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

January 30, 2007

Research Question 1: Key Sex and Gender Differences

1. What key sex and gender differences exist in the current healthcare system structure, including reimbursement mechanisms, content of and access to appropriate care, quality of care, and patient safety?

Our first research question focuses on the key sex and gender differences in the current U.S. healthcare system. For this section of the report, we focus on these differences as they pertain to access to care and quality of care (which includes content of care and patient safety).¹

While access differences are relatively easy to quantify, it is more difficult to accurately catalog sex- and gender-based differences in quality. There is a growing body of research discovering sex-based differences in treatment approaches, responses to treatment, speed and accuracy of diagnosis, etc., as well as possible improvements to the healthcare system to account for known sex- and gender-based differences. However, very few studies have extensively investigated how the practice of gender-specific medicine affects care outcomes.

In this section of the report, we begin with a discussion of sex- and gender-based differences in access to healthcare, and follow this with a discussion of sex- and gender-based differences in healthcare quality.

To begin our examination of research question 1, we look at the fundamental question of access. How do women and men access medical care in the United States? What barriers impede access to care? How do patients enter the medical system, and do these access points fragment² care?

We first describe research as it relates to sex- and gender-based differences in access to care, and then look briefly at how race, socioeconomic status, geographic location, and other factors may impact access. As will be described in more detail in this section and elsewhere, access to the U.S. healthcare system is exceedingly complex. The U.S. has no centralized system of care, and thus patients enter care through multiple entry points that change as patients age, switch jobs or insurance plans, or lose coverage. Women's access to medical care is affected by their use of gynecological services, which tends to additionally fragment care.

Among U.S. adults, non-emergency care is typically initiated by patient decisions. While the factors motivating care-seeking vary by individual and situation, our review uncovered one model that offers a helpful description to conceptualize this process. Developed by Currie and Wiesenberg (2003), the model was based on women in Third World countries but is generally applicable. It has three phases:

1. Do I have a health problem? This includes perception and knowledge of illness.
2. Can I seek healthcare? This includes financial resources, time, empowerment to seek care, and personal mobility.
3. Will I seek healthcare? This includes ease of accessing care, perceived quality of care, and attitudes about healthcare workers.

The model outlines basic issues related to access to care, and provides a starting point to look for gender differences in care-seeking.

As described in Currie and Wiesenberg's (2003) model, the recognition of illness is not sufficient to motivate care. Many barriers can impede access to care, and some of the most frequent of these barriers are described in this section of the report.

One of the primary barriers to access is lack of insurance coverage.³ Studies have shown that women with insurance coverage are more likely to use healthcare services. For example, "women who had insurance coverage for hospitalization and prescription medication were significantly more likely to use these services. ... Similarly, near-elderly women whose insurance covered some cost of their dental care were 2.4 times more likely to report a dental visit over the past 2 years" (Xu, Patel, Vahratian, & Ransom, 2006). Uninsured women are more likely to report that they have not visited a provider in the past year, that they have delayed care, that they were not able to see a specialist, and that they did not fill a prescription (Salganicoff, 2004). This lack of access also has been tied to worse medical outcomes:

Lack of health insurance coverage is associated with reduced access to health care and poorer medical outcomes. Several studies show that uninsured persons are less likely to have a regular source of care, less likely to receive preventive and primary care, less likely to receive required preventive services, and more likely to delay needed medical care than insured persons. Another study found that uninsured persons are more likely to be in poor health and to die earlier, even when analyses controlled for other sociodemographic factors (Freeman & Lethbridge-Cejku, 2006).

Ability to pay is yet another driver of access to care, and women tend to have lower incomes than men, making this a larger barrier for women (Xu & Borders, 2003). A recent study found that among low-income women one-third "delayed or went without needed care in the prior year because they didn't have insurance. Half (52%) of poor women and 38% who are near poor (100% to 199% of poverty) report they delayed or did not get needed health care because of the cost. ... Two-thirds of uninsured women (67%) report delayed/forgone care due to costs" (Salganicoff et al., 2005).

Indeed, cost is a frequently cited barrier. A study of women's mental health needs found that "cost is one of the most frequently reported barriers" to care-seeking, and this is true especially among the uninsured (Sherbourne, Dwight-Johnson, & Klap, 2001).

Other barriers to care include the following:

• Illness or disability: Those in poor health face additional challenges accessing care, "including costs, lack of insurance, and limited access to specialists," and "these additional barriers may further worsen health" (Salganicoff et al., 2005). This problem is especially pronounced among those with disabilities. For example, one study found that "although women with disabilities place a high value on preventive health screenings, they often do not participate in them because of the difficulties associated with living with a disability" (Mele, Archer, & Pusch, 2005).
• Caregiving responsibilities: Women are more likely than men to be caregivers (9% of women vs. 4% of men) (Donelan et al., 2001), and caring for another makes it more difficult to access care for oneself. A comprehensive study found that women caregivers "are 50% more likely than other women to report they experienced difficulties in getting needed medical care services," and twice as likely to not get care when they need it or to not fill a prescription because of costs (Donelan et al., 2001). Women with young children also mention the absence of childcare as a barrier (Khoury & Weisman, 2002).
• Lack of time: Time to access care is an important barrier for all people, but women may be especially pressed for time because of their multiple roles. For example, one study found that "younger women (ages 18 to 34) often reported extreme role overload (due to parenting as well as working or schooling) and therefore wanted health care providers who could see them promptly when they needed care, did not keep them waiting, provided services efficiently, and were willing to provide information or prescriptions by telephone" (Scholle et al., 2000).
• Wait times and inconvenient locations: Long wait times are a barrier to care, especially for men: "Waiting times of 30 minutes or longer in a physician's office sharply reduced the likelihood of a man's having visited a doctor" (Xu & Borders, 2003), and "waiting longer than one week for an appointment for regular or routine care is a strong negative predictor of men's ratings of their health plan and of men's getting needed care, but this does not hold for women" (Weisman, Henderson, Schifrin, Romans, & Clancy, 2001). Another study (this one based in Canada where coverage is universal) found that "waiting times were the most highly mentioned barrier followed by the geographic location of doctors' offices relative to respondents' residences" (Wellstood, Wilson, & Eyles, 2006).
• Language or cultural barriers: Poverty, birth outside the United States, and lack of citizenship all have been linked to lack of access to healthcare (Freeman & Lethbridge-Cejku, 2006). Additionally, factors such as "the availability of translator and other culturally sensitive services" help to facilitate care (Khoury & Weisman, 2002).

Men and women tend to access the healthcare system in markedly different ways. Three of the primary drivers of these differences are women's use of obstetrical and gynecological services, women's greater tendency to have a primary care provider and/or a usual source of care, and women's greater willingness to seek preventive care.

Obstetricians and gynecologists are key sources of medical care for women, and "gynecologists are also the exclusive providers of primary care to substantial segments of women in the United States, especially women in their prime reproductive years" (Lentz, Ayala, & Eckert, 2006). One study of low-income women found that nearly half "identified the ob/gyn clinic as PCP [primary care physician]," and this was true even while these women reported a "similar burden of physical and mental health problems" to other women in the study (Scholle & Kelleher, 2003). Additionally, "nearly one half of all obstetricians and gynecologists consider themselves to be primary care providers" (Lentz et al., 2006). Thus, gynecologists are a critical source, and sometimes the only source, of primary care for women. Moreover, there is some evidence that women relying on a gynecologist for primary care during pregnancy may not replace this source of care after the birth of a child (Feinberg et al., 2006), further highlighting the importance of these providers as a means to reach women.

It is worth noting that use of ob/gyns varies by age. Elderly women are more likely to use a regular care provider (and not to see a gynecologist), whereas women in their reproductive years are more likely to rely on a gynecologist and not to see a regular care provider (Salganicoff et al., 2005).

