Interview with Meredith RosenthalQI and Health Disparities
Meredith Rosenthal: Putting P4P Into Perspective
Despite all the enthusiasm surrounding the potential that pay for performance (P4P) programs offer for quality improvement, Harvard School of Public Health professor Meredith Rosenthal, Ph.D., remains skeptical that it will have as dramatic an effect as some are counting on. Rosenthal’s skepticism stems from her extensive research on P4P trends and outcomes over the past five years, with a variety of Harvard colleagues including Arnold Epstein, Richard Frank, and Bruce Landon, that have appeared in such journals as Health Affairs, the Journal of the American Medical Association, the New England Journal of Medicine and Medical Care Research and Review. Through this research, some of which was supported by PHSI, Rosenthal has shown that while efforts to pay providers to reach certain performance targets are growing rapidly, the evidence on the effects on those efforts on patient care is ambiguous. In fact, a June 2007 study from researchers at Duke University found that paying hospitals extra money did not significantly improve quality of care or medical outcomes for patients, conflicting with an earlier study from the federal government conducted by Premier. Rosenthal, who is leading a team at PHSI to study real-world examples of performance measures in hospitals and systems, tries to set the record straight on P4P and its long-term potential for quality improvement.
PHSI: What have you learned about the utility and promise of pay for performance?
MR: First, the notion of paying based on measures that reflect quality and value resonates with almost everyone. Even the most skeptical physicians agree that pay for performance makes sense conceptually. They may worry that the measures are not quite right but they understand the merits of it paying for good care.
We also have found that P4P is very complicated to implement. It is very difficult to figure out and reconcile other political concerns with the best way to pay for high quality performance. When we look at the field, P4P programs are generally weak – they are not designed to maximize their effect and they are constrained by a number of factors. Thus, it is not surprising that most evaluations of P4P have found only modest effects on quality improvement.
PHSI: Then is P4P just a fad or is it becoming an entrenched form of paying for health care?
MR: It clearly has the properties of a fad as there is irrational enthusiasm for it. But I would say P4P is here to stay. There has been very, very broad uptake of P4P. A study I just published in Health Affairs with other Harvard colleagues of the earliest adopters show that they are still behind P4P but they are evolving in how they apply it. They are shifting from a focus on processes of care to one that emphasizes patient outcomes, keeps costs efficient, and promotes use of health information technology. In other words, they are now pursuing more meaningful measures to capture health status and predict whether someone is at risk for an adverse health outcome.
In another national survey, we found that about half of HMOs are using P4P and it seems to have embedded itself in the U.S. health care payment system. It is also creeping into Medicare, which will be significant.
Despite this activity, I do think the enthusiasm will die down when people realize that the problems with the health system are bigger than P4P will address and what we really need are broader payment and delivery reforms. While it is not entirely clear what the next stage will be for P4P, I do expect we will continue to see it as a key element of payment for health care. I certainly don’t see us moving backwards. On the other hand, I don’t think P4P will ever represent 50 percent of payment. It will be more like 10-15 percent.
PHSI: If P4P is just a transitional payment approach, then what is the next step to address the issues of costs, efficiency, and value for the health care dollar?
MR: We are still overpaying for things we don’t think are worthwhile. Maybe the next step is to start reducing payments for services like some diagnostic tests that we think are too expensive relative to their value or moving away from a system based on fee for service. There is growing sentiment to move toward some form of capitated payment or an enlightened version of capitation.
PHSI: In his new book Better, your colleague Atul Gawande talks about how patients and doctors rarely know who the high performers are or how they might imitate them. What role will P4P play in fostering that?
MR: We don’t see market forces driving quality to efficient levels. The idea behind P4P is to make the payment system resemble what would happen in a functioning market. But so many aspects of the market are constrained in medicine. We don’t encourage physicians to charge more because they are good doctors. Instead, we have fee schedules, which generally have little or nothing to do with quality. There is no way for market forces to adjust and provide incentives for quality so P4P can help foster that.
PHSI: Would making things more transparent help?
MR: I have less confidence that greater transparency of quality information is necessarily the way to go, particularly for consumers. Health care is a different kind of commodity. But I don’t object to transparency. I find merit in the claim that providers out of a sense of pride will respond to publicly reported information even though consumers won’t. We have seen a lot of public reporting to date and little evidence that consumers find information useful and informative. No matter how good information is, the fact is that when people make choices about health and health care, they don’t behave like a traditional consumer. They prefer to delegate decisions to a professional. So, I am very doubtful that health care consumerism is going to be substantial enough to move the market.
PHSI: But why should payers – or patients for that matter – pay providers more for doing what they should be doing already?
MR: This is a very common but unproductive argument. We should think about this as paying differently, not paying more. P4P is just about making incremental payment related to quality and outcomes. Fee for service was set up as the best alternative we had at the time. It makes eminent sense to pay for what is valuable. What is exciting is that this is starting to happen. For example, HealthPartners in Minnesota is no longer paying for services related to health events that should never have occurred such as the wrong blood type getting transfused. This is a great first step.
