Robert Berenson, M.D., is an Institute Fellow at the Urban Institute and an expert in Medicare and health care policy. He served as the head of Medicare payment policy for the agency now known as CMS in the late 1990s and as an assistant director of the White House domestic policy staff in the Carter administration. More recently, he was vice chair of the Medicare Payment Advisory Commission.
Robert Berenson, M.D.
Berenson is a board certified internist who spent 20 years in practice, the last 12 in a group practice in the District of Columbia. He organized and managed a successful preferred provider organization in the Washington metropolitan area and is the co-author of two books and numerous articles on health care policy.
AAFP News Now recently sat down with Berenson to ask him about payment reform, the emergence of new payment models for health care, and the role played by primary care and family medicine in an evolving health care system.
Q. Why does the term capitation carry such negative connotations?
A. Capitation was done poorly in the 1990s. There was no risk adjustment. In my own primary care practice, the patients who were the most ill with diabetes were referred to us, but there was no adjustment for the fact that their health care costs were inevitably going to be higher. So technically, there were a lot of problems with capitation.
Also, primary care physicians sometimes were at risk for all health care spending. It was a black box with withholds determined through non-transparent decisions by health plans. Physicians had legitimate complaints about the operational side of capitation.
There are, however, physician organizations in many parts of the country, not just in California, that prefer capitation -- and they call it capitation. In places like Denver and Portland, Ore., capitation is not a dirty word.
Capitation merely refers to a per-member, per-month (PMPM) payment. That could just mean capitation for the primary care of a patient and nothing else. That is how it works in the United Kingdom -- a physician is paid a capitation rate for the primary care services for a population. Global payment is a form of capitation in which the PMPM is for all or most health services, not just the primary care services. The physician is only at risk for his or her own services.
Q. You have said there are two different concepts being talked about in terms of value-based payment in the public and private sectors. Can you explain each of these?
A. Some people think we are going to be able to measure the overall performance of a physician or a hospital or an accountable care organization (ACO), meaning we can literally measure the performance and base a substantial amount of payment on that performance.
I am quite skeptical that we have satisfactory performance measures to permit that type of calculation. So I am quite skeptical that we are going to have a value index for every physician, a provision mandated by the Patient Protection and Affordable Care Act by 2015. In my view, we simply don't have the robust set of measures and the ability to enable that.
The second form of value-based payment is adopting payment methods that research shows are associated with better performance, at least on the spending part of the value equation. But this does not mean measuring performance in real time.
This includes moving from a fee-for-service to a global payment system -- adopting a population-based payment method that we think produces better results. This would have an important role for measurement, but it is not largely dependent on measuring performance to determine what the payment would be. Rather, the incentives for controlling spending are embedded in the payment model. They are not some extra assessment based on the measured performance.
Q. How does primary care and family medicine fit into the new payment models?
A. In two different ways. Some are trying to develop new payment models based on fee-for-service that target primary care and family medicine. The idea is to pay fee-for-service but then have an additional payment for care management because a lot of care management cannot be paid for on a fee-for-service basis. This would entail some kind of hybrid payment model, which has become standard in Denmark and recently was put in place in the Netherlands. It is combined fee-for-service and capitation -- a monthly fee.
Some of the more ambitious payment models, like global payment or shared savings, assume that primary care would be a core of a larger organization. That whole organization would be paid in a new way, and primary care would be viewed as more important than it is today because a strong primary care base is needed to improve and hold down costs and meet quality metrics.
In fee-for-service, organizations don't necessarily need a strong primary care base because you are not rewarded for what primary care can do in prevention, early detection of illness and care coordination.
Q. Do you consider the blended payment model an appropriate one for the patient-centered medical home?
A. Yes, I do. I do not believe you can reimburse on a fee-for-service basis for a lot of the activities that happen outside of face-to-face patient encounters. This includes the activities that family physicians do related to the myriad phone calls, the e-mails, the conversations with other physicians, with social service agencies, with visiting nurse associations and with family members. All of those activities cannot be paid for efficiently with a fee-for-service payment model.
You cannot pay on a fee-for-service basis for a 45-second phone call for a patient who didn't come in for a visit but for whom you are giving some advice to on the phone. Trying to pay for that on a fee-for-service basis would be terribly inefficient. You are better off rolling that all up into a monthly fee that you are going to pay the practice.