Unfortunately, women who utilize a gynecologist as their primary source of care do not receive the same care as women seeing both an internist and a gynecologist. For example, one study of women attending the same women's health clinic found that gynecologists were "practicing as specialists and providing little primary care medicine" (Lentz et al., 2006). It appears that women receive the best prevention screening when they receive care from both a generalist and an ob/gyn: "Seeing two physicians for regular care (a generalist and an ob/gyn), compared with seeing a generalist alone, is consistently associated with nonelderly women's receipt of more recommended clinical preventive services" (Henderson, Weisman, & Grason, 2002).

This use of two doctors is growing more frequent among women. A significant portion of American women (one-third to one-half) already receive such care (Bean-Mayberry, Chang, McNeil, & Hudson Scholle, 2006). One way to both promote and coordinate such dual care is to have women attend comprehensive clinics that "combine ob/gyn and general medical care services in a single site" (Scholle & Kelleher, 2003). Some research suggests "these centers achieve improved rates of preventive care and satisfaction" (Scholle & Kelleher, 2003). Another system that improves access to multiple care providers is managed care plans. Such plans have "facilitated access to ob/gyns and reinforced the pattern of seeing both a generalist and a reproductive health specialist for regular care" (Henderson et al., 2002). Women in such plans "report significantly greater use of two physicians" (Henderson et al., 2002).

In general, women are "more likely to have a usual primary care provider than men" (Agency for Healthcare Research and Quality, 2004), and are "more likely to enter treatment through primary care [as opposed to emergency services]" (Hauenstein et al., 2006). Having a regular source of care also is an important predictor of the use of other services. For example, "women with a connection to the health care system are more likely to have their mental health needs met" (Sherbourne et al., 2001).

Having a regular source of care is associated with other positive benefits as well: "Women who have a usual source of care (a place they usually go when they are sick) are more likely to receive preventive care, to have access to care (as indicated by use of a physician or emergency room, or not delaying seeking care when needed), to receive continuous care, and to have lower rates of hospitalization and lower health care costs" (Health Resources and Services Administration, 2005b).

Thus, women appear more likely than men to enter the healthcare system, and are more likely to have a regular point of access through which to enter the system. One of women's key points of access is gynecologists, although women receive better care if they visit both a gynecologist and a generalist.

Because of the use of multiple providers, women's care tends to be fragmented. While this fragmentation can improve care, as described previously, it also complicates the coordination of care and increases redundancies. Women tend to use three main types of doctors who provide overlapping services: "family physicians, internists, and gynecologists" (Gonen, 1999).

Fragmentation of care can create problems with knowing where to receive care. For example, "there is no clear specialty ownership of problems of the breast. The breast has traditionally been a sort of orphan organ, and it seems that unless it is a seriously diseased breast, no specialty really wants to take responsibility for it. As a result, primary care clinicians' lack of expertise in the management of common breast problems is out of proportion to what is known" (Carlson, 2001). Additionally, fragmentation creates measurement problems, as it is not always clear what services are being offered where, making tracking of services or treatment challenging (Carlson, 2001). Finally, when care is fragmented, it becomes more difficult for specialists to collaborate on care (Unutzer, Katon, Sullivan, & Miranda, 1999).

Men and women exhibit different patterns in the types of services they use. Overall, women use more services than men. For example, "in 2002, females of all ages made almost 530 million physician office visits, compared to only 361 million visits made by males" (Health Resources and Services Administration, 2005b, data from National Ambulatory Medical Care Survey, 2002). In general, "males (65 percent) were less likely than females (78 percent) to have an office or outpatient visit" (Ro, Casares, Treadwell, & Thomas, 2004).

As previously noted, women are more likely to obtain preventive services, with almost one in five women's visits for prevention purposes; the most preventive visits are made by women between the ages of 25 and 44 (Health Resources and Services Administration, 2005b). Additionally, "ninety-four percent of women received preventive health care during the year" (Taylor, Larson, & Correa-de-Araujo, 2006). Consistent with this emphasis on prevention, women may be more likely to seek "medical care earlier in their disease stage" and to seek "healthcare on their own volition rather than relying on referral from a healthcare provider" (Box, Olsen, Oddone, & Keitz, 2003). Despite this increased desire to seek care, "women may have slightly more problems getting referrals, equipment, and assistance" (Cleary, Zaslavsky, & Cioffi, 2000).

Type of health insurance plan is a key component of use of services. Some data indicate that "enrollment in plans with richer coverage for office visits is significantly associated with greater use of Papanicolaou smears and mammograms, and that this association is more pronounced among hourly workers for whom there is a greater difference in out-of-pocket expenses for office visits between PPO and the FFS plans" (Friedman et al., 2002). For women, coverage by a managed-care plan (as opposed to a fee-for-service plan) leads to more gender-specific preventive services: "Women in managed care are significantly more likely than women in other types of health plans to report receiving a physical breast examination and Pap smear in the past 12 months. Women aged 40 years and over in managed care plans are more likely than women in traditional fee-for-service plans to have received a mammogram" (Weisman & Henderson, 2001).

Just as sex and gender are key influencers on healthcare, access patterns also vary by factors such as race, ethnicity, socioeconomic status, geography, and sexual orientation. Each of these variables can interact with others, which further complicates analysis. In this section of the report, we briefly describe access trends as they relate to these five main variables. Our intention is to provide a broad outline of where these factors might interact in important ways with sex or gender.

Across the board, U.S. minority populations face greater barriers to healthcare access. These barriers include the following:

• More limited health insurance coverage: "Almost one-third of married Hispanic or Latino women did not have health insurance coverage compared with 7.5% of married non-Hispanic white women and 11.6% of married non-Hispanic black women" (Freeman & Lethbridge-Cejku, 2006).
• Insurance coverage is worse among those born outside the United States: "Foreign-born Hispanic or Latino women were 2.3 times as likely as U.S.-born Hispanic or Latino women to lack health insurance coverage. For foreign-born Hispanic or Latino women, lack of health insurance coverage decreased with the number of years they resided in the United States" (Freeman & Lethbridge-Cejku, 2006).
• Fewer usual places of care: "Overall, the percentage of Hispanic or Latino women who had a usual place to go for health care was 80.2%, significantly lower than non-Hispanic white women (91.7%) and non-Hispanic black women (89.5%)" (Freeman & Lethbridge-Cejku, 2006).
• Hispanics are less likely to have a regular doctor: African American and white women with insurance are more likely to have a regular doctor than Hispanic women (Salganicoff et al., 2005), while non-Hispanic white or Asian men with higher socioeconomic status are also more likely to have a regular doctor (Ro et al., 2004).
• Less preventive care: "Compared with non-Hispanic white women, Hispanic or Latino women were more likely to receive routine or preventive care in a clinic or health center and less likely to receive such care in a doctor's office. Overall, 56.4% of Hispanic or Latino women did not get routine preventive care anywhere compared with 41.4% of non-Hispanic white women and 40.9% of non-Hispanic black women" (Freeman & Lethbridge-Cejku, 2006).
• Fewer physician visits: "Overall, most women (87%) have visited a doctor in the past year; however, there are differences among subgroups of women. Nearly a quarter (24%) of Latina women have not had a physician visit, compared to 14% of African American and 11% of white women" (Wyn, Ojeda, Ranji, & Salganicoff, 2004).
• More barriers to care: "Nearly one in five Latinas (18%) and one in 10 African American women delayed care because of transportation problems, compared with 5% of white women. Latinas (15%) are also more likely to report that problems in obtaining childcare resulted in delayed or unmet care" (Wyn et al., 2004).

