If you're black and you're having a heart attack in an American hospital, the care you will receive is now almost as good as if you were white. There's an 86 percent chance you will be undergoing cardiac catheterization within 90 minutes of arriving to the emergency department—which is the standard of quality care. If you were white, that chance would be 92 percent, a mere 6 percent difference.

If you're black and you have pneumonia, there's a 94 percent chance you'll be given antibiotics within six hours of arriving at the hospital. That's almost as good as the 96 percent chance if you were white.

The chance you'll be given a flu shot is 90 percent. Pretty much the same as the 93 percent chance if you were white. Pretty much the same.

These numbers were just published as part of a massive study in the New England Journal of Medicine, led by Amal Trivedi, an associate professor of medicine at Brown University. The researchers reviewed more than 12 million hospitalizations nationwide between 2005 and 2010 and describe the results as "heartening."

They are heartening in that the numbers represent improvements. In 2005, the difference between white and black heart-attack patients getting timely cardiac catheterization was more than double what it was in 2010. Other metrics improved similarly. Disparities in medical care provided to white and black patients in U.S. hospitals over the five-year period narrowed by 8.5 to 11.8 percent across various metrics, and by 6.2 to 15.1 percent between white and Hispanic patients. Trivedi and colleagues attribute their findings to more equitable care for patients treated in the same hospital, as well as performance improvements among hospitals that disproportionately serve minority patients.

So this is technically improvement, and it will be reported as improvement, but it's the kind of improvement that distances itself from the designation. It is a reduction in injustice, not cause for celebration. These numeric gains look big, but these metrics—what is done for an imminently sick person when they are in the hospital—are a poor proxy for the deepest, insidious health disparities across races in the United States. The healthcare system is one that, as recently as World War II, would not mix donations in blood banks from white and black people, despite the empiric fact that the blood is biochemically identical. And the race problem in American healthcare is a problem for every American.

* * *

The New England Journal today simultaneously published a study that is much less heartening. It showed that racial disparities in preventive care—measures taken to keep people healthy and out of the hospital—persist, largely unchanged between 2006 and 2011. John Ayanian, a professor and director of the Institute for Healthcare Policy and Innovation at the University of Michigan, and colleagues found that elderly black Medicare enrollees were substantially less likely than their white counterparts to have adequate control of their blood pressure, cholesterol, and blood sugar.

The researchers noted that as of 2008, life expectancy was 5.4 years shorter for black men than white men in the United States, and 3.7 years shorter for black women than for white women. Among men, 38 percent of that racial gap is accounted for by cardiovascular disease and diabetes. Among women, the number is even larger, 54 percent. Poorly controlled blood pressure, cholesterol, and blood sugar are understood to underlie much of that mortality, and much other morbidity.

The studies together show that we have been able to improve general quality of care for a limited number of conditions in hospital settings, but the more important metric, health outcomes, remains largely unchanged. According to Marshall Chin, an internal medicine physician and professor in healthcare ethics at the University of Chicago, we still have a long way to go before we actually improve outcomes and reduce disparities.

What Trivedi's study examined were limited measures that are easier to improve than long-term health outcomes. In the case of Trivedi's study, Medicare required hospitals to monitor and publicly share their performance statistics on the metrics that the researchers examined. Medicare linked hospital reimbursement to how well hospitals did on those metrics. So they had clear incentives to do better. But right now, according to Chin, there aren't many incentives to reduce disparities.

Specialties like preventive care and primary care, where these disparities are best addressed, are poorly reimbursed. Doctors in those fields earn much, much less than their specialist peers. And on a larger scale, there is very little incentive for healthcare organizations to work on addressing social determinants of health: the social and economic drivers of disparities, starting with creating healthy communities.

"Unless we have further targeted interventions, we won't make as much progress as we can for reducing disparities," said Chin. "Healthcare systems and hospitals do what is incentivized, so we need to incentivize reducing disparities."

External incentive models like public reporting of outcomes and pay-for-performance programs may ultimately, though, exacerbate healthcare disparities—particularly if they inherently penalize institutions that disproportionately treat underserved and underprivileged populations.

Ayanian's study looked at cardiovascular risk factors that require efforts from patients and clinicians over years, as opposed to just prescribing medications. Caring for things like blood pressure and blood sugar require thinking about how people can live healthy in a global sense: how to eat well and exercise well and sleep well. These are much harder and more important outcomes than, say, "Were you given an aspirin in the hospital?"

A relatively small number of hospitals care for the majority of racial and ethnic minority patients in the United States, referred to as "safety net" hospitals, end up taking care of the uninsured and people with poorly-paying insurance like Medicaid.

"We need to be careful that if we develop this pay for performance model, we don't unfairly penalize these safety-net hospitals," said Chin. For example, if your hospital cares for a relatively large number of homeless patients, and other people who can't afford their medications, odds are that patient won't do well as an outpatient. They will bounce back to the hospital and be readmitted, and that will look like a failure. "So we need to judge the hospitals by the quality of the care they provide," said Chin, "but take into account that if they have a more difficult patient population to care for. You don't want to create incentives for people not to care for those patients."

Careful design of a pay-for-performance system happens by providing more resources to safety net hospitals—an attempt to level the playing field, in some ways. The National Quality Forum also recommends that programs adjust for the socioeconomic status of patients when scoring outcomes, to avoid comparing hospitals in affluent areas directly to safety-net hospitals. It can also be beneficial to reward improvement, like reductions in disparities, as opposed to simply overall outcomes.

"Until we go from a fee-for-service system that rewards volume of procedures to something that rewards preventive and primary care and organizations that address the social determinants of health in the community," said Chin, "we can only go so far."

Trivedi's article may be, more than anything, evidence that physicians and healthcare systems will do what is incentivized.

"I know a lot of people with a pessimistic sense," said Chin, "looking at the Ayanian study and seeing that health outcomes didn't improve. The thing is, though, compared to 10 years ago, we actually know a lot about how to reduce disparities. At this point it's basically about having the national will to make reducing disparities a priority."