The Women's Health Initiative-Failure or Success?

What to do about women? This question is at the heart of the Women's Health Initiative. Perhaps the appropriate conclusion is that medicine has a lot left to learn. The results have challenged our assumptions about women's health, perhaps as importantly as the methods we use to find answers. Rather than highlighting our understanding, the study has exposed our ignorance. Women are both confused and angry regarding the recommendations being received from clearly conflicting data.

This study is a large, multicenter, double blind placebo controlled trial. The results are in conflict with what has been anticipated from epidemiologic cross-sectional observations. Medicine tends to exalt the trials and denigrate the observations. Doing this helps to resolve the sense of conflict. The recent discontinuation of the Prem/Pro arm of the study is distressful to many women, perhaps appropriately. Rather than dismiss the conflict, we should investigate the significance of these disparate conclusions, rather than simply dismiss the observational findings as inferior.

An important difference between controlled vs. observational studies can be found in the involvement of the women, including their subjective experience as an influence on the outcome. Observational studies are biased by the decisions of the participants as to the perceived benefit, which becomes part of the decision about continuing the therapy. This information is excluded in the trials through blinding and subjective experience is controlled by providing a placebo. The bias of the trial is then to exclude subjectivity as an influence on the outcome. Is it possible that what has been identified in these studies is the value of subjectivity? The observational studies identify a benefit to HRT in many aspects of health, while the trials seem to negate the observations for issues of coronary artery disease and stroke. An advantage of large studies is the identification of relatively small effects, with the assumption that small differences will become larger over time. However, the observational studies are already considering the question of length of treatment and reach the opposite conclusion that coronary artery disease is reduced.

The issue of control relates to who has the control. In the trial, it is the investigators; with the observational study it is the subjects. It should not be surprising that when women are not given the opportunity to include their subjective experience in the course of their health care that things do not go well.

Rather than ignore the significance of observational studies, we should make an effort to resolve the discordance through understanding of individual variance, both physiological and psychological. (The distinction between these two disciplines may be artificial and shrouded in our ignorance as well.) The Women's Health initiative highlights the need for approaches that can prospectively identify the differences between women and include the value of subjective experience. Rather than being controlled, women should be understood. Although the trials may be "double blind", it appears that only the investigators are also deaf because the women are listening and seem to hear the problem. We must somehow identify what is right for each person, not give "cookie cutter" advise to all women.

What can we do with the data obtained from the women's health initiative? It cannot be ignored. While recognition that the source of confusion may be inherent in the methods, we must make decisions with imperfect understanding. Those who prefer to act out of concern for the public health will argue that medroxy progesterone should not be used. Others will see the problem as generalizable to all HRT, perhaps motivated by self-interest. It is possible that many women will lose trust in advice from the medical establishment.

We must recognize the limitations of the data, while working with women to understand how the results apply to their situations, and ultimately strive for better knowledge. That knowledge could lead to better understanding, particularly in the context of careful observation at the level of each individual. The problems of women's health are characteristic of the conflict between public health vs. individualized care. This problem pervades our health care system, particularly in the arena of prevention.

The Women's Health Initiative can help us to recognize the problems encountered when attempting to apply knowledge acquired in studies of heterogeneous groups is applied to individual people. The Women's Health Initiative calls attention to the principle that women's health has more to do with understanding differences than identifying similarity. It is unfortunate that we will not have the definitive answers that were promised. We should be grateful to those women who are participating, not only for their contribution to the clarification of women's health, but also for the opportunity to more critically evaluate the limitations of the large double blind placebo controlled trials. The translation of medical science to individual cases requires the careful consideration of many factors including data, experience, and impressions of both the patients and providers. If we can learn this from the Women's Health Initiative, it should not be considered a failure, but a grand success.

Dr. Robert S. Fredericks
July 2002