Educational level (which corresponds to income) also has an influence on the use of healthcare services. Taylor, Larson et al. (2006) report the following:

• "Women with more than 12 years of education had a higher probability of using any medical care (93%) and ambulatory care (84%), but a lower likelihood of an inpatient hospital stay (9%) or home health service use (2%) when compared to women with less than 12 years of education."
• "Women with higher education (97%) were also more likely to report using preventive health services than women with 12 years of education (92%) and those with less than 12 years of education (90%). In addition these women reported fewer average numbers of prescriptions (14) compared to women with high school graduation (16) and women with less than 12 years of education (20)."

Geographic area is another key influence on access to care. Rural residents typically have lower incomes and more limited education, and are more likely to be uninsured (Larson & Correa-de-Araujo, 2006). In addition, rural areas have the following characteristics:

• More limited access to overall healthcare services: "There are fewer health care providers per capita in rural areas than in non-rural areas. Although 20% of Americans live in rural areas, only 9% of the Nation's physicians practice in rural areas" (Agency for Healthcare Research and Quality, 2004).
• More limited access to specialty health services: "Both men and women living in the most rural areas receive less specialty mental health treatment than do those living in MSA [Metropolitan Statistical Area] counties" (Hauenstein et al., 2006).
• More structural barriers: "Rural residents also face structural barriers, including a finding that they are more likely to travel long distances to obtain health care than their urban counterparts" (Larson & Correa-de-Araujo, 2006).

These additional barriers mean that rural care may be worse overall. For example, "rural women appear to be less likely to report cholesterol checks and mammograms during the previous 1 and 2 years, breast exams during the past 1 year but not 2 years, and dental exams during the past 2 years when compared to women from large MSAs" (Larson & Correa-de-Araujo, 2006).

There is some research that suggests that lesbians and gay men may be less likely to seek preventive care: "Recent studies indicate that lesbians continue to have lower rates of preventive care including cancer-screening services such as mammography or Papanicolaou (Pap) tests than do women in the general population. Other studies indicate that gay men are also less likely to seek preventive health care than their heterosexual counterparts" (Bonvicini & Perlin, 2003). Lesbians (but not gay men) may also be less likely to have health insurance, to see a doctor, and to have a usual source of care:

In analyses controlling for all background characteristics, women living with a woman had significantly reduced odds of having health insurance (odds ratio, 0.6), of reporting a health care visit in the previous year (0.7) and of saying that they had a usual source of care (0.5); their odds of having forgone medical care in the past year because of cost issues were nearly twice those of women living with a male partner (1.9). Men who lived with a male partner had significantly elevated odds of having seen a health care provider in the last year (1.6); they had marginally elevated odds of reporting a usual source of care, and did not differ from men living with a female partner with regard to health care coverage or unmet medical needs (Hollander, 2006).

Having considered access issues, we now turn to the question of gender- or sex-based differences in the quality or content of healthcare. This is the most basic question in this literature review. Does one sex enjoy better treatment or outcomes? Are diagnosis or referral rates different between the sexes? Are observed differences between the sexes clinically significant?

To begin, we look at sex-based models of interactions with the healthcare system, as well as research issues related to examining these trends. We next present our findings pertaining to differences in healthcare quality between the sexes, looking in particular at differences in outcomes, diagnoses and referrals, and overall quality of care. Next, we identify current expert opinion on these differences. Finally, we highlight areas where race, ethnicity, socioeconomic status, and geographic or other factors may impact these overall trends.

In our review, we uncovered three models that describe some of the key differences between men and women and their relationship to the healthcare system. We briefly describe each model. While each model is unique, all point to potential sources of differences regarding effective and efficient treatment for male and female patients.

The first model was developed by Kasle, Wilhelm, and Reed (2002). It was based on women's views of what constitutes wellness. The main finding was that "women defined health and well-being largely in terms of relationships." Thus, "the appropriateness of the use of a purely biomedical model for understanding women's health can be questioned because it conceptualizes the individual as if in isolation from relational and social contexts. In contrast, our data suggest a model in which health is inseparable from interpersonal and social circumstances" (Kasle et al., 2002).

The second model was developed by Bekker (2003) and is called the "Multi-Facet Gender and Health Model." It was developed because "the relationships between sex, gender, and health have many facets ... resulting in sex differences in health including prevalence of illnesses, illness behaviour, treatment outcome, and implications of illnesses." The model shows how sex/body type interacts with gender, and how these two factors interact with sex differences in daily life, coping experiences, and differential diagnostics and treatment to lead to overall differences in health and outcome. Her conclusion is that "the relationships between biological sex, gender (masculinities and femininities), and health are multiple and highly complex, and, thus, generally, not reducible to one (type of) factor" (Bekker, 2003).

The final model was developed by van Middendorp, Geenen, et al. (2005) in a study about rheumatoid arthritis. The authors were looking at differences in emotional regulation between men and women suffering from this condition, and used a structural equation modeling approach to understand potential differences. The authors' conclusion is that "relationships between emotion regulation and perceived health were more frequent and stronger for women than men. This held especially for the affective dimension of health" (van Middendorp et al., 2005). The observation that "women are more emotionally oriented than men" in regards to their care needs offers support for gender-based health interventions (van Middendorp et al., 2005).

These three models suggest multiple ways in which care for women may differ from care for men. Two of the three models specifically describe how emotional components of health are more important to women than to men, and all three models describe how health differences between the sexes are more complex than matters of biology or physiology, and must take into account each sex's definitions of what constitutes overall health.

One of the continuing problems in trying to examine sex- and gender-based differences in healthcare is a lack of available data to study possible differences (see question 5 for a more detailed discussion on this topic). As Correa-de-Araujo and Clancy (2006) noted, "in-depth analysis of gender differences in the quality of health services provided in the United States is still limited." Partly, this is because men still outnumber women in many research trials (e.g., "in heart disease trials, the male population outnumbers the female at a ratio of 3.66-1" (Azad & Nishtar, 2005)), but the difficulties extend beyond an availability of data on women.

A historical deficit of information on women as medical research subjects means that the unique biology and physiology of women is not yet well-understood. For example, "although women have been included in clinical trials during the past 5-10 years, analyses of the data to address questions in women, men, and various racial/ethnic groups are lacking. Compounding factors are small numbers of women, women not included in early phase clinical trials, and weight or body mass index (BMI) not being considered" (Anthony, 2005). An additional factor to consider is women's unique biological rhythms:

Biological processes and functions in women are well organized in time, as evidenced by the expression of ultradian (high frequency), circadian (~24-hour), circamensual (~monthly), and circannual (~yearly) rhythms and by the changes that occur with menarche, reproduction, and menopause. Attributes of women's circamensual structure have been explored in depth, particularly with regard to fertility/infertility and birth control. However, the role of 24-hour and other rhythms in health, disease, and treatment has been little studied (Smolensky, Hermida, Haus, Portaluppi, & Reinberg, 2005).

Such differences need to be accounted for in well-designed research studies.

Another issue is that researchers are not yet accustomed to systematically collecting data to look for such differences. The establishment of routine reporting measurements for healthcare which include gender would

not only help identify and monitor differences in care but could also help plans and researchers disentangle factors underlying gender differences in care. For example, differences in care may reflect provider attitudes, gender differences in patient preferences, or unique barriers to certain aspects of care faced by each gender. By more closely comparing the characteristics and practices of enrollees, health care providers, and health plan policies with little or no differences to those with substantial differences, or those in which women appear to receive better care, a great deal could be learned about factors that reduce or promote gender differences across systems of care. Similarly, instances in which outcomes differ by gender but processes of care appear to be equivalent may reflect important gender differences in behavior or possibly physiology. Because causal factors and effective interventions may vary across settings, the availability of such data will also be crucial to allowing quality improvement teams to identify the specific factors underlying gender differences in health outcomes and plan performance and to initiate appropriate interventions within their respective organizations. Sharing knowledge between provider organizations about best practices for improving quality of care for women can help accelerate progress. (Bird, Fremont, Wickstrom, Bierman, & McGlynn, 2003)

Others also have called for detailed reporting to assist in identifying differences: "Few diseases have been studied with specific regard to presentation in or effect on gender. Limited information about treatments exists, since pharmacologic and clinical studies historically have focused on male subjects. Thus, an extensive amount of work is needed to identify gender variations" (Fanelli, 2004). Routinely collecting and reporting information on gender in all medical studies, as well as making such data available in treatment records, would enable more specific research on this topic.

