Literature Review on Effective Sex- and Gender-Based Systems/Models of Care

January 30, 2007

Research Question 5: Sex- and Gender-Based Healthcare Practices

5. How effective are sex- and gender-based healthcare practices? What measures have been used to determine the effectiveness of these practices? What barriers exist to translating/adapting these approaches into clinical practice?

This research question looks at how sex- and gender-based approaches have been applied to healthcare practices, and how effective these approaches have been. We begin with a focus on women-specific programs, including overall trends in women's health centers and studies of how effective these centers have been. We then discuss more specific examples of women's health programs, including the Centers of Excellence (CoEs), Community Centers of Excellence (CCOEs), similar programs sponsored by other Federal agencies, other women's health programs in the United States, and other programs around the world. For each program, we look at measures of effectiveness as well as barriers to implementation. Next, we discuss the transition from an emphasis on women's health to one on gender-based health, including initial findings in this area. Finally, we conclude with a discussion of data quality issues and approaches to track the effectiveness of sex- and gender-based healthcare.

The number of women's health centers has been steadily growing. These centers appear to be meeting a need for a population of women seeking care in a more female-friendly setting. Women's health centers are particularly effective at providing preventive care for women, and studies have found that women who choose to attend such centers tend to be younger and are more likely to prefer female doctors.

The number of women's health centers has been steadily increasing since 1990, "probably based on market analysis indicating that women-oriented services would attract women and their families to new and coordinated services, resulting in better health" (Giardina et al., 2006). Women's health centers have become so predominant that they are now "a common fixture in most academic health centers and community hospitals. ... With 3,600 women's health centers in the United States serving nearly 15 million women, it appears that women's health centers are here to stay" (Gharib & Manson, 2001). Despite their increase in number, however, we did not find much data related to the composition of such centers. One study suggests that about a quarter of such centers operate for-profit, with for-profit centers being more likely to be located in suburban areas (Khoury, Weisman, & Curbow, 1997). We did not find any data indicating what kind of services or care these centers provide.

With their growth, such centers also have become more prestigious. In particular, the creation of the CoEs did much to elevate the status of women's healthcare, as well as to legitimize and broaden research in this field. Center directors reported that before the creation of CoEs,

[w]omen's health was often viewed simply as reproductive health, and it was aligned with obstetrics and gynecology. With the development of a CoE, its clinical component provided care to women in a more comprehensive way and, with its companion components of research, education, community, and leadership, the CoE viewed health among women more broadly to include such aspects as cardiovascular care and other health risks as they present uniquely in women. Such a comprehensive focus led to the expansion of the concept of women's health as a discipline (Goodman et al., 2002).

In other words, the creation of CoEs "acted as a catalyst for institutional change in expanding the field of women's health" (Goodman et al., 2002). The prestige associated with the CoE designation allowed the centers to develop new curricula and expand clinical operations and research (Goodman et al., 2002). CoEs thus have had a positive impact on the entire field of women's health, helping to establish women's care as a new academic discipline.

Women's health centers also are meeting the needs of a different type of patient. Comparisons among women choosing to attend such a center as opposed to a general practice have found that they are "younger, better educated, and have higher levels of physical function, though perhaps more psychosocial problems" (Carlson, 2000). They may also be more likely to be single, employed, to have fee-for-service insurance, and to choose their doctors based on referrals from friends (Harpole, Mort, Freund, Orav, & Brennan, 2001). Such women also are more likely to prefer a female provider and to demand more information from and easier communication with their physicians (Gharib & Manson, 2001).

Multiple studies have compared the care received by women attending women's health centers vs. other healthcare settings. The initial findings from these studies are quite positive. Women's centers seem to perform better on preventive services for women's issues, and comparably on most other general health issues. Women's care may also lead to higher satisfaction scores and potential cost savings. However, care may be worse for a small number of issues. Findings include the following:

• Improved preventive screenings and satisfaction: "Women who use specialized women's health settings have comparable or higher quality measures (e.g., preventive screening) and higher satisfaction compared to women in other community settings with traditional primary care" (Bean-Mayberry et al., 2006). These results are based on a study of women veterans receiving care in women's clinics or traditional primary care at ten Veterans Affair (VA) medical centers.
• Improved preventive screenings: "The delivery of basic preventive health services ... was as likely to be received by patients in women's health centers as by those in traditional general internal medicine practices. However, patients who received their care at women's health centers were more likely ... to receive their breast exam and Pap smear from their primary care provider, suggesting a potential for increased efficiency in the provision of gender-specific preventive services at primary care women's health centers. Preventive counseling services such as discussions about the use of hormone replacement therapy and dietary calcium were also more likely to be received by patients who obtained their care at women's health centers" (Harpole et al., 2001). These results are based on a study of quality of care in women's health centers compared to traditional internal medicine practices in academic medicine.
• More recommended gender-specific services: Women who attend women's health centers "are more likely to receive some recommended gender-specific preventive care, and have lower utilization rates for medical services. Thus, preliminary data suggest that women's health centers may add value through effects on delivery of preventive services, patient satisfaction, and resource utilization" (Carlson, 2000). These results are summarized from a study of quality of care in women's health centers compared to traditional internal medicine practices in academic settings.
• Some recommended gender-specific services: "The women's health center appeared to perform significantly better on mammography and modestly (though not significantly) better on other gender-specific measures. On non-gender specific measures, the performance of the women's health center was varied-while doing somewhat better on cholesterol screening, it performed significantly less well than the general medical clinic on colon cancer screening" (Phelan et al., 2000). These results are based on a study of quality of care at a hospital-based women's health center compared with care at a general medicine clinic.
• Improved preventive screenings and satisfaction with provider communication: "Satisfaction with communication with the physician was higher in the women's health centers, though overall levels of satisfaction were similar. Delivery of some gender-specific preventive services (hormone replacement therapy and dietary calcium counseling) was significantly higher at women's health centers; women were more likely to receive such services, as well as breast examination and Papanicolaou smears, from the primary physician. Less educated, lower income, and minority women attending women's health centers received benefits consistent with the entire cohort" (Carlson, 2000). These results are based on an analysis of three studies involving multiple care sites.
• Potential cost savings: "Women's health centers, if indeed fulfilling the promise of 'one-stop-shopping,' might offer a more efficient model of care with increased convenience and potentially lower cost" (Gharib & Manson, 2001). These results are based on an analysis of two studies involving multiple care sites.

