Summer 2024 - Vol. 19, No. 2 Editor’s note: The following is a transcript of the Laurence E. Carroll, MD 2024 Lecture, held on April 1, 2024, at Penn Medicine Lancaster General Health. The Laurence E. Carroll, MD Lecture Endowment was established by gifts from friends and family of Dr. Carroll to honor his memory, legacy, passion, and lifelong commitment to medical ethics and continuing medical education. To make a gift to the endowment in Dr. Carroll’s memory, call 717-544-7126.

In 1999, concerns about patient safety and health care quality rose to the nation’s consciousness. In September of that year, the Institute of Medicine released a report called, “To Err Is Human.”¹ This report made front-page news across the United States, partly because it quantified for the first time the degree to which we have a problem with medical errors in this country.

The report estimated almost 100,000 deaths annually from medical errors — that’s more than twice the number of deaths from the next most common cause of accidental deaths, which is vehicles. By providing a number to this problem of medical errors, the report brought to attention a problem that many people — not only policymakers, health care workers, and administrators, but also the public — didn’t even know we should be worried about. It began an important policy discussion around the frequency of deaths from medical errors.

Around the same time, there emerged the beginning of a reckoning over unequal health care treatment in the United States. Again, a report from the Institute of Medicine, called “Unequal Treatment,”² drew attention to the fact that quality-of-care in the United States is not uniformly low; rather, it is differentially low across certain racial and ethnic groups. The report demonstrated that Black and Hispanic patients have higher rates of uninsurance. It also demonstrated that Black and Hispanic patients have decreased access to care compared to white patients and higher rates of death.

Since that time, a significant body of research has documented the problem we have with low and variable quality-of-care in this country and has prompted a discussion about how to improve quality-of-care.

Many quality improvement efforts have focused on restructuring the way we deliver quality care, often using financial incentives aimed at providers. It has often been assumed that these efforts will result in better quality-of-care for everybody and might decrease health care disparities, but our experience has demonstrated that that is not the case. Policies aimed at improving quality-of-care rarely also reduce disparities and may in fact worsen disparities.

To understand why disparities might worsen, it is helpful to first understand why quality-of-care is low. Health care suffers from a problem that economists call “asymmetric information.” Health care is a technical field. It requires a specialized knowledge, and patients don’t typically have complete knowledge about their medical condition or what the best treatment is, and so, understandably, they rely on their physicians and other providers to make decisions on their behalf.

However, patients can’t easily observe the quality of their physicians, so they don’t always choose a physician that may be highest quality or best able to make those decisions. As a result, physicians may be unresponsive to patients’ demand for high-quality care.

This problem is exacerbated by the way we have historically paid for care in this country, through a fee-for-service payment system. Physicians, for many years, have been compensated for the quantity or intensity of care they provide, rather than the quality of that care. This often results in problems like overuse of care, unnecessary care, and high costs of care.

To correct the problem of asymmetric information leading to low quality-of-care, the solution, according to economists, is to change the incentives: high-quality care should be financially rewarded, regardless of whether patients can directly observe it. There have been two general approaches to try to implement this in practice — using targeted payment incentives and global payment incentives.

TARGETED PAYMENT INCENTIVES
Going back to the 1990s, Medicare led the way in changing payment in a very targeted way, using public-reporting of quality (report cards) or using “pay-for-performance,” which is a more direct way to tie payment to quality.

Public-reporting offers a straightforward way to improve quality through selection. The idea is that, if we give patients more information about the quality of medical providers, they can use that information to preferentially select high-quality providers. This can shift patients from lower quality providers to higher quality providers, but alone it doesn’t do anything to incentivize the low-quality physicians to improve quality-of-care.

That’s where the second pathway — the “change” pathway or the “quality improvement” pathway — comes in. Here, two things happen.

First, when we measure quality and report it to the public, we also let physicians know what their quality is. Just as patients can’t easily observe quality-of-care, physicians often don’t know their own level of quality or how it compares to their colleagues and peers.

Next, and perhaps more importantly, making information about quality-of-care available to the public gives physicians the incentive to compete on quality, so they might then put more effort into improving their quality-of-care in an effort to maintain or improve their market share. This is a very economic view, but it is, I think, a simple and elegant solution to the problem of low quality-of-care in the setting of asymmetric information. If only it worked that well.

In reality, we’ve often found that public-reporting alone is insufficient to change the behavior of physicians, in part because the incentives are too weak. It relies on consumers being able to find and use the information. There are a lot of competing reasons that patients choose a physician, and it’s not always because of the grade they get on a report card. And so, without consumer choice, providers may not have as much motivation to improve.

Thus, soon after public-reporting started, Medicare and other payors began to add pay-for-performance, which directly adjusts fee-for-service payments to be higher when higher quality-of-care is provided. It can also include penalties for providing poor care.