Thus, the study of sex- and gender-based differences in healthcare continues to be limited by a historical deficit in data on women, an emerging understanding of unique issues about women's health, and the absence of standardized reporting mechanisms to collect and analyze data to facilitate research on possible differences. These are all characteristics of an emerging field, and it is expected that progress will be made in addressing these issues in the years to come. For now, however, such issues make it more difficult to determine where and when sex- and gender-based differences in health emerge, as well as to determine the clinical significance of observed differences.

The remainder of this section of the report comments on what is known to date about healthcare quality differences between the sexes.

One of the ways to look for differences in quality of care is to look for differences in outcomes between men and women suffering from the same conditions or receiving the same treatments. While this is not a true measure of quality differences, it is a reasonable proxy. Our review documented several such instances, in almost all cases showing women with worse outcomes than men. Generally, the data are not available to explain the underlying cause(s) of these outcomes. For example, are women more susceptible than men, or are women receiving different or worse treatment than men? Regardless of the cause, the observed pattern of differences suggests that women may be disadvantaged in terms of quality of care for certain diseases. Some of these diseases are discussed below.

Women appear to have worse outcomes after strokes: "Even after adjustment for age [of onset of stroke], women have poorer outcomes, such as greater pre- and poststroke disability, a higher likelihood for admission to nursing facilities, and greater mental impairments than men" (Bushnell et al., 2006). This finding suggests that women may require different treatment for stroke, and that their current stroke-care may be less than optimal. Bolstering this notion is the finding that "men and women may respond differently to stroke prevention strategies, as shown in 2 different gender-specific cohorts" (Bushnell et al., 2006). Additionally, women are more likely to report non-traditional stroke symptoms than men, potentially delaying their care (Labiche, Chan, Saldin, & Morgenstern, 2002). Indeed, "women with acute stroke wait longer in the ED [emergency department] for physician evaluation and have worse functional outcome and mortality" (Washington & Bird, 2002).

Diabetes is another disease in which women may be at higher risk. One study reported that "clinicians should be aware that ... men and women with Type II diabetes appear to suffer from a wide variety of symptoms and complications and that gender differences in these conditions exist. ... Females with Type II diabetes may suffer disproportionately from the disease compared to males" (Summerson, Spangler, Bell, Shelton, & Konen, 1999).

Women do worse both physically and mentally after major trauma events than men, "independent of injury severity and mechanism" (Holbrook & Hoyt, 2004). The reasons for this difference are not understood, but the outcome gap suggests that further research is necessary to improve trauma-care for women (Holbrook & Hoyt, 2004).

Heart disease is one of the most studied conditions in which sex- and gender-based differences in outcomes have been observed. Most of these differences favor men, and men appear to have a better long-term survival rate than women. Complicating the study of these outcomes is that women are usually older and more likely to have diabetes when they have their first coronary event; therefore, they don't get invasive procedures as often and are more likely to die (Boccardi & Verde, 2003). Notwithstanding these differences, the data do suggest that women may receive worse care:

• "The time to presentation was longer and diagnostic mistakes more frequent in women than in men, whereas medical treatment and the rate of revascularization procedures were similar" (Boccardi & Verde, 2003).
• "After coronary bypass surgery, operative mortality is higher for women ... in part due to excessive rates of congestive heart failure. Similar findings are observed in angioplasty registry data. Additionally, of those with heart failure symptoms, women are more likely to have a lower quality of life with more frequent depression when compared with men" (Bean-Mayberry et al., 2006).
• "Although advances in technology have contributed to improvements in the diagnostic and therapeutic approaches to myocardial infarction, women do not seem to benefit from such advances, as they continue to suffer inequalities in the utilization of such procedures and to have worse outcomes from AMI [acute myocardial infarction] than men do. It is possible, therefore, that the worse outcomes for AMI observed among women compared to men are caused by inequalities in the quality of care they receive. Such inequalities were observed in our study with disappointing frequency" (Correa-de-Araujo, Stevens et al., 2006).
• "The in-hospital mortality among women with acute myocardial infarction tends to be greater than in men" (Boccardi & Verde, 2003).
• "There is a significant difference in the number of observed/expected cases of PA [pulmonary artery] rupture, with females being greatly overrepresented and males underrepresented" (Kaczmarek, Liu, & Gross, 2003).
• "The results of this study support results found in the literature that women do not have as favorable outcomes after CABS [coronary artery bypass surgery] as men, particularly on the physical outcomes. ... Overall, women in this study seemed to perceive that they experienced poorer outcomes after CABS compared with men preoperatively and at both the 1-month and three-month postoperative periods. The women reported lower levels of activity, vigor, and overall health rating and higher levels of SOB [shortness of breath] and depression" (Keresztes, Merritt, Holm, Penckofer, & Patel, 2003).

The final category of outcome differences we uncovered relates to adverse drug reactions. Women are more likely than men to suffer such reactions (Anderson, 2005; Miller, 2001; Rademaker, 2001), and such reactions may be "50 to 75% more likely in women than men" (Rademaker, 2001). Reasons why women may be more vulnerable to adverse drug reactions include the following:

• Women use different drugs than men: Women use a different range of drugs, especially related to contraception, menopause, and pregnancy (Rademaker, 2001).
• Women take more drugs than men: "Women take drugs for contraception, and a woman's reproductive function increases the need for medical treatment at an earlier age than for men. Drug usage information shows that women use about 60% of all medications" (Miller 2001). Women are more likely to have a chronic condition than men (Salganicoff et al., 2005), and this may also increase their drug usage.
• Women have different body compositions: There are different body weights between males and females, as well as different percentages of body fat, all of which influence drug reactions (Anderson, 2005). Additionally, "genetics, age, height, weight, lean-fat ratio, diet, exercise, concurrent disease, smoking and alcohol, and the administration of concomitant drugs all contribute to antipsychotic drug response, as does end-organ sensitivity. Together, these factors can account for a 10-fold variability in the dose needed for effective response. Men and women show differences in all of these variables, either as a result of the action of sex-specific hormones or of divergent gender roles" (Seeman, 2004).
• There are sex-based differences in drug responses: "Sex-based differences in drug response can be due to the drug itself, resulting from the activity of transporters or metabolizing enzymes. It can also be due, in some part, to the interaction of the physiological differences between the sexes and the site of drug release and subsequent absorption controlled by the delivery system. The combination of sex-based physiological differences with the fixed-release rate of these oral delivery systems leaves open a myriad of possibilities for differences in pharmacokinetics and pharmacodynamics to occur" (Donovan, 2005).
• There are differences in drug absorption and disposition: "Several mechanisms relevant to drug absorption and disposition have been shown to exert gender-specific activity differences, and for some drugs these have the potential to result in clinically relevant differences in pharmacological response" (Meibohm, Beierle, & Derendorf, 2002). Additionally, "all or some of the pharmacological dynamics of a medication (e.g., stereochemistry, drug delivery systems, pharmacokinetics, pharmacodynamics, pharmacogenetics, chronopharmacology, and modulators [ions]) can differ between women and men" (Correa-de-Araujo, 2005).