Women's centers tend to have more female physicians, and some studies have indicated that female physicians offer more preventive services (see research question 6 for more information). However, "above and beyond the effect of physician gender, women's health centers offer an advantage to patients with respect to gender-specific preventive services" (Harpole et al., 2001). Thus, women's health centers do seem to improve care for women in some aspects, although the studies that support this finding are not comprehensive.

Similar to the overall trends related to women's healthcare centers, CoEs have been found to serve a unique subset of the overall female population and to provide greater preventive care services. In this section, we discuss these trends, as well as barriers CoEs have encountered in implementing care specifically for women.

A comprehensive study compared CoE patients to those in a nationally representative sample conducted by The Commonwealth Fund (CWF). Similar to the findings reported for all women's health centers, CoE patients have a higher income, are better educated, and are more likely to be employed. The study found the following:

Patients in the CoE clinical sample include a higher percentage of nonwhite women (African-American and Hispanic), were somewhat less likely to be unemployed (37% versus 41% in the CWF), had a higher proportion of college graduates (56% versus 21%), and were more likely to be in the higher income categories (e.g. 11% versus 4% reported an annual household income of $75,000 to $100,000). ... CoE patients were less likely to be uninsured than those in the CWF survey (4% versus 14% in the CWF) and more likely to be covered by private insurance (67% versus 53%). The proportions of respondents with public-sponsored insurance (Medicare and Medicaid) were similar. More striking differences between the CoE clinical sample and the CWF sample were found for enrollment in a managed care plan (78% versus 48%) and having a female physician (92% versus 24%) (Anderson et al., 2002).

The CoE model has been shown to increase preventive screenings for women:

Adjusting for region, age, education, perceived health status, and managed care enrollment, a statistically significant higher proportion of women in the CoE clinical sample report receiving all six screening tests compared with women in the CWF [Commonwealth Fund] sample. Likewise, women in CoEs were more likely to report counseling for smoking cessation, exercise, alcohol or drug use, domestic violence, and sexually transmitted disease. Effect sizes above 0.15 were found for Pap tests, physical breast examination, mammogram, and smoking cessation counseling (Anderson et al., 2002).

Similar findings have been reported elsewhere (Henderson, Hudson Scholle, Weisman, & Anderson, 2004).

Women using CoEs also report higher satisfaction with their healthcare: "A higher proportion in the CoE clinical sample were highly satisfied with their health care, compared with women in the CAHPS [Consumer Assessment of Health Plans] sample (81% versus 73%), and a higher proportion of women in the CoE clinical subsample were highly satisfied, compared with women in the community comparison sample (86% versus 79%)" (Anderson et al., 2002). This effect is especially pronounced among women who use a CoE as their first point of contact with the healthcare system (Anderson et al., 2002).

Women who attend CoEs thus experience a higher quality of care and greater satisfaction with this care. What is unclear, however, is why the CoE model works, and if "there are specific (replicable) aspects of the CoE model, which led to the higher quality performance, or whether the entire model as implemented is necessary" (Collins, 2002).

A possible explanation for the quality of care differences is that women visiting a CoE may experience a more comfortable and welcoming setting to discuss sensitive topics:

Counseling on sensitive topics such as domestic violence, sexually transmitted diseases, and sexual function or concerns is higher in CoE settings. The woman-friendly model of care in CoEs may help patients to feel comfortable talking about their concerns. Settings that are not explicitly dedicated to women's health care may be perceived by women as less able to address sensitive and gender-related concerns. The providers who work in these settings may be more attentive to psychosocial circumstances that affect women's lives and health by virtue of their exposure to research and the CoE model (Henderson et al., 2004).

Another possible explanation is that the co-location of multiple specialty services "enhances communication among clinicians and promotes integration of multiple specialty perspectives through informal encounters (hallway or lunchtime conversations) and formal exchanges (case conferences)" (Carlson, 2000). In addition, the co-location of services can "improve quality by increasing ease of access. ... Current standards of care for women stipulate periodic testing not routinely provided by primary care physicians, such as mammography and bone densitometry. Providing core services at the same place and, when possible, at the same time, increases the likelihood that women will obtain needed care" (Carlson, 2000).⁵

CoE directors believe that the strength of the CoE model is "leadership development such as mentoring; community outreach with an increased focus on community needs and services; clinical services that were expanded and improved; educational leadership that concentrated on gender-specific health issues and increased diversity; and research with a greater focus on women's health issues, coordination, and institutional support" (Goodman et al., 2002).

