In developing the agenda for the workshop, the steering committee identified a wide range of factors that could influence the differences in U.S. women’s health outcomes in comparison with outcomes in other countries. Those potential factors include several institutional factors that were discussed at the workshop: access to health care, possible biases in the delivery of health care, and the quality of health care for cardiovascular disease and diabetes.
In her opening remarks, Alina Salganicoff (Henry J. Kaiser Family Foundation) argued for incorporating gender-stratified analysis in all work on public health issues. The gender-specific analysis that has been done illustrates the importance of learning about why there are significant differences in mortality and in well-being between women in the United States and women in other countries and also within different groups of women in the United States. Her presentation included findings from unpublished surveys by the Kaiser Family Foundation.
Access to medical health care is one of the key components in health, Salganicoff noted, but she said that other health care services to women—beyond what actually takes place in doctors’ offices—are also important to health outcomes. An examination of access issues would include insurance coverage, the range of services that women need, and how they use the available service. It would also include how primary
care, preventive services, reproductive and sexual health care services, mental health, chronic conditions, and cancer contribute to or diminish women’s health.
Access should be examined over the life span, Salganicoff said. Though the report on international comparisons discussed by Woolf cover the mortality of women before age 50, it is important to consider issues of access for women over the age of 50. For those older women, a key issue is one’s quality of life.
Salganicoff said that progress is being made in the area of health insurance. The Affordable Care Act (ACA) has affected coverage rates for both women and men in the United States. From 2010, before the ACA was enacted, to 2014, the uninsured rate for women fell from 19 percent to 11 percent (National Center for Health Statistics, 2015). Over this period, there have been in increases both in enrollment in Medicaid and in private insurance. The increase in Medicaid enrollment has taken place even though only about half the states expanded Medicaid to include their poorest residents, she noted.
The uninsured rate for men is higher, largely because men have not historically qualified for Medicaid. The larger proportion of female enrollees is evident in all states, not just those in which Medicare has been expanded. Salganicoff’s conclusion is that women place a priority on having health coverage. She speculated that the larger proportion of female enrollees may be due to their past experiences with the health care system.
But many women still are not insured. She referred to a 2014 survey conducted by the Kaiser Family Foundation (unpublished) that looked at why the uninsured remained uninsured. For purposes of this workshop, she stratified the results by gender. Even after the expansion of Medicaid, 47 percent of women responded that the available insurance was too expensive: 13 percent who said they were unemployed or not eligible through work; 8 percent who were told they were ineligible; 7 percent who said they were not eligible because of their immigration status; and only 4 percent of women who said they do not need insurance. (The survey did not include information on the other 15% of women who reported that insurance was too expensive.) Salganicoff said that these data indicate that women overwhelmingly want to be insured, but they do not have a pathway to get coverage.
Coverage is important because it enables access to a usual source of primary care, she noted. In turn, having a usual source of primary care is an important determinant for successful health outcomes.
Type of Providers
Other survey data show that women use the health care system differently than men. For example, one study (Salganicoff et al., 2014) found that women are much more likely to have multiple providers. One-quarter of women said that one of their routine providers is an obstetrician/ gynecologist, but that changes over their lifetimes; and as they age, they rely on other specialists and so become accustomed to relying on multiple providers. The penchant for using multiple providers introduces issues of communication and care coordination. The effect on health outcomes of using multiple providers needs more research, she said.
The type of insurance coverage also determines the type of providers that women use. Women with private insurance are much more likely to use a private doctor’s office or a health maintenance organization (HMO). Women with public insurance tend to use clinics, although with the advent of ACA, a significant share of them are now using HMOs or doctors’ offices. Women who are uninsured obtain their usual care largely in clinics, or, for 16 percent, in emergency rooms. Salganicoff suggested that a research agenda on women’s health consider the effects of these different sites of care on outcomes, considering how they shape the type of care that is provided.
Another impact of the ACA has been to broaden coverage of preventive services. Because of the impotence of preventive services, the U.S. Preventive Services Task Force, the Advisory Committee on Immunization Practices, and the Health Resources and Services Administration of the U.S. Department of Health and Human Services recommend or require that all private insurance plans cover several services without cost sharing. Many of these services, such as Pap tests to screen for cervical cancer and mammograms to screen for breast cancer, as well as blood pressure, cholesterol, and colon cancer screening, are critical to women’s health outcomes.