When drugs are specifically tested to look for potential differences, such differences appear to be at least somewhat common. One study found that "the Food and Drug Administration (FDA) reviewed 300 new drug applications between 1995 and 2000. Of the 163 that included a sex analysis, 11 drugs showed a >40% difference in pharmacokinetics between males and females, which was listed on the product label, yet no dosing recommendations were made based on sex" (Anderson, 2005).

This suggests that there is a need to do additional such testing for differences, especially in Phase I and II clinical trials where potential side effects and reactions are first observed (Mastroianni, Faden, & Federman, 1994).

Another indicator of possible differences in quality of care between men and women is differences in diagnosis or referral rates for treatment of the same condition when there is evidence that prevalence rates for men and women are similar. In other words, if one sex is more likely to be correctly diagnosed or referred to effective follow-up treatment, this is a possible indicator of higher quality of care. However, it is possible that other factors unrelated to quality could cause these discrepancies (e.g., age of presentation). Some of the studies we found on this topic control for such factors, and results from these studies should be viewed as more robust.

As in our examination of differences in outcomes, most of the data we found suggest that men receive better care than women. However, there are several examples in which women appear to receive better care, and these are noted as well.

The following are among the numerous examples of men appearing to receive better care (based on a higher percentage of cases diagnosed or more referrals or follow-ups):

• Women receive less health counseling: A study of the Department of Veterans Affairs care found that women were less likely to be prescribed nicotine patches to assist them in quitting smoking (Sherman, Fu, Joseph, Lanto, & Yano, 2005). Also, physicians "are less likely to counsel women than men about diet, exercise, and weight reduction" (Agency for Healthcare Research and Quality, 2004).
• Women receive fewer high-technology treatments: White men are 2.5 times as likely as white women to be recommended for kidney transplants for treatment of end-stage renal disease, even controlling for clinical factors (Agency for Healthcare Research and Quality, 2002). Women also are "less likely to receive expensive, high-technology services, such as dialysis and transplantation, timely diagnosis of lung cancer, and specific diagnostic and therapeutic interventions for heart disease" (Bird, Shugarman et al. 2002).
• Women may have a more difficult time receiving basic care: A Medicare managed-care survey found that "women may have slightly more problems getting referrals, equipment, and assistance"; women also were less likely to get a flu shot (Cleary et al., 2000).
• Women receive worse diabetes care: Women may receive fewer secondary preventive services for diabetes; e.g., "women were less likely to have a lipid profile or receive retinal eye and foot exams" (Correa-de-Araujo, McDermott, & Moy, 2006). Additionally, women with diabetes "were less likely to use lipid-lowering medications and to receive a lipid profile" to assess and help manage cardiovascular disease risk (Ferrara, Williamson, Karter, Thompson, & Kim, 2004).
• Women are screened less often for colon cancer: Older women "may not be getting the most effective colorectal screening as often as older men" (Kosiak, Sangl, & Correa-de-Araujo, 2006). In general, women of all ages may be less likely to undergo screening for colon cancer, and this problem is exacerbated by women receiving primary care from gynecologists who may be less inclined to screen for this disease (Woods, Basho, & Engel, 2006).
• HIV-infected women may receive less aggressive treatment: HIV-infected women are "less likely than [HIV-infected] men to receive HAART [highly active antiretroviral therapy] (78% versus 82%, p < .001), receive PCP [Pneumocystis jiroveci pneumonia] prophylaxis (65% versus 75%, p < .0001), or have their hepatitis C virus status known (87% versus 88%, p < .02) despite being seen more regularly (69% versus 66%, p < .04)" (Hirschhorn et al., 2006).
• Women receive worse cardiology care: Women undergo procedures "such as angiography or angioplasty, less frequently than men" (Penco et al., 2003), and are "less likely to undergo diagnostic and therapeutic cardiac procedures" (Bird et al., 2003). Women are "40% to 50% less likely to have an angiogram than men even after adjustment for multiple clinical factors, including the results of exercise testing" (Daly et al., 2006). Women also may experience longer delays in the administration of thrombolysis (Grace et al., 2003). Women experience worse treatment for AMI [acute myocardial infarction], with fewer secondary prevention treatments and less rehabilitative therapy than men (Penco et al., 2003). Finally, men are "more likely to receive cardiac catheterization, PTCA [percutaneous transluminal coronary angioplasty], and CABG [coronary artery bypass graft] than women" (Jiang, Elixhauser, Nicholas, & al., 2002).

We found a smaller number of examples in which women appear to receive better care than men (based on a higher percentage of cases diagnosed or more referrals or follow-ups). These include the following:

• In a diagnostics study based on case scenarios, depression was more likely to be recognized in older female patients (vs. older male patients). Depression is more common in women, and this may be one reason it was more likely to be recognized (Stoppe, Sandholzer, Huppertz, Duwe, & Staedt, 1999).
• Men "receive fewer blood pressure measurements and cholesterol measurements. ... It appears that some of the discrepancy in receipt of preventive services between men and women younger than 50 results from the difference in the number of visits to healthcare providers [women make more visits]" (Viera, Thorpe, & Garrett, 2006).

In addition to observed differences in outcomes, diagnoses, and referrals, we found several studies with strong enough controls that we felt they represented significant differences in overall quality of care between men and women. As before, most of these differences favor men over women.

The first such difference has to do with the issue of cardiovascular care. Cardiovascular care has been extensively studied for gender differences, which means that increasingly sophisticated techniques are being utilized to determine where care differences are significant. An example is a study by Weisz, Gusmano et al. (2004), which involved patients in the United States, France, and England. They found a consistency of results across these three countries that suggests that their results are quite robust. They reported that

Despite differences in health system characteristics, including health insurance coverage, availability of medical resources, and medical culture, we found consistent gender differences in rates of percutaneous transluminal coronary angioplasty and coronary artery bypass grafting across the 9 spatial units. The rate of interventional treatment in women with CAD [coronary artery disease] was less than half that in men. This difference persisted after adjustment for the prevalence of heart disease. ... This finding is consistent with the results of clinical studies suggesting that gender disparities in the treatment of CAD are due at least in part to the underdiagnosis of CAD in women (Weisz et al., 2004).

Thus, it appears that women do receive worse care for cardiovascular disease, at least for some measures. Indeed, the U.S. Agency for Healthcare Research and Quality recently reported that "significant gender differences in cardiovascular care have been demonstrated" (2004), and Cabana and Kim (2003) reported that "the gender differences in cardiovascular care have been well documented." This is consistent with the findings we reported previously that women are less likely to receive certain cardiovascular treatments.

Even where clinically significant differences in treatment are observed, however, correcting these differences remains complex. Continuing with the example of care for heart disease, women may do worse with some kinds of treatment because of their smaller vessel size. Thus, even as the medical profession becomes aware of these differences, it is still difficult to effectively address these issues:

Although recent technological advances have improved the success rate of PTCA [percutaneous transluminal coronary angioplasty] in women, reaching the same immediate and long-term results as in men, greater cardiac and peripheral procedural complications are still more frequent in women. This is mainly due to a more advanced age and more comorbidities - in particular diabetes -, smaller vessels and, probably, some sex-related factors. ... PTCA may become the preferable revascularization procedure for women (Presbitero & Carcagni, 2003).

As discussed previously, another disease in which well-documented differences exist between men and women is diabetes. An extensive literature review on this topic found that

women with diabetes, regardless of menopausal status, have a 4- to 6-fold increase in the risk of developing coronary artery disease (CAD), whereas men with diabetes have a 2- to 3-fold increase in risk. Women with diabetes have a poorer prognosis after myocardial infarction and a higher risk of death overall from cardiovascular disease than do men with diabetes (Legato et al., 2006).

The investigators also reported that most research on this subject has not effectively distinguished between the sexes, thus complicating the job of examining sex-based differences in care or response to treatment (Legato et al., 2006).

