Pay for performance (P4P) makes sense intuitively. Instead of paying for quantity, we pay for quality. Hospitals that do better get bonuses, hospitals that don't are penalized. The incentives align with outcomes we want.
Reality, though, isn't that simple. One of the arguments many (including me) have made is that if you just set metrics, and then measure them, under-resourced hospitals could get penalized. Here's how:
Let's say there's a hospital that services an affluent suburb where everyone is insured and has a primary care doctor. When you discharge a patient from that hospital, they have great follow-up, they fill their prescriptions, and they don't get readmitted. Let's say there's another hospital that services a poor, urban area where insurance is spotty, many don't have cars to get to appointments, and fewer have a regular source of care. When you discharge a patient from that hospital, they may not be able to afford a prescription or see a primary care doc immediately. They're more likely to get readmitted.
So if readmissions are your quality metric, hospital #1 will do great and hospital #2 will do poorly. Then if you tie payment to that metric, hospital #1 (which was already doing well) gets more money, and hospital #2 (which was already struggling) gets less. In other words, pay for performance could penalize the hospitals that need the most support.
Don't believe me? Cue the NYT:
The panel found that existing payment policies unintentionally worsen disparities between rich and poor by shifting money away from doctors and hospitals that care for “disadvantaged patients.”Measures of health care quality and performance — widely used by Medicare and private insurers in calculating financial rewards and penalties — should be adjusted for various “sociodemographic factors,” the expert panel said. The panel was created by the National Quality Forum, an influential nonprofit, nonpartisan organization that endorses health care standards.
“Factors far outside the control of a doctor or hospital — patients’ income, housing, education, even race — can significantly affect patient health, health care and providers’ performance scores,” said Dr. Christine K. Cassel, the president of the organization.
By the way, this panel was commissioned by the Obama administration, which thinks P4P is a great idea. What will they do with this information?
The Obama administration commissioned the study, but is not entirely comfortable with the recommendations, officials acknowledged. The existing policies of the National Quality Forum and the government say performance scores should generally not be adjusted or corrected to reflect differences in the income, race or socioeconomic status of patients...The Obama administration has championed the idea of pay for performance, with financial penalties for hospitals where deaths, readmissions or complications occur at rates above the national averages.
The administration has said adjusting the data for social or demographic factors would be equivalent to accepting a double standard, with lower expectations for the care provided to low-income patients.
It gets worse:
When reporting on performance, the government often takes account of clinical factors, like the presence of diabetes or the severity of a patient’s illness. But the Obama administration said it “currently does not adjust quality outcome measures for patient socioeconomic status” because it does not want to lower the bar for services to low-income patients.
Unfortunately, it's often harder to provide services to low-income patients:
But research by the Medicare Payment Advisory Commission, which advises Congress, tends to support the National Quality Forum.“Lower-income patients have higher readmission rates,” the commission said, and major teaching hospitals, which serve large numbers of indigent patients, face the highest penalties.
Dr. Atul Grover of the Association of American Medical Colleges, a member of the expert panel, said: “Teaching hospitals take care of the poorest, sickest, most vulnerable patients. They should not be penalized for factors outside their control.”
Measuring quality is hard. It really, really is. P4P requires that it be reasonably easy. They need metrics they can get from everyone without the gathering of data taking too much time or money. But those factors sometimes lead us to measure the things we can easily measure, not the things that matter.
I remain skeptical about pay for performance. My prior posts on P4P can be found here, here, here, here, here, here, and here.