Women's Trust in and Use of Information Sources in the Treatment of Menopausal Symptoms
Article Outline
Background
Frequent shifts in expert opinion over whether or not women should use hormone therapy (HT) or another menopausal treatment have left women in a difficult position; they must determine where they can obtain trustworthy menopause information. In this study, conducted 10–12 months after the Women's Health Initiative study first published results, we identified sources women use for information about menopause treatments, identified how trust-related dimensions (trustworthiness, knowledge, helpfulness, bias, and vested interest) influenced use of these sources, and determined how these trust dimensions varied with women's HT use status.
Methods
A total of 765 women >45 and <61 years of age from a Midwestern managed care organization responded to a survey. Trust dimensions regarding family, friends, physicians, pharmacists, other health care providers (HCP), and various media as sources of menopause information, and use of these sources for menopause information were examined.
Results
Women ranked physicians significantly higher than pharmacists and other HCP for trustworthiness, helpfulness, and knowledgeability. Perceived bias, knowledgeability, and helpfulness showed the strongest influence on trust in physician as well as nonpharmacist HCP. Menopause stage, HT use status, and helpfulness influenced use of physician for information. Helpfulness and knowledgeability influenced trust in the pharmacist, whereas actual use was predicated on having used HT, trust, and helpfulness.
Conclusions
Women who had never used HT trusted HCP less. HCP have important roles in providing menopause information to perimenopausal women. HCP may need to reach out and initiate these important discussions with their patients.
Background
Many middle-aged women are health conscious and are increasingly taking responsibility for their health care decisions. One decision women facing menopause make is whether or not to use hormone therapy (HT) or some other treatment for menopausal symptoms (Guillemin, 1999, Kennedy, 2003). Both expert and public opinion on how and when to treat menopausal symptoms have fluctuated repeatedly (McCrea, 1983, Thompson, 1995, Watkins, 2001), moving from the strongly pro-HT stance of the 1990s to the limited use position of the early 2000s (Grady, 2002, McIntosh and Blalock, 2005, Udell et al., 2006, Writing Group for the Women's Health Initiative Investigators, 2002). Coverage of the Women's Health Initiative's (WHI) surprising 2002 report of increased cardiovascular disease risk with HT use was particularly intense (Ingebretsen, 2002, Shojania et al., 2002). A recent reanalysis of the WHI suggesting HT did not increase coronary heart disease risk in younger women confuses the picture again (Manson et al., 2007), and with new trials underway the controversy will undoubtedly continue.
Women obtain information from a variety of sources when deciding about menopause treatments, including health care providers (HCP) such as physicians, pharmacists, and others, the media, friends and family. Several studies have focused on where women get their information about menopause (Andrist, 1998, Clinkingbeard et al., 1999, Griffiths, 1995, Griffiths, 1999, Hoffman et al., 2005). In a 1999 survey study, 76% of the Midwestern US participants (n = 665) cited women's magazines as their most frequent source of information about menopause, followed by HCP (68%) (Clinkingbeard et al., 1999). In regard to HT specifically, Swedish women were significantly more likely to obtain new information from newspapers (43.8%) and television or radio (31.7%) than HCP (18.3%) in 2003 (Hoffman, Hammar, Kjellgren, Lindh-Astrand, & Brynhildsen, 2005). One reason the media has been the most frequently cited menopause information source may be due to the frequency and prominence with which it has addressed the topic (Andrist, 1998, Watkins, 2001). However, this does not mean the women actually trust and use the information.
Trust has been suggested as an important reason for whether or not women use a source of information, and also helps to determine the influence of information on behavior (Frewer et al., 1996, Mechanic, 1996, Mechanic, 1998, Slovic, 1992, Tio et al., 2007). For instance, a recent qualitative Australian study found that trustworthiness and knowledgeability were among the most frequently mentioned factors influencing perceptions of (general) medical information source reliability (Tio et al., 2007). Frequent shifts in expert opinion leave women in a difficult position because they must determine where to obtain trustworthy information about the risks of menopause treatments. Evidence suggests women have become less trusting of medical advice and physicians because of the publicity surrounding the WHI (McIntosh and Blalock, 2005, Schonberg et al., 2005). This study adds to the literature by helping menopause information providers to understand factors influencing how trustworthy they appear to their audience, an important issue that has not yet been adequately explored.
Trust is reached after evaluating the likelihood of another's actions being for the trustor's benefit, and determining “that the other's interests are best-served by cooperating with you” (Gilson, 2003). Trust is also based on perceptions of character and motives, which are inferred from behaviors such as demonstrating concern and helpfulness (Korsgaard et al., 2002, Whitener et al., 1998).
Patients do evaluate sources of information on several dimensions (Frewer et al., 1996, Suter et al., 2007, Tio et al., 2007). Frewer et al. found 2 major components explained 89% of trust in information about food safety (Frewer et al., 1996). The first component included “trustworthiness,” “knowledgeability,” and “bias.” The second component focused on the source's self-interest, or “vested interest.”
