the "worst case scenario" incentiveschemes about which we worry most. Second, even with strong evidenceof the existence of negative consequences (such as patient dumping), it isnot necessarily correct to conclude that the overall effect of P4P would beto make the situation worse. Finally, it is compellingly clear that paymentmatters for the orientation and organization of the delivery system (witness,for example, the rise of specialty hospitals), so that a policy of delayingpayment reforms until more integrated and sophisticated practices emergeas the authors suggest might further slow needed progress in the capacityfor coordinated, prospectively oriented care.
P4P IN PRACTICE IN THE UNITED STATES
Although advocacy for P4P may sound frighteningly naïve and blind topotential unintended consequences, the implementation of these programsis often much more nuanced. For example, while the Centers forMedicare & Medicaid Services is lookingto profile and reward individualphysicians in Medicare, many commercialhealth maintenance organizations(57%) have implemented P4Ponly at the medical-group level.4 In addition, thedesigners of payment incentives in both the publicand private sector have recognized the need to targetimprovement in care for vulnerable populations. Inthis vein, Blue Cross Blue Shield of Massachusettshas incorporated cultural competence training as arequirement of its P4P program, while recentMassachusetts legislation requires that the Medicaidprogram incorporate measures to reduce racial andethnic disparities into a new P4P effort. This is not to suggestthat P4P programs are uniformly enlightened, but that thereare working examples that demonstrate the opportunity todo better, and perhaps even do good, with P4P.
NET EFFECTS OF PERFORMANCEPROFILING AND PAYMENT
Demonstrating that incentives can and do have perverseeffects has been an important part of the research and policydiscussion around quality measurement, reporting, and payment.But evidence of the existence of unintended consequencesalone, even sizeable ones, is not enough to prove thatP4P is a bad idea. Whether P4P is good policy depends on thetotal net benefit, a function of the costs, including negativeconsequences of the incentives, and the benefits. Moreover,although negative effects such as upcoding, avoidance ofhigh-risk patients, and excessive focus on measured processesand outcomes cannot be eliminated, they can be minimizedby thoughtful design. It is important to reinforce, however, thepoint made strongly by McMahon et al that the effects of P4Pmight be very different for different populations so that netbenefits need to be examined not just on average but forpotentially vulnerable groups.
THE CHICKEN AND EGG OF PAYMENTAND ORGANIZATION
As McMahon et al point out, perhaps the biggest dilemmawith P4P is an assumption that incentives will somehowimprove quality and efficiency of care, even if the essentialsystems for providing care are lacking. Nonetheless, urgingpolicy makers to focus on addressing the lack of coordinatedsystems of care before introducing financial incentives ignoresthe need to provide an economic rationale for the developmentof such systems. Without performance incentives ofsome kind (including prospective payment in this category)there is little or no economic motivation for creating populationhealth management tools, such as disease registries andcare teams. Because it would be reasonable to conclude thatthe fragmented delivery system we have is perfectly designedto take advantage of the largely volume-based reimbursementsystem, it might also be reasonable to suggest thatchanges in payment need to pave the way for a more highlyevolved delivery system.
Sorting out whether and how to P4P is a complicated matterabout which informed and reasonable people could easilydisagree. The lackluster results from some recent, high-profileinitiatives5 combined with troubling reports of unintended consequencesin similar programs have dampened some of what wasprobably excessive enthusiasm for the idea. But the need to promotea different approach to healthcare delivery, with a focuson improving outcomes and efficiency, remains critical.Financial incentives surely have a role to play; there is no betterevidence for their power than our own dysfunctional healthcaresystem. P4P is obviously not a complete solution, andadditional payment reforms supported by improvements ininfrastructure, training, and quality measurement will be needed.Progress is likely to occur if these changes are coordinatedwithin communities, designed based on evidence and with forethoughtto protect against potential adverse effects, and evaluatedsystematically to allow continuous improvement.
From the Harvard School of Public Health, Boston,Mass.
Meredith B. Rosenthal, PhD, Department ofHealth Policy and Management, Harvard School of Public Health, 677Huntington Ave, Rm 405, Boston, MA 02115. E-mail: email@example.com.
The author reports no relationship or financial interestwith any entity that would pose a conflict of interest with the subject matterdiscussed in this commentary.
Am J Manag Care.
1. McMahon LF, Hofer TP, Hayward RA. Physician-level P4P—DOA? Canquality-based payment be resuscitated? 2007;13:233-236.
N Engl J Med.
2. Casalino LP. The uninteneded consequences of measuring quality onthe quality of medical care. 1999;341:1147-1150.
Wall Street Journal Online.
3. McGinley L. Medicare plan would give bonuses for superior care.Available at: http://online.wsj.com/article/SB105397331989909600.html. Accessed April 24, 2007.
N Engl J Med.
4. Rosenthal MB, Landon BE, Normand SL, Frank RG, Epstein AM.Pay for performance in commercial HMOs. 2007;355:1895-1902.
N Engl J Med.
5. Lindenauer PK, Remus D, Roman S, et al. Public reporting and payfor performance in hospital quality improvement. 2007;356:486-496.