PHSI: Do you then see any evidence that P4P has done much to improve care? Studies seem to contradict themselves on that count.
MR: It probably has in a small way. But part of the reason the change has not been dramatic is that payers have been reticent to take too bold a step. I am hopeful that as it becomes more routine, payers will be more ambitious when it comes to applying P4P and we will start to see more effects on health outcomes. The continued debate on how P4P can focus more on patient outcome measures rather than process measures may lead us to an important step in that direction.
PHSI: Your recent study in Health Affairs found an increasing number of P4P plans using cost efficiency measures as a target for rewarding physicians and hospitals. Did that surprise you?
MR: I am surprised at the extent to which the earliest adopters of P4P have moved to outcomes and cost measures. The programs we looked at had all gotten bigger, broader, deeper and more focused on cost and outcomes. I do think policymakers need to understand that P4P is evolving from a system based purely on quality improvement to something more complicated that is trying to simultaneously address cost control and quality improvement. What is not clear is whether P4P is the right tool for reaching all of these goals.
PHSI: So if P4P isn’t the magic bullet some think it is, what else is out there that is exciting for pushing payment policy to embrace value?
MR: I am most excited and interested in payment reform proposals more broadly and understanding both payment systems and the organization and market systems in which payment reform can take place. I’m focusing now on reform efforts organized by coalitions of payers, including Medicare and the often-dominant Blues plans, because having a concentration of patients is critical for payment policy reform to work effectively.
PHSI: Why is this next frontier so exciting? What will it offer in terms of pushing the quality movement forward?
MR: Even the best-designed pay-for-performance programs are inherently limited – for example, by the data and measures available and by the pull of the incentives generated by the underlying payment system. When pay for performance is implemented by an individual payer, which most are, its effects are further diluted. So coalition-based efforts to question and reform the way we pay doctors and hospitals at a more fundamental level at least begin with all the right elements in play. Whether or not these efforts find appropriate and feasible solutions remains to be seen.
PHSI: In that vein, you and some other Harvard researchers recently looked employer use of value-based purchasing strategies. What did you find?
MR: We surveyed executives at 609 of the largest employers in the U.S. and we found that as a whole, employers on their own do not appear to be using their purchasing power with health plans and providers to improve the quality of health care.
PHSI: So what does that mean in terms of quality improvement?
MR: A couple of things. The results support the argument that the costs assigned to developing such things as health plan report cards erect barriers to wider adoption. This is particularly true among smaller employers. We found among that group that the benefits of measuring quality and using the information for contracting purposes do not appear to be enough to justify to the cost. That is why I think coalitions may be critical because they can more easily undertake QI efforts than individual employers can.
back to table of contentsIs The Quality Improvement Field Overlooking Health Disparities?
University of Chicago Associate Professor of Medicine Dr. Marshall Chin has been examining what kinds of quality improvement interventions are needed to reduce racial and ethnic disparities. Although there has been a flurry of evidence documenting the problem of health care disparities, there has been a paucity of research exploring how to reduce those disparities through QI activities. As Director of The Robert Wood Johnson Foundation “Finding Answers: Disparities Research for Change” National Program Office, Dr. Chin has begun to challenge the research community to look for and test solutions for reducing racial and ethnic disparities through QI efforts. Findings from a review of some 200 interventions to reduce racial and ethnic disparities in health care were discussed during an October 2007 panel briefing organized by Dr. Chin’s group. Leaders in the field offered promising strategies for health care organizations, providers and payers to reduce disparities in their own organizations. For more information, visit www.solvingdisparities.org/article/213706?o202353=&lpid=1484. Seven articles summarizing these findings were published as a supplement to the October 2007 edition of Medical Care Research and Review. Dr. Chin recently presented his emerging research questions at a PHSI panel discussion on disparities. PHSI talked to Dr. Chin about his project and why the QI movement should concentrate its efforts on the racial and ethnic disparities problem.
PHSI: Your research has focused on quality improvement (QI) and its relationship to minority health. Has the QI field essentially stayed away from addressing racial and ethnic disparities?
MC: Even though there is a lot of room for overlap, we have mostly seen only parallel development of these two fields. The Institute of Medicine included the goal of narrowing racial and ethnic health care disparities within its six different domains of quality of care. But in many ways, this issue has been like a second or third child. People have tended to focus on quality improvement but the equity issue has gotten short shrift. In contrast, in the disparity field, the vast bulk of work has been on describing disparities and pointing out that they exist but there has been relatively little work done on the solutions or what interventions we need to apply to reduce them.
PHSI: In the quality improvement arena, there has been a lot of attention paid to public reporting and pay for performance. Where does the racial and ethnic health care disparity issue fall when it comes to P4P?
MC: Given the national interest in P4P, it is amazing how little attention has been paid to its potential effect on racial and ethnic disparities in care. Most studies related to pay for performance have been done in predominantly white, well-insured populations. We do not know if P4P will improve or worsen disparities, or whether we need to build in safeguards such as rewarding relative improvement in quality in addition to absolute levels of quality. These are all new terrains that have to be actively investigated.