The 2013 Kaiser Women’s Health Survey (unpublished) found that rates of Pap tests, mammograms, blood pressure, and cholesterol screenings were already quite high for women with both private and public insurance programs. However, fewer than one-half of the insured had colon cancer screening and the rates for the uninsured were well below the rates for the insured. Screening is an important step, but more research is needed on what happens after screening. For example, African American women are more likely to get a Pap test than non-Hispanic white women, but they have a higher incidence of cervical cancer and a higher likelihood of mortality from cervical cancer.
Reproductive and Sexual Health
Salganicoff discussed another issue of importance to women’s reproductive and sexual health—contraception. She stressed that reducing the rate of unintended pregnancy and increasing the rate of intended pregnancy could make a really huge difference in women’s quality of life and their health and well-being. One in five women of reproductive age is not using any contraception, which is the reason for the high unintended pregnancy rate in the United States relative to other countries. About one-half of pregnancies are unintended pregnancies, and they tend to result in poorer outcomes, perhaps starting with delayed initiation of prenatal care. The consequences appear to differ by race and ethnicity. It would be especially important for research and analysis to focus on American Indian and Alaskan Native populations.
A research agenda on women’s health should take into account the improvement in the use of long-acting reversible contraceptives, such as intrauterine devices, implants, and injections. Such contraceptives are now covered by insurance, and it is likely that usage will increase: the effect on the unintended pregnancy rate should be measured, Salganicoff said. It is important also to understand the reason for the long-standing dramatic declines in teenage birthrates: Is the decline due to reduced sexual activity, increased access to contraception, or both?
The rates of preterm births, which are declining, are another topic worthy of further research. Infant mortality is extensively researched, but the United States also has unacceptably high maternal mortality, with particularly high rates for African Americans, Salganicoff noted.
Violence in women’s lives is related to both well-being and health. It can be as blatant as the sexual assaults on college campuses, or it can be related to issues of sexual identity, sexual orientation, and the health disadvantage of bisexual women, which is emerging as a topic in research. The issue is persistent, and Salganicoff contended that this is an area that is ripe for exploration.
Despite the high prevalence of violence in women’s lives, it is not a topic that is often discussed between women and their doctors or other health care providers. The 2013 Kaiser Family Foundation women’s health survey (unpublished) found that only about one in four women have ever discussed, or have discussed in the past 1-3 years, any level of dating or sexual violence.
An associated topic is mental health, which is an important area for researchers. There are differences in the types of mental health conditions that affect women and men. In addition, little is known about the
relationship between mental health and other health conditions. Likewise, depression, anxiety, and postpartum depression need study. Research has shown that women experience double the rates of serious psychological distress of men. Yet only 41 percent reported that they discussed a mental health issue, such as depression or anxiety, with their health care provider in the past 3 years (unpublished Kaiser Family Foundation survey).
Costs and Time
Costs and time play a role in obtaining health care, Salganicoff said. The 2013 Kaiser Family Foundation survey (unpublished) found that 65 percent of the uninsured said they delayed or went without needed care because of cost, which is not surprising. But surprisingly, 16 percent of those with private insurance and 35 percent of those on Medicaid reported that costs were a barrier to obtaining health care, citing increasing deductibles and premiums as reasons. Over the past 5 years, workers’ earnings have increased by about 10 percent while deductibles now cost 67 percent more than in 2010, and premiums have increased by an average of 27 percent.
Logistical issues also loom large for women. Often, they just cannot find the time to obtain care. In the 2013 Kaiser Women’s Health Survey, about one in four women said that they delayed or went without needed care because they could not find time. This included one in four women, particularly low-income women, who said they could not take time off of work. Many women in low-wage jobs do not have sick pay leave, and they forfeit wages when they go to the doctor. In addition, one in five low-income women said that they could not take time to obtain health care because they had problems getting child care. Postponing or delaying care can affect women’s health and well-being.
Long-term care is a women’s issue because it is hugely expensive, and women have higher rates of chronic conditions and limited functional status than men at every age and thus are more likely to need long-term care. Moreover, Salganicoff said, when women get ill, they are much less likely to have social support than men. For example, 41 percent of women today who are over 65 are widowed, compared with 13 percent of men. One-third of those widows are living alone, and 39 percent live on incomes of less than $20,000 a year. Moreover, women are much more likely than men to serve as caregivers, which can affect their own health and well-being.