Q. What is the role for fee-for-service in the health care system?
A. There are parts of the country that will not easily form organizations that can take risks under a global payment system. I think we will continue to have fee-for-service, and I think fee-for-service can work pretty well if the incentives are structured properly, better than today.
People make the assumption that fee-for-service means being paid for more volume. But that is not true -- if the payments for fee-for-service bear a close relationship to the underlying cost of production, there would be fewer rewards for generating excess volume. If we pay more for office visits and less for interpreting EKGs, we would probably get more office visits and fewer unnecessary EKGs. I believe fee-for-service can absolutely affect behavior if we get the prices right or at least closer to right.
Q. Where does primary care fit into ACOs?
A. A lot of physicians envision hospital-specialist ACOs. But if you have ACOs that assume financial risk, any rational group of health professionals will understand that primary care is essential as the base for population-based health care.
I am actually more sanguine that primary care will be recognized as essential to ACOs -- at least the ones that are not formed by hospitals with their active medical staffs. One thing I am encouraged by is that many of the ACOs that have shown up to participate in the Medicare ACO program are physician led and not hospital led. I think it is worth testing both physician- and hospital-led ACOs.
One of the concerns I share with primary care physicians about hospital-led ACOs is that primary care physicians no longer have relationships with hospitals. As a result, primary care physicians will not be sitting on hospital governance committees and won't be part of the decision-making process. I think enlightened hospitals will understand this is a problem, and they will reach out to primary care physicians. I am aware of several hospital systems that are trying to do that. I don't know how successful that kind of collaboration will be.
I am more optimistic that physician-led organizations will provide a very prominent role for primary care in ACOs. This has been demonstrated by a number of independent practice associations.
Q. Why do you have concerns about the ACO shared savings model?
A. What I have seen bothers me in the following ways. There is a real difference between the ACO payment model and traditional capitation. In traditional capitation, the payment rates were determined by community averages. Actuaries would convert fee-for-service spending into PMPM equivalents. If you were a high spender, you were in a sense penalized, and if you were a low spender, you were rewarded because the capitation was based on the community average.
Some forms of shared savings are just the opposite. They accept your baseline spending and reward you if you reduce your spending in the subsequent year. Shared savings simply takes your base spending, compares it to your next year's spending, and that is the only basis for determining whether you have earned a bonus. That gives me a lot of problems. What I have seen locally is that a medical group that had remarkably low hospital utilization and high-patient satisfaction scores received a zero bonus. They already were doing as well as they could do and were not able to generate additional savings to share.
Another group that had much higher hospital rates, much higher emergency room use and higher use of other ancillary services was given huge bonuses because they reduced what was terribly wasteful to only wasteful. Good for them, but you shouldn't be penalizing your good practices because they have been good practices. We can be a little more sophisticated about this.
For the good practices, we can give them a target for spending that recognizes that their baseline is excellent, and that they get some rewards for that. Similarly, the poor performers don't receive full bonuses for being better.
Q. Where do you think we will end up in five to 10 years with these payment models?
A. I think we are going through a lot of intermediary payment forms that are not going to be terribly useful. I don't think bundled payment episodes are going to get us very far, and I don't think shared savings is going to get us very far. In the end, we want to make life a little simpler rather than having a raft of alternative payment models. I think we'll have two basic payment models: reformed fee-for-service with pressure on payment rates and risk-bearing provider organizations receiving some form of global payment. Blue Cross and Blue Shield in Massachusetts is testing that model, and that is the right model in my view.
In 10 years, I think we are going to want primary care physicians -- with hospitals and without hospitals -- bearing financial risk and being accountable for cost and quality for a population. That is where I think we need to be. I think during the next five years, we will test all of these alternatives, figure out if any of the others are viable, and then make a decision where we are going -- make that destination very clear and specify the roadmap for getting there.
We need to make it clear to physicians and other providers that if you want to stay in fee-for-service, there is going to be continued financial pressure on fee-for-service. But it will be there at least for a while. We also want to present an alternative so you are rewarded based on your own performance. That is what global payment permits.
Related ANN Coverage
Primer on Payment Reform: Rewarding Value Over Volume