When Ben Harder and his team of health analysts at U.S. News & World Report developed a suite of health equity measures for America’s hospitals, they expected to find some level of disparity, but nothing prepared them for the shocking magnitude of inequity they uncovered.
“I think the scale of it, I would say, was a surprise to me,” Harder, managing editor and chief of health analysis at U.S. News said to The Cancer Letter. “To see that four of five hospitals treated a patient population that was disproportionately white was a surprise.”
The new health equity metrics were designed to assess representation of racial and ethnic minorities among patients who access common services at hospitals across the United States. As it turns out, at 80% of hospitals that were evaluated, racial and ethnic minorities are underrepresented among patients, relative to the demographics of the hospitals’ service areas.
These measures are one way of navigating the labyrinthine nature of health disparities, Harder said. For instance, it’s important to examine whether the quality of care offered to racial and ethnic minority patients is contributing to disparate patient outcomes.
“Differences in the quality of hospitals that tend to treat different racial and ethnic populations, that’s potentially a big problem right there,” he said. “There have been a number of studies that have looked at racial differences in where patients get treated—the literature and the evidence shows that people of lower income status and of racial and ethnic minority background tend to be treated at lower quality hospitals, however you define lower quality, whether you use the U.S. News rankings, or you use some proxy like volume or academic status.
“Unless two hospitals have equivalent outcomes, you have a disparity simply because there’s a difference in hospital quality and they’re not treating the same population.”
If U.S. News decides to use health equity measures to inform the Best Hospitals rankings, these rankings may change, with many top-tier institutions sliding downward.
“To be honest, the hospitals that excel in many different areas, advanced clinical care, are not the same ones that we see indexing high on representation of nonwhite groups,” Harder said.
“To what extent that’s because of something the hospital is doing or not doing, we don’t know. This is really just descriptive, but it does reflect differences in access to that care for that entire community.”
Harder spoke with Matthew Ong, associate editor of The Cancer Letter.
Matthew Ong: This year’s Best Hospitals rankings found that racial and ethnic minority patients are underrepresented at four out of five hospitals. That’s 80% of hospitals. Has this been documented before?
Ben Harder: We found a number of studies that looked at specific procedures or specific areas of care that had similar findings. So, this is certainly consistent with what scientists have identified in the past.
But I could not find any study that looked as broadly at elective care as we have. We really included everything in our analysis, every type of procedure that is generally performed on an elective basis. And so, that included things like colon cancer surgery, lung cancer surgery, various heart procedures, knee replacement, hip replacement and so on.
This is also the first time, to my knowledge, that data like these have been reported for each individual hospital, not just the nation as a whole.
So, this is an analysis of health equity—or rather, inequity—on an unprecedented scale.
BH: I think that’s fair, yes.
I see that these new health equity measures, although separate, would now be available alongside hospital rankings. What prompted you and your team to embark on these measurements?
BH: The issue of health equity is central to health care in the U.S.
A lot of preventable or manageable conditions become as serious as they do because of inequities in access to health care, in the social determinants that influence people’s health and the outcomes that the medical community can achieve for different populations. And so, that is really the impetus for us looking at this.
We want to understand which hospitals are providing the care that their population needs and achieving good outcomes for everyone in their community, and not just for those who may find it easier to access care or may be diagnosed earlier.
Was it a surprise for you to find that this is a problem across 80% of U.S. health care? I ask this, because I can almost hear some people I know say, “Well, but we already knew disparities and underrepresentation are systemic issues. It’s an evergreen societal problem,” in that nasal Washingtonian know-it-all undertone.
BH: I think the scale of it, I would say, was a surprise to me. We weren’t surprised that nonwhite patients were underrepresented among the important services, these are therapeutic services.
Given everything that we know about the access to care and the limits on access, the obstacles to access for many racial and ethnic minorities and for lower income patients, we unfortunately expected to see some disparity here.
But to see that four of five hospitals treated a patient population that was disproportionately white was a surprise.
Could you describe your methodologies for assessing health equity?
BH: We looked at eight different measures of health equity. And I want to just be candid, this is not a comprehensive look at health equity.
We are at the beginning stages of this initiative, which I anticipate extending many years into the future, but we focused on some measures that we felt would help us understand who is accessing care, hospital care, specifically, and for what.
And so, several of those measures look at elective care and assess if the hospital’s elective-care population represents the residents in its community. That was one piece.