So, what do we know about public-reporting and pay-for-performance and how effective they are in improving quality? One study completed shortly after these incentives were put into place demonstrates that pay-for-performance has a stronger effect on improving quality-of-care compared to public-reporting.³ While this is encouraging, over the years we’ve found that this effect is quite variable and that the effect of pay-for-performance is often quite small: physicians will improve their quality-of-care for the discreet things for which we are paying them, but often not for other related items. For instance, hospital outcomes or patient mortality rates may not improve.

There are also other shortcomings that limit the effectiveness of pay-for-performance. The incentives are often too small, the rewards are received long after care is delivered — so not salient to providers — and the biggest rewards tend to go to providers who were already doing well. Thus, around the time that the Affordable Care Act was passed, there was an increased interest in using strategies that more fundamentally altered the payment system.

GLOBAL PAYMENT INCENTIVES
The second approach, “global payment incentives,” moves away from the pay-for-performance piece-rate system. Global payment shifts the focus to managing health populations or an episode of illness. It holds providers accountable for the costs of care across the episode or population, and often includes shared savings or shared risks. This means, if a clinical provider can manage the health of their population for less than a benchmark set by the insurers, the provider can keep some of that savings for themself.

This doesn’t just reduce costs; providers are also held accountable for the quality-of-care. The idea is to achieve high-quality care at lower cost.

A bundled payment is a common type of global payment incentive. While fee-for-service pays hospitals, radiologists, surgeons, and anesthesiologists separately for the care of each patient hospitalized for surgery, in a bundled payment, a single payment is divided across all the different providers for all the different services. This incentivizes individual providers and organizations to work together to provide high-quality care at a lower cost, therefore providing high-value care.

Another common approach to global payment is through accountable care organizations. Accountable care organizations bring providers together to agree that they’re going to care for a defined population — for example, nursing home patients — to manage the care across space and time. They hold themselves accountable for the overall quality and costs for this defined group of members. Yet, when we look at whether these approaches are working, we see mixed results.

EFFECT OF VALUE-BASED PAYMENT ON HEALTH EQUITY
If we take a step back and think about how these value-based payment (VBP) programs affect disparities, we recall there’s been a general assumption that improving quality will help everybody without exacerbating existing inequities. The idea is that rising tides lift all boats. And it is certainly possible that the lowest quality providers may respond to these financial incentives; as a result, it may improve or reduce health care disparities.

The flipside is, VBP programs may not do that. Providers are relatively risk averse and may avoid complex patients if their salary is tied to the outcomes of high-risk patients.

VBP may also worsen disparities due to differences in provider resources. Safety-net hospitals and clinics operate with low financial margins; they don’t have a lot of financial resources because the payment systems don’t provide them. They also typically have lower quality-of-care in the absence of quality improvement incentives — that may be related to the complexity of the patients they care for, which in turn may be related to their patients’ poor financial conditions.

VBP disproportionally penalizes safety-net hospitals, taking resources from organizations that need them the most while giving to organizations that need them the least. Pay-for-performance and other payment models tend to be rolled out in a cost-neutral way, so there’s a fixed pot of money distributed based on quality or value. For example, Medicare may withhold 2% of payment and then reallocate that money to hospitals based on performance scores.

If you are a safety-net hospital and have lost 2% of your payment when you don’t meet quality benchmarks — because that payment is now going elsewhere — you may find yourself in a cycle of poverty. This cycle can result in decreased quality and may lead to hospital closure. This has, in fact, occurred frequently in rural areas, reducing access to care for at-risk patient populations; it probably ultimately worsens patient outcomes.

A lot of evidence collected over the last decade suggests that VBP programs are not helping health disparities and, in many cases, seem to be worsening them. The question remains, how can we redesign payments to advance the goal of health equity?

REDESIGNING PAYMENT TO ADVANCE EQUITY
There’s been significant attention focused on the problem of payment incentives and health care disparities in the last five years. Medicare, to their credit, has taken the evidence produced by researchers very seriously, and has tried to advance an agenda that will address and reduce racial disparities. Medicare and the Center for Medicare and Medicaid Innovation have made an explicit goal of addressing disparities and increasing participation of safety-net providers in VBP programs.

There are four approaches to modifying the current VBP program to more purposefully advance health equity. The goal of these is to not just avoid harm, but to improve disparities. I’ve ordered them here from what I think are the weakest incentives, or the least likely to work, to the strongest incentives, the most likely to advance health equity.

Create Accountability for Equity
The idea here is to modify existing VBP programs to include metrics focused on disparities and equity. Instead of just measuring admission rates or the rate of use of home dialysis, we can also include measures of disparities and equity. Other approaches in this area are to meaningfully reward providers for reducing disparities and achieving equity, and also ensure that health equity performance — and payment for that performance — represent a significant percentage of a provider’s overall quality score.

