Pay-for-performance (P4P) is the latest rage among health wonks as to how to improve the health care system. But does P4P really improve quality?
Mullen, Frank and Rosenthal (2008) hope to answer this question. One would initially believe that paying physicians to perform certain medically necessary tasks will improve quality. Further, some P4P involves structural rewards for physician practice EMR. If a physician installed an automated system to remind diabetic patients to get their A1C test, the quality benefits likely will extend to non-diabetic patients as well who will get more regular check-ups.
However, there are some reasons to believe that P4P may not work as expected. Until recently, P4P in the U.S. has been a very small component of a physician’s compensation. In the UK, P4P payments are large [the average GP practice earned $133,200 in P4P payments from the NHS] but since the P4P payments in the U.S. are a small fraction of the physicians income, it likely has little behavioral impact. Further, P4P may not result in better medical care, but better documentation of the care that has always been given. Most importantly, “providers may shift resources toward rewarded dimensions of quality at the expense of unrewarded dimensions, which may result in a decline in overall quality of patient care.”
Data
In order to test whether or not P4P is improving quality, Mullen, Frank and Rosenthal collected data from Pacificare. Pacificare experienced a wave of P4P innovations in the early 1990s until today.
- 1993: Pacificare begins collecting quality information on providers.
- 1998: Pacificare makes these quality reports public.
- 2003: Pacificare starts to pay bonuses based on the provider quality reports. this is named the Quality Incentive Program (QIP)
- 2004: Pacificare joins Integrated Healthcare Association (IHA) inauguration. IHA is a P4P program that has ten times the bonus payments as QIP and included six of California’s largest health plans.
The authors have Pacificare’s quality measures before QIP, during QIP and during IHA. As a control group, the authors use Pacificare provider quality measures in Oregon and Washington which were not subject to any P4P incentives.
The authors hypothesize the P4P will improve quality for the service for which providers receive bonuses. For non-P4P services, there will be quality improvements when a performance metric shares commonality in production with many other medical services. For instance, bonuses based on identification and scheduling will improve overall quality. On the other hand, bonuses who’s quality metric depends on physician time or effort will likely decrease quality on non-measured dimensions.
Results
“Of the six measures initially rewarded by IHA, only cervical cancer screening showed consistently positive returns.” Chlamydia screening also improved after it was added to the IHA list. “On the other hand, appropriate asthma medication rates actually decreased…when P4P was introduced in California…Preferred antibiotic usage, which was rewarded by the small-scale QIP but ignored by the larger IHA effort also declined…”
Conclusion
P4P is a very blunt instrument. In some cases it works fairly well but in other cases it does not. One problem with P4P is that it that it confuses three aspects of medical care: 1) getting patients who need medical care into the doctors office, 2) getting physicians to provide the correct care to the patients in the office and 3) documenting the care. The first problem is one of outreach and a reminder recall system, but also depends on the demographics of the physician’s patient based. The second aspect could be better measured with a missed opportunity metric. And the third would be made easier with EMR.
Is P4P rewarding doctors who have rich patients that see the doctor often? Is P4P rewarding doctors who document care more accurately? Or is P4P really improving the quality of medical care?
Because P4P is such a blunt instrument, policymakers and insurance companies cannot answer these questions.
Recent comments