The goals of this study were to 1) identify sources women use for information about menopausal treatments, 2) identify the dimensions of trust influencing the use of these sources, and 3) determine how these dimensions of trust vary with women's HT use status.
Methods
This project was part of a larger study conducted by a major Midwestern university 10–12 months after initial publication of WHI results. It focused on factors related to use of HT in women with and without diabetes from a not-for-profit managed care organization (MCO). Both the university's Institutional Review Board and the MCO's Research Committee approved the study. The larger study is described below briefly, followed by specifics of this report.
Phase I of the larger study involved 6 focus groups; 1 focus group each of women with diabetes and women without diabetes who currently used HT, previously used HT, or never used HT. HT- and menopause-related perceptions and issues were explored. Data were qualitatively analyzed and results used in developing a survey instrument. After pilot testing, the survey was mailed to 1,886 women identified from the MCO. Return of the survey indicated consent. All women had access to a comprehensive prescription medications benefit program.
Sample selection
Inclusion criteria for participating in the survey included: female gender; age >45 and <61 years; and willingness to participate in the study. Women with diabetes were oversampled because of potential HT underuse among women with diabetes, which was reported several times during the 1990s despite its purported cardioprotective benefits (Keating et al., 1999, Robinson et al., 1996, Stafford et al., 1997). Results from the larger study, including comparing women with and without diabetes, are reported in Huston (2003). Women with diabetes were identified from the MCO's diabetes registry, which followed the Health Plan Employer Data and Information Set (HEDIS) Comprehensive Diabetes Care 2004 measures, excepting the continuous enrollment requirement. Women without diabetes (n = 1,065) were randomly selected from the approximately 13,000 women in the appropriate age range. Women with diabetes who were surveyed (n = 821) comprised the entire population of women with diabetes meeting additional enrollment criteria enrolled in the MCO. Exclusion criteria were: 1) endometrial cancer or 2) breast cancer during the previous 5 years, and 3) a history of thrombophlebitis or thromboembolic disorder, because they are contraindications to the use of HT (Ayerst Laboratories 2002).
Measures
Survey items were developed from information gathered from the focus groups and previously published work. The survey consisted of several parts, including demographic (population group, age, education) and health history information (diabetes status, self-assessed menopause stage, HT use status), and for this study, the use and perceptions of information sources about menopause treatments. Women assessed their own menopause stage as 1) without any sign yet, 2) just beginning, 3) in the middle, 4) near the end, or 5) completed (Garamszegi et al., 1998). Women in the first 2 stages were considered as being in “early” menopause, and women in the last 3 stages were considered in “late” menopause. These 2 groups were compared because women with little or no experience with menopause might be expected to see things differently than women with such experience. A more complete description is provided in Huston, Kirking, and Shimp (2006).
The women were asked to evaluate family, friends, physicians, pharmacists, other HCP, television ads, print ads, television programs, print (books, magazines, etc.), and the Internet as sources of information about menopause treatments. Respondents were asked to rate each source on 5 dimensions of trust: 4 (trustworthiness, knowledge, bias, and vested interest) adapted from a previous instrument focusing on trust-related dimensions of information sources (Frewer et al., 1996) and 1 (helpfulness) as an action contributing to perceptions of benevolence and caring, seen as “critical” to building trust (Whitener et al., 1998). The addition of helpfulness was supported by findings from the preparatory focus groups. A 5-point scale was used, with 1 being most negative or lacking in the characteristic and 5 most positive or having most of the characteristic. For example, trustworthiness ranged from not trustworthy to completely trustworthy. Respondents were also asked which sources they actually used to get information about treatment of menopause.
Data collection
Dillman's Tailored Design method was used, which involved 5 contacts mailed directly by the MCO personnel (Dillman, 2000). Stamped, self-addressed envelopes were provided. Surveys were returned to the research team at the College of Pharmacy. Numerical coding ensured confidentiality.
Data analysis
Means and standard deviations were calculated for continuous data, and frequency distributions and percents for categorical data. Between-group comparisons were made using 1-way ANOVA, the Student t-test, or χ2 analysis as appropriate. Principal components extraction with varimax rotation was performed with SPSS 15.0 on the 10 different sources of information for the 5 trust components. Ordinary least-squares regression analysis run in SPSS was used to analyze relationships between perceived trustworthiness of HCP, use of HCP as an information source, and demographic and information variables.
Results
Response rate
Of the 1,886 surveys mailed, 44 could not be delivered and 30 of those completed were not eligible. The overall usable survey response rate was 765 or 42.2%. Previously, the MCO experienced approximately 30% misclassification of diabetes using the HEDIS diabetes registry. Approximately 22% were misclassified by HEDIS, based on women's self-declared, physician-diagnosed diabetes status. The gross response rate adjusted for percent misclassified for women without diabetes was 43.4% and for women with diabetes was 36.2%. There were no significance differences between the 3 response waves on general demographic characteristics (age, educational, income, population group) or health status variables (prior oral contraceptive use, hysterectomy status, self-assessed menopause stage, or HT use status), indicating sample representativeness (Churchill & Iacobucci, 2004). A total of 689 women for whom complete menopause stage and HT use data were available were included in this analysis. Women in the early stage of menopause who currently or had previously used HT were excluded because there were too few for analysis. These excluded women did not differ significantly from included women in terms of education, income, or population group.
