The term bleeding disorders refers to a group of conditions in which abnormalities in hemostasis can lead to spontaneous or prolonged bleeding.¹ Hemophilia, an X-linked genetic blood disorder that affects boys and men primarily, is the most well known. Women who carry one copy of the abnormal gene also can experience heavy bleeding symptoms and have mild hemophilia.² Other bleeding disorders, such as von Willebrand disease (VWD), are a result of inherited abnormalities on autosomal chromosomes and can affect both men and women.³ Symptoms resulting from these disorders can be more severe for women because of their associated effects on menstruation, pregnancy, and childbirth.^2,4,5

Approximately 1% or as many as 2 million women in the U.S. might have an undiagnosed bleeding disorder,⁶ yet most of these disorders remain undetected for years, and many are never diagnosed despite serious gynecologic and obstetric challenges experienced by these women throughout their lives. A 2009 CDC study⁷ showed that the prevalence of VWD or other blood factor deficiencies among women with menorrhagia (excessive menstrual bleeding) was much higher than U.S. gynecologists generally perceived it to be. A systematic review⁸ of published studies estimated that 5%–24% of women with menorrhagia might have undiagnosed VWD.

Women with bleeding disorders are at increased risk for complications such as anemia, bleeding during pregnancy, and postpartum hemorrhage and may undergo unnecessary hysterectomy and other uterine surgeries that can lead to life-threatening complications.^4,9 Furthermore, undiagnosed bleeding disorders affect women’s quality of life and well-being by limiting participation in education, employment, social, and leisure activities.^9,10 Studies^2,4,10 have shown that complications can be decreased or prevented and quality of life can be improved by early diagnosis and appropriate management.

Since 1996, CDC has been committed to collaborative research to determine the prevalence of bleeding disorders, assess provider awareness of these conditions, and evaluate treatment and management options to improve care for women with bleeding disorders.¹¹ Likewise, the National Hemophilia Foundation (NHF) has been working to raise awareness that women are affected by bleeding disorders.

In 2009, CDC and NHF entered into a cooperative agreement to improve the health of people with bleeding disorders through health education. Specific program activities focused on increasing awareness of bleeding disorders among undiagnosed women and health professionals. These activities are consistent with Healthy People 2020 objectives Blood Disorders and Blood Safety (BDBS)-14 and BDBS-15, which focus on appropriate referral of women who are symptomatic for timely and accurate diagnosis and treatment of bleeding disorders.¹²

To better understand young women’s knowledge, attitudes, and perceptions about bleeding disorders, the NHF commissioned an online assessment conducted by Harris Interactive in 2010 to (1) assess women’s knowledge about bleeding disorders; (2) identify how women aged 18–25 years perceived menstrual problems within their personal, self-care, work/school, home, and social domains; (3) determine if young women were more likely to take action—such as seeking health care for heavy bleeding—because of the perceived benefits of taking action or because of the negative consequences of not taking action; and (4) determine the best strategies for developing and presenting key messages about menorrhagia and bleeding disorders to young women aged 18–25 years.

In spring 2010, the NHF convened a panel of experts from NHF, CDC, and Harris Interactive to develop a survey instrument to assess the knowledge, attitudes, health behaviors, and menstrual experiences of young women. Over the course of 3 months, the group used an iterative process consisting of brainstorming activities and nominal group techniques to develop the survey instrument. Questions were developed to capture knowledge of menstruation and bleeding disorders; menstrual experiences within personal, self-care, work/school, home, and social contexts; health care–seeking behavior; preferred sources of health information; and demographic information.

After several rounds of review and revisions, items were finalized and agreed upon by the panel. In addition, the panel constructed a scenario about a young woman with symptoms of a bleeding disorder that was embedded in the survey and used to determine the preferred messaging strategy for presenting credible health information to young women. Owing to budget constraints, no formal pilot testing was done. However, members of the expert panel independently reviewed the questions and conducted a usability test before the survey was launched.

In September 2010, a national survey of women aged 18–25 years in the U.S. was conducted using a web-assisted personal interview. The survey was determined to be public health practice (non-research) by CDC on July 10, 2010. Therefore, an institutional review was not required. The survey was administered in English and took approximately 15 minutes to complete.

Respondents were selected from the Harris Poll Online (HPOL), a multimillion-member panel of cooperative online respondents. The panel has been used in other public health research studies.^13–18 An invitation with a link to the online survey was e-mailed to 59,124 female members of the HPOL. A total of 1,243 women aged 18–25 years participated in the survey, with an overall response rate of 4.8%. Of those participants, the responses of 1,239 were used in the study analyses; the responses of four participants were excluded from the analyses because they had never had a menstrual period.

