A host of methodological flaws form the backdrop of the published medical literature on menopause. Common flaws include: (a) vague or nonexistent operational definitions; (b) lack of baseline data; (c) problems in establishing dose-response relationships; (d) lack of control groups; (e) overgeneralization; (f) retrospective reporting; (g) use of biased symptom checklists; (h) lack of context; (i) pejorative language; and (j) neglect of issues of diversity. A brief discussion of some of these is provided below; more detailed commentary is presented elsewhere (Rostosky & Travis, 1996). These methodological flaws are the mechanisms by which details of a distorted science are constructed into established facts that form the basis for medical practice.
Even the most elemental aspects of established scientific methodology were missing. For example, studies vary greatly in the operational definitions applied, and they frequently consist of only the most casual specifications (e. g., a general age range without any reference to the status of individual women). Under such arrangements, a 55-year-old woman on hormone replacement therapy could be labeled menopausal, whereas a 35- year-old woman with a complete hysterectomy and no hormone replacement could be placed in a control group.
It is not uncommon for studies to include menopausal and perimeno- pausal women, implying that if a woman is not actually menopausal, she may soon become so; and also implying that women falling within a broad age range are potentially at risk for menopause symptoms. These studies may encompass 20 or more years of women’s lifespan (e. g., 36-61 years; Huerta, Mena, Malacara, &. Diaz-de-Leon, 1995) or 33—56 years (Cutler, Garcia, & McCoy, 1987). When menopausal and perimenopausal age ranges are combined with the premenstrual and perimenstrual phases, fully 30% of a woman’s lifespan can be discounted as “under the influence” of some hormonal upheaval. Throughout, one finds terms reflecting decline, deficiency, and disease, for example pelvic atrophy, ovarian dysfunction, and estrogen deprivation. One study went so far as to refer to “these problem
women” (Garnett, Studd, Henderson, Watson, Savvas, &. Leather, 1990, p. 918).
Occasionally studies do code individual women with respect to key variables of interest, but these efforts often involve retrospective reporting. The respondent women may be asked to remember their daily fluctuations and physiological conditions from two years previously. To assist memory, researchers may present symptom checklists. However, these typically have an overwhelmingly negative slant (e. g., joint pain, nervousness, and short temper).
In any case, conditions and experiences with specific relevance to women are often strongly characterized by their biological correlates. Such language and focus imply that the experience is fundamentally determined by biology. It is constructed as a biological event rather than a social or political event. For example, Masters and Johnson (1966) noted a variety of physiological changes in menopause related to sexual response and attributed these to “steroid starvation,” although they reported no data about levels of circulating hormones. Such characterization typically ignores important contextual variables and conveniently locates the problem in the individual woman.
We were particularly disturbed that many studies supposedly designed to assess the efficacy of a particular clinical practice would not logically allow such inferences because comparison baseline data was not available. In a related problem, many studies failed to include control groups. Thus, for example, although the responses of menopausal women to symptom checklists (e. g., joint pain, short temper) are seen as informative, these are never compared with responses of men in the same age groups. The lack of comparative or baseline data leaves a question about the nature and size of differences in the experiences of menopausal women and women with other hormonal statuses, or even in comparison with men, who may also experience such symptoms. This lack of baseline or reference data can render a general picture of menopause as a stressful experience plagued by numerous discomforts.
More troubling than the distortion of descriptions of menopause itself is the opportunity for the construction of claims regarding the curative powers of medical interventions, especially the use of hormone replacement to alleviate various conditions that may be secondary to menopause. For example, Lobo and colleagues reported that 75% of the women in the study experienced no dyspareunia following hormonal replacement therapy, but they never indicated what percentage had reported dyspareunia before treatment (Lobo, McCormick, Singer, & Roy, 1984). This kind of methodology is equivalent to the use of testimonials to prove the benefits of hormone replacement. If the same methods were used to advocate the merits of a health elixir, it would be recognized as merely advertising rather than promoted as science.