Demographic results
Respondents were predominantly well-educated Caucasians with relatively high incomes. See Table 1 for a description by the 4 major analysis categories, based on self-assessed stage of menopause and HT use status. The 4 categories were: 1) women in the early stage of menopause who had never used HT, and women in the late stage of menopause who had 2) never used HT, 3) previously used HT, or 4) were currently using HT. Most women had used oral contraceptives (81.3%). Two hundred fourteen (31.2%) women had diabetes. Women with diabetes were significantly less likely to have completed college or to have done postgraduate work (38.4% vs. 58.0%; p < .001), to earn <$50,000 per year (38.0% vs. 19.5%; p < .001), or to have undergone a hysterectomy (25.9% vs. 18.1%; p = .01) than women without diabetes. Women with diabetes were also more likely to be in a later stage of menopause and to have ever used HT (a current or previous HT user) than to have never used it. A more detailed comparison between women with and without diabetes is presented by Huston (2003).
Table 1. Demographic Information
| HT use status | Self-assessed Menopause Stage | |||
|---|---|---|---|---|
| Early | Late | |||
| Never | Never | Previous | Current | |
| N | 226 | 161 | 175 | 127 |
| Mean ± SD | ||||
| Age (yrs)∗∗∗ | 49.6 ± 2.5 | 53.4 ± 3.4 | 55.0 ± 3.4 | 54.0 ± 3.3 |
| No. of menopausal symptoms experienced in the previous 6 months∗∗∗ | 2.0 ± 3.0 | 3.8 ± 4.3 | 3.7 ± 4.2 | 3.2 ± 4.2 |
| N (column %) | ||||
| Education | ||||
 Less than high school | 3 (1.3) | 2 (1.2) | 2 (1.1) | — |
| High school graduate | 28 (12.4) | 28 (17.4) | 28 (16.1) | 16 (12.6) |
| Some college | 69 (30.5) | 53 (32.9) | 61 (35.1) | 35 (27.6) |
| College graduate | 55 (24.3) | 27 (16.8) | 36 (20.7) | 34 (26.8) |
| Postgraduate work | 71 (31.4) | 51 (931.7) | 47 (27.0) | 42 (33.1) |
| Income ($) | ||||
| <25,000 | 6 (2.9) | 7 (4.6) | 8 (5.0) | 3 (2.6) |
| 25,000–49,999 | 46 (22.2) | 35 (23.0) | 32 (19.9) | 23 (19.7) |
| 50,000–74,999 | 44 (21.3) | 38 (25.0) | 42 (26.1) | 28 (23.9) |
| 75,000–99,999 | 45 (21.7) | 28 (918.4) | 37 (23.0) | 27 (23.1) |
| ≥100,000 | 66 (31.9) | 44 (28.9) | 42 (26.1) | 36 (30.8) |
| Population group | ||||
| African-American | 10 (4.5) | 11 (6.9) | 6 (3.5) | 8 (6.4) |
| Caucasian | 202 (90.2) | 137 (86.2) | 159 (92.4) | 115 (92.0) |
| Other | 12 (5.3) | 11 (6.9) | 7 (4.1) | 2 (1.6) |
| Diabetes status∗∗ | ||||
| Yes | 53 (23.6) | 55 (34.2) | 69 (40.1) | 37 (29.1) |
| No | 172 (76.4) | 106 (65.8) | 103 (59.9) | 90 (70.9) |
| Hysterectomy∗∗∗ | ||||
| Yes | 17 (7.6) | 19 (12.0) | 42 (24.0) | 58 (45.7) |
| No | 208 (92.4) | 139 (88.0) | 133 (76.0) | 69 (54.3) |
| Prior oral contraceptive | ||||
| Yes | 177 (78.7) | 127 (78.9) | 143 (81.7) | 104 (81.9) |
| No | 44 (19.6) | 33 (20.5) | 32 (18.3) | 23 (18.1) |
| Couldn't remember | 4 (1.8) | 1 (.6) | ||
| Used a non-HT therapy∗∗ | 63 (27.9) | 61 (37.9) | 75 (42.9) | 42 (33.1) |
Principal components analysis results
Principal components analysis revealed 3 distinct factors for all 5 dimensions of trust. Sample size for the 5 dimensions varied between 646 and 661 women. Only 6 of the 50 loadings were between .7 and .8, of which 4 loaded onto “Internet.” The remaining 88% of the loadings were >.8, with 19 having values >.9. Loadings of ≥.7 are considered excellent (Comrey & Lee, 1992). Factor 1 included family and friends; factor 2 included HCP (physicians, pharmacists, and others); and factor 3 comprised the media sources of television ads, print ads, television programs, print, and Internet. There was no cross-loading. These results indicate that the 3 major information source categories across all 5 of the trust dimensions are family and friends, HCP, and the media. Use of each of these 3 categories for information is presented in Table 2.