In closing, Salganicoff emphasized that insurance coverage matters: things are improving but still far short of where the nation needs to be, and women experience barriers for all types of services. She suggested the need for a gender-sensitive approach to research on health services and public health, as well as clinical research.
In response to a question from the floor, Salganicoff said that women sometimes do not avail themselves of care for addictions because of concern for their children. They fear that, once the police and the courts get involved, there is a possibility of losing their children. That fear makes a huge difference in the likelihood of their seeking care.
A participant noted that issues of substance abuse and mental health disorders are extremely complex situations. There is a lack of capacity in treatment programs that can accommodate children and pregnant women, and there is a stigma attached to substance-abusing women. For female drug users, particularly those who have children, that stigma is a huge barrier to seeking services.
The subject of bias in health care delivery as a factor in explaining the relative differences in women’s health was addressed by Paula Johnson (Harvard Medical School and Brigham and Women’s Hospital). As background, she first introduced the Social Progress Index, developed by a collaboration of a number of health institutions and the Harvard Business School. The index looks at the foundations of well-being with data from about 130 countries, which represent 95 percent of the world’s population.
In this index, as in the study discussed by Steven Woolf (National Research Council and Institute of Medicine, 2013) earlier in the workshop, the United States is ranked low—35th with regard to the foundations of well-being and 68th in health and wellness. The Social Progress Index found that the factors driving the country’s low rankings are premature deaths from noncommunicable disease, obesity, disparities with regard to the environment, and violence.
As further background, Johnson reminded participants of a few of the factors that affect women’s health. She noted that the disadvantage of U.S. women is not only in terms of mortality but also in terms of morbidity. Depression is the number one cause of women’s disability in the United States, as it is around the world. She said that women are 70 percent more likely to be diagnosed with depression across their lifetimes, and they are misdiagnosed somewhere between 30 and 50 percent of the time. Johnson also noted that the higher rates of cancer for U.S. women are associated with cigarette smoking: lung cancer is the number one cancer killer of women. It is important to note that although smoking rates are plateauing, women develop lung cancer with less smoking than do men.
Women who are nonsmokers are more likely to be diagnosed with lung cancer than male nonsmokers, but Johnson observed that the current criteria for screening for lung cancer are the same for women and men.
She suggested that they should be different. Screening should take into account that there are powerful combinations of risk factors that predict lung cancer in nonsmoking women that usually occur at a much younger age for women than for men.
A major health system issue in the United States, Johnson stated, is the fragmentation of health practice and service. At the patient level, this translates into a lack of integration across providers. At the physician level, it translates into a lack of integration of gender-specific information into practice. There is also a failure to take a life-span approach, to look only at women’s health at specific times or for specific episodes of life.
Johnson then introduced three discrete examples to illustrate the topic of bias in health care delivery—caregiving, violence, and the need for a model for care delivery that is more integrated for women of reproductive age.
Caregiving is a major women’s health issue. Johnson reported estimates that 25 percent of women caregivers have health problems as a result of their caregiving activities (Family Caregiving Alliance, 2016). In addition, the reported health issues of caregivers are more severe than for women generally: 25 percent of female caregivers reported fair to poor health, in comparison with 12 percent of women more generally. Other differences in the study noted by Johnson were that women caregivers were twice as likely not to fill a prescription because of cost than noncaregivers, and they tended to have mammograms less often than recommended.
Johnson reported other risks to a woman’s health if she is a caregiver. The relative risk of the development of cardiovascular disease in women caring for an ill or disabled spouse for more than 9 hours a week is nearly twice the average risk—1.82. And the relative risk of death from coronary heart disease is about 2.5 times higher than the average for women. And women who provided 36 or more hours of care weekly to a disabled spouse were almost 6 times more likely than noncaregivers to experience depressive symptoms. Caregivers were also 2.5 times more likely than non-caregivers to live in poverty and to receive supplemental security income (from the Social Security Administration).
Johnson next discussed an ongoing study at the Brigham and Women’s Hospital that is looking at transitions of care by the Patient-Centered Outcomes Research Institute. The survey asked people who were being discharged from the hospital about receiving care and caregiving responsibilities. The preliminary findings are revealing: 24 percent of female patients reported that they were cared for by their husbands
while 48 percent of the male patients reported that were cared for by their wives; 29 percent of the women were a caregiver for someone else, compared with 17 percent of the men were also caregivers. Of the caregivers, 61 percent of the women were full-time caregivers; only 35 percent of the men were full-time caregivers.