While all of these possible explanations could be related to a higher quality of care at CoEs, it is difficult to determine with certainty the reasons for this success because patients attending CoEs are self-selected. Thus, "some effects ... could have arisen because highly motivated patients sought out women-focused care" (Anderson et al., 2002). Additionally, patients may be seeking out other aspects of CoE care settings, such as the co-location of services. Studies are unable to control for such patient preferences. Another potentially confounding factor is that CoEs contain a higher proportion of female physicians, so observed effects could be physician-gender effects (see question 6 for a discussion of such effects). Studies have controlled for this difference, however, and the effect of CoEs continues while controlling for physician gender:

The hypothesis that CoE effects observed in the original evaluation might be attributed to physician gender was not upheld for breast cancer screening and detection, and counseling on domestic violence, STDs, sexual functions or concerns, and family or relationship concerns, or for satisfaction with care. These findings support the view that characteristics of the CoE settings, and not just the gender of the physicians in these settings, are responsible for higher quality of care for some measures (Henderson et al., 2004).

Thus, more work remains to determine which aspects of the CoE model have been most crucial to its success and can and should be replicated. However, the initial results are promising enough that "extending the benefits of the CoE model to more women, including women in traditionally underserved areas such as rural communities, deserves to be a high priority for women's health policy and health care quality improvement" (Henderson et al., 2004).

The CoEs, like any model of healthcare, face certain barriers to successful implementation. Key barriers identified by CoE directors include "the CoEs' acceptance, greater collaboration, and limited resources" (Goodman et al., 2002). In particular, a significant barrier to developing a one-stop shopping model is "financial and space constraints ... the financial pressures placed on these stand-alone facilities can be significant and thus a deterrent to institutions adopting this model" (Milliken et al., 2001). If CoEs attempt a center-without-walls approach, then they are subject to geographic problems for both providers and patients, an increased struggle to build a common identity, and increased infrastructure costs (Milliken et al., 2001). CoEs also may face problems with turf-guarding and demands on faculty time (Goodman et al., 2002).

As noted previously, OWH also has created Community Centers of Excellence (CCOEs). These community centers were designed using an "integrative approach that links existing activities and resources, rather than creating new ones, with a community-focused organization as the nucleus for operationalizing the new model to deliver comprehensive services to women" (Office on Women's Health, 2004).

We found one evaluation report for these CCOEs (Office on Women's Health, 2004). This report found that CCOE clients have a high degree of trust for CCOE providers (91%). In addition, CCOEs are meeting or exceeding their programmatic goals related to comprehensive and integrated delivery of women's healthcare services, training for lay and professional healthcare providers, public education and outreach, community-based research, leadership development for women, and technical assistance and replication of the CCOE model (Office on Women's Health, 2004). Thus, early results suggest that the model has been a success.

In addition to the CoE and CCOE models, other Federal agencies support women's health through a variety of approaches. These include clinical approaches to care as well as support for the study of women's health issues. We describe some of these programs and their effectiveness for agencies such as the Department of Veterans Affairs (VA), the Centers for Disease Control and Prevention (CDC), and the National Institutes of Health (NIH).

The VA's approach to women's healthcare was described in some detail in our discussion of question 4. In this section, we provide available data on the effectiveness of the VA's approach to care for women.

More than half of all VA healthcare sites have a women's clinic or program (Bean-Mayberry et al., 2003). Women using these women's clinics "are more satisfied than those in general primary care" (Goldzweig, Balekian, Rolon, Yano, & Shekelle, 2006). Another study found that such women are more likely to "report excellent overall satisfaction (odds ratio, 1.42). ... Multivariate models demonstrated that receipt of care in WCs [women's clinics] was a significant positive predictor for all 5 satisfaction domains (i.e., getting care, privacy and comfort, communication, complete care, and follow-up care)" (Bean-Mayberry et al., 2003). The study authors recommend that "since these WCs show better quality in terms of satisfaction, other quality indicators should be evaluated. If WCs reduce fragmentation and improve health care delivery, the model will be applicable in VA and non-VA outpatient settings" (Bean-Mayberry et al., 2003).

Thus, the VA model, like the CoE model, may improve the overall care of women by treating them in gender-specific clinics. Note, however, that the data we found on the treatment of women in the VA system (one of the most comprehensive and well-developed), mostly includes improvements in process measures (e.g., cholesterol checks) and satisfaction, not intermediate outcomes (e.g., blood pressure control among individuals with hypertension).⁶ Thus, while these early data on women's care are encouraging and suggest that this kind of approach should continue, continuing attention must be given to the issue of data collection and analysis to allow for detailed evaluations of this type of approach (additional issues related to data collection are discussed in more detail later on in this section).

The WISEWOMAN program also was described in some detail in question 4. In this section, we discuss the effectiveness of WISEWOMAN as a woman-specific intervention program.

One evaluation of a WISEWOMAN project in North Carolina "showed that women who received lifestyle counseling through the enhanced intervention reported less fat in their diets at follow-up than did women who received the minimum intervention" (Will et al., 2004). The WISEWOMAN project has been successful in reaching "financially disadvantaged, uninsured, and multiethnic women" (Will et al., 2004). WISEWOMAN projects have successfully increased physical activity and nutrition, but "have been less effective in influencing physiological measures (e.g., blood pressure, lipid levels, and anthropometric measures)" (Will et al., 2004). WISEWOMAN suggests that critical barriers (e.g., structural barriers) still need to be overcome, and is "now planning to supplement the current approach with a broader societal approach to improve health behaviors" (Will et al., 2004).