Account for Social Risk in Performance Measurement
The incentives we have used for the past couple of decades have differentially penalized providers who take care of a high number of patients at social risk. Another approach is to level the playing field in a more meaningful way and not financially penalize those providers.

This is an excellent idea in theory, but it’s harder to implement. It may be done by rewarding providers for improving their performance and stratifying performance to compare providers with others who have a similar makeup of patients. Another option is adjusting performance measures for social risk — that is, using statistical techniques to “level the playing field” and account for difference in social risk across providers.

One very recent study seems to suggest that, if implemented today, adjusting performance based on social risk would result in a substantial increase in the likelihood that safety-net hospitals and minority-serving hospitals receive a bonus. That would be progress in the right direction.

Financially Support Under-Resourced Providers
Another step in the right direction is for VBP programs to direct financial support to under-resourced providers serving low-income patients through upfront payments that are not tied to performance or equity. Sometimes called “equity pools,” these payments support capacity building and practice transformation. Though there are barriers to widespread implementation like this — to many politicians and policymakers, the unrestricted funds can be perceived as being politically unpalatable — such programs are likely to have a larger effect than the other approaches, which are less about providing additional funding and more about improving quality-of-care.

Address Drivers of Inequities
But if we really want to improve health equity, we need to take a step back and think about what’s driving health inequity. It’s a myriad of things, such as economic opportunity and having access to healthy food, a safe environment, insurance, and health care providers. Some of these are being addressed through VBP programs, but fundamentally the ability of those programs to have a meaningful impact on health inequities is low.

But there are approaches I’m optimistic about. In work I’ve been doing over the past year, my colleagues and I have been thinking about how to implement VBP programs in a way that makes sense in community health centers. We’ve been spending time talking to community health center providers to understand the barriers they face in participating in VBP programs and the modifications that would be needed for those programs to be successful in that setting.

What we hear from them is that the activities they engage in and the care they provide are often so upstream to the health care delivery being targeted by VBP programs, it’s hard for them to fit into current program frameworks. These organizations often prioritize finding stable housing, addressing food insecurity, job training, and stable sources of health insurance for their patients. We know these activities affect the health and long-term outcomes of patients, yet we don’t have a system in place that rewards providers for addressing them.

I am optimistic that if we find a path forward that meaningfully addresses these drivers of health inequities, we can make meaningful progress to advance the goal of health equity. I’m less optimistic that there’s anything we can do to the existing VBP structures that’s going to meaningfully address these inequities.

It’s a hard problem to solve, and it’s not surprising in some ways that in the past 20-plus years we’ve been working on this, we’ve made very little progress. The good news is, we are doing much better in terms of reducing the harms that come from VBP programs.

There are things we can do to make the existing system work better for everybody, putting patients at the center of what we do. Stakeholder input — including patient input — throughout the policy process, would be helpful. We should be building equity directly into the policy process.

It’s a complex health system that we work in — it often responds in ways that are surprising and unintended. All new policies need to be monitored for their impact on equity. This is something that didn’t used to happen routinely; it happens relatively routinely now.

We need to build in the ability to pivot policies quickly as more data become available. It’s hard for large insurers like Medicare to do that; it’s easier in small settings. But the inflexibility that’s built into traditional Medicare makes this kind of pivoting very hard, which is why it takes decades to try to get to these solutions.

These kinds of activities are being adopted more and are important next steps in the goal of taking the existing system and improving it to decrease disparities. Thinking more about how we can more meaningfully affect these drivers of health inequities would have a longer term, more valuable impact on quality-of-care.

ACKNOWLEDGEMENTS
The editors thank Janet Carroll, Serena Chamness of LG Health Continuing Medical Education, and Jennifer L. Groff of the LG Health Foundation for helping to bring this article to fruition.

REFERENCES
1. Institute of Medicine (US) Committee on Quality of Health Care in America, Kohn LT, Corrigan JM, Donaldson MS, eds. To Err Is Human: Building a Safer Health System. Washington (DC): National Academies Press (US); 2000.
2. Institute of Medicine (US); Board on Health Sciences Policy; Committee on Understanding and Eliminating Racial and Ethnic Disparities in Health Care; Smedley BD, Stith AY, Nelson AR, eds. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington (DC): National Academies Press (US); 2003.
3. Lindenauer PK, Remus D, Roman S, et al. Public reporting and pay for performance in hospital quality improvement. N Engl J Med. 2007;356(5):486-496. Rachel M. Werner, MD, PhD, is executive director of the Leonard Davis Institute of Health Economics at the Perelman School of Medicine and the Robert D. Eilers professor of health care management and economics at The Wharton School of the University of Pennsylvania