Table 2. Use of Information Sources According to Self-assessed Menopause Stage and HT Use Status
| HT use status | Menopause Stage | ||||
|---|---|---|---|---|---|
| Early | Late | ||||
| Never | Never | Previous | Current | Total | |
| N | 226 | 161 | 175 | 127 | 689 |
| HCP | N (column %) | ||||
| Physicians∗∗∗ | 133 (58.8) | 131 (81.4) | 163 (93.1) | 122 (96.1) | 549 (79.7) |
| Pharmacists∗∗∗ | 15 (6.6) | 18 (11.2) | 41 (23.4) | 39 (30.7) | 113 (16.4) |
| Other HCP∗∗∗ | 48 (21.2) | 41 (25.5) | 50 (28.6) | 52 (40.9) | 191 (27.7) |
| Media | |||||
| Television ads | 14 (6.2) | 14 (8.7) | 15 (8.6) | 11 (8.7) | 54 (7.8) |
| Print ads | 23 (10.2) | 29 (18.0) | 33 (18.9) | 22 (17.3) | 107 (15.5) |
| Television programs∗∗ | 27 (11.9) | 19 (11.8) | 40 (22.9) | 28 (22.0) | 114 (16.5) |
| Print | 110 (48.7) | 96 (59.6) | 102 (58.3) | 76 (59.8) | 384 (55.7) |
| Internet∗∗ | 49 (21.7) | 52 (32.3) | 65 (37.1) | 35 (27.6) | 201 (29.2) |
| Family and friends | |||||
| Family | 89 (39.4) | 78 (48.4) | 80 (45.7) | 52 (40.9) | 299 (43.4) |
| Friends | 116 (51.3) | 90 (55.9) | 95 (54.3) | 67 (52.8) | 368 (53.4) |
Frequencies, means, and ANOVA results
A total of 41.4% of the women who discontinued HT did so primarily because of news reports, 19.7% because their physician told them to, 14.7% because of problems from menopause or HT, and 24.2% for other reasons. Most women used multiple sources of information, with the most frequently cited numbers being 3 (21.8%) or 4 (16.2%). Sources of information according to menopause stage and HT use status are presented in Table 2. The most frequently cited information source in all 4 groups was the physician.
There were significant differences between menopause and HT use status groups in use of all 3 categories of HCP. Post-hoc analysis (Tukey's Honestly Significant Difference) showed that women who had ever used HT (current and previous users) were significantly more likely (p < .001) to have used both physicians and pharmacists than women who had never used HT. Women currently using HT were significantly more likely (p = .001) to use other HCP than were other respondents. Women who currently or had previously used HT were significantly (p < .01) more interested in television programs, whereas previous HT users used the Internet significantly more often (p < .01) than women in the early stage who had never used HT.
Table 3 presents the mean ratings on the 5 trust dimensions for HCP by self-assessed menopause stage and HT use status. All 4 groups of women ranked physicians significantly higher than pharmacists and other HCP with regard to trustworthiness, helpfulness, and knowledgeability. No significant differences were detected for bias or vested interest.
Table 3. Mean Ratings of Physicians, Pharmacists, and Other Health Care Providers on Trust Dimensions by Menopause Stage and HT Use Status
| HT use status | Menopause Stage | ANOVA Results | ||||
|---|---|---|---|---|---|---|
| Early | Late | |||||
| Never | Never | Previous | Current | Overall Significance level | Post hoc comparisons | |
| Mean ± SD | ||||||
| Trustworthiness | ||||||
| Physician | 4.1 ± .8 | 4.0 ± .9 | 4.2 ± .8 | 4.5 ± .6 | ∗∗∗ | Early/Never vs. Current∗∗∗ Late/Never vs. Current∗∗∗ |
| Pharmacist | 3.6 ± .9 | 3.5 ± 1.0 | 3.8 ± .9 | 4.0 ± .9 | ∗∗∗ | Early/Never vs. Current∗∗ Late/Never vs. Current∗∗∗ |
| Other HCP | 3.6 ± .9 | 3.5 ± .9 | 3.7 ± .9 | 3.9 ± .8 | ∗∗ | Early/Never vs. Current∗∗ Late/Never vs. Current∗∗ |
| Knowledgeable | ||||||
| Physician | 4.3 ± .8 | 4.3 ± .7 | 4.4 ± .8 | 4.5 ± .6 | ∗ | Early/Never vs. Current∗ Late/Never vs. Current∗ |
| Pharmacist | 3.9 ± .9 | 3.9 ± .8 | 4.0 ± .8 | 4.3 ± .8 | ∗∗∗ | Early/Never vs. Current∗∗∗ Late/Never vs. Current∗∗ |
| Other HCP | 3.8 ± .8 | 3.8 ± .8 | 3.9 ± .9 | 4.0 ± .8 | ∗ | Early/Never vs. Current∗ |
| Helpful | ||||||
| Physician | 4.2 ± .8 | 4.1 ± .9 | 4.3 ± .8 | 4.6 ± .6 | ∗∗∗ | Early/Never vs. Current∗∗∗ Late/Never vs. Current∗∗∗ Late/Previous v. Current∗ |
| Pharmacist | 3.7 ± .9 | 3.6 ± 1.1 | 3.9 ± .9 | 4.1 ± 1.0 | ∗∗∗ | Early/Never vs. Current∗∗ Late/Never vs. Previous∗ Late/Never vs. Current∗∗∗ |
| Other HCP | 3.7 ± .9 | 3.6 ± 1.0 | 3.8 ± .9 | 4.0 ± 1.0 | ∗∗∗ | Early/Never vs. Current∗∗ Late/Never vs. Current∗∗∗ |
| Bias | ||||||
| Physician | 2.5 ± 1.2 | 2.5 ± 1.2 | 2.4 ± 1.2 | 2.3 ± 1.2 | ||
