Assessing P4P’s Performance

The healthcare industry started with a theory: If payments to providers are tied to performance, overall quality of care will improve.

It has been a while since my last science class, but from what I remember about the scientific method, the next steps after you've developed a hypothesis are roughly: 1) conduct experiments, and 2) measure the results against your original prediction.

From Medicare's Physician Quality Reporting Initiative and Physician Group Practice Demonstration to a variety of private-payer P4P plans, the idea has been tested plenty of times at this point. So how does it hold up against the initial goals?

Well, that depends on who you ask. There certainly is evidence that physicians and hospitals will change behavior when working under a performance-based system, although some attribute that in part to an "observer effect" (i.e., physicians tend to change behavior when they know they're being measured, regardless of the financial consequences).

But many physicians remain skeptical. There seem to be three major problems with P4P in its current form:

1. Insufficient incentives. CMS boasted last month that 10 groups earned $16.7 million in incentive payments under its P4P demonstration program. However, some have raised concerns that that's not enough to really offset the costs of participating. Even in private-payer P4P programs, incentives rarely exceed 2-3% of a doctor's overall compensation. To really have an impact, that percentage will have to increase.

2. Administrative burdens. Even a small bonus might be considered worthwhile if the P4P programs weren't such an administrative headache. Doctors are frustrated. In an MGMA survey released this week, 63% of PQRI participants reported difficulty capturing and submitting data, and a whopping 93% had difficulty accessing their reports. The entire premise of P4P hinges on data collection and feedback. If that isn't happening efficiently, the program won't work.

Some of these problems are just initial bugs that can be fixed. The American College of Physicians offers an example of a doctor participating in PQRI who was forced to submit a one-cent charge per code so that Medicare's carrier would register the submission, and he then had to go back and manually write off the charges afterward. These administrative hassles—as well as actual costs associated with collecting data—can effectively cancel out bonus payments if they aren't large enough.

3. Unintended consequences. Writing in the New York Times this week, Sandeep Jauhar, MD, argues that "whenever you try to legislate professional behavior, there are bound to be unintended consequences." If you define performance based on whether antibiotics are administered to a pneumonia patient within six hours, will doctors rush their diagnoses or begin overprescribing antibiotics? If physicians are rewarded based on outcomes, will they be tempted to cherry pick patients who can comply with the measures? These are legitimate questions that will only become more important to answer as the size of incentive payments grows.

Most of the criticisms have to do with the execution of the programs rather than their broader goals. There still isn't enough evidence to rule out the original hypothesis—that tying payment to performance can improve quality. So despite all of the flaws, we shouldn't necessarily scrap the entire concept.

But this also isn't the time to roll P4P out to the entire industry. Many believe the PQRI and other demonstration programs are trial balloons for a mandatory federal pay-for-performance system in the near future. Momentum is taking us in that direction, but CMS and other P4P architects should slow down and pay attention to the feedback offered by physicians and address the serious concerns that have been raised.

We need more experiments before jumping to any serious conclusions.