A pay-for-performance program in which the metrics were selected by the participating physicians seemed to pay off in better quality, according to a federally-funded study that tracked the results.
The Palo Alto Medical Foundation Research Institute tracked the care delivered by 167 primary care physicians at the Palo Alto Medical Clinic in 2007. Researchers looked at what individual doctors did and how they did it, and then distributed bonuses based on the findings, using physician-specific incentives.
Primary care physicians at two affiliates, the Camino Medical Group and Santa Cruz Medical Group, were also tracked, but in a different way, says Harold S. Luft, PhD, a health economist and director of Palo Alto Medical Foundation Research Institute.
“We focused primarily on Palo Alto partly because that medical group decided to bring the pay-for-performance incentives down to the physician level,” says Luft. “At Santa Cruz, for example, if the pediatricians were getting very high scores, then extra funds were given to the pediatrics department. At Camino, if the overall group got extra money, it was distributed across the board. But at Palo Alto, if Dr. Jones got high scores, Dr. Jones would get a bonus.”
The Palo Alto physicians actually devised the P4P program, developing it from evaluation programs such as the National Committee for Quality Assurance’s HEDIS (Health Plan Employer Data and Information Set) measures. The Palo Alto Medical Foundation Research Institute, in partnership with the University of California — San Francisco, was awarded a federal contract to assess the effect of the incentive program. They submitted their report to the Agency for Healthcare Quality and Research on April 16. The institute is the research arm of the Palo Alto Medical Foundation, the not-for-profit organization that contracts with the three medical groups.
“We as a research group said to the Palo Alto Medical Foundation and its medical groups, You are doing this incentive approach and that sounds pretty interesting. Can we study it?” says Luft. He and his team got the Palo Alto doctors to agree to one modification: “Half of the doctors, randomly chosen, got a check in the mail each quarter based on their performance. Half were simply told how much they would get, and at the end of the year they got the check.”
Did getting a check every quarter make a difference? “There was no significant difference in changes in quality scores or anything else for those doctors who got the check each quarter versus those who just got the report and got the check at the end of the year,” says Luft.
The maximum a doctor could have earned was about $5,000 a year in bonuses, but only one came close to that amount. Most earned about $2,400.
“Picture one doctor getting a report and getting a $600 check and another getting a report that says that at the end of the year, you will get an additional $600 for this quarter’s effort and then, based upon next quarter, you get another check depending on how you do then,” says Luft.
They are still mulling over some of the findings. “A doctor who might be good at getting patients with diabetes to have very good blood sugar control was sometimes not particularly good at getting good blood pressure control for those same patients. That’s sort of strange and we don’t know why. I would have started out thinking that there were good docs and not so good docs, and instead, high scores on one measure do not seem correlated with high scores on other measures. Maybe it’s something else that’s going on in the data, or that clinically it is easier to get change in some measures than others.”
Another aspect of the pay-for-performance program was that the doctors got to see each other’s grades online. “They had been getting these kinds of reports for several years now,” says Luft, “so they are comfortable with them. They can see not only their own scores, but the scores of all the other doctors, by name. They could see where they stood. They could see what their scores were. They could drill down and see exactly which patients were involved in their own scores.”
Physicians are notoriously competitive, but Luft describes this reporting effort as an informational tool, rather than a grade contest. “It is not to show that they are better than anybody else,” says Luft. “It is to understand why somebody else is doing better than they are.”
Primary care was the focus. Besides the 167 PCPs at Palo Alto, there were approximately 125 PCPs at Camino and about 85 at Santa Cruz. To arrive at a score, the P4P program used a numerator (patients who received recommended care) and denominator (patients who were eligible for the recommended care) for each measure.
The study states: “In the electronic ‘quality workbook,’ each physician could review the list of his or her patients in the numerator and denominator for each metric to validate their patient list for each quality metric. Before the initiation of the physician-specific incentive program, all the physicians reviewed instructions on how to use and interpret the quality score workbook.”
“We found that quality was improving in a bunch of different measures,” Luft says. “In some cases, it seemed to be improving a bit more rapidly in 2007 than it had in 2006.”
What helped is that Palo Alto physicians had been using electronic health records since 2000. “So they really thought that they had a sense of ownership of the data,” says Luft. “They felt comfortable with that. They were now looking at all these measures as helping them improve quality of care for their patients.”
Also, Palo Alto Medical Foundation participates in the Integrated Healthcare Association’s pay-for-performance program. That California-wide program is funded by various health insurers and pays medical groups based on their performance scores for IHA plan patients. The PAMF program was a variation of what the doctors were used to.
So rather than play it safe and focus just on what IHA was encouraging, the PCPs in the study tried to improve the level of care, says Luft. The study states: “The specific definitions of the measures used ... were somewhat different, reflecting the group’s organizational goals, high standard of quality of care, and information technology capacity. For example, the thresholds for the diabetic control indicators were stricter than IHA’s (e.g., the target for glycemic control was HbA1c<7% rather than <8% and LDL < 100 rather than < 130 mg/dL).”
Of course, for some categories, such as frequency of prescribing asthma controller when needed, there wasn’t much room for improvement. “In some cases the scores were in the high 80s to low 90s already and it is going to be hard to get it much better than that,” says Luft. “And in one case, chlamydia screening, it is just very hard to get those young women tested and it is not clear what the denominator is. In general, quality seems to be good and getting better.”
Perhaps what is most interesting about what is going on with P4P at Palo Alto came after the study. At the end of 2007, the three medical groups — Palo Alto Medical Clinic, Camino Medical Group, and Santa Cruz Medical Group — merged into the Palo Alto Foundation Medical Group. The study states: “The merger led to an effort to harmonize physician payment across the three geographic divisions. Thus, continuation of the physician-specific P4P, which had been implemented only in the Palo Alto division using a division-specific quality monitoring system, was not a feasible option.”
Nonetheless, Palo Alto is ready to take P4P in a yet another direction. It is looking at instituting what it calls a comprehensive value-based system. The study states that “Such a system will not focus on physician-specific incentives, but the incentives will be given to a ‘team.’ Thus, one of the ideas behind physician-based P4P incentives — bringing financial rewards closer to the day-to-day practice level — seems to be alive and well.”
It will be interesting to see how this plays out.