These issues are not being adequately addressed in the health care system, Johnson said. She advocated for acknowledging that caregiving is a major issue at the individual and population levels. She emphasized the need for community and public health strategies in both ambulatory and inpatient settings that address caregiving as a risk to health.
The experience of violence is the highest correlate of chronic disease in women. Violence is associated with depression, obesity, cardiovascular disease, substance abuse, and pain. It brings an increased risk of depression. Johnson said that research shows that women who have experienced domestic violence are 80 percent more likely to have a stroke, 70 percent more likely to have heart disease, 60 percent more likely to have asthma, and 70 percent more likely to drink heavily than women who have not experienced intimate partner violence (Centers for Disease Control, 2008).
Recognizing this issue of violence, a new concept called trauma-informed care, is emerging. The key elements of trauma-informed care are (1) realizing the prevalence of trauma, (2) recognizing how trauma and exposures to violence, especially cumulatively, affects people, and (3) responding by putting this knowledge into practice.
Care for Reproductive-Age Women
Johnson asserted that reproductive health and cardiovascular health live in separate silos, in spite of what is known about their relationship to each other. Women who experience preeclampsia, gestational diabetes, preterm delivery, or have a low birth-weight baby will in later years have twice the risk of cardiovascular death as that of women who did not experience those conditions or outcomes. Research is finding that children of these women will also be at higher risk for negative health outcomes.
There is a racial disparity in pregnancy-related conditions, Johnson noted. Although rates of preeclampsia and gestational diabetes are increasing among white, black, and Hispanic women, the rate is increasing significantly more for black women. For all groups, the increase underscores the importance of looking across the life span at diseases that affect women’s health and the ultimately the health of their children.
Johnson acknowledged that there is a lack of evidence-based research on how to decrease the risk of pregnancy-related conditions. Along with such research, there is also a need for improved measurement. Such important measures as the Healthcare Effectiveness Data and Information Set (HEDIS), which are used to evaluate health care systems by a set of standard measures, are not stratified by gender. Although some of the measures are female specific, most are aggregated. She noted the new Triple Aim Initiative, which was instituted by the Institute for Healthcare Improvement to measure improvements in the individual experience (quality and satisfaction) in the health of populations, along with reductions in per capita costs of health care.1
In closing, Johnson observed that the advancement of the idea of precision medicine in the health care delivery system requires sensitivity to gender issues.
A workshop participant inquired about the status of the important research on caregiving as a risk factor. It would be important to segregate the impact of caregiving by educational level, race, and culture. Johnson replied that some studies are under way, but the data are preliminary, and sample sizes have been too small to support conclusions. She stated that there is a need to get caregiving on the agenda for funding and research organizations and agencies and to organize larger multicentered trials so the full depth and breadth of the problem can be examined.
Another participant turned to the issue of physician effects. He expressed a concern that male physicians (who are twice the number of female physicians) may not be as aware of some of these women’s health issues. There is a need for outreach to male health care providers.
In response, Johnson underscored the need to implement evidence-based structural change in the health care system so that recognition of the issues becomes part of the delivery system. For example, questions about caregiving should be part of the normal conversations between health care providers and women patients. It is also important to identify and deal with the stress that accompanies caregiving. A participant asked if it would be possible, within the context of electronic health records, to include prompts that are relevant to gender so that information that will help guide solutions is made readily available. With regard to bias in the delivery system, necessary changes could be incorporated in medical training. Johnson agreed that incorporating gender-specific prompts
1For detailed information on the Triple Aim framework, see http://www.ihi.org/engage/initiatives/tripleaim/Pages/default.aspx [January 2016].
on electronic medical records would be helpful, but would not be the whole answer to the problem. Changes in the means of identifying treatments based on evidence are also needed. For example, medical personnel already screen for domestic violence because it is a requirement of the Joint Commission on Accreditation of Health Care Organizations, but in many cases, checking the box is not followed by an understanding of the history of the violence and its cumulative effect.