Preliminary analyses were conducted in 2011 and final analyses were completed in 2013. Data were weighted, using a rim-weighting algorithm,¹⁹ to be representative of the U.S. population of women aged 18–25 years. Variables used for weighting were region, age, education, household income, race and ethnicity, and a propensity score. Propensity score–matching methodology has a strong statistical lineage in situations of nonrandom assignment to groups.^19,20 The propensity score is used to correct for potential attitudinal and behavioral biases that can arise with the use of an online panel that cannot be corrected through weighting of demographic characteristics alone.

These potential biases are associated with panelists’ nonrandom choices to be active online, join a panel, and respond to a particular survey. The propensity score is designed to measure differences related to the way respondents share and use information and participate in or engage with society, two key dimensions associated with their nonrandom choices. To validate this approach, Harris Interactive has conducted numerous parallel tests comparing results from online research with results from probability-based samples.²⁰

After completing items about the symptoms of bleeding disorders and menstrual health, all respondents—with the exception of those diagnosed with a bleeding disorder—were presented a scenario about a young woman who had symptoms of a bleeding disorder (n=1,205). Respondents were assigned randomly to react to messages from either a gain-framed (n=603) or loss-framed (n=602) perspective.

The gain-framed messages stressed that positive results would be gained by taking action to seek treatment from a healthcare professional. The loss-framed messages focused on the negative consequences of not taking such action. Women were asked to rate, on a scale of 1 to 4 (1 being not at all and 4 being a great deal), how motivating a series of statements would be in influencing them to seek treatment from a healthcare professional if they experienced similar symptoms.

Using weights provided by Harris Interactive, distribution of responses was calculated using SAS, version 9.3. Chi-square statistics were calculated to assess associations between variables. Significance tests were reported for probabilities <0.05. Women were stratified into two categories: being at risk and not being at risk for a bleeding disorder. Those at risk were defined as either having periods lasting 8 days or longer, often or always bleeding through a pad or tampon in 1 hour or less, or often or always feeling a sense of flooding or gushing. These definitions are consistent with expert opinion as being predictive of bleeding disorders.^21,22

A total of 1,239 women aged 18–25 years participated in the survey. In the sample, 58% self-identified as white, 15% as black or African American, 15% as Hispanic, 8% as Asian or Pacific Islanders, and 5% as other (this survey did not use official Office of Management and Budget race and ethnicity categories to collect these data); 34% had completed high school and 45% had completed some college or had an associate degree; and 44% were unemployed, whereas 38% were employed part-time. Each geographic region of the U.S. was represented in the sample (Table 1).

The results of this study suggest that lack of information about bleeding disorders is a problem because many women were unsure about the definition. Lack of knowledge also was evident in that fewer than half of respondents recognized that periods lasting 8 days or longer should be cause for concern. Women at risk perceived their stress to be higher during the heaviest days of their periods and believed that heavy bleeding interfered with activities of daily living, which in turn affected quality of life. Educating women regarding symptoms, particularly soaking through a pad or tampon in 1 hour or less (which they might have considered normal), could prompt them to initiate a conversation with a healthcare professional.

Health messages can be worded based on the perceived behavioral gains or losses of taking actions that affect health. This approach is referred to in the health communications literature as gain- or loss-framed messaging.²³ Some studies^24,25 have shown that gain-framed messages are effective in situations in which the risk associated with the health behavior is low. Conversely, when the risk is high, loss-framed messages might be more effective.

Meta-analyses^26,27 of health message framing indicate that gain-framed messages might be more persuasive than loss-framed messages when promoting prevention-oriented behaviors. Jonathan van’t Riet and colleagues²⁸ postulate that message framing might have implications for increasing acceptance of health-promoting messages. This is because “information acceptance might be of critical importance” and “people seeking information on the Internet are most likely to select information that seems relevant, interesting, and trustworthy.”²⁸

Data from this study suggest that messages to influence women who have symptoms of a bleeding disorder to seek health care should focus on gain-framed statements. The messages that survey respondents found most compelling were those that focused on knowing that there is treatment available, that care for the disorder can prevent pregnancy complications and miscarriages, and that care could result in a work and school life that is not disrupted by heavy bleeding. The NHF is using the findings from this study to revise existing loss-framed health communication materials with new gain-framed messaging.

Respondents were asked to select their preferred channel for receiving health information from a list of traditional and digital media. Roughly half indicated that none of the channels listed were a preferred way to receive information about menstruation, heavy periods, and bleeding disorders. Perhaps, as health topics became more personal, women were more likely to select “none of the above” because they may prefer to receive information from individualized communication channels such as healthcare providers, family members, or friends. Unfortunately, these response choices were not included for this question. Further study is needed to determine the best dissemination channels for presenting key messages about these topics to young women.