One area where women appear to receive more frequent care than men is in mental health treatment: "Overall men receive less mental health treatment in both the general health and mental health specialty sector than do women" (Hauenstein et al., 2006). This advantage for women has been found consistently and across numerous studies, including studies that controlled for multiple factors. For example, one such detailed study reported: "We found that gender differences in overall use of outpatient mental health services persist despite control for type of mental disorder and social and economic factors. Having a mood and/or anxiety disorder appears to be a direct determinant of use for men and women" (Rhodes, Goering, To, & Williams, 2002). In addition,

the persistence of the gender difference in use is compelling since this study overcame a number of the limitations of prior studies. ... At issue is the question of whether the remaining gender difference in use represents an unmet need for care for men and/or inappropriate use for women. Although the magnitude of the difference is not large (adjusted OR=1.72), the absolute number of individuals affected is large. ... Inappropriate use may lead to worse health outcomes for some women. On the other hand, some women may be seeking help at an earlier, less severe stage of their illness, which may lessen the impact of illness (Rhodes et al., 2002).

There is an entire category of differences in care that we have not yet discussed. These are observed differences that may or may not have any clinical significance. In other words, care is known to be different between men and women, but no evidence yet suggests whether these care differences are meaningfully related to health outcomes. To be fair, many of the studies we have reported on thus far might reasonably be put into this category. However, we have made what we believe are reasonable classifications regarding the findings we have presented thus far, and we believe these assumptions are likely to be correct. For example, it seems a fair assumption that a delay in a correct diagnosis for a cardiac condition would negatively impact care, or that women's greater difficulty getting referrals would harm their overall care.

Nonetheless, there are other differences we found reported in the literature that do not yet suggest a direction of clinical significance favoring one sex over the other. We include some examples of them here as a means to elucidate the difficulty and complexity of this field:

• Coronary artery disease: "CAD [coronary artery disease] affects a demographically different group of women compared with men. Women develop CAD at an older age, present with different symptoms and comorbidities, and respond to these symptoms differently than men. Once age, comorbidity, and contraindication are taken into account by multivariable analysis, gender differences in treatment are rendered small and, for the majority of treatments, insignificant. The exception is in the use of procedures, but the clinical significance of this different treatment pattern is unclear" (Gold & Krumholz, 2006).
• Disasters: "There is some evidence showing that women and men may suffer different negative health consequences following a disaster. It is not clear whether this is because of biological differences between the sexes, because of socially determined differences in women's and men's roles and status or because of an interaction of social and biological factors" (World Health Organization, 2002).
• Clinical visit time usage: The use of clinical visit time differs between male and female patients: "Female gender was associated with a higher proportion of the visit time spent on physical examination, structuring the interaction, patient questions, screening, and preventive services, such as Pap smear, and family and emotional counseling. Male gender, on the other hand, was associated with a greater proportion of time spent on procedures and on health behavior counseling" (Tabenkin, Goodwin, Zyzanski, Stange, & Medalie, 2004).
• Asthma: "Women report worse asthma symptoms and perceived quality of life due to their asthma while reporting better asthma care measures. Several studies have shown that although women report worse asthma symptoms or more asthma admissions, they have the same or better pulmonary function measures when compared with men. Better pulmonary function measures would suggest that asthma severity is lower for women; however, women consistently report more asthma symptoms" (Sinclair & Tolsma, 2006).

Much work remains to be done to determine whether such differences are clinically meaningful, and, indeed, to determine the importance of various factors in driving these differences; e.g., in the asthma example, it is not known whether these differences are caused by differing experience of the disease between men and women, differing perceptions of symptoms, differing treatments, or the need for sex-specific instruments to measure asthma symptoms. Some or all of these factors could be influential.

The evidence presented in this section thus far paints a splintered picture of quality, generally favoring men as the recipients of better quality across a wide number of health conditions. Here, we attempt to discuss the larger picture, drawing together the opinions of numerous experts in this field.

This discussion must begin with overall morbidity and mortality figures. In the United States, women live longer, "but they do not necessarily live those extra years in good physical and mental health. On average, women experience 3.1 years of disability at the end of life" (Agency for Healthcare Research and Quality, 2006). Women have "lower age-adjusted death rates than men for 13 of the 15 leading causes of death. However, women are more likely than men to report having arthritis, asthma, autoimmune diseases, and depression" (Agency for Healthcare Research and Quality, 2004). Denton, Prus et al. (2004) nicely summarize this literature when they report that

while women generally experience poorer health than men, the pattern of gender differences in health is varied. Women have lower rates of mortality but, paradoxically, report higher levels of depression, psychiatric disorders, distress, and a variety of chronic illnesses than men. However, the direction and magnitude of gender differences in health vary according to the symptom/condition and phase of the life cycle. Female excess is found consistently across the lifespan for distress, but is far less apparent, even reversed, for a number of physical symptoms and conditions.

In other words, the pattern of gender differences in overall health across the lifespan is complex, with women living longer but suffering more from certain diseases and conditions.

The significance of this difference is debatable. While women certainly are disadvantaged in some areas, this disadvantage may be smaller than previously assumed and appears to be shrinking. Two recent summaries come to a similar conclusion in this regard:

The idea that gender differences in health may be not as simple as they seem was recently addressed in a special issue of Social Science and Medicine (1999) that re-examined the generalization of women's greater morbidity. The studies yielded considerable variability in gender differences in health. For example, some reported that women's poorer health was marked for mental and somatic symptoms, but less so for long-term illness and self-rated health. Others observed a weak female excess for malaise symptoms and minor psychiatric morbidity and no gender differences in physical symptoms. Still others found that, despite the greater prevalence of long-term illness, physical symptoms, and psychological distress among women, the magnitude of the differences was often small (McDonough & Walters, 2001).

Recent examinations of gender differences in physical health suggest that women's disadvantage may be smaller than previously assumed, varying by health status measure and age. Using data from the 1997 & 2001 National Health Interview Surveys, we examine gender-by-age differences in life-threatening medical conditions, functional limitations, and self-rated health and consider whether potential mediating mechanisms (e.g., socioeconomic status, behavioral factors) operate uniformly across health measures. The results show that the gender gap is smallest for life-threatening medical conditions and that men do increasingly worse with age. For self-rated health, men are more likely to report excellent health at younger ages, but with increasing age this gap closes. Only for functional limitations do we find a consistent pattern of female disadvantage: Women report more functional limitations than men, and the gap increases with age (Gorman & Read, 2006).

Another factor to consider is women's use of the healthcare system compared to men: "Women use health services more than men and consume more than twice as many drugs, due in part to the fact that they have more comorbidities, such as diabetes and hypertension, and are more likely to suffer from the chronic pain such conditions often entail. In contrast, men are more likely to have acute illnesses" (Kosiak et al., 2006). Women also are more likely to "establish routine care-seeking behavior early in life by undergoing regular Pap-smear tests, receiving prenatal care, and managing well-child care in the family" (Asch et al., 2006). Women thus may be larger consumers of preventive and chronic care, while men use the system more often for acute care, which is still the focus of the U.S. healthcare system (Kosiak et al., 2006).

Consistent with this pattern of use, one study found that "women had higher scores than men for preventive care (57.8 percent vs. 50.1 percent, p < 0.001) and chronic care (57.9 percent vs. 54.5 percent, p = 0.01) but lower scores for acute care (51.9 percent vs. 58.4 percent, p < 0.001)," and "women had higher scores than men for screening (56.7 percent vs. 42.9 percent, p < 0.001) but lower scores for treatment (56.0 percent vs. 59.3 percent, p = 0.001)" (Asch et al., 2006).