The NIH Office of Research on Women's Health (ORWH), along with other Federal partners, developed the Building Interdisciplinary Research Careers in Women's Health program (BIRCWH) (Bierman, 2003). The program offers "career development and mentoring awards for junior faculty who have recently completed or are soon to complete clinical training or postdoctoral fellowships" (Pinn, 2004). The goal is to

promote the conduct of interdisciplinary research and the transfer of findings to benefit women's health by increasing the number and skills of investigators at awardee institutions. This mentored research and career development experience is designed to lead junior researchers to independent interdisciplinary scientific careers addressing women's health. Thus, this program simultaneously promotes interdisciplinary research and the ability of investigators at awardee institutions to become independent career scientists in women's health (Pinn, 2005).

The program is relatively new, but BIRCWH awardees have proved to be successful scholars, publishing "more than 630 publications and 525 abstracts" and receiving 40 NIH grants (Pinn, 2005).

Unfortunately, these programs are not available to those studying these issues from a social sciences background. This excludes a potentially important group of scholars, and undermines the full effect of supporting interdisciplinary work.

NIH also has established the SCOR (Specialized Centers of Research on Sex and Gender Factors Affecting Women's Health) initiative. SCOR is "another form of support for careers in women's health, engaging investigators at sites nationwide in studies involving a gamut of concerns, from pharmacokinetics in women to incontinence, pain sensitivity, irritable bowel syndrome, and the roles of sex and gender in drug abuse" (Pinn, 2004). SCORs help to "increase the transfer of basic research findings into clinical practice" by "conducting interdisciplinary research focused on major medical problems affecting women and comparing gender contributions to health and disease" (National Institutes of Health, 2006b). Their goal is:

to promote innovative institutional interdisciplinary research in areas important to women's health. The centers address ORWH research priorities for women's health, which are determined through annual reviews of current advances and continuing gaps in scientific knowledge by representatives from NIH institutes and centers and from the broader scientific, health professional, and advocacy communities. SCORs were envisioned as a national resource associated with one or more major medical complexes and dedicated to furthering the research effort on women's health or sex/gender factors to translate basic research to clinical application. Each SCOR is expected to have: (1) a theme for an overall research program for a disease or health topic that addresses a critical area of opportunity to advance knowledge and patient care; (2) an interdisciplinary team of basic and clinical researchers to provide a mutually supportive interaction in meeting the research objectives; and (3) both basic and clinical research projects that develop new approaches, significant hypotheses, and improved strategies for clinical advances in women's health or gender-specific medicine (Pinn, 2005).

As in the BIRCWH program, support is not available to those studying these issues from a social sciences background. Additionally, SCORs require a disease-based approach, so they exclude many potential applicant institutions, teams, and types of work.

Both the BIRCWH program and SCORs are relatively new. However, both programs already are contributing positively to interdisciplinary research and forging new relationships among researchers working in these areas (Pinn, 2005).

We were able to locate evaluation data for one comprehensive approach to women's health, the Columbia University Center for Women's Health. Background information on this center was provided in question 4.

The Columbia University approach is in many ways similar to the CoE approach. Not surprisingly, quality of care and patient satisfaction with this model also are similar. Findings are based on a patient survey conducted in 2003:

The patients (ranging in age from 21 to 90 years) reported that their satisfaction was high: 82% appreciate the integrated preventive services and recommended the center to friends and family, 69% received care for over a year, and 65% rated their care as the "best care possible." Moreover, the majority indicated they received annual preventive examinations, including physicals, Pap smears, breast examinations, and cholesterol screenings, and preventive issues were discussed, including smoking, diet, exercise, hormones, vitamins, and calcium. These findings are comparable to a report that women seen in comprehensive programs of CoEs were more satisfied with care and received significantly more screening and counseling services than women in other settings. However, areas that were identified for improvement, such as greater availability for new and walk-in visits and access to physicians by e-mail, also were noted (Giardina et al., 2006).

Women's health models are being implemented in many nations. Such models tend to be comparable to the U.S. model, with creation of women's centers predominating. As in the United States, such centers tend to attract a slightly different segment of the female population and seem to increase satisfaction among women patients. Some data are available to document differences in quality of care between women receiving care in a woman-only setting and care in a general setting. We report on findings from Canada, Thailand, Brazil, Australia, and the Netherlands.

As noted previously, Canada created and implemented a model of national women's health centers that is similar to the CoE model (Health Canada, 1996). Like their U.S. counterparts, these centers have been successful in leveraging funds. For example, the British Columbia center was initially funded with $2 million (Canadian) and has been able to leverage an additional $7.5 million (Canadian) (British Columbia Centre for Excellence in Women's Health, 2001). The same center was rated quite positively by Health Canada for its capacity-building for women's health as well as for creation of new knowledge (British Columbia Centre for Excellence in Women's Health, 2001).