PHSI: Why is it important to single out disparities? Shouldn’t we just be trying to improve quality generally and won’t benefits accrue to all from that effort?
MC: This raises the question: Can a general quality improvement technique improve care for everyone? We have found evidence that generic quality improvement techniques can help improve care but culturally tailored approaches appear to be more successful. We have not done a good enough job tailoring approaches in the real world or individualizing to your particular patient population or organizational context. What may be appropriate for a well-insured white population may not work well in an African American population who lives on the Southside of Chicago. In addition, minority groups are not monolithic and if you apply a one size fits all technique to every population in every part of the country, you are not going to do as well as if you tailor it to specific patients and providers and health care organizations.
PHSI: But that seems like a very expensive, complicated and daunting task?
MC: It would be wonderful if there was a magical single bullet solution but the world is complicated and within health care there is growing recognition of the importance of patient-centered care. One of the cruxes is this whole idea of individualization and actively engaging patients in their care. Culturally tailored care as one of the solutions to disparities is one example within the broader patient-centered care movement. It demonstrates that people realize we have to individualize care if we are going to provide the highest quality of care within the context of practice guidelines and appropriate treatments.
PHSI: Then what steps do we need to take to implement and sustain QI strategies to reduce racial and ethnic disparities?
MC: First, we need to better define effective solutions that are tailored to patients. Second, we need to better understand how to work with providers and administrators on how to successfully implement these solutions. It may be a wonderful model program but the question is ‘how do you adapt it for your setting?’ This is a managerial challenge. Also the issue of sustainability is a huge one. Things have to make sense financially from the individual organizational perspective to work and from the policy and reimbursement context. The system has to be aligned in a way so people can do the right thing and are given appropriate incentives to do the right thing. Chronic diabetes care is a good example. Performing foot exams and getting patients to see the eye doctor or get blood tests – these activities will cost health providers more money if they have poorly paying insurance mixes. There needs to be ways to reform the system so there are financial incentives to provide high quality of care, rather than if by providing high quality of care, you are financially harmed.
PHSI: So how do you overcome that possibility?
MC: There is always a danger of gaming. One issue here is that our most vulnerable patients are potentially put at risk under a P4P model. The pressures inherent in P4P could create perverse incentives for providers to dump their minority patients with limited insurance or resources because they are having a harder time attaining goals for doctors. It is usually more challenging to take care of patients with fewer resources.
PHSI: So how we do we address this? How do you create financial incentives?
MC: I believe most providers would like to do the right thing in terms of providing equitable care to all. However, no matter your heart, unless you have financial streams that can support the mission to provide care to all, it’s difficult to do over the long-term. We need to enact policies that create appropriate incentives and financial streams to spur this.
PHSI: Is there enough research or evidence to make the business case to use QI in disparities reduction?
MC: There is a small amount of literature on this question. From a societal perspective, many types of QI initiatives can be cost effective while reducing disparities. For example, preventive care for patients with chronic disease can lead to better health and be cost effective, and even save money in systems where you have a defined captive patient population. However, for chief executive officers in stand-alone health centers, there are few quality initiatives that actually save money, especially for uninsured or underinsured patient populations.
PHSI: So, how do you convince them to care about this issue?
MC: First, it’s the right thing and I think that people would love to do the right thing. I think a major message in the disparities/equity field is this should not be some marginalized activity. It is among the Institute of Medicine’s six pillars of quality. Many QI efforts are pretty much generic QI. To address disparities, a lot of hospitals will do community outreach or a marginalized effort like a special screening at a local health fair. While these activities have value, what we are arguing is that disparities should be an integral component of all QI efforts. If you want to provide the best care, you have to individualize care.
The second thing, there probably will be over time, an increasing set of more formal factors which are going to encourage this such as the incorporation of disparity measures into HEDIS or Medicare performance benchmarks that organizations are judged on economically or for publicly reporting purposes.
PHSI: Given all the attention that disparities are getting generally, do you see things shifting in this direction already?
MC: Yes, in some ways quality and disparities have been together for awhile, but in the past three years we have seen wider and more prominent recommendations on how closely intertwined they are. Both are critical. It is a very exciting time. There is now growing national consensus that we need to use quality improvement techniques to reduce disparities and equity is a fundamental component of quality and quality improvement.
PHSI: So what is your message to employers, purchasers of health care?
MC: The message is that it is good business to contract with health care organizations that are cognizant of the importance of reducing disparities and providing quality care to all and individual care. We are increasingly seeing a diverse racial and ethnic workforce and if you want to be attractive in employee recruitment and retention, the smart employer is going to look carefully at the health plan they contract with.
PHSI: What would you say then are the top three questions researchers need to consider when looking at this issue?
MC: First, what are effective interventions to reduce racial and ethnic disparities in care? Second, how can we effectively implement these interventions and tailor them to specific circumstances of a given health care organization and their patient population? And third, how can we make these interventions sustainable financially and what kind of policy reforms do we need to create this type of environment and incentives for people to do the right thing?
Project Updates