Several other measures looked at, essentially, preventable hospitalizations for conditions that are sensitive to the outpatient care, the ambulatory care that a person receives, including preventive care, disease management.
We wanted to understand things like diabetes and heart failure and COPD—how equitable is the management of that disease, and the preventive steps that can stave off the need for hospitalization?
And unfortunately, for those measures, we found overwhelmingly that minority patients, specifically Black patients, were overrepresented. We focused those measures on looking at Black versus non-Black.
The Black residents were more frequently hospitalized for preventable causes, which we interpret as a sign they are not getting the same degree of access to preventive services and to disease management within the community.
It’s not too complicated. So, the elective care measure, it’s the same measure concept across five different populations. One is all nonwhite patients, and then we subdivided it into Black patients, Hispanic patients, Native American patients and Asian and Pacific Islander patients.
There’s sort of that roll-up measure of all nonwhite patients, and that’s what the four of five is—a reference to how patient population at each hospital of nonwhite patients compared to the proportion of residents in the community who are nonwhite, but the findings were broadly similar across each of those subgroups, with particular focus depending on the community.
I think the scale of it, I would say, was a surprise to me. To see that four of five hospitals treated a patient population that was disproportionately white was a surprise.
There are some areas where there’s a large Black population, but most of the hospitals did not see as many Black patients in elective care services as resided in the community.
The other three measures are related to this other concept called the ambulatory care sensitive conditions.
What we looked at was the rate at which patients were hospitalized for these conditions, patients who were Black versus patients who were non-Black residents, because it’s looking at a community-level rate where Black residents in the community are hospitalized at a similar rate to non-Black residents, or were they in fact hospitalized at a higher or a lower rate.
In about seven in 10 communities that we evaluated, Black patients were hospitalized at a higher rate, and in most of those communities, their rate was not only higher than non-Black patients locally, but also higher than the national average.
I noticed that cancer surgery is one of the services in which minorities are underrepresented. Do you have numbers specific to cancer surgery?
BH: Unfortunately, I don’t at this time, although that’s a good question and something we can try to look at in the future. We defined this panel of procedures as elective care procedures and pooled all patients across those procedures.
My teammates Tavia Binger and Ronan Corgel, who performed most of our equity analysis, have looked at, in the past, several different services that were more specific in building up to this work.
They were looking specifically at colon cancer disparities that have been identified by the research community as an issue, particularly related to the efficacy of screening rates in different populations.
They found, for example, that both Black and Hispanic Medicare patients were less likely than their white peers to undergo surgery after a diagnosis of colon cancer.
What are your sources of data for these health equity measures? Are they the same as the ranking data?
BH: Yes, for the most part they are the same, which is to say that we looked at Medicare data where we have tens of millions of records of individual encounters that patients have had with hospitals.
If Medicare paid the bill, we have a record of the visit and it’s all anonymized. So, we don’t know the names or anything else about the individuals. We just know which community they live in and what particular hospital they were treated at, but that’s it.
Their privacy is secured, but from the perspective of what we need, we can identify which hospital they were treated at, where they’re from broadly within that community. And so, that was the data that we used for most of this analysis.
For the measures, looking at ambulatory care and preventable hospitalizations, we actually used data that was generated also from Medicare data, but it was published by the Dartmouth Atlas, which is a long running academic project at Dartmouth.
And so, Tavia used data from that project that was publicly available to supplement the data that we had directly from Medicare.
At this point in time, you don’t include Medicaid or private claims data; right?
BH: That’s right. And some would say that that’s a limitation of our analysis, because obviously insurance differences have a massive impact on access to care and what care people feel they can afford.
On the other hand, I think it’s also a strength of our analysis in that by focusing on just the Medicare population, we take the effect of not having insurance out of the equation.
So, you would expect, since everyone we’re looking at has the same insurance plan, that we should not see disparities in access. Yet, we do.
That is most striking. Did you find specific regions, states, or even hospital systems in which these disparities are most obvious or severe?
BH: We found differences, for sure, regionally, and we focused much of our analysis on both the hospital and looking at that hospital in the context of its community.
So, we weren’t necessarily comparing disparities around the elective care, say, this community’s in worse shape than that one. We defined “community” as the “hospital service area,” which is actually a concept that was developed by researchers years ago.
Each hospital is assigned to a hospital service area. In suburban and rural parts of the country, it is often the case that there’s only one hospital in each hospital service area.