A Thailand project to implement a more gender-sensitive care model at rural women's health clinics was successful (more details on this project were provided in question 4). The project involved intensive staff training, and improved the quality of care received by patients as well as the patient/staff interactions:

Health staff who participated in the action-research ensured that they took extra time in providing advice, being more careful in recording medical history and getting permission from women before asking sensitive questions. Respect is given during gynaecological examinations, and explanations on the results of diagnosis and treatment are provided. Although clients wait longer, they are happier because they receive better care. Although clients are still shy to discuss sexual health problems with the staff, they are encouraged to do so by the staff. The staff learned to provide women-friendly services by being friendlier in their behaviour and choosing their words more carefully so as to build rapport with the women (Boonmongkon, 2000).

The clinics have found that they are now providing more services to women, and that women are reporting high overall satisfaction with the program (Boonmongkon, 2000).

Another women's care model that we previously described (see question 4) is a comprehensive approach to women's care implemented by the reproductive healthcare services at the Perola Byington Hospital in Sao Paulo, Brazil. The program made greater use of nurse-assistants to screen for the most common gynecological problems to increase the number of patients seen per doctor per hour (Pinotti & Tojal, 2001). Each doctor worked with four nurse-assistants, both improving the content of care and increasing the efficiency of care: "In the traditional model, care is given only according to symptoms presented and a physician can see on average ten patients per four hours of medical work. In the Perola Hospital model, integrated care was given and coverage per four hours was 40 patients, four times greater than in the traditional model" (Pinotti & Tojal, 2001).

Australia has a series of Women's Health Centers (WHCs) that provide care to patients in a woman-only environment. Patients at these WHCs tend to be disadvantaged and see the WHCs as a safe setting (Broom, 1998). This is especially important for women who have previously been victims of abuse. Women patients like the longer appointments provided at the centers and tend to prefer female providers; women patients are extremely "enthusiastic" about the care they had received in such settings (Broom, 1998). In addition, the care women received at the centers was viewed as more holistic (Broom, 1998). This study did not report any findings pertaining to quality of care received at these centers; however, the results do suggest that the centers are meeting a need for some women.

Aletta is a women's health center in the Netherlands that was started in 1984. Its medical philosophy is similar to that of many U.S. health centers: "consideration of the patient's personal and social factors, respect for the patient, encouragement of self-responsibility, and demedicalization" (Carlson, 2000). A study was conducted to compare care at Aletta with that for other Dutch female patients (as reported in Carlson, 2000). The study found that Aletta patients were younger, better-educated, and less likely to be unemployed (Carlson, 2000). In addition, women reported more independence and individualism and "were less likely to seek help for minor illnesses" (Carlson, 2000):

Important measures of health status (perceived state of health and number of chronic health conditions) did not differ significantly between women receiving care at the women's health center and those receiving care elsewhere. However, patients at the women's health center reported more psychosocial problems and recent stressful life events. Utilization of medical care resources by visits to the general practitioner did not differ between the groups. However, mental health and alternative health care services were more frequently utilized by patients at the women's health center.

No data were reported on quality of care differences between the two samples.

To date, we have been reporting primarily on care that concerns one gender (i.e., care for women). This is because the first movements toward gender-based medicine started with attention to women's healthcare as a means to correct for previous deficits. As the women's health movement has matured, however, it has brought with it the discovery that men also have specific health needs that have been neglected by traditional medical approaches:

Over the past 30 or so years, women's health advocates have argued-rightly, and with some considerable success-that women have been ill-served by a male-dominated health service, and that policies and practices need to change to reflect the specific needs of women. It is now becoming clear that men have also suffered a disadvantage from the way the health service has been structured and decisions have been made (Banks, 2004).

As a result of these findings, the research focus is slowly shifting from women's health or men's health to gender-based medicine.

Gender-based medicine is, in many ways, the natural progression from women's health. Gender-based medicine capitalizes on what research has discovered is unique about women and applies that knowledge to expand our understanding of both sexes. As one researcher in this field has said, "to continue to consider the health of women in isolation from the wider world of human health is to miss a spectacularly important opportunity to improve the health of all" (Legato, 2003). Another predominant expert in this field described the progression this way:

"Women's health" is expanding into the larger concept of gender-specific medicine. Women's health is no longer an isolated phenomenon, divorced from mainstream medicine and regarded as a political or feminist issue. Instead, thoughtful scientists now see women as important sources of new information that will correct essentially male models of normal function and the pathophysiology of disease (Pinn, Begg, Rudick, & et al., 2006).

The mission of gender-specific medicine is "to promote research that uses biological sex as an important variable in constructing scientific protocols" (Legato, 2006c). This movement, while a natural and perhaps inevitable progression, is still very much in its infancy:

The idea of establishing gender-specific medical practices, even as a pilot experiment, to test whether or not the new knowledge when applied to clinical care impacts the quality or length of life has not even been seriously approached. In fact, principles of best practices for the sex-specific care of patients are only now being developed. Gender-specific medicine still means women's health to all but the indoctrinated few, and many continue to view women's health as a feminist, commercial, or boutique issue more suited to marketing teams for hospital centers than to serious practitioners of medicine (Legato, 2003).

