The three preceding chapters discuss the committee's findings about and recommendations for each component of the Women 's Health Initiative (WHI). In this chapter the committee would like to address the WHI as a whole by reflecting upon the questions posed in the IOM Statement of Task, as well as interconnected concerns that relate both to the societal context of the WHI as well as to the IOM committee assessment.
RESPONSE TO THE STATEMENT OF TASK
appropriate topics considered? Should other topics be included?
NIH has provided a justifiable rationale for the diseases selected for study in the WHI. These diseases are important causes of morbidity and mortality among women in the United States; they merit further research to test the efficacy of preventive measures and to develop effective programs that educate and motivate women to adopt proven prevention strategies. Heart disease and cancer are the leading causes of death among women, and the high incidence of breast cancer is appropriately a matter of concern. Osteoporotic fractures, which occur with much higher frequency among women than men, are a leading cause of disability among older women. Research leading to prevention of these conditions, or at least postponement of the age of onset, would be expected to improve substantially the quality of life of older women.
NIH could have chosen other diseases or risk factors for ill health as the focus for the WHI, and those choices could also be defended. There are many unanswered questions regarding prevention of illness, enhancement of well-being, and delay of morbidity and mortality while maintaining an acceptable quality of life.
Because the WHI is among the most complex studies ever undertaken, and because the WHI Clinical Trial (CT) is already under way, it would not be appropriate to suggest other topics to be included. This might overburden an already complex set of studies, could incur additional costs, and would almost certainly lead to further delays. Moreover, as noted above, it was not the committee's task to redesign the WHI.
Are the methodologies and the study designs appropriate to address the stated research objectives? Is the size of the study population appropriate?
The methodologies selected—a clinical trial, an observational study, and community studies—are appropriate for the study of the efficacy or effectiveness of certain interventions to improve women's health, for generating further hypotheses, and for implementing and evaluating community intervention strategies. Because details of the Community Prevention Study (CPS) have not been decided, little can be said about the appropriateness of the specific methods and designs of the study. The majority of NIH staff time and effort has been devoted to designing the CT; therefore, it is the focus of most of the comments and recommendations in this report.
Three primary hypotheses are the basis of the CT:
A low fat dietary pattern will reduce the incidence of breast cancer and colorectal cancer;
Hormone replacement therapy will reduce the incidence of cardiovascular disease; and
Calcium and vitamin D supplements will reduce the incidence of osteoporotic hip fracture.
The design of the trial, the number of women who will participate, and the methods used to determine health status all flow from these hypotheses. Some of the design decisions are based on evidence obtained from earlier studies, and some decisions are based on assumptions. Assumptions must be made regarding many aspects of any clinical trial, and examples include the strength of the protective effect, the number and timing of the endpoint events (occurrence of heart attacks, hip fractures, breast cancer), the ability to recruit and retain the study participants, and the ability to maintain adherence and behavior change in the intervention groups.
To determine whether the proposed methods for the CT were appropriate for the stated research objectives, the committee analyzed the assumptions underlying the methods for each hypothesis and identified the major threats to the successful achievement of each hypothesis. The committee then weighed the risks of an unsuccessful trial against the potential benefits of additional information to be learned from a successful trial.
Using this approach, the primary hypothesis that a low fat dietary pattern will reduce breast cancer and colorectal cancer incidence was judged by the committee to be inappropriate for emphasis as a primary hypothesis in the trial. The primary hypotheses of the HRT branch and the CaD branches were judged to be appropriate.
Several secondary hypotheses are also included in the design of the CT. Most prominent is the hypothesis that a low fat dietary pattern will reduce the incidence of cardiovascular disease. Using the same approach of weighing the threats of an unsuccessful
trial against the benefits of a successful trial, the committee judged this secondary hypothesis to be appropriate for inclusion as a primary hypothesis in the trial.
These conclusions led to specific recommendations about the design and conduct of the CT, which are presented in chapter 2.
Are the costs accurately estimated and sufficient to reach the desired objectives?
Cost estimates are fairly well developed for the CT, deducible for the Observational Study (OS), and absent for the CPS. To a large extent, this reflects the stage of development of each component. It is therefore difficult to say with certainty that overall costs are sufficient to reach the objectives of the WHI.
The CT appears to be very tightly budgeted. Although the approximately $1,000 per participant per year costs are within the range of costs (unadjusted for inflation) NIH incurred for other clinical trials conducted over the past 20 years, they are very low relative to the costs of more recently conducted comparable trials.
As noted earlier, the committee believes that the total costs of the CT will be greater than the $625 million provided by NIH. NIH and Vanguard Center representatives have indicated that the additional funds necessary for successful completion of the trial will be covered by the institutions in which the Vanguard Centers are based. This reliance on institutional support may be reasonable in the case of the Vanguard Centers, but the committee felt it is unlikely that an additional 29 institutions can be identified that have both the experience to carry out the tasks of high quality research and the ability to provide additional resources. If all participating institutions honor their agreements to provide additional support for the CT, and if there are no unanticipated problems, the budgeted amount might be sufficient. However, it is very unlikely that there will be no problems.
Will the study produce sufficiently reliable results to justify the costs?
The committee concluded that valuable scientific information could be obtained in the redesigned study. However, the question of whether the investment is justified is, in part, a question about cost-effectiveness, and that involves consideration of alternative study designs. The committee thought it likely that much of the information could be obtained in better designed, smaller, more focused studies that could have a greater chance of success and probably be less costly. However, it recognized that a study in a broad-based population can have merit and proponents as well. The committee did not consider its task to be a consideration of alternative designs, and did not do so. Thus, the committee admits some skepticism about the merits of this particular investment, but it cannot offer a definitive conclusion about cost-effectiveness.