In a follow-up question, a participant wondered about the practical implications of some of the research that was just reported. Once you have determined the importance of a certain potentially harmful factor, such as caregiving, the question becomes what to do about it. Health care professionals need access to support systems—social support services, peer navigators, and other support systems—to help with the economic and health care challenges that individuals face. These systems need to be developed. The research agenda needs to include evaluations to see what those interventions should look like.
Johnson responded that it is important to think in terms of implementation science around these particular issues. The implementation would include not only social supports, but also traditional approaches to such issues as the risk factors for cardiovascular disease. Such an approach requires a merging of two worlds of medical care.
Chloe E. Bird (RAND) reported on current research that is mapping differences in quality of care using the HEDIS measures that focus on cardiovascular disease and diabetes to understand why women receive poorer quality care than men. She prefaced her remarks with the observation that, in any other area of routine care except cardiovascular disease and HIV, women have higher quality of care than men. The data show that if women have insurance and access to care, they get screened for various conditions at higher rates than men. They also generally get their prescriptions and take them, and they often have better outcomes, except for cardiovascular disease.
Bird addressed attempts to determine if cardiovascular disease is actually different in men than in women. Although the 1993 Revitalization Act (amended in 2001) directed the National Institutes of Health (NIH) to include appropriate proportions of women and minorities in research, gender-based analysis is not required. If one wants to understand women’s health, she argued, it is important not only to have women in the research but also to analyze the data by gender.
Bird referred to research that shows that women have different types
of heart disease than men—they are much more likely to have blockage in the small vessels of the heart, microvascular disease (Handberg et al., 2006). This becomes a problem if angioplasty is not properly placed.
Secondary prevention is another example of gender differences in cardiovascular care, Bird noted. Statins are generally prescribed, but they do not seem to be tolerated as well in women as they are in men. Women often report musculoskeletal pain from statins. So women are less likely to be getting care from a cardiologist, but, when they are, they are also less likely to be getting titrated on a statin, thus able to control high cholesterol. These differences have tended to shift the survival curve in favor of men.
California has seen a shift in the heart-disease survival curve for women. In 2013, for the first time, the state saw the same rates of women dying from heart disease as men. To understand this trend, RAND mapped gender gaps in care based on data from the California health plan (Bird et al., 2014). The study focused on people with cardiovascular disease and whether the quality of care and screening for high levels of LDL cholesterol would help to explain this trend. In seven of the eight regions in California in the study, women were not getting as high-quality care as men. In three of the regions, the care gap was more than 5 percentage points. Bird said the same gap in care is true for diabetes, a disease that traditionally affects women more than men. Diabetes affects the risk of cardiovascular disease, and it has a bigger effect on women than on men: for men, diabetes doubles the risk of cardiovascular disease; for pre-menopausal women, it quadruples the risk. In a study at the county level in California using data from one health plan, the same relationship was found: 79 percent of the counties had gaps in care that favored diabetic men at risk for cardiovascular disease (Bird et al., 2014). Recent research shows that women have 40 percent greater likelihood of having a heart attack if they have diabetes than do men (reported by Diabetologia, 2015).
These findings relating to the gender gap and its effects have been building over many years. RAND studies have looked at gender, race, and socioeconomic disparities in the quality of care for cardiovascular disease and diabetes over more than a decade (see Bird et al., 2007). The studies found gender gaps across all of these measures.
Quality of care plays a role in outcomes, too, Bird noted. The mapping analysis indicates that HMOs are better at hitting quality indicators than preferred provider organizations (PPOs). In routine aspects of care, HMOs do a good job in comparison with PPOs. This finding may be related to the fact that HMOs are designed to emphasize routine, inexpensive aspects of screening, treatment, and also hitting intermediate outcomes. Also of importance, the mapping study found that there were no gender gaps on average among the HMO patients, but there were gaps for PPO patients (Bird et al., 2014).
Bird concluded that the mapping approach is showing promise as an analytical tool. It has been extended to using the data from the Centers for Medicaid & Medicare Services, looking at all of the measures that are available across the entire Medicaid managed care sample and focusing on measures for which there was a gap of 5 percentage points or more. To summarize the research, for all of these measures except the control of high cholesterol, women did better than men. Other indicators tended to show that differences in cardiovascular outcomes were much larger for gender than for socioeconomic status. The gaps for younger women were more pronounced than those for older women.
Although mapping holds promise, there are challenges, Bird said. For one thing, analysis of the gender gap by race and ethnicity faces an obstacle in that health plan data do not generally include data