| Pharmacist | 2.6 ± 1.1 | 2.5 ± 1.1 | 2.4 ± 1.1 | 2.5 ± 1.1 | ||
| Other HCP | 2.6 ± 1.1 | 2.6 ± 1.1 | 2.4 ± 1.1 | 2.5 ± 1.2 | ||
| Vested Interest | ||||||
| Physician | 3.0 ± 1.3 | 2.9 ± 1.3 | 3.0 ± 1.4 | 2.8 ± 1.4 | ||
| Pharmacist | 3.0 ± 1.2 | 2.9 ± 1.3 | 3.1 ± 1.3 | 2.8 ± 1.3 | ||
| Other HCP | 3.1 ± 1.1 | 2.8 ± 1.3 | 3.0 ± 1.3 | 2.7 ± 1.3 | ||
In comparing across menopause stage and HT use categories, current HT users ranked all HCP (physicians, pharmacists, and other) significantly higher with regard to trustworthiness than did never HT users, although the magnitudes of the differences were small. Current users found physicians to be significantly more helpful and knowledgeable than did women in the other 3 groups. Current HT users also found both pharmacists and other HCP significantly more helpful than did never users.
Detailed results from both the family and friends group and the media group are not presented because there was only 1 significant difference: Early never users rated family (p < .01) and friends (p < .01) significantly higher with regard to helpfulness than current users.
Regression results
Significant differences were found between the 4 groups of women (early never users, late never users, late previous users, and current users) for both use of and trust in HCP. Predictors of perceived trustworthiness and use of physicians, pharmacists, and other HCP as menopause treatment information providers are presented in Table 4. Perceptions of vested interest, age, education, hysterectomy, prior oral contraceptive use, and use of a non-HT therapy were not significantly different, and were not included in the final equations. Neither the total number of symptoms nor the experience of hot flushes were important in explaining either trust in the HCP or their use for information. Perceptions of bias, knowledgeability, and helpfulness seem to have the strongest influence on trust in the physician, whereas menopause stage and HT use status, along with helpfulness, were most influential on whether or not a physician was actually used as a source of information. Having diabetes also had a significant, positive impact on trust in the physician.
Table 4. Variables Predicting Trust in and Use of HCP for Menopause Treatment Information
| Dependent Variable (Adj. r2) | Independent Variables∗ | |||||||
|---|---|---|---|---|---|---|---|---|
| HT Use Status | Diabetes† | Trust-Related Dimensions | ||||||
| Late/Never | Previous | Current | Trust | Knowledge | Helpful | Bias | ||
| Physician | ||||||||
| Trust (.50) | ||||||||
| b (SE) | .12 (.05) | .33 (.04) | .38 (.04) | −.11 (.02) | ||||
| B | .07 | .30 | .39 | −.11 | ||||
| p | .01 | <.001 | <.001 | <.001 | ||||
| Information Source (.21) | ||||||||
| b (SE) | .23 (.04) | .38 (.04) | .37 (.04) | .05 (.02) | ||||
| B | .26 | .43 | .37 | /11 | ||||
| p | <.001 | <.001 | <.001 | .03 | ||||
| Pharmacist | ||||||||
| Trust (.47) | ||||||||
| b (SE) | .40 (.04) | .38 (.04) | ||||||
| B | .36 | .40 | ||||||
| p | <.001 | <.001 | ||||||
| Information Source (.15) | ||||||||
| b (SE) | .15 (.04) | .21 (.04) | .05 (.02) | .06 (.02) | ||||
| B | .17 | .22 | .13 | .16 | ||||
| p | <.001 | <.001 | .01 | <.01 | ||||
| Other HCP | ||||||||
| Trust (.43) | ||||||||
| b (SE) | .41 (.04) | .30 (.04) | −.05 (.03) | |||||
| B | .37 | .33 | −.07 | |||||
| p | <.001 | <.001 | .04 | |||||
| Information Source (.07) | ||||||||
| b (SE) | .17 (.05) | .07 (.03) | ||||||
| B | .14 | .14 | ||||||
| p | <.01 | <.01 | ||||||
∗Values included only if p < .05. The default condition is early self-assessed stage of menopause and never used HT. |
†Dichotomous variable: 0 = no, 1 = yes. |
Perceptions of helpfulness and knowledgeability were most influential in predicting trust in pharmacists, whereas use of pharmacists for information was predicated on having used HT (late stage current or previous user), trust in, and helpfulness of the pharmacist. Helpfulness influenced the use of pharmacists both directly and indirectly through trust. Perceptions of helpfulness and knowledgeability also influenced trust in other HCP, along with perceived bias. Trust and being a current HT user were the most influential variables for explaining use of other HCP, although only 6% of the variance was explained.