And so, if a hospital is the only hospital in the service area and it’s one of the four to five that has the disproportionately white elective care population, then that pretty much means that hospital service area, that everyone living in that community, we’re seeing less access, less utilization of elective care by nonwhite patients.
And to what extent that’s because of something the hospital is doing or not doing, we don’t know. This is really just descriptive, but it does reflect differences in access to that care for that entire community.
In other communities, predominantly some more densely packed urban communities, there may be multiple hospitals in some cases, even a dozen hospitals or more in a community, in a hospital service area.
And in those parts of the country—particularly where we see large minority populations in both cities in the South and in cities that received either large immigrant populations at some point or were destinations of the Great Migration—what we see is that there are, in many cases, one or few hospitals that serve a population that is reflective of the demographic makeup of the community, or even treats a disproportionately large nonwhite population.
And there are many other hospitals in that community, in some cases, that treat a disproportionately white population. And so, what we see is that there appears to be hospitals that tend to treat nonwhite patients and hospitals that tend to treat white patients.
Even within the same community we’ll see both types. That won’t surprise anyone in health care.
Many cities have a safety net hospital where lots of uninsured patients go, where lots of Black and Hispanic patients go—and where lots of people who’ve historically been disadvantaged by society go, whether it’s structural racism or linguistic barriers—those safety net hospitals, in general, we see are the ones that are indexing higher on our measures.
They are treating more minority patients, racial and ethnic minorities than reside in their community. Other hospitals, particularly those specialty hospitals that focus on surgical care, tend to treat more white populations.
The remaining 20% of hospitals with good representation of the demographics in their service areas—are these hospitals mostly located in urban centers, and are they mostly hospitals that provide care to a disproportionate share of minority residents? Who are they, and where are they?
BH: That’s a great question. I would not say that there’s a clear pattern there at this point.
We see some of those hospitals in rural areas or in areas where there’s only a single hospital in the HSA and they have a population that is pretty representative of their community.
But we also see them especially in these urban areas: the safety net hospital, or one of a few hospitals where patients from racial ethnic minority groups tend to go. In some cases, most of the other hospitals that are near them are at the far end with a disproportionately white population.
No, I wouldn’t say that there’s a clear urban/rural distinction or based on population size. It really does vary.
You’ve talked about expanding your measurements of health equity in the future. What might those expansions entail?
BH: We think of these measures that we’re publishing this week as measures of access, and they look at when, where and for what patients in different demographic groups are utilizing hospital care, but there are many other facets of health equity that we’d like to look at.
In addition to access, we’ll look at outcomes as a second domain of health equity. Are patients who are Black or Hispanic, or economically disadvantaged achieving the same outcomes at a hospital as patients who are from the majority group? And then social determinants being the third domain.
Those three domains are not mutually exclusive—there’s overlap and there are things that may be worth measuring that don’t necessarily fit neatly into one of these. But each represents an important category of equity.
So, between access, outcomes, and social determinants—there are different terms for this concept, social factors or social drivers—what are the actions that hospitals are taking to meet the needs of the population, of everyone who they could serve, before those people need hospital care? And ideally that’s what prevents illness rather than restoring it when they are sick enough to need hospitalization.
Indeed, you can look at so many variables—education, socioeconomic status, etc.
BH: Exactly, right. Food insecurity is one that frequently comes up when I’m talking with hospital leaders. Homelessness, housing insecurity, all of those things, and environmental threats, whether it’s triggers for asthma, air pollution, and so on.
Will these health equity measures be factored into overall ranking metrics in the future?
BH: It’s possible. That’s not something we’ve decided on. We are looking forward to a healthy dialogue with many different stakeholders.
I mean, we’re publishing these measures this week. This will be the first time that many people have seen them. We look forward to having conversations with equity researchers, patient advocate groups, with healthcare organization leaders, about what we have measured and what remains to be measured, and how those measures might fit into our future rankings.
So, I think that that future remains unwritten, but we look forward to having a far-ranging discussion about what the right way is to advance social health justice.
What do you hope your audience, and perhaps even policymakers, would do with your findings? What’s the urgency here?
BH: I think the time is now. The time has been now for a long time when it comes to disparities in health care.
These measures may not be actionable in and of themselves, but we think they’re part of the solution in the sense that both patients and the public and healthcare leaders need to understand these disparities if they’re going to address them.
