Compounded Bioidentical Hormone Therapy: Identifying Use Trends and Knowledge Gaps Among US Women

JoAnn V. Pinkerton, MD; Nanette Santoro, MD


Menopause. 2015;22(9):926-936. 

In This Article


Survey Design

The market research firms Harris and Rose each administered a population-based, cross-sectional Internet survey on menopause and HT to US women. The source populations were drawn from Web-based nonprobability consumer panels maintained by Harris and Global Market Insite (GMI). Members were recruited to their respective panels primarily through Internet recruitment drives (eg, advertisements and newsletters) and opted to join the panels. Privacy policies were fully disclosed to each member at registration, at which time they submitted a profile containing their E-mail address, name, home address, age, and other demographic information. An e-mail was sent to the e-mail address recorded at registration, with a link for panel applicants to confirm their desire to join the panel and to sign a confidentiality agreement.

The Harris survey was conducted between June 24 and July 10, 2013, and the Rose survey was conducted during three consecutive weeks in April 2014. The sample population was invited via e-mail to take the survey during the defined time frame. The e-mail provided an encrypted link to the survey, which is housed on a secure database and must be completed in a single sitting. Participation was voluntary, no medical procedures were conducted, and risk to participants was considered minimal. Although a written informed consent form was not formally obtained, respondents were advised— before they began the survey—that their opinions regarding medications or products that they might be taking for health were being sought, and they were assured that their answers and identifying information would remain confidential.

Responses were captured electronically and stored on a secure private server. Survey administrators applied several techniques to exclude duplicate or fraudulent responses. Respondents' digital fingerprints were compared with their registered profile, and surveys completed in less than two fifths of the median time estimated for completion were excluded. Only aggregated deidentified data were provided for analyses. As a reward for completing the survey, panel members received sweepstakes entries and/or points redeemable for approximately US$10 in cash or merchandise.

Harris and GMI are members of The World Association for Opinion and Marketing Research and comply with the International Chamber of Commerce/World Association for Opinion and Marketing Research International Code on Market and Social Research. Harris also conforms to the American Association for Public Opinion Research Code of Professional Ethics and Practices, the Health Insurance Portability and Accountability Act, and other US privacy regulations and guidelines.

Sample and Inclusion Criteria

For the Harris poll, women aged 45 to 60 years who were currently going through menopause or had experienced menopause were eligible to participate. With the goal of accruing 800 completed surveys, Harris invited 10,781 women aged 45 to 60 years from its consumer panel to take a survey on an unspecified topic. Invitations were balanced by US Census Bureau statistics for age, race, geographic location, household income, and education levels; invitees were systematically sampled from among panel members matching each target demographic. Quotas were established according to the desired number of completers per demographic attribute, and additional invitations were sent in batches in blinded fashion until each quota was met. These steps were taken to help mitigate sample biases and to ensure reliable and projectable survey results.

For the Rose survey, women aged 40 years or older who (1) confirmed current or former use of an HT product from a list of generic and branded drugs approved by the FDA to treat menopausal symptoms or (2) identified themselves as ever users of any product for "hormone therapy replacement or supplement" were eligible to participate. Women who did not indicate current or prior use of an HT product on the list or who answered "no" when asked whether they had ever used any product for "hormone therapy replacement or supplement" were excluded. In all, 90,120 women aged 40 years or older who belonged to the GMI consumer panel were invited, with the intent of obtaining 2,000 completed surveys. As with the Harris poll, the Rose survey used systematic sampling of panel members to ensure that the number of women invited per demographic attribute comported with US Census Bureau figures. Additional responses were sent in waves until response quotas for all demographic categories were satisfied.

Throughout this report, women who responded to the invitation are referred to as "respondents." Those respondents who were deemed eligible after answering all screening questions and who completed the remainder of the survey are referred to as "completers."

Survey Instrument

Questionnaires were developed by TherapeuticsMD in conjunction with survey administrators. Before the Rose survey was fielded, it was tested for face validity among a sample of 100 women and adjusted as needed. Each of the finalized surveys took about 15 minutes to complete. Most questions were multiple choice and allowed participants to select only one answer. Answer choices were rotated randomly for each participant to minimize potential bias, and a response was required before proceeding to the next question.

Harris respondents were asked whether they had ever experienced menopausal symptoms and whether they were currently experiencing them. Both Harris and Rose completers were asked about age at menopausal symptom onset, whether they had ever experienced specific menopausal symptoms, and severity of any symptoms. Completers in both surveys were also asked about HT use (type, where obtained, duration, and effectiveness) and specifically about CHT use. Because CHT is described in different ways, pertinent questions attempted to define CHT using terminology associated with how this treatment is prescribed. Harris completers were given a list of treatments of menopause symptoms and asked to indicate any treatment they had tried, including "bioidentical hormone replacement therapy (HRT) from a specialty pharmacy (personalized specifically for you by a special compounding pharmacy—local or mail order/Internet)." Ever users in the Rose survey were asked, "Was your prescription for hormone therapy specifically formulated, personalized, or compounded specifically for you based on your hormone levels?" This question was added to the survey after 273 of the eventual completers had already taken the survey to permit the collection of more complete information on CHT use. Only Harris completers were asked, "Do you believe that bioidentical hormone therapies compounded at a specialty pharmacy are FDA-approved?"

Data Analyses

Weighting by US Census Bureau statistics was applied to the final pool of Harris respondents to ensure that demographic attributes of age, race, and geographic distribution were proportional with the general population of US women. The Harris survey also weighted results to the US population by age, income level, educational status, and race, which did not materially affect outcomes. The pool of Rose respondents yielded a population of completers well-matched to the general population on target demographic variables and did not require weighting. Both surveys were slightly underweighted for race (80%-84% of completers were white). In analyzing data for eligible completers, we considered only women who answered 100% of the survey questions posed; partial interviews were excluded.

For the Rose survey, responses were analyzed for the entire pool of completers and stratified by age (40-44, 45-49, 50-54, 55-59, 60-64, 65-69, 70-74, 75-79, and ≥80 y). Subset analyses were conducted for women aged 50 to 64 years, which is the age group more likely to experience menopausal symptoms. In addition, responses were stratified by region, income level, ethnicity, current or former use, and hysterectomy or menopause status (perimenopausal and postmenopausal vs premenopausal).

There is no consensus for calculating response rates or estimating sampling error for a sample drawn from a nonprobability panel.[20] As recommended by the American Association for Public Opinion Research[20] in accordance with terms defined in the International Organization for Standardization report ISO 26362:2009, we reported the participation rate and cooperation rate, calculated as follows:

We also calculated the qualification rate using the following formula:

To estimate US trends in CHT use and annual spending on CHT, we used Rose survey information on rate of current HT use, number and cost of HT products used, and duration of use; the 2012 census estimates; and a summary report of HT prescriptions filled by US women aged 18 years or older. The report was generated using PHAST 2.0 (Symphony Health), a database of prescription information collected weekly from retail, mail order, and specialty pharmacies across the United States. (Calculations are detailed in "Results.")