Trust in television programs and the Internet was also regressed on menopause and HT use status group, diabetes status, and the appropriate measures of bias, knowledge, and helpfulness. Use of television programs and the Internet was regressed on the same factors plus the appropriate trust measures. Factors significantly influencing trust in television programs (47.0% variance explained) were perceived bias (p < .001), which had a negative relationship, knowledge (p < .001), and helpfulness (p < .001) of the program. Only 11.0% of variance in television program use was explained, with trust (p < .001), helpfulness (p < .01), and being a late previous HT user (p < .001) having the greatest influence. Factors significantly influencing trust in the Internet (50.8% of variance explained) were perceived knowledgeability (p < .001) and helpfulness (p < .001). Actual use of the internet for information (15.6% of variance explained) was influenced most strongly by being a late stage never user (p < .01), a previous user (p < .001), trust in the Internet (p < .001), and Internet helpfulness (p < .01).
Discussion
Women in the current US study, 10–12 months after publication of the initial WHI results, were significantly more likely to seek information about menopausal treatments from HCP, primarily from physicians, than from the media, which is consistent with the Australian findings regarding general medical information. It is possible that trust in medical advice has declined, but in the United States it seems that trust in the media has declined more.
This study confirmed that women use several different sources for information about menopausal treatments and that use varied with self-assessed menopause stage and HT use status. Physicians were the most important source for all groups. Never users were significantly less likely to discuss the topic with HCP or watch television programs on the topic than either current or previous users. It is unlikely this lower discussion level is due to lack of access, because all women in the current study belonged to the same MCO. Neither did they seem to seek information from friends and family or from other media more often than women who had ever used HT. Never users did not seem to experience fewer symptoms than previous users or current users during the preceding 6 months, although they were more likely to use a non-HT therapy. These women may feel less comfortable discussing non-HT therapies with their HCP. As 1 woman in a recent qualitative study said, “Some physicians are more open-minded, but some are just definitely medical model, not willing to look at alternatives or best alternatives, so I mean when you go there, that's the perspective you're gonna get” (Suter et al., 2007). HCP should consider initiating discussion about menopause and menopause therapies with their patients when appropriate, making it clear they are willing to discuss alternatives as well as HT.
Trust is significantly associated with where women obtain menopause information, particularly for pharmacists and other HCP. This is supported by findings from Tio et al. (2007), who also found trust important in determining use of information sources. Never users rated physicians, pharmacists, and other HCP as significantly less trustworthy than did current users. This may reflect current HT users' positive experiences with these HCP, perhaps owing to more frequent visits or longer relationships. It could also result from never users deciding against HT use without a physician visit, thereby missing an opportunity to build trust. Although previous users' trust levels were not as low as those of never users, neither were they as high as those of current users. Some erosion of physician trust among some women may have occurred as a result of the WHI publicity, because more than twice as many women in this study discontinued HT because of news reports than because their physician told them to. This finding suggests that menopausal women should not be treated as a homogeneous group. After accounting for trust, knowledge, helpfulness, and bias, women with diabetes were significantly more likely to trust the physician than were women without diabetes, but having diabetes was not related to trust level for pharmacists and other HCP. Women with diabetes may trust physicians more than do women without diabetes because they visit their physicians more frequently, affording them the opportunity to build stronger, more intimate trust relationships over time.
Perceived knowledge and helpfulness were important predictors of trust for all 3 HCP: physicians, pharmacists, and other. Women who had never used HT considered physicians both less helpful and slightly less knowledgeable about menopause remedies than did current HT users. The magnitude of the difference with regard to knowledgeability was quite small; perceptions of helpfulness may be more important.
Perceived helpfulness had a direct relationship with the use of both physicians and pharmacists for information, but did not for other HCP. Unfortunately, HCP have not always been perceived as helpful. A total of 49% of women discussing menopause with their HCP responding to a 1999 survey indicated their questions were not answered (Clinkingbeard et al., 1999). Physicians and pharmacists who wish to develop a menopause-focused practice especially need to make clear to patients their availability and willingness to help. Women who had never used HT considered health providers to be significantly less helpful than do women who had used HT.