There may be certain things that we see in our data that don’t necessarily require action, and maybe other things that require immediate remediation. So, it’s not for me alone to say which of these observations require the highest level of urgency.
But I think one thing that concerns many people, because we’re not the first ones to observe this or think of this, is the fact that when different hospitals treat different populations—if one hospital treats this disproportionately white population, and one treats this disproportionately Black or Hispanic population, or a low-income population—unless two hospitals have equivalent outcomes, you have a disparity simply because there’s a difference in hospital quality and they’re not treating the same population.
There have been a number of studies that have looked at racial differences in where patients get treated—the literature and the evidence shows that people of lower income status and of racial and ethnic minority background tend to be treated at lower quality hospitals, however you define lower quality, whether you use the U.S. News rankings, or you use some proxy like volume or academic status, or what have you.
Differences in the quality of hospitals that tend to treat different racial and ethnic populations, that’s potentially a big problem right there. So, if we can’t achieve equitable outcomes within each hospital, there’s an issue, but also we need to be able to achieve equitable outcomes across the population.
We have a second-order challenge if different demographic groups are being treated at different hospitals. So, I think just highlighting that is a really important aspect of what we’re doing, because to be honest, the hospitals that excel in many different areas, advanced clinical care, are not the same ones that we see indexing high on representation of nonwhite groups.
And that’s very telling?
BH: It is, yes.
I do realize that there are other announcements in this year’s rankings. Could you describe them briefly? What should we be paying attention to?
BH: I think that the broad theme is that, as it has been our mission for years, we want to provide decision support—essentially data-driven information that patients can use to make more informed decisions, that they and their doctors together can use to make more informed decisions and provide that to as broad a swath of patients as possible.
And from a clinical perspective, what that means is evaluating different services, so that a patient who has a heart failure can identify a hospital that can meet her needs, and a patient who has lung cancer and needs a resection can identify a hospital that can meet his needs.
We’ve expanded the portfolio of services that we evaluate hospitals in actually pretty substantially this year. And this is our largest single-year expansion since just around the time I joined U.S. News in 2007, that was the year we created the Best Children’s Hospitals rankings.
This year, we’ve added ratings in seven different procedures and conditions that we previously didn’t evaluate, and those are part of really some big ones. Pneumonia is one of them, which is more important in the middle of a respiratory pandemic than ever, but it’s always been an important component of hospital care.
These measures may not be actionable in and of themselves, but we think they’re part of the solution in the sense that both patients and the public and healthcare leaders need to understand these disparities if they’re going to address them.
Other new ratings this year include heart attack, stroke, kidney failure, in which there were also major racial and economic disparities. We evaluated hospitals in back surgery, spinal fusion. We’ve also evaluated hospitals in managing hip fracture cases, and in diabetes, a very important and highly prevalent disease.
Those are all new ratings that we’ve added this year. For some of those, we’ve evaluated 4,000 or more hospitals, pretty much every hospital in the country that provides care to patients with that condition or who need that procedure. And so, we’ve been able to add those facets to the total picture of hospital care that we provide.
All of those ratings are based entirely on objective data. So, we’re looking at risk-adjusted outcomes, at things like staffing and the availability of specific therapies, patient experience and so on.
In addition, we revamped one of our specialty ranking methodologies—we evaluate hospitals in medical rehabilitation. That had been one of the very few specialties that we evaluated based entirely on a physician survey, because we didn’t feel in the past that we had good objective measures to bring to bear on that particular specialty.
This year, we’re deploying an objective methodology that still uses physician opinion as a component, but we’ve added 11 objective measures as well.
Our results didn’t change that much, actually. I guess it goes to show that when hospitals excel on one measure of quality, even a subjective one like expert opinion, they often do excel in the objective measures as well.
But it’s an important step, I think for us and for the community of patients who need rehabilitation in a hospital that we’re able to now evaluate that in a rigorous, objective way.
It definitely sounds like you’ve outdone yourself in the middle of a pandemic.
BH: We’ve been busy. I got more done when I didn’t have to commute!
Working from home is great. Did we miss anything?
BH: I think that that really covers the highlights. I think with the expansion of our services, the other major thing that we’ve really been focused on in addition to equity this past year, is our hope that that serves a broader population of patients, so there are more patients who can identify the services they need.
There are still services that we don’t evaluate, including some areas of cancer care that I’d love to be able to address and shed light on in the future. There’s still work ahead.
Thanks for taking the time to speak with me.
BH: Thank you.
