The Alan Guttmacher Institute   search  
home home about contact e-lists support agi buy help
publications article archive state center media center tablemaker
Perspectives on Sexual and Reproductive Health
Volume 36, Number 2, March/April 2004

Source and Perceived Quality of Information Are Linked To Users' Odds of Discontinuing Hormone Therapy

In the 6-8 months after publication of findings from the Women's Health Initiative on the risks associated with postmenopausal hormone therapy, more than half of women in a large California health plan who had been regularly using hormone therapy tried to stop, even though two-thirds overall did not know what the study's main findings were, according to a survey of female members of the plan.1 The odds of trying to stop were significantly elevated among women who were sent a letter about the study's findings, reported receiving high-quality information from the media or correctly answered at least four of five questions about risks related to hormone therapy. Most women who tried to stop using hormones succeeded, but one in four resumed use.2 The 30% of women who had troublesome withdrawal symptoms had sharply increased odds of resuming use; odds were also elevated for women who had had a hysterectomy, had received hormones from a provider other than a gynecologist or perceived their risk of hip or spine fracture to be above average.

Knowledge and the Decision To Stop Using Hormone Therapy

Shortly after the July 2002 publication of the Women's Health Initiative findings, the health plan sent a letter to members who were using combined estrogen-progestin therapy, describing the study's findings and suggesting either that they stop using hormone therapy or that they consult with their provider about continuing. From January through March 2003, researchers conducted a telephone survey of all health plan members aged 50-69 who had used hormone therapy regularly during the year before publication of the study results. The women were asked about their key sources of information on the study's findings and about the quality of information. They also answered a multiple-choice question about the study's overall findings and five true-false questions on the risks of hormone therapy.

Analyses were based on 670 women, whose average age was 59. Roughly three-fourths were white, and more than a third had completed college. Small proportions (1-7%) had ever had a heart attack, breast or colon cancer, venous thromboembolism or osteoporosis, and fewer than a third believed their risk of these conditions to be above average. Forty percent of the women had undergone a hysterectomy. While 39% used only estrogen, 61% used both estrogen and progestin; half of the women had been using hormone therapy for at least nine years. The main reasons women cited for starting to use it were relief of hot flashes or excessive sweating (36% of women), relief of other types of symptoms (26%), hysterectomy (21%) and health promotion (17%).

Although 93% of women said they had heard about new findings on hormone therapy since July 2002, only 57% considered the information they had received high-quality (i.e., gave it a score of at least eight on a scale of 0-10). The latter proportion included 21% of those who had gotten information from the mass media and 32-34% who had gotten it from the health plan or their provider.

Sixty-four percent of women did not know what the study's overall findings were. Another 23% chose a correct answer (hormone therapy is bad for women overall or makes no difference), while 6% answered incorrectly (it is good for women overall) and the rest were unsure. The proportion of women answering correctly was higher among those who had been sent a letter than among those who had not (30% vs. 12%). Only 30% of women got four of the five questions on the risks of hormone therapy right; the proportion was again significantly higher among those who had been sent a letter than among others (37% vs. 19%).

Fifty-six percent of women had tried to stop using hormone therapy since July 2002. Among those who continued to use hormones, 26% cited relief of hot flashes as the reason, 19% protection against osteoporosis and 47% relief from other symptoms; 6% had not considered the issue, and 3% were waiting to discuss it with a provider.

In univariate analyses, a number of factors pertaining to women's background characteristics, self-perceived health status, hormone therapy use, and information on and knowledge about research findings on hormone therapy were related to whether they had attempted to discontinue use. A multivariate analysis revealed several independent associations. The odds of trying to stop hormone therapy were about three times as high among women who had been sent letters as among those who had not (odds ratio, 2.7). Women's odds were doubled if they considered information from the media to be high-quality, if their main reason for starting to use hormone therapy was health promotion, if they used a standard dose of estrogen instead of a lower dose and if they got at least four of the five true-false questions correct (1.9-2.1).

The investigators note that despite the media's extensive coverage of research findings on hormone therapy, women may not have "the requisite knowledge to make informed decisions' and may not be satisfied with the information they get. "Efforts are still required to provide women with adequate information about risks and benefits of hormone therapy," they conclude.

Predictors of Difficulty in Stopping

In a second study, using data from the same survey, the researchers assessed factors associated with difficulty in stopping hormone therapy among the 377 women who had tried to stop. The social, demographic and other characteristics of these women were generally similar to those of the study group overall. Whereas 74% of women had succeeded in stopping hormone therapy, 26% had resumed use. Overall, 72% had tried to stop abruptly, while the rest had tapered the dose or frequency of use; the success rate did not differ between these two groups.

The majority of women (70%) had no or only mild withdrawal symptoms after stopping hormone therapy, but 30% had troublesome symptoms. In half of women, these symptoms began within a week of stopping hormone therapy; the incidence did not differ by whether women stopped abruptly or gradually.

In a multivariate analysis, women who experienced troublesome withdrawal symptoms had dramatically elevated odds of not being able to stop hormone therapy (odds ratio, 8.8). Odds were also elevated for women who had undergone a hysterectomy (1.9), had received hormone therapy from a provider other than a gynecologist (2.2) or believed that their risk of hip or spine fracture was above average (1.4). Women with different types of troublesome symptoms had similarly elevated odds of stopping unsuccessfully (odds ratios, 2.2-2.4).

Further analysis revealed a set of three factors that clinicians can easily assess to identify women who are likely to have trouble discontinuing hormone therapy: having had a hysterectomy, having started hormone therapy for reasons other than health promotion and having used hormones for at least a decade. In a multivariate analysis, the odds of being unable to stop using hormone therapy were seven times as high among women with these three factors as among women with none of them.

Health care providers, the investigators note, can counsel women that most women do not experience troublesome symptoms when stopping hormone therapy, and that for those who do, behavioral measures may provide relief. Nonetheless, the investigators conclude, "it is important to identify effective measures to help users who would like to stop but are unable because of withdrawal symptoms."

—S. London


1. Ettinger B et al., Effect of the Women's Health Initiative on women's decisions to discontinue postmenopausal hormone therapy, Obstetrics & Gynecology, 2003, 102(6):1225-1232.

2. Grady D et al., Predictors of difficulty when discontinuing postmenopausal hormone therapy, Obstetrics & Gynecology, 2003, 102(6): 1233-1239.

home :: about :: contact :: e-lists :: support AGI :: buy :: help
publications :: article archive :: state center :: media center :: tablemaker