Perceived bias was also an important factor influencing trust in physicians and other HCP, but not pharmacists. Some HCP may push a narrow medical viewpoint, seeming uninterested or unwilling to discuss complementary or alternative therapies (Suter et al., 2007). Presenting balanced information and acknowledging the validity of alternative viewpoints may be helpful in reducing perceptions of bias. Because bias did not significantly influence trust in pharmacists, it may seem that pharmacists are more disinterested than either physicians or other HCP. This may reflect perceived role differences, however. Physicians and other HCP are more likely to be involved in the initial decision to use or not to use HT, whereas pharmacists are more likely to provide information once a prescription has been written.
It is interesting that women who had previously used HT cited a higher level of Internet and television program use related to HT than the other groups. Hoffman et al. (2005) also found that women who had discontinued HT used the Internet more frequently than either current or never HT users. It may be that they were stimulated to search for information from a quickly accessible source because of the WHI publicity. Increased Internet use could also be associated with a search for information about alternative therapies. Evidence suggests that some women have low confidence in physicians' ability to provide information about complementary and alternative medicines (Ma, Drieling, & Stafford, 2006). HCP should consider discussing the reliability and validity of Internet and television program content with their patients, because many women obtain information from these potentially unreliable information sources. HCP may wish to identify reliable Internet information sources for their patients. It would also be useful to teach patients how to properly evaluate websites for reliable medical information.
The limitation of a cross-sectional study is mitigated by inclusion of women in both early and late stages of menopause, and inclusion of previous as well as current and never HT users. As with any cross-sectional study, however, causality could not be assessed. Although the response rate was modest, demographic characteristics and health variables were not statistically significantly different across response waves, which suggests respondents do represent the sampling frame (Churchill & Iacobucci, 2004). The sample represents a well-educated, primarily Caucasian population, generally of higher socioeconomic status, which could reduce generalizability because this population has access to health resources that others may not.
Another limitation of this study is that neither physicians nor other HCP were differentiated into practice type or specialty. Future work should consider doing this. Future work should also determine if menopause symptom severity influences use of HCP as sources of information.
Conclusions
This study confirms that women use HCP as major sources of information about menopause treatments. HCP should be prepared to discuss this topic with their patients as appropriate. Women do seek information from numerous sources, however, and HCP should understand that and be prepared to help women identify and evaluate reliable print, Internet, and other media sources.
With regard to the trust dimensions, sources of information about menopause therapies fall into 3 clear categories of HCP, the media, and family and friends. Women who had never used HT seemed to have lower levels of trust for HCP. HCP interested in providing menopause-related services to these women should consider making special efforts to emphasize their willingness to help and knowledgeability on the subject, and to disseminate this information to women not currently taking HT.
HCP should also consider making explicit their willingness to discuss complementary and alternative medicine menopause therapies, because some women may not feel comfortable initiating the discussion. HCP willing to discuss complementary and alternative medicine may appear less biased than those not willing to discuss complementary and alternative medicine. HCP may wish to consider whether or not the content or style of their recommendations gives an appearance of bias, as this seems to negatively influence patient's trust for these care providers.
Perceived knowledge and helpfulness played an important role in influencing trust in HCP. Although helpfulness played a role in whether or not women discussed menopausal treatments with physicians and trust played a role in determining discussions with other HCP, both trust and helpfulness influenced discussions with pharmacists. Pharmacists and other HCP who wish to build trust with patients in this area may seek contact in nontraditional settings, perhaps by giving educational talks or seminars at churches or local exercise facilities. This would provide them an opportunity to build relationships, and display both their knowledge and willingness to help.
Physicians, pharmacists, and other HCP have important roles in providing information about menopause and its therapies to perimenopausal women. Women clearly do look to their HCP for this information. HCP may need to reach out and initiate these important discussions with their patients.
Acknowledgments
The authors thank Thomas Spafford for his logistical assistance.