This finding of lower scores for women on "acute care and treatment-related care [is] consistent with the literature on disparities in health care between men and women" (Asch et al., 2006). Indeed, "many women's health needs are inadequately addressed" (Agency for Healthcare Research and Quality, 2004), and differences remain "between men and women in the receipt of evidence-based health care" (Correa-de-Araujo & Clancy, 2006). These differences are in many cases subtle, but the ability to detect and address them is increasing as medical knowledge of gender- and sex-based differences increases. If the trend continues, further study will negate the importance of some differences that are now believed to be significant, while highlighting new discrepancies of great importance.

If a quality gap in care does exist, it is important to understand why it exists. Numerous theories have been offered, and explanations include variations in the healthcare system, gender-related differences in work patterns and resources, and different stressors and disease patterns between men and women. Short summaries of two frequent theories appear below:

• Structural, behavioral, and psychosocial forces: "The pathways through which structural, behavioural and psychosocial forces influence health are different for men and women. ... While we found that there is considerable commonality in the social predictors of health, there are indeed important gender differences. First, age, family structure, main activity, education, income, occupation and social support are important structural determinants of health for both genders, although their effects differ for men and women. Second, smoking, alcohol consumption and physical activity are more important determinants of health status for men and body weight is more important for women. Third, the impact of childhood/life events, chronic stressors and psychological resources play an important role in determining health, but their effects are generally stronger for women than men" (Denton et al., 2004).
• Constrained choices: Bird and colleagues (Rieker & Bird, 2005; Ross & Bird, 1994) have argued that gender differences in health extend beyond disparities to include other differences in men's and women's opportunities to pursue a healthy life. Their earlier work identified "gender difference in labor and lifestyles" as critical (Ross & Bird, 1994). For example, "women are less likely to be employed, and are more likely to work part-time, have lower incomes and more economic hardship, and to do more unpaid domestic labor than men. ... Women also have more distress and fewer subjective work rewards, both of which are associated with poor health" (Ross & Bird, 1994). Their more recent work has focused on the notion of constrained choices: "Constrained choices impact on men's and women's stress levels as they experience competing demands on their time and other resources, which can in turn affect their psychological and physical responses to stress. Such effects could occur through gender differences in exposure to role-related chronic stressors. ... Constrained choices may also impact positive and negative health behaviors and coping styles that affect both psychological and physical functioning. Consequently, we contend that gender differences in the constraints contribute to health disparities both directly and indirectly by affecting both men's and women's choices and their cumulative biological risk" (Rieker & Bird, 2005).

Just as sex and gender are key influencers on healthcare quality, quality of care also varies by factors such as race, ethnicity, socioeconomic status, geography, sexual orientation, and disability status. Each of these variables can interact with others, which further complicates analysis. It is important to note that the pathways through which these variables impact health are complex, including "health care affordability, geographic access, transportation, education, knowledge, literacy, health beliefs, racial concordance between physician and patient, patient attitudes and preferences, competing demands including work and child care, and provider bias" (Fiscella, Franks, Gold, & Clancy, 2000).

In this section of the report, we briefly describe quality trends as they relate to race and ethnicity, socioeconomic status, geography, sexual orientation, and disability status. Our intention is to provide a broad outline of where these factors might interact in important ways with sex or gender.

It is worth noting, however, that sex and gender disparities in healthcare differ in substantial ways from racial, ethnic, or socioeconomic disparities. Most importantly, gender cannot be treated as a single-directional "cause" of disparities, because the interaction patterns are complex (Rieker & Bird, 2005). Gender differences in health are influenced by biology and physiology, with both sexes having unique needs. In comparison, nongender disparity models assume that the influences of discrimination and inequality tend to "go in the same direction for both biological and social factors" (Rieker & Bird, 2005).

Numerous healthcare disparities related to race and ethnicity have been documented for any number of issues, including provision of screening, aggressiveness of treatment, and overall outcomes. Whites almost always receive better care and are frequently the reference group used in studies. Quality of care varies among blacks, Asians, Hispanics, and American Indians/Alaska Natives, with the most frequent differences in quality of care occurring between whites and blacks or Hispanics. This may be because "evidence suggests that institutionalized forms of racism are linked to both access and quality of care for all people of color, but are particularly detrimental to African American and some Latino subgroups" (Zambrana, 2001).

A comprehensive report produced by the Institute of Medicine (Smedley, Stith, & Nelson, 2003) found that "racial and ethnic minorities tend to receive a lower quality of healthcare than non-minorities, even when access-related factors, such as patients' insurance status and income, are controlled." The authors report that these disparities are complex, involving past and ongoing discrimination, cultural and other barriers to care, the characteristics of the clinical setting, and the complexity of the healthcare system.

Specific examples of racial or ethnic health disparities include the following:

• Black patients "consistently received poorer health care relative to that of whites across all measures. ... Blacks with diabetes were less likely than whites were to have an annual eye exam or to have HbA1C testing. Blacks were also less likely to have an ambulatory/preventive care visit in the measurement year" (Virnig et al., 2002).
• Hispanics are "less likely than whites were to have a mammogram, LDL cholesterol testing after an acute cardiovascular event or as part of their diabetes care, or an ambulatory or preventive care visit with a doctor" (Virnig et al., 2002).
• Native Americans are "less likely to receive a mammogram, [and] to have their LDL cholesterol checked as part of their diabetes care" (Virnig et al., 2002).
• "Women of color are more likely to report that they are in fair or poor health. One-fifth of African American women, 29% of Latinas, and 13% of white women assess their health status as fair or poor. African American women are more likely to have a physical condition that limits routine activities such as participating in school or work or conducting daily housework. ... Over half (57%) of African American women ages 45 to 64 have been diagnosed with hypertension, twice the rate for white women (28%) of the same age. African American women (40%) are also significantly more likely to have arthritis than Latinas (33%) and white women (32%). African American (16%) and Latina (17%) women both experience higher prevalence of diabetes compared to white women (9%). However, African American women (4%) are less likely to have osteoporosis compared to Latinas (12%) and white women (10%)" (Wyn et al., 2004).
• "Black women are less likely than other women or men to have access to lifesaving therapies for heart attack" (Agency for Healthcare Research and Quality, 2002).
• "Black women have higher death rates than white women due to heart disease, cancer, and stroke" (Agency for Healthcare Research and Quality, 2004).
• "Mammography was received less often by black, Asian, and AI/AN [American Indian/Alaska Native] women compared with white women, [and] by Hispanic women compared with non-Hispanic white women" (Agency for Healthcare Research and Quality, 2004).
• "African American Medicare patients were less likely than whites to undergo catheterization, angioplasty and bypass surgery, and Latino Medicare patients were less likely than whites to undergo catheterization and angioplasty" (Lillie-Blanton, Evadne Rushing, & Ruiz, 2003).
• "Latina, Asian/Pacific Islander and American Indian/Alaska Native women are less likely to be screened for breast and cervical cancer than are white and African American women" (Lillie-Blanton et al., 2003).
• "As with breast and cervical cancer, mortality rates from colon and rectum cancer are higher among African Americans than among whites, despite comparable screening rates" (Lillie-Blanton et al., 2003).
• "Racial/ethnic minority groups are less likely than whites to have had a dental visit in the past year, regardless of age" (Read and Gorman 2006).
• Provider behavior may be a critical component by which African Americans receive worse care. For example, one study found that "African American women as a whole may be vulnerable to prejudice within reproductive health care contexts" (Thorburn & Bogart, 2005).

While race- and ethnicity-based discrepancies in quality of care are well-documented, controls for potentially confounding factors (e.g., socioeconomic status) often reduce or eliminate such disparities. In at least one study, such controls actually indicated that blacks had better quality of care scores than whites (Asch et al., 2006). Thus, socioeconomic status is often a better predictor of disparities in care, although institutionalized racism remains a key concern and should not be overlooked. As one author noted, "health and social scientists are moving toward consensus that health disparities between racial and ethnic minority populations ... arise from psychosocial and cultural factors related to the social definition of population groups [e.g., internalized racism], as opposed to genetic differences at the population level" (Hogue, 2002).