It is important to note that the modified study that the committee felt could be justified scientifically is quite different in its aims from the study proposed by NIH. The
study proposed by NIH has as a primary focus the test of whether DM will reduce the risk of breast cancer. The modified study focuses on CHD. If NIH decides to accept the recommendations of the committee about modification of the study, it will need to consider whether the methods and study population proposed to test interventions on the risk of both CHD and breast cancer should be the same as those for a test of effects on the risk of CHD. Moreover, NIH will need to consider whether the study objectives that can be achieved in the modified study are worth the investment.
Whatever the merits of the WHI, the committee has no doubt about the need for a substantial investment in research on women's health.
The committee contemplated the societal context of the WHI throughout its more focused discussions of study rationale, design and methods, and cost, specifically the following issues:
efficacy and effectiveness,
public health choice versus individual choice, and
considerations beyond science.
Efficacy and Effectiveness
In assessing whether a study is successful, one looks at whether it added more evidence toward answering a question. Obviously, whether the question is posed appropriately affects this assessment. It is not clear to the committee whether the WHI CT is designed as a test of efficacy or a test of effectiveness. And therein lies much of the difficulty in interpreting criticism of the CT.
Framed as a test of the efficacy of a low fat dietary pattern in reducing the incidence of coronary heart disease, for example, the CT will be plagued from the start. If there is no difference in coronary heart disease incidence between intervention and control groups, does one conclude—as an efficacy trial is set up to do–that the intervention is not associated with a change in outcome? Or might one state that, as followed, the dietary intervention did not have an effect? To be able to state the efficacy finding with assurance, one would need evidence that the dietary intervention in fact took place. Yet, as critics have pointed out, there is no available biomarker for fat intake and state-of-the-art measures of food intake are inexact, depending on respondent memory, honesty, and ability to report what they have eaten.
Similar questions exist with the hormone replacement intervention. Is the HRT branch testing whether 0.625 mg of conjugated equine estrogen daily, with or without 2.5 mg of medroxyprogesterone, affects the risks of cardiovascular disease, osteoporotic fractures, and breast and endometrial cancers? Or is it testing whether the availability of HRT, with all of its perceived and actual side effects, and the related achieved adherence, are related to the measured outcomes?
NIH officials seem unsure of whether the CT, at this point in its development, is a test of efficacy or effectiveness. The committee suggests that it is somewhere in between. Given the extensive, though imperfect, data collection over the course of the study, it is possible that investigators will be able to acquire some information on efficacy, but it is less likely that they will acquire enough information to answer the basic question. The CT will be more likely to assess whether efforts to make major lifestyle and drug interventions (HRT and DM) are effective in changing certain risks of morbidity and mortality.
That test of effectiveness is still of major public health importance, but the committee reiterates a point it made earlier: no one study is sufficient. Other, smaller studies must explore the biological and behavioral intricacies of each larger lifestyle intervention, in order to determine which elements are efficacious and why. Then, additional studies on the effectiveness of lifestyle interventions must develop methods generalizable beyond clinical trial volunteers. The committee assumes that NIH would continue to fund such a range of studies as part of its overall research agenda.
Public Health Choice Versus Individual Choice
Individuals want to be able to make choices. Some choose to let their doctors choose; others want more personal control. For some people, length of life is most important; for others, quality is the determining factor. For most, it is an unenunciated, ill-defined combination.
Meanwhile, NIH faces a dilemma. It seeks to examine multiple endpoints but needs to make a binary decision–to stop or continue the trial. How can it resolve that dilemma? Total mortality is a useful initial methodologic approach, as well as a useful public health endpoint. It will inform some decisions that must be made for a community as a whole. Total mortality is not sufficient, however, for public health decisions or individual decisions about health care. Total mortality includes increased mortality from some diseases and decreased mortality from others. It ignores increased or decreased disabilities, pain, deformity, ease of life—insum, the quality of life. It may also mask increased risk for some subgroups. Individuals, having different levels of comfortable risk-taking, fears, desires, and personal circumstances, might make different choices for themselves than a well-meaning government or science council might for the group as a whole. Thus, whatever the outcome of the CT, choices for individual women need to be informed by their own baseline risk, fears of adverse outcomes, desire for potential benefits, and willingness to seek alternatives.
Considerations Beyond Science
The committee's charge was to examine the meaningfulness of potential results from the WHI in terms of rationale, study design and methods, cost, and overall justification. However, the committee was constantly faced with broader considerations. Colleagues, advocacy organizations, executive and legislative branch representatives, all approached the committee with reasons why the WHI should proceed, and why canceling it would unleash disasters. If it is cancelled, some argued, no one would ever trust the government again; it would prove that the government does not really care about women's health; it would so greatly disappoint the community which has gathered around this project that it will be impossible to ever galvanize them again; it would be seen as an unwelcome political blow to those in Congress who have pressed for more women's health research.
Others argued that it would be wrong and harmful to women's health to spend $625 million and find after 14 years that little in the way of useful information had been learned. It would not help women's health and women's health research to publicly announce and then misuse an enormous amount of money.
The committee tried to keep these considerations out of its judgment on the merits of the science, but they are important issues. With the committee recommendation that the WHI could proceed—though with significant changes of duration and focus—the discussion of these issues should continue.