References
- . The impact of media attention, family history, politics and maturation on women's decisions regarding hormone replacement therapy. Health Care for Women International. 1998;19:243–260
- . Premarin (conjugated estrogen tablets, USP) Patient Package Insert. Philadelphia: Wyeth-Ayerst Laboratories; 2002;
- . Marketing research: Methodological foundations. 9th ed.. Florence, KY: South-Western; 2004;
- . Women's knowledge about menopause, hormone replacement therapy (HRT), and interactions with healthcare providers: an exploratory study. Journal of Women's Health and Gender-Based Medicine. 1999;8:1097–1102
- . A first course in factor analysis. 2nd ed.. Hillsdale, NJ: Lawrence Erlbaum Associates; 1992;
- . Mail and Internet surveys. The tailored design method. 2nd ed.. New York: John Wiley & Sons, Inc; 2000;
- . What determines trust in information about food-related risks? Underlying psychological constructs. Risk Analysis. 1996;16:473–486
- . Menopausal status: Subjectively and objectively defined. Journal of Psychosomatic Obstetrics and Gynaecology. 1998;19:165–173
- . Trust and the development of health care as a social institution. Social Science & Medicine. 2003;56:1453–1468
- . Cardiovascular disease outcomes during 6.8 years of hormone therapy. Heart and estrogen/progestin replacement study follow up (HERS II). JAMA. 2002;288:321–333
- . Women's decisions about whether or not to take hormone replacement therapy: Influence of social and medical factors. The British Journal of General Practice. 1995;45:477–480
- . Women's control and choice regarding HRT. Social Science and Medicine. 1999;49:469–481
- . Managing menopause: A critical feminist engagement. Scandinavian Journal of Public Health. 1999;27:273–278
- . Changes in women's attitudes towards and use of hormone therapy after HERS and WHI. Maturitas. 2005;52:11–17
- . Understanding women's decisions about hormone replacement therapy. Ann Arbor: University of Michigan Press; 2003;
- . Use, intentions, and beliefs about hormone replacement therapy in women with and without diabetes. Maturitas. 2006;55:58–68
- Ingebretsen, M. (2002, December 30). Top health stories of '02: the strange and important. The Wall Street Journal [online version]. Available from: http://online.wsj.com/article/0,SB104117561424046953,00.html.
- . Use of hormone replacement therapy by postmenopausal women in the United States. Annals of Internal Medicine. 1999;130:545–553
- . Patients are expert in their own field. British Medical Journal. 2003;326:1276–1277
- . Trust in the face of conflict: The role of managerial trustworthy behavior and organizational context. Journal of Applied Psychology. 2002;87:312–319
- . US women desire greater professional guidance on hormone and alternative therapies for menopause symptom management. Menopause. 2006;13:506–516
- Estrogen therapy and coronary-artery calcification. New England Journal of Medicine. 2007;356:2591–2602
- . The politics of menopause: the “discovery” of a deficiency disease. Social Problems. 1983;31:111–123
- . Effects of media coverage of Women's Health Initiative study on attitudes and behavior of women receiving hormone replacement therapy. American Journal of Health-System Pharmacy. 2005;62:69–74
- . Changing medical organization and the erosion of trust. The Milbank Quarterly. 1996;74:171–189
- . Public trust and initiatives for new health care partnerships. The Milbank Quarterly. 1998;76:281–302
- . Can postmenopausal hormone replacement improve plasma lipids in women with diabetes?. Diabetes Care. 1996;19:480–485
- . After the Women's Health Initiative: Decision making and trust of women taking hormone therapy. Womens Health Issues. 2005;15:187–195
- Shojania, K.G., Mariani, S.M., & Lundberg, G.D. (2002). MedGenMed's selection of the top 10 medical/health stories of 2002. Available from: http://www.medscape.com/viewarticle/446310. Accessed December 24, 2002.
- . Perceptions of risk: Reflections on the psychometric paradigm. In: Krimsky S, Golding D editor. Social theories of risk. Westport, CT: Praeger; 1992;p. 117–152
- . Low rates of hormone replacement in visits to United States primary care physicians. American Journal of Obstetrics and Gynecology. 1997;177:381–387
- . Women's approaches to evaluating complementary and alternative therapies for menopausal symptoms. Canadian Family Physician. 2007;53:84–90
- . Estrogen replacement therapy in practice: Trends and issues. American Journal of Obstetrics and Gynecology. 1995;173:990–993
- . Ascertaining consumer perspectives of medication information sources using a modified repertory grid technique. Pharmacy World & Science. 2007;29:73–80
- . Effect of the Women's Health Initiative on osteoporosis therapy and expenditure in Medicaid. Journal of Bone and Mineral Research. 2006;21:765–771
- . Dispensing with aging: Changing rationales for long-term hormone replacement therapy, 1960-2000. Pharmacy in History. 2001;43:23–37
- . Managers as initiators of trust: An exchange relationship framework for understanding managerial trustworthy behavior. Academy of Management Review. 1998;23:513–530
- . Risks and benefits of estrogen plus progestin in healthy postmenopausal women. Journal of the American Medical Association. 2002;288:321–333
Sally A. Huston, PhD, is an Assistant Professor of Pharmacy at the University of Georgia. She received her Ph.D. from the University of Michigan. Her research interests include women's menopause decisions, and children's disease self-regulation.
Rebekah M. Jackowski, PharmD, is a Clinical Assistant Professor of Pharmacy at the University of Arizona. She received her PharmD from the University of Michigan and completed a Community Pharmacy Practice Residency at the University of Iowa/Osterhaus Pharmacy.
Duane M. Kirking, PharmD, PhD, recently retired from the University of Michigan College of Pharmacy and School of Public Health. During his faculty career, he co-founded and directed the Center for Medication Use, Policy, and Economics.
Funded by the Upjohn Research Fund at the University of Michigan.
PII: S1049-3867(09)00003-6
doi:10.1016/j.whi.2009.01.004
© 2009 Jacobs Institute of Women's Health. Published by Elsevier Inc. All rights reserved.