However, important progress in this area is being made. Recent progress toward the adoption and expansion of gender-specific medicine includes the following:

• "In 2001, the World Health Organisation's 'Madrid Statement' urged member states to 'actively integrate gender mainstreaming into public policies that determine health'"(Banks, 2004).
• "This year [2006] the Committee of Experts on the Inclusion of Gender Differences in Health Policy was set up by the Council of Europe, in Strasbourg. The key aims of this committee are determining what is good practice and drawing up clear recommendations on how to integrate the biological, cultural and social dimension of gender into the Council of Europe Member States' public health programmes (prevention, diagnosis, care), as well as, integrating the gender dimension in training programmes for the health professionals" (Rieder, 2006).
• In 2006, an announcement was made that a new global center, "The International Association for Gender-Specific Medicine," is forming (Legato, 2006a). The Center's founders are from Berlin, Sweden, and the United States, and the Center is hoping to increase international collaboration on these issues.

Thus far, data on the effectiveness of gender-specific medical interventions are extremely limited. We uncovered one report on such interventions that was produced by the Interagency Gender Working Group (IGWG) of the United States Agency for International Development (USAID). This report was an attempt to evaluate the success of reproductive health programs that have integrated gender, although the programs described are mostly interventions for females.⁷

Nine programs were evaluated for the report. Many programs were in settings where women had limited autonomy, and a goal of several of the programs was to help women learn to make their own health choices. Other programs offered training "in more than reproductive health, in areas such as literacy, employment skills, legal rights, parenting, child health, and social mobilization. Others focus on strengthening women's voice in community planning, in bringing issues to government officials and donor organizations, and educating men on the importance of women's reproductive health and family planning" (Feldman-Jacobs, Olukoya, & Avni, 2005). The interventions primarily accommodated, as opposed to challenged, gender differences:

For example, a health centre may decide to add a small childcare facility so that mothers can visit health care facilities even when they do not have anyone with whom to leave their children. Some health services have tried to make their settings more male-friendly, for example, by offering evening hours of operations, allowing the presence of male partners during childbirth, or introducing couple counselling. Still others, recognizing women's lack of economic resources, lower fees to make services more accessible (Feldman-Jacobs et al., 2005).

The conclusion of this report is that "integrating gender into reproductive health programmes has a positive impact on achieving reproductive health outcomes" (Feldman-Jacobs et al., 2005). The report offers the following four recommendations (Feldman-Jacobs et al., 2005):

1. "Integrate gender into programmes to achieve positive reproductive health outcomes. With very few exceptions, the programmes that integrated gender achieved positive reproductive health outcomes."
2. "Focus on community participation or community empowerment strategies when seeking to transform gender relations. Gender-integrated components of reproductive health programming are often embedded in participatory or community empowerment initiatives."
3. "Provide support for more rigorous evaluation of interventions that integrate gender. Further efforts in refining evaluation methods and indicators, especially gender indicators, are necessary to generate a solid body of evidence regarding the value of gender integration. Gender indicators are still being developed and tested. Researchers and programme implementers need to be trained in effective use of these measurement tools. Funding for evaluation needs to be allocated, scaled up, and long-term. Sufficient time and funding are needed for programmes to show effects."
4. "Recognize the unique contribution that gender integration brings to the success of STI/HIV interventions, achieving sustainable health results and greater gender equity."

Any effort to improve healthcare to accommodate sex- and gender-based differences will need to include an effort to track and evaluate the success of interventions. High-quality data are needed to either demonstrate or refute the effects of various approaches so that interventions can be tested and targeted. Data need to be collected for both men and women to allow for valid statistical comparisons. For example, "stratifying quality indicators by gender would allow [healthcare] plans to specifically evaluate the effectiveness of improvement interventions among women and men" (Bird et al., 2003).

In this section, we discuss issues that need to be resolved to ensure the collection of high quality data, discuss approaches to data collection, and highlight recent progress in this area.

There are two primary issues with collecting quality data to detect sex- and gender-differences in disease occurrence, diagnosis, or treatment. The first is that data on women may not be available. Women are still underrepresented in some medical trials (Azad & Nishtar, 2005). Additionally, data may not be collected in sufficient quantity to allow for subgroup comparisons between men and women. For example, a major source of care data is the HEDIS (Health Plan Employer Data and Information Set), which is collected by the NCQA (National Committee for Quality Assurance):

To minimize burden on plans, for some measures the NCQA only requires participating plans to collect data from a sample (411 cases) of eligible enrollees. Though this sample size was selected to allow comparisons of performance rates between plans, the number of cases does not provide sufficient statistical power to compare performance between subgroups, such as men and women. The small denominator problem is not limited to HEDIS measures that combine administrative and chart data (i.e., "hybrid measures") but also can occur for measures that plans can compute solely from administrative data (Bird et al., 2003).

Moreover, data may not be available because sex is not part of the reporting requirements. For example, although quality measurements are widely used in health plans and are "crucial to improving quality of care in general and reducing gender differences in care ... few [health] plans currently stratify such measures by gender" (Bird et al., 2003).

A similar, but more severe, shortage of data is common among those attempting to study potential racial, ethnic, or socioeconomic disparities, because these categories comprise even smaller segments of the population (compared with about half the patient population being male or female), and data on race, ethnicity, or socioeconomic status is collected less frequently than data on gender (Long, Bamba, Ling, & Shea, 2006). Finally, the small numbers problem makes it especially difficult to do an analysis of interactions between gender, race, ethnicity, and socioeconomic status.

In addition to problems with insufficient data, this field also suffers from a lack of reporting on gender similarities or differences even when data are available. Data may not have been recorded in such a way as to allow for gender- or sex-based analysis (Anthony, 2005). For example, Legato, Glelzer et al. (2006) conducted an analysis of available data on the treatment of diabetes, looking specifically for information on sex- and gender-differences:

Often in the diabetes literature (as in the literature on other diseases), even when both men and women were included in studies, no attempt was made to assess the impact of sex-specific differences on the data. For optimal prevention and treatment of diabetes, these differences must be acknowledged, planned for, and factored into an effective treatment regimen, which will differ significantly as a function of the patient's sex.

Thus, insufficient data on both sexes to allow for comparison (either because one sex was underrepresented in research studies or because collected data do not allow for analysis on the basis of sex) poses a challenge to study of sex- and gender-based differences in healthcare. The result is that while several studies provide data on the care of one sex, few studies have "directly examined gender differences in performance on a variety of objective quality measures" (Bird et al., 2003). As another researcher team put it, "lack of evidence ...plagues women's health. ... Data on medical processes and outcomes in women has lagged well behind that on men" (McKinley et al., 2002).

One possible solution to gathering needed data is to pool data from various global sources (Legato, 2006b). Another approach is described by Bird, Fremont et al. (2003):

These challenges can best be addressed through a concerted effort. For example, it would be reasonable to encourage stratified reporting of selected measures such as a lipid check among diabetics which can be calculated solely from administrative data and generally have sufficient cases for subgroup comparisons. Furthermore, if the differences in measures requiring hybrid data are large enough, those comparisons should be made as well.

Additionally, useful data might be pooled from multiple sites to create a dataset that is sufficient for analysis. For example, the Department of Veterans Affairs collected data for analysis on "high-prevalence, high-cost, high-impact conditions among women veterans, as well as conditions with disproportionate burden among women (e.g., obesity, incontinence, osteoporosis) or with distinct clinical presentations in women (e.g., coronary artery disease)" through an effort to pool data from multiple VA sites:

Over 15 research centers responded to our requests for gender-specific analyses of existing data, demonstrating both the capacity and commitment to furthering the VA Women's Health Research Agenda. While the results of these secondary analyses are too numerous to cover here, subsequent priority-setting was informed by the most prevalent diagnoses (e.g., post-traumatic stress disorder [PTSD], arthritis, chronic low back pain, hypertension, chronic lung disease, depression), most commonly prescribed drugs (e.g., simvastatin, levothyroxine, lisinopril), and gender comparisons in patient satisfaction, quality, and costs of care. This process highlighted that these data sources had been underutilized in the past, demonstrating substantial opportunities for additional analyses (Yano et al., 2006).

Once data are available for analysis, the next consideration is how to identify significant sex- and gender-based differences. The first part of this task is to define what needs to be measured. The measurement of quality in healthcare is complex, requiring the measurement of numerous variables. While there is "no consensus on a definition of quality in women's health care," "organizations such as the National Committee for Quality Assurance (NCQA) have begun to develop and utilize indicators for process variables relevant to women's health care" (Carlson, 2000).

Carlson (2000) suggests that any quality measure should include both process and outcome variables, including the following:

1. Process variables
1. Adherence to guidelines for screening and preventive care (examples include screening for cervical and breast cancer, sexually transmitted diseases, depression, violence, and osteoporosis; counseling at menopause; and assessment of risk for diabetes and heart disease)
2. Adherence to guidelines for management of conditions such as cervical dysplasia, osteopenia, diabetes, breast complaints, and chlamydial infection, including timely follow-up of positive screening tests
3. Access to care
4. Interpersonal aspects of care, such as the amount of information exchanged during an office visit and the active involvement of the patient in the decision-making process
5. Utilization of health care resources in the ambulatory setting, including number of office visits, use of diagnostic tests, time spent with providers, and changes in source of care over time
2. Outcome variables
1. Functional status, including general health, mental health, and social functioning
2. Clinical status, including condition-specific measures when available
3. Changes in patient behavior, such as smoking cessation or initiation of exercise
4. Patient satisfaction
5. Professional satisfaction

Another attempt to define what should be measured was conducted by the National Committee for Quality Assurance's Women's Health Measurement Advisory Panel (MAP). The MAP was established because "the Health plan Employer Data and Information Set (HEDIS) is limited in its scope of women's health-related performance measures"; thus, the MAP was asked to help "expand and develop HEDIS measures to better represent women's health issues" (McKinley et al., 2002). HEDIS measures at the time focused "only on a limited number of reproductive issues (prenatal care and childbirth) and on breast and cervical cancer screening" (McKinley et al., 2002).

This extensive review process led to the following results:

The Women's Health MAP initially prioritized a list of clinical conditions affecting women, based upon the prevalence of the conditions, the morbidity and mortality resulting from these conditions, and the ability of health care systems to affect care relevant to these conditions. The top conditions included cardiovascular disease, unintended pregnancy, osteoporosis, breast cancer, mental health, violence, lung cancer, cervical cancer, and obesity/eating disorders. Within each of these categories, the MAP reviewed scientific evidence and existing measures and then outlined potential new measures. ... The MAP systematically reviewed existing literature, deliberated issues at multiple meetings, and included discussions with external content experts to support the identification and evaluation of potential performance measures in several priority areas in women's health. Over 30 indicators were considered across the prioritized list of clinical conditions. ... These high-priority indicators frequently lacked one or more key attributes necessary for successful deployment as a comparative performance indicator in women's health (McKinley et al., 2002).

