Ten years after its scathing report on the National Institutes of Health's failure to include women in clinical research, the General Accounting Office (GAO) has concluded that the NIH is doing much better. Women are clearly taking part in clinical studies--in even greater numbers than men. And the amount of money devoted to diseases, such as breast cancer and depression, that disproportionately afflict women has risen steadily, outpacing increases in the NIH's overall budget. But NIH-supported researchers aren't always putting their data on women subjects to use.

The 2 May GAO report, amplified by a study in the June Journal of Women's Health & Gender-Based Medicine, shows that only a small fraction of publications based on NIH-funded research report a sex analysis of the data. "It's important to have women in clinical trials," says Phyllis Greenberger, head of the Society for Women's Health Research (SWHR) in Washington, D.C., "but not for the hell of it. The point is to do the gender analysis."

Such analyses are crucial, she and others note, because women respond differently to some drugs, carry a higher risk of certain diseases, and can present with disease symptoms different from men's. Without a comparison of the sexes, both men and women miss out on sensitive diagnostics and tailored treatments. The director of NIH's Office of Research on Women's Health, Vivian Pinn, acknowledges the importance of gender-based analyses but points out that NIH has no control over whether grantees carry out and report them. "We don't dictate editorial policies for journals," Pinn says.

Researchers had historically been reluctant to include women subjects, says molecular biologist Regina Vidaver of the SWHR, because they didn't want to deal with potential birth defects or variability in responses due to hormonal changes during the menstrual cycle. The problem was exacerbated in 1977 after the Food and Drug Administration barred women of childbearing age from participating in early clinical trials because of fear of birth defects. In 1985 the U.S. Public Health Service pointed out the obvious repercussion: The lack of information could seriously compromise health care for women. The NIH, in response, urged researchers to include women in their clinical studies.

Not much had changed by 1990, according to a GAO study, which along with outrage over studies showing the benefits of exercise and of aspirin for preventing stroke--conducted in men only-- prompted NIH to act. By the end of the year, the NIH began to require the inclusion of women and minorities in research. In 1993, Congress passed the NIH Revitalization Act that established guidelines for accomplishing that.

By many measures, the past decade's efforts have succeeded. Women constitute 62% of all subjects in clinical studies funded by NIH grants to outside researchers. For phase III protocols, the last stage of clinical trials before a treatment is approved for widespread use, 75% of subjects are women. Even when sex-specific studies, such as those focusing on ovarian or prostate cancer, are excluded from the analysis, more than half the remaining subjects are women.

The GAO report cautions that NIH bookkeeping methods preclude a detailed analysis of funding for women's health research. But funding for some conditions that disproportionately affect women grew steadily between 1993 and 1999. Research expenditures went up 78% for osteoarthritis, 59% for breast cancer, and 73% for depression and mood disorders. For comparison, the overall NIH budget rose by 29% during that time. The NIH also collaborates with other federal agencies in the Women's Health Initiative, a study of 164,000 postmenopausal women that is examining the effects of hormone replacement therapy, diet, and vitamins on cardiovascular disease, breast cancer, and the bone-thinning disorder osteoporosis.

The problem, Vidaver points out, is that NIH-funded researchers still aren't breaking down their data by sex. Her team examined hundreds of NIH-funded, non-sex-specific studies published in The New England Journal of Medicine, The Journal of the American Medical Association, the Journal of the National Cancer Institute, and Circulation between 1993 and 1998. Of those, 80% included women. But only 25% to 30% of the studies with women subjects reported a gender analysis, which may have been as brief as a statement that there were no significant sex differences. That share remained constant over the period, Vidaver adds.

NIH's Pinn says that the analysis is premature, because the studies would have been funded before the 1993 law was implemented. But Greenberger counters that the NIH began urging the inclusion of women starting in 1986, adding that "we thought at least we'd see a trend" toward more gender analyses.

The society isn't waiting for NIH to make things happen. On 25 May, it wrote to the editors of 32 leading journals, calling on them to revise publication guidelines to require a sex analysis.