Ashish Jha, M.D., director of the Harvard Global Health Institute, discusses how to improve physician pay-for-performance methods during a recent discussion hosted by the Alliance for Health Reform.
Even as some physician pay-for-performance plans have struggled to succeed, a group of experts said more effort should be devoted to making the concept work before consigning it to failure.
Panelists at a recent Alliance for Health Reform briefing(www.allhealth.org) here noted that any movement away from a payment method that rewards volume over value involves a slippery slope because improved quality often requires more of a physician's time, which may not be tied to financial or other tangible incentives.
Initial pay-for-performance initiatives have been criticized as being ineffective, but Ashish Jha, M.D., M.P.H., director of the Harvard Global Health Institute in Cambridge, Mass., argues that not enough programs have been undertaken to pass judgment. Rather than viewing pay-for-performance as a circumscribed, static entity that can be accurately assessed by simply asking whether it works, he said, "A better question is, 'How do we get pay-for-performance to work?'"
- A panel of medical experts recently said more efforts should be made to allow physician pay-for-performance plans to work before labeling them failures.
- Some panelists said effective pay-for-performance initiatives could include incentives other than financial rewards, as well as multiple smaller incentives in place of one larger reward.
- Questions remain about how to evaluate performance for tasks not associated with specific CPT codes.
Jha said performance evaluations should include a smaller number of measures that are easier to track rather than the typical dozen or more that are included in many programs. And that's not the only thing that needs to change, he added.
"Do we need bigger incentives?" Jha asked. "Yes."
In his view, the medical community needs to be more aggressive in experimenting with different incentive programs. He cited Amazon as an example of an entity that is continually changing its home page and its marketing strategy to customers in an attempt to sell more products.
"They are experimenting all the time," Jha said. "That's what health care needs to do. There is no right or wrong way to do pay-for-performance. You just need to get the incentives aligned so that organizations can make them work."
Some of the best solutions to encourage a change in behavior will not necessarily require more money. Ateev Mehrotra, M.D., an associate professor at Harvard Medical School, said behavior studies indicate that individuals are generally risk-averse and that providing modest regular incentives offers stronger encouragement than a single reward even if the two are of the same value.
Transition to Performance-based Pay Is Tricky, Group Finds
Wholesale physician payment reforms remain a work in progress, as one physician group in Minnesota discovered. The Fairview Medical Group, which operates more than three dozen primary care and multispecialty clinics in the Minneapolis-St. Paul area, attempted to transform its payment system from one relying solely on relative value units (RVUs) to one that combines quality measurements with patient satisfaction and physician productivity assessments.
Patrick Herson, M.D., the group's president, explained during a recent panel discussion hosted by the Alliance for Health Reform that when the new payment system launched in 2009, the group hoped to move away from fee-for-service within three years. But that goal was not achieved, and the group shifted back to using RVUs after productivity dropped and physicians responded that improved quality requires a greater time commitment with patients that was not adequately rewarded.
In a follow-up survey of the group's physicians, 80 percent said RVUs were highly important in determining compensation; patient satisfaction was only considered very important by 28 percent. Using the survey as a guide, payment was recently changed whereby 90 percent of a physician's salary is calculated using RVUs and 10 percent is based on patient panel size.
"We're kind of circling back into the neutral zone to take another run at this," Herson said.
Consumers will travel to a store that is out of their way to save $5 on a minor product such as a clock radio, but they will not make the same effort if they can save $5 on a flat-screen television costing $1,000, he said. He said that 10 separate payments of $10 each can encourage behavior changes more than a single payment of $100.
"Smaller and more frequent incentives can be more powerful and can motivate behavior change," Mehrotra said. "Other incentives can be just as powerful as money. Physicians work in a complex environment where a lot of factors motivate their behavior. Money isn't the only one."
Another difficulty with performance measures is they do not always account for physicians who work in hospitals or other large medical facilities where they are able to exercise only minimal authority.
For example, said Mehrotra, "You could profile individual physicians in a hospital, but they may not have the ability to make decisions that affect the AMI (acute myocardial infarction) mortality rate."
One question that remains is how to evaluate performance for tasks that are not associated with a CPT code, such as telephone or electronic communication with patients.
"I hear people say, 'You just need to generate a CPT code and an RVU (relative value unit) for all the things that are considered high-value care,'" Mehrotra said. "It's said that a lot of care can be provided on the phone or you can do a telemedicine visit, so all we need to do is create a CPT code for that. Then nirvana will be reached and we'll have high-value care."
A possible downside of such an approach, Mehrotra warned, is that it could simply lead to a greater volume of telephone calls and telemedicine visits without necessarily improving care.
Measuring effective end-of-life care is possibly the most difficult aspect of performance evaluation, Jha observed. The medical community does not do an effective job of determining whether a terminal patient received proper care in a hospital, was transferred to hospice too soon or simply did not want to receive any more treatment, he said.
"The challenge of developing metrics for that is we haven't figured out what good care at the end of life looks like," said Jha.