Many studies document health care disparities -- the lamentable failures of the medical establishment to deliver the best available care to women, minorities and other groups.
But there is a flip side: When it comes to brand-new drugs and medical devices, it's sometimes good to be left out. That's because therapies that perform well in clinical trials often prove less effective -- and sometimes more dangerous -- when put to widespread use.
One reason, critics say, is another type of disparity: a lack of diversity in early clinical trials, which tend to enroll relatively healthy white males but few children, women, minorities and older patients.
As a result, the research ignores biological differences that cause people of varying ages, genders and ethnicities to react differently to the therapies. "We are extrapolating from a clinical-trial population that looks nothing like the general population," said Rita Redberg, a cardiologist at the University of California, San Francisco.
Consider the implantable cardiac defibrillator (ICD), a pacemakerlike device designed to quell dangerously erratic heart rhythms and thus prevent deadly heart attacks.
A recent Duke University study found that white men were more likely than black men or women of all races to receive the devices. That wasn't a surprise, but this finding was: The defibrillators appeared to do little to extend the lives of many patients.
"The bad news may not be for women and minorities, but for white men who are undergoing a procedure that ... has not been shown to extend their lives," wrote Redberg, in an editorial accompanying the study.
If the number of clinical-trial patients in a particular demographic group is too small, experts said, researchers might not be able to produce any statistically significant findings about whether a treatment actually benefits the group.
"When a problem occurs rarely, and you do the study in a couple of hundred people, you're not going to see the problems until later," said Dr. Richard Lange, a professor of cardiology at the Johns Hopkins School of Medicine.
When those problems do come to light, they often make headlines.
Merck's pain drug Vioxx, for instance, caused an uproar in 2004 after a study found it increased patients' risk of heart attack. The effect was found by analyzing a database of 1.4 million Kaiser Permanente members, including 26,748 who took Vioxx -- a diverse group compared to the participants in the early clinical trials on the drug.
Last spring, another study brought into question the effectiveness of drug-eluting stents, devices used to prop open clogged coronary arteries. While the stents performed well in early clinical trials, the newer study, which involved a more diverse group of patients, found they were no better than medications for treating patients with mild coronary artery disease.
Last month, a Food and Drug Administration advisory panel ruled there was no evidence over-the-counter cold medicines work for children, because most of the research was done in adults. In light of reports of children dying after taking the remedies, the panel advised they not be given to kids under age 6.
The defibrillators now in question are designed to prevent deadly heart attacks brought on by erratic heartbeats, or arrhythmias -- which account for about 150,000 deaths a year. When a patient's heart rhythm becomes dangerously erratic or rapid, the ICD delivers electrical shocks to the heart to jolt it back on track.
Early clinical trials indicated ICDs were more effective than drugs at treating patients whose heart problems put them at high risk for sudden cardiac death. Use of the devices expanded in 2004, after two clinical trials found they also extended the lives of healthier cardiac patients who had not suffered an arrhythmia.
That persuaded the federal Medicare program -- whose patients are generally older than 65 and increasingly likely to be female -- to broaden its coverage of the devices. In the first year after that happened, the program paid for 25,000 implants at $30,000 to $40,000 each.
But the studies on which that expansion were based included mostly white men who were relatively young and healthy. In one trial, 16 percent of the participants were women, the average age was 64, and no results were reported for black patients. In the other trial, 23 percent of the participants were women and 24 percent were nonwhites.
By way of contrast, the recent Duke study investigated the devices' effectiveness in a much broader population of Medicare beneficiaries.
The study focused on racial and gender disparities among those who received the devices, finding that white men were more likely to receive the devices than women and minorities.
"Sudden cardiac death is the leading cause of mortality in the U.S.," said Dr. Lesley H. Curtis, an associate professor of medicine at Duke and lead author of the paper. "But sometimes life-saving devices aren't used equally."
The study found ICDs extended the lives of high-risk patients who had suffered a cardiac arrest or an arrhythmia. But lower-risk patients, who had never experienced an arrhythmia, saw no benefit. Yet this was the very class of patients Medicare expanded its coverage to serve in 2004.
Curtis, who has been a paid consultant for ICD manufacturers, emphasized the study was not designed to rigorously address the issue of which patients benefit from ICDs. But other doctors said many patients might be getting the devices implanted needlessly.
"When it is applied to the general population, the outcomes are not as good," said Lange.
If future research confirms that's the case, experts said, low-risk patients might be better off sticking to medications instead. ICDs require an invasive operation to implant and have been known to deliver a shock to the heart at inappropriate times.
"People describe it as being kicked in the chest by a horse," Redberg said.
On Oct. 15, Medtronic said faulty wires known as "leads" in its newest ICD models might result in ill-timed shocks. The leads, which connect the device to the heart and were implanted in about 235,000 patients, were linked to five deaths.
Critics of testing procedures said the announcement highlights the importance of follow-up studies after a device -- and accompanying parts -- has gone to the market.
In addition to catching malfunctions or adverse reactions, they said, follow-up studies provide important data on whether drugs and devices work for different types of patients.
Why are there so few women and minorities in clinical trials? Experts said the reasons are complex.
Women in their childbearing years may be absent because of concerns about the effects of a treatment on a developing fetus, they said.
Experts say African-Americans tend to be generally more distrustful of the medical establishment than other groups -- the long-term fallout from incidents like the Tuskegee study, a 40-year experiment begun in 1932 in which researchers withheld syphilis treatment from black farmers.
Older patients are often excluded because they may suffer from a number of age-related conditions that might complicate the research. This, in turn, may result in fewer female participants, since women tend to be older than men when they develop certain illnesses.
Lange said the lack of diversity in studies puts the FDA, Medicare and private insurers in the difficult position of using narrowly focused data to decide which drugs will be used in a broad population of patients.
But the alternative -- only offering treatment to white males -- is equally undesirable, Lange said. "You can imagine the outcry," he said, "if Medicare said we're not going to give drugs to women and minorities because we don't have the data."
To address the issue, Congress passed a law in 1993 that encourages federally funded clinical trials to include women and minorities. The FDA also began encouraging companies to report race- and gender-specific trial results.
"The message is that we encourage diversity in research," said Terry Eigo, director of the FDA's Office of Special Health Issues.
She cited a 2001 FDA study of 185 new drug trials between 1995 and 2000 that found about half of the participants were women. But the same study found that the proportion of blacks in clinical trials declined from 9 percent in 1995 to 6 percent in 1999. Few other minorities participated, and Hispanic numbers were particularly low.
Eigo said the FDA stops short of forcing companies to include patients from every demographic, because it would slow medical progress. "We don't have mandates," she said. "If you required substantial numbers of people in every subgroup, the size of the trials would be massive. It's likely that the benefits would not be worth the cost."