Medical journals do not ask if sex affects results

Our contributing health editor continues her three-part report on the failure of medical researchers to analyze their results by gender.


(Second of three-part series)

The editors of medical journals fail to note whether researchers publishing results in their pages have followed federal law requiring data be analyzed by gender, researchers at the Society for Women's Health Research said last week at a news conference. If they did, the editors could make a substantial contribution to women's health.

A study published by the society in its June Journal of Women's Health and Gender-Based Medicine examined 865 articles in four leading peer-reviewed medical journals. The authors found that 21 percent of the 865 studies failed to include women, despite federal regulations requiring researchers to do so.

And the study showed only 27 percent of the 674 studies that included women analyzed the data by the sex, another violation of federal policy.

This study follows on the heels of a General Accounting Office report that found “scant” evidence researchers who received funds from the National Institutes of Health had analyzed their results by sex-also a violation of federal law. (To read part one of this series: Despite Law, Gender Ignored by Medical Research by Melinda Voss, use this link.

Findings from both the studies run counter to efforts that began in 1986 to include women in clinical trials. In 1993, a federal law took effect requiring researchers to include women in clinical trials and to analyze gender differences, said Regina M.Vidaver, the lead author of the study and the society's scientific programs manager.

Analyzing gender differences is critical to improving health care for men and women, because so little is known about whether the genders respond differently to various diseases and treatments, the society contends.

Research has already ferreted out some differences, however. For example, women are more apt to feel abdominal, jaw or back pain before a heart attack, rather than the severe chest pain that men tend to experience.

The medical journal researchers looked at results published in 1993, 1995, 1997 and 1998 in the New England Journal of Medicine, the Journal of the American Medical Association, the Journal of the National Cancer Institute and Circulation.

The authors focused on articles about federally funded research of medical conditions that occur in both males and females, such as Alzheimer's disease. So, studies of such sex-specific conditions as ovarian cancer were not included in the analysis.

“Unfortunately, the study results show little or no increase in the proportion of studies that report sex analysis since the first recommendation appeared in 1986,” said the society's Vidaver.

The study's authors minimized one possible shortcoming of the study-that is, that many of the studies they reviewed could have began before the 1986. To determine if that was a problem, the authors contacted 18 investigators of studies published in 1998. Only one researcher said the study in question began before 1986.

In light of these findings, the Society of Women's Health Research sent letters to editors at 32 medical journals asking them to strengthen their publication criteria regarding women's involvement in clinical research.

“We believe that questions about the availability of women subjects and analysis must become part of peer review and editorial decision-making,” said Phyllis Greenberger, the society's executive director.

At least one journal editor agreed.

Dr. Catherine DeAngelis, editor of the Journal of the American Medical Association told Women's Enews earlier this month that she checked the last 20 issues of her journal for gender analyses because the study's findings surprised her.

Of the three clinical trials reported in those 20 issues, she said, none analyzed data by gender.

“When I looked at them, I said, ‘Gee, it would have been nice to have it separated that way.',” she said. As a result, she said, “From now on, we will ask authors to analyze the data for women separately from men to see if there's a difference provided there's a large enough number in the study.”

In fact, DeAngelis said, researchers should analyze their data to “every extent possible, not only for women, but for different racial groups, for different age groups. The problem is to get sufficient numbers of any one of these groups.”

Dr. Gregory Kurfman, deputy editor of the New England Journal of Medicine, took a more cautious approach. “Journal editors don't do the research and we don't design the studies. We select publications based on their merit. We're as interested in the study of women as we are in the study of men,” he said, noting the journal has published parallel papers looking at men's and women's health questions. “The idea that there are differences in men and women is widely appreciated,” he said. “I think we need to be looking for these differences and designing studies to address them for sure.”

Editors at the other two journals, Circulation and The Journal of the National Cancer Institute, could not be reached for comment.

The society's Vidaver acknowledged that including women in sufficient numbers for gender analysis can be tricky, even for the best-intentioned researchers.

“But we don't want this to be a back burner issue,” she said. “We want this to be the first thing researchers think about when they're recruiting study participants.”

Next week: Third part of this series: What obstacles prevent women from participating in clinical trials and what can be done to overcome them.

Melinda Voss has her master's degree in public health. A former reporter for The Des Moines Register, she freelances and teaches journalism at the University of Minnesota.