Getting What We Pay for in Pay-for-Performance Programs - Annals of Internal Medicine: Fresh Look Blog


Wednesday, January 17, 2018

Getting What We Pay for in Pay-for-Performance Programs

In a research article published in Annals of Internal Medicine, investigators studied performance in Medicare’s Value-Based Payment Modifier program (VM)—a prominent pay-for-performance (P4P) program that adjusts providers’ Part B payment rates on the basis of their performance on a set of quality and cost measures (1). The authors were interested in 2 questions: Is engagement in VM associated with changes in provider performance on quality or cost measures? Does the program potentially impact health disparities?

The concept of P4P is straightforward: financially reward providers for high-quality care and penalize them for low-quality care. However, those of us familiar with the payment approach know that implementation is far more complicated. The study’s headline demonstrated results that, like prior P4P efforts, the VM was not associated with improvements in the quality or cost-efficiency of care. Also troubling was the fact that the program could also negatively affect health disparities.

Although Medicare sunsetted the VM program in 2016, policymakers have used the program’s framework to implement a new P4P initiative in the Merit-Based Incentive Payment System (MIPS). A mandatory program that affects many physicians and Medicare beneficiaries around the country, MIPS extends approaches from the VM to pay and penalize providers on the basis of performance on quality and cost measures. 

What should we in the medical community make of the lackluster results from such P4P programs as the VM, particularly given the potential implications posed by MIPS?

Some have argued that P4P should be scrapped altogether as an approach to provider payment. For example, authors of an accompanying editorial didn’t mince their words, noting how the Annals study is damning to current policy and “should be the final nail in the coffin of the current generation of P4P” (2). They decry MIPS as currently designed and offer other value-based payment models, such as bundled payments and accountable care organizations, as better ways forward.

Others note that with improvements in policy design, P4P could potentially be salvaged. For example, existing P4P programs could benefit from 1) transparent, straightforward designs; 2) focus on a few actionable measures for which physicians can change behavior and practices, and 3) large enough incentives to motivate behavior change among physicians and care organizations. With these modifications, some believe programs like the VM and MIPS could be harnessed to improve quality and contain costs.

As a health system administrator and policy researcher, I believe both cases have merit and should be thoughtfully considered by policymakers. However, as a physician, I take issue with a more implicit, fundamental assumption highlighted by P4P: that physicians are primarily motivated by financial reasons.

I do not believe this is true. Physicians certainly consider financial reimbursement and their livelihoods in decision making and behavior. However, physicians are also motivated by many nonfinancial considerations, including the desire to do well at their work and serve as patient agents, guarantors of social good, supporters of professionalism and professional standards, and good citizens of the workplaces and organizations.

Not only does the current generation of P4P fail to preserve or emphasize such motivations, one might argue that by tying payment to a wide range of care delivery processes, existing P4P programs can unintentionally counteract these motivations. As an internist, I have many colleagues who work tirelessly to improve their quality of care by screening vulnerable patients for depressive symptoms or engaging them in chronic disease management. My peers do not do these things to receive bonus payments and supplement their income. They do them because of the belief that these things are right and consistent with their professional ideals.

As the Annals study highlights, there is a great deal at stake in reforming how providers are paid, and in some ways, we have gotten what we’ve paid for in P4P. I hope that whether we fashion a new generation of P4P or scrap the approach in favor of others, future policies will be designed to honor both physicians’ financial and nonfinancial motivations. I believe the payout is worth the (likely significant) effort. Without it, we risk financial incentives confounding meaningful nonfinancial motivations as vital aspects of our profession.

  1. Roberts ET, Zaslavsky AM, McWilliams JM. The Value-Based Payment Modifier: program outcomes and implications for disparities. Ann Intern Med. 2017. [PMID: 29181511] doi:10.7326/M17-1740 
  2. Frakt AB, Jha AK. Face the facts: we need to change the way we do pay for performance. Ann Intern Med. 2017. [PMID: 29181535] doi:10.7326/M17-3005


  1. Great points in the fight between P4P and other value-based care initiatives. As a general principle in quality work, changing the "targets" will not result in changes in the underlying process, rather it might lead people managing the process to distort the reporting of results to meet the desired metrics.

    I hesitate to agree with the authors that the final nail in the coffin has been shut for P4P b/c while it hasn't worked out quite as we'd hoped, we all should acknowledge that from the outset we knew this would be a painful experience with many failures before we got closer to getting it right. The question is are we at that point? I don't think so, particularly because we haven't leveraged the full capabilities of our healthcare system to try to successfully deploy P4P. Many places aren't fully utilizing our EHRs to provide interpretable and actionable real-time electronic clinical quality measures to help providers improve. I think we should keep pushing for P4P, while deploying other value-based initiatives (ACOs) elsewhere so as to not mix results from each and allow for appropriate study of each initiative individually.

    1. Thank you for these comments. I agree that it is important to ensure we are testing appropriate hypotheses with policy interventions and appropriately evaluating the results, and also that prior P4P demonstrations have not necessarily been ideal in terms of incentive desire and behavioral principles. I also think testing multiple approaches is important to identifying which payment models are most effective in different circumstances. Nonetheless, because the data we do have appears pretty consistent, what would be potential options for exploring P4P further?


By commenting on this site, you agree to the following Terms & Conditions of Use.