Socioeconomic status exerts a powerful influence on quality of care. Higher socioeconomic status has multiple positive effects on health. For example, "socioeconomic status affects women's health directly through access to resources and indirectly through psychosocial factors and social roles. Lower socioeconomic position is associated with lower levels of perceived control and self-esteem, both of which are associated with greater levels of depression and poorer self-rated health" (Read & Gorman, 2006). Socioeconomic status also is linked to greater access to insurance, but the effect of being low in socioeconomic status is more than an insurance effect (although having insurance contributes to visit continuity and coordination of specialty care): "Issues such as poor financial access (i.e., out-of-pocket costs for the primary care office visit), availability of high-quality practitioners, or other unmeasured factors may contribute to the failure of some lower SES women to attain ‘quality' primary care" (O'Malley & Forrest, 2002). In addition,

women with less education have been found to be more intimidated by physician authority, and less likely to ask questions. ... Low socioeconomic status is closely associated with less access to and knowledge about favorable health care practices, more trusting or less trusting attitudes toward health care providers, less skills to successfully navigate the health care system, and less ability to afford the costs of the health care system (Zambrana, 2001).

Thus, higher socioeconomic status is linked to many variables that improve access to care, use of care, and overall health.

Healthcare related findings associated with socioeconomic status include the following:

• "Poor or near poor women are more likely to report fair or poor overall health; limitations of activity; and having anxiety or depression, arthritis, asthma, diabetes, hypertension, obesity, and osteoporosis" (Agency for Healthcare Research and Quality, 2004).
• Pap tests and mammograms are less common among women with lower socioeconomic status (Agency for Healthcare Research and Quality, 2004).
• "At the highest income level, both black and white women have a 5-year advantage in life expectancy over their male counterparts. ... Race, income, and gender all make independent contributions to disparities in health. ... White women with the highest level of income have the highest life expectancy at age twenty-five (58 years), while low-income black males have the lowest (42 years)" (Williams, 2005).
• "Low-income women, who are less mobile or empowered than other women, may have a narrower choice of health care settings or be more likely to encounter sites that are lagging in professional practices, which are demanded by more affluent patients" (Anderson, 2003).
• "Lower income and minority women not only have shorter life spans, but also experience a higher burden of illness during their lives" (Bierman, 2003).

Because of its multiple effects on health, controlling for socioeconomic status often corrects for other disparities. For example, one study on acute myocardial infarction found that "once a comprehensive set of covariables (including sociodemographic background, personal clinical history, and evidence-based medications) were accounted for, ... sex and racial/ethnic disparities no longer existed" (Iribarren et al., 2005). In another example, disparities in the rates of visiting a doctor were eliminated for African Americans when income was controlled for (Read & Gorman, 2006). Finally, "for self-rated health, the gender gap diminishes or disappears for all racial/ethnic groups once age and socioeconomic status are considered, a finding that appears to mirror those of prior studies" (Read & Gorman, 2006).

There is some evidence that rural residents experience worse care than urban residents, including overall worse reports of health, a greater frequency of chronic conditions, and higher death rates (Agency for Healthcare Research and Quality, 2004). Other factors that may vary at the geographic level also influence health. For example, work is emerging on the impact of social capital on health (e.g., see Browning & Cagney, 2002, for an example). Likewise, the "political, economic and cultural context of a community" all affect health (Gomez & Muntaner, 2005). These factors may influence health both on a larger regional level and at smaller neighborhood levels; e.g., studies have associated the receipt of health services with residing in specific ZIP codes (Elston Lafata, Cole Johnson, Ben-Menachem, & Morlock, 2001). Of interest to this review, the affect of these factors on health may differ in important ways between men and women. However, research to date has not examined the role of gender related to these issues.

Lesbians, as a subpopulation of women, may be at increased risk for adverse health outcomes. This is because of several factors, one of the most important of which is discomfort disclosing sexual orientation in a primary care setting. Lesbians who disclose their orientation are more likely to seek preventive care (Bonvicini & Perlin, 2003). Lesbians also seem to underutilize health services and present for healthcare at later stages of illness:

The infrequency of primary care visits by lesbians has been linked, in part, both to the mistaken belief that they are at low risk for gynecological diseases, and to [lesbians'] discomfort with health care providers. Lesbians also present for health care later in the course of their illnesses than heterosexual women; they are more likely to orient health care around a specific crisis rather than routine evaluation. Thus, a principal health care risk for lesbians may be delayed diagnosis due to lack of seeking appropriate care or to inadequate screening (Barbara & Quandt, 2001).

Lesbians also may have more health risk factors:

It has ... been reported that smoking is more prevalent in lesbians than in straight women. In addition, according to Hall, lesbians are more likely to abuse alcohol. As a result of these behaviors, Harrison and Silenzio reported lesbians were predicted to have a higher risk of morbidity and mortality from breast, ovarian, lung, and endometrial cancers than their straight counterparts ... one of the most significant medical risks for gays and lesbians is that they avoid routine health care (Bonvicini & Perlin, 2003).

Similarly, Diamont and Wold (2003) report "higher rates of smoking and alcohol use among lesbians and bisexuals, greater prevalence of overweight, and higher use of mental health services. In addition, negative experiences and interactions that lesbians and bisexuals have had within the healthcare system appear to discourage many from seeking needed medical care."

Finally, we consider the quality of care received by women with disabilities. This is a research topic that has been "largely ignored until recently" (Mele et al., 2005). Disability status may pose special equipment needs for provision of even routine care. For example, for women with spinal cord injuries and disorders, "pelvic examination for cervical cancer screening ... may be more difficult (e.g., difficulty getting on the examination table) and, for that reason, may be neglected. In terms of breast examinations, although proper positioning for standard mammography is achievable, often the assistance of a well-trained technician or use of such equipment as a standing wheelchair is needed" (Lavela, Weaver, Smith, & Chen, 2006). Such difficulties mean that "health care facilities are often inaccessible and ill equipped to serve people with disabilities" (Mele et al., 2005). The result is that "commonplace procedures like CBEs [clinical breast exams], Pap smears, and mammograms are overlooked or omitted for women with disabilities because of inaccessible office environments" (Mele et al., 2005).

In this section of the report, we examined differences between men and women related to access to healthcare and quality of healthcare. Our findings suggest that women remain disadvantaged compared with men in both these regards.

Because women have fewer resources in terms of time, money, and insurance coverage in their own names, it is more difficult for them to access the healthcare system than it is for men. However, women are more likely to seek preventive care than men, perhaps because they have been socialized to seek care before a health crisis occurs. Additionally, a higher proportion of women than men become more accustomed to seeking routine medical care because of women's use of reproductive medical services. Gynecologists remain a key access point to the healthcare system for women, especially among women of reproductive age.

Additionally, women appear to receive worse quality of healthcare than men do. This appears to be especially true for acute conditions. However, women, as more frequent users of preventive care, may actually receive better preventive care than men. The fact that women's higher rates of healthcare utilization and preventive care does not carry over into better treatment in general is particularly striking and points to the need for new approaches to improve men's and women's healthcare.

Despite significant gains in recent decades, women remain relatively understudied as medical research subjects. As a result, medical treatments may not be as appropriate for women as they are for men; women are more subject to adverse drug reactions than are men.

A similar pattern of discrepancies in access and quality of care occurs for minority populations and for those low in socioeconomic status. Thus many of men's and women's relative disadvantages in healthcare are exacerbated for minority and low SES groups.

Current as of March 2007