Note that, depending on where the healthcare system data are collected from, some of these measures can be measured electronically, some are measured via satisfaction or other surveys, some are measured from chart reviews, some are computed or derived, and some are hybrid measures derived from multiple sources. Thus, even after the data to be analyzed are well-defined, the collection of data itself remains a challenge.

Not surprisingly because of the numerous problems with data availability and collection, very few examples exist of gender-based quality measurements in healthcare. In their comprehensive review of quality of care data, Bird, Fremont et al. (2003) found just two examples of managed-care systems that are beginning the process of reporting quality results by gender:

Although we did not find any plans that were directly comparing objective quality measures by gender, several plans are working on addressing this issue and we did find a couple of examples worth noting. PacifiCare, based in California, recently began calculating performance on various quality measures, including several related to cardiovascular care, separately by gender. However, the data were used to prepare gender-specific reports of quality of care, rather than to compare men's and women's quality of care or to develop a quality improvement program aimed at women's care. In other words, the report on the quality of women's health only included performance scores for women enrollees in each plan. The other intriguing example comes from Cigna, which rather than stratifying existing quality measures has hired an expert consultant to develop gender-specific quality measures possibly with different parameters for men and women. Their intent is to ensure that quality measures are meaningful and reflect appropriate care for men and women and are based on the latest evidence showing differences in risk factors and physiology by gender.

In sum, much work remains to be done to collect, analyze, and report on sex- and gender-based medical differences. The MAP HEDIS effort concluded:

Health services researchers and funding agencies must now focus their efforts on identifying and overcoming the measurement challenges reported here: 1) better defining the scope of women's health, 2) recognizing and utilizing the expanding research base on women's health, 3) exploring gender differences in existing measures that are not presently reported by gender, 4) overcoming the problems of small sample sizes, 5) acknowledging and incorporating the slow diffusion of clinical practice guidelines, and 6) responding to the burden of reporting created by expanding measurement sets. These are large, but not insurmountable challenges, and they will require the development of more creative approaches to quality measurement, and a continuing focus on the quality of care provided to women through our nation's managed care organizations (McKinley et al., 2002).

While there has been an historical shortage of data (on women in particular), some recent progress is worth noting. NIH's Office of Research on Women's Health (ORWH), for example, is helping to facilitate the reporting of sex-based data from clinical trials:

NIH was mandated to initiate and support the necessary outreach efforts to recruit and retain women as volunteers for clinical studies. Fueled both by the Act and by ORWH's existing mandate and established policy, a tracking system to monitor inclusion of women in research was initiated, in collaboration with several components of NIH including the Office of Extramural Research. For the first time in NIH history, it became possible to identify women's participation in human subject research, to accurately determine numbers of participants, and to productively analyze that data. .... The ORWH experience continues to demonstrate that strength, value, and utility rise from our differences. The understanding of variations and specificities in health and disease among women and men has fueled fresh insight far more than it has minimized either research arena. Women's health research has amply demonstrated its ability to add to the pool of scientific knowledge while shedding a sometimes unexpected light on men's health issues (Pinn, 2004).

Thus, while many barriers exist, progress is being made, and will continue to be made, to facilitate the collection and reporting of data on gender-based medicine.

This research question concerned the effectiveness of sex- and gender-based healthcare practices. In answering it, we must start with the caveat that most of our answer concerns the effectiveness of women's healthcare practices. Gender-based healthcare practices-that is, medical care that effectively addresses differences in men's and women's healthcare needs-is still in its infancy. As such, almost no data exist on the overall effectiveness of such approaches. We suspect, as do others (Legato, 2006b), that when gender-based medical practices are fully implemented, they will be highly effective. However, this assumption is speculative because this promising area of evidence based medicine is itself in its infancy.

The women's health movement, however, is beginning to produce effectiveness results, and these are quite promising. For now, women's care has primarily meant the treatment of women in a women-only setting. Accordingly, women's health clinics have arisen in the United States as well as in other countries. Universally, the women who attend such clinics rate their overall satisfaction as quite high. In addition, however, studies are showing that these women are receiving better quality care⁸, especially better preventive care on traditional women's issues (e.g., screening mammography).

CoEs clearly are leaders in this kind of woman-specific care, and the CoE model, as a women's healthcare model, can only be viewed as a success. This assessment, however, is primarily based on research undertaken by the CoEs themselves. Thus, while CoE-based studies demonstrate a variety of successes, additional comparative work is needed to show the relative and absolute contribution of these centers to women's healthcare above and beyond what may be occurring in healthcare delivery.

The primary barriers to implementing this kind of women's care model are inertia (that is, a lack of extrinsic motivators to change the way care has always been delivered), as well as insufficient funding and space. Having a national recognition program like that provided by the CoEs is very helpful in overcoming these barriers.

Finally, we must note that data to demonstrate the effectiveness of sex- and gender-based approaches to care are extremely sparse. Women are still at a deficit in terms of their inclusion in medical research trials. Likewise, existing reporting systems both fail to report on gender and do not collect adequate sample sizes to allow for gender-based analysis. These issues are slowly being addressed, however, and we expect the data available to analyze potential sex- and gender-based differences in healthcare to continue to improve. For now, though, we must conclude that data are insufficient to answer the question of whether sex- and gender-based approaches to healthcare are effective. As noted, we suspect that such approaches will be effective, based on the documented differences in men and women that appear to have clinical significance. The resolution to this question, however, requires additional research.

Current as of March 2007