Erasing Racial Data Erased Report’s Truth

Do black Americans receive poorer healthcare than whites?

Two years ago, a National Academy of Sciences panel on which I served concluded that the answer was yes. At the behest of Congress, we had reviewed hundreds of research studies, gathered diverse views and issued a report documenting widespread racial disparity in dispensing medical care.

The Bush administration promised to eliminate this inequity, and federal researchers drew up a report card on “prevailing disparities” in healthcare. This assessment, ordered by Congress and completed last summer, confirmed that racial and socioeconomic disparities were “pervasive in our healthcare system,” and that minorities received poorer care and were more likely to die avoidable deaths from cancer, cardiac illness, AIDS, asthma and other diseases.

But this report was never published. In its place, the Department of Health and Human Services issued a cheery rewrite touting administration successes and asserting that claims of minority groups receiving worse care than whites were unproved.

What happened?

Health and Human Services Secretary Tommy G. Thompson said it was a “mistake.” He told a House Ways and Means Committee hearing last week, “Some individuals [in the department] took it upon themselves, that thought they were doing the right thing. They wanted to be more positive.... “ He said that when the matter came to his attention “a couple of weeks ago,” he ordered his staff to “put ... out the original report just the way it was.”

But according to sources in the department and internal correspondence, Thompson twice refused to approve versions containing the findings on racial disparities in healthcare. Senior department officials objected that these findings were “inappropriate and misleading.” Rewrites were ordered in July and again last fall, according to these sources.

A side-by-side comparison of the original version, which I’ve obtained, and the approved report that was released two days before Christmas reveals just how stunning the makeover was. All findings of racial disparities were omitted in the Dec. 23 report. Although conceding that the healthcare Americans receive varies according to race and class, the revised document rejected the “implication” that these differences “result in adverse health outcomes” or “imply moral error ... in any way.”

To make its case, the report cherry-picked isolated examples of better medical outcomes among minority groups and sidestepped overwhelming evidence that blacks and Latinos received poorer care. Some of this evidence was downplayed; much more was simply ignored.

These wholesale changes shrink the responsibility of healthcare institutions for racial disparities in the delivery of medical care. And therein lies the key to how the rewrite came about. Among conservative health-policy thinkers, belief in personal responsibility for health runs deep. Claims of racial inequity in health and medical care are anathemas because they tend to point blame away from patients and toward doctors, hospitals, health plans and the government. To these conservatives, talk of racial disparity in the health sphere -- and talk of public- or private-sector initiatives to reduce it -- distracts from the work of inspiring citizens to take care of themselves. As President Bush has said, “Better health is an individual responsibility.”

The “individuals” who Thompson said “took it upon themselves” to rewrite the report shared this outlook. This made them deeply skeptical of the original report’s conclusions. They contended that findings of a link between race and quality of healthcare were unjustified because he report failed to show that race mattered by itself, apart from social class and insurance status. They noted that African Americans and Latinos, on average, had lower incomes and less generous insurance than whites. These factors could explain the racial differences gleaned from the researchers’ raw data, they said.

Suppose for the moment that racial differences vanished when these factors were fully taken into account. Would the racial disparities seen in the raw data then be unfit to report? Yes, if the link between being black (or Latino) and being poorer (and less well-insured) is understood as part of the natural order of things and not a social problem. But no, if racial disparities in healthcare delivery are wrong when they arise from racial gaps in wealth and insurance status.

Science can’t resolve this dispute. It’s a political and moral issue beyond the reach of statistical methods. But Congress has given its answer: In telling Health and Human Services to track “prevailing disparities” arising from both race and class, it defined racial disparity as a social problem -- whether it stems from class differences or not. Congress instructed the department to report on it, and that’s what the researchers did.

The senior officials who ordered the rewrites disregarded this instruction. And in highlighting “good news” (examples of minorities doing better) and submerging the prevailing “bad news” (poor and minority populations doing worse), the administration violated its statutory duty.

Moreover, there’s plenty of evidence that racial disparities persist even after social class and health insurance are taken into account. The panel on which I served, under the auspices of the Institute of Medicine, a part of the National Academy of Sciences, reviewed many studies that separated out the effects of income, education and insurance. The studies concluded that race mattered independently. They were the basis for our finding to this effect, a finding endorsed by a dozen peer reviewers.

The HHS researchers who compiled the report card embraced this finding as definitive. Rather than retesting it, they developed a large number of quality-of-care benchmarks to track racial disparities. According to my Johns Hopkins colleague Darrell Gaskin, a leading disparities researcher whom I asked to review the original version, this approach reflected state-of-the-art understanding in the field.

But the officials who ordered the rewrite would have nothing of it. They wanted the researchers either to retest with their own data the conclusion that race matters independently -- something they were not equipped to do -- or to refrain from drawing it. This broke with standard scientific practice, which builds on previously published work.

Those who sought the rewrites feared political blame for the inequities the researchers found. They missed an opportunity to take credit -- by laying out the bad news and responding with initiatives to remedy it. Health plan executives have said they’d act if demand arose. A high-profile federal report card on healthcare disparities could prompt patients to insist on progress.

Individual responsibility can also make a positive difference in health policy, more so than some activists acknowledge. The proper scope of personal responsibility is a question at the heart of our politics, and politically appointed officials are entitled to press their views on this question. Yet, there must be limits to the role of politics in science if science is to remain relevant to the social questions that beset us. Researchers mustn’t ignore facts that don’t fit their premises. Beyond this, the value choices that nudge scientific judgment should be open to professional and popular scrutiny. Scientists must be free to state their views and to challenge others without worrying about political retaliation. And when Congress asks federal research agencies to report on scientific questions, the law should guarantee these agencies more independence than the HHS researchers received.