Same hospital, different quality
Medicaid beneficiaries have more patient safety events than privately insured patients at the same facilities. A researcher talks about why that might be and what hospitalists can do.
It can be comforting to blame the health care system as a whole for disparities. When studies show that patients of lower socioeconomic class or Black race have worse outcomes, as they often do, the easy answer is that the patients had less access to care, were treated at lower-quality hospitals, or faced other systemic barriers to good health—still problems, but not ones for which physicians are likely to feel responsible.
However, a recent study disproves those explanations. An analysis of inpatient records from 26 states found that Medicaid beneficiaries had significantly higher rates of eight of 11 adverse patient safety indicators than privately insured patients in the same age range (19 to 64 years). For example, respiratory failure and sepsis developed in 2.9 and 2.5 more Medicaid patients per 1,000, respectively, compared to privately insured patients.
Controlling for patient factors failed to eliminate most of these differences, as did adjustment for which hospital treated the patients. “This indicates that the overall differences in the quality of care that Medicaid patients and privately insured patients receive is related to unequal treatment at the same site of care,” stated the study, which was published by the Journal of Patient Safety on Aug. 28.
ACP Hospitalist recently spoke to the study's author, Anuj Gangopadhyaya, PhD, an assistant professor of economics at Loyola University Chicago.
Q: What motivated this study?
A: Quality is hard to measure in the health care setting generally, but patient safety is a more tangible and objective measure that we can assess. Adverse safety events are never events. They shouldn't be happening in hospitals.
I led several analyses on the topic of patient safety, mostly surrounding differences in care delivered to Black and White patients. … Black patients experience higher rates of adverse safety events relative to White patients. We tend to see Black patients accessing hospitals that have higher rates of adverse events overall, but also, when in the same hospital as White patients, facing higher adverse safety events.
Trying to disentangle what the driver of that disparity might be took me into this analysis. One of the theories was that it could be linked to the ultimate payer for each group. Black patients have higher rates of Medicaid coverage relative to White patients. Medicaid, as a payer, reimburses providers less than private insurance and Medicare.
Q: What did you think of the findings? Were you surprised by any?
A: I expected, based on prior literature, to see higher rates of adverse safety events among the Medicaid population. What was fairly surprising is that, at least in terms of payer disparities, the evidence points to differential care within hospitals as the main driver of differences. It's not a story in which Medicaid patients are having trouble accessing safe facilities. Your payer type isn't necessarily playing a big role in steering you towards better or worse facilities. Outcomes tend to correlate with differential care within the same hospital given to people with different coverage. In this paper, I don't tackle exactly what that mechanism is, but if you're trying to shape policy, you're going to have to look at the way care is delivered within the hospital.
Q: Do you have theories about the causes of differing outcomes within the same hospital?
A: I can tell you what it can't be. The analysis controls for Medicaid versus private insurance patient differences in age, sex, racial/ethnic background, state of residence, comorbidities, and diagnostic conditions at the point of admission. When you rule the patients' characteristics out, the question becomes how are care teams being allocated to patients? Is there anywhere in our resource allocation that payer type is coming into play? How are different kinds of doctors being resourced to different kinds of patients? Do their own quality measures differ based on the patient that they're working with? That can open up entire questions about physician biases and discriminatory care, but it can also be systemic-, structural-, institutional-level aspects.
Q: What advice would you give physicians about their role in this?
A: That's a harder question. I've talked to physicians about this after those earlier studies. The first people who reach out to me are not saying, “Hey, this is eye opening.” It's more a defensive take. I understand that. I could stand here and advocate for nondiscriminatory care. That's an easy thing to say. Leveraging that on the ground level is hard. You can talk about policies that seem to [incorporate] good practices: Make sure that your hospitals have a chief equity officer, for example. Or proactively asking why are conditions occurring differently for certain subgroups in our population?
I can't speak to why the problem exists, but I do think that doctors can see whether it exists in their own hospitals—on the ground, which subgroups have higher rates of things that we shouldn't expect? That's the first step. Most policies aimed at promoting quality of care in the United States target overall quality at a hospital level, and I think we're starting to understand that's probably not enough. The policy should be leveraging carrots and sticks around the quality delivered by subgroup, whether by race and ethnicity, by sex. This is not a new problem in American health care, but I want it to be less of a problem.
Q: There's been increasing attention in recent years to disparities in care by race and ethnicity. Have we made any progress on those?
A: When you expand to race and ethnicity, the conversation becomes a little different from the payer one. In previous research I've seen [on race disparities], the aspect of being able to access higher-quality institutions does seem to emerge, whereas I don't see that in the Medicaid versus privately insured populations. Why certain patients get into specific hospitals is a tougher question, and it takes a whole different set of policies to address. Since the Affordable Care Act in 2014, we've done a really good job in reducing disparities in coverage. That, I thought, would reduce a lot of disparities in access. I'm not seeing a lot of differences in quality outcome disparities despite these massive changes in coverage. Dancing around the edges with those policies doesn't seem to be doing much to narrow disparities in quality of care received. We almost need to open up the box again and ask ourselves what can we do?
Q: There's also been a focus on penalizing hospitals for adverse events. Is that working?
A: I think those penalties for quality, withholding Medicare payments, are fine and good. It goes back to the idea that you might want to leverage those penalties on not just overall care, but also take into account differences in the levels of quality delivered to subgroups of patients within each institution, to the extent that you can measure those.
Q: What additional research are you planning on this topic?
A: I recently partnered with The Leapfrog Group to further conduct analysis in disparities in care by overall hospital care. They apply a letter grade to each hospital, based on overall rates of patient safety and hospital-acquired conditions. We took the underlying discharge data for a large share of those hospitals and asked, “Grade A hospitals, do they deliver equitable care to each of their patient groups?” This was an analysis by race and ethnicity. The answer was no, the Black patient accessing an A-level hospital can only expect to receive better care relative to a Black patient accessing a B-level hospital. There's this consistent premium that we're seeing that White patients are able to access within institutions, even high-grade ones, that Black patients don't get at those same institutions. The ultimate goal of this analysis is to keep pushing CMS into discussing and thinking about ways that we can shape policy to reform the inequities we see.
Q: Do you think we could potentially see hospitals being rated or reimbursed based on the equity of the care they provide?
A: At least as an experiment, to see if that's an effective way of remedying some of the differences. You can also ask private insurers, given that patient safety events are going to run up their bills, too. How can private payers get into this game? I hope successful payment models on the private side can generate spillover reform on the public side as well.
Q: Is there anything else that you want hospitalists to take away from your work?
A: Look closely at the study. It's ruled out a lot of potential causes, [including] patient characteristics or steering towards institutions. So it really becomes a focus on why is care being delivered differently within your institutions? It could be a resource allocation question or a care team allocation question. Just asking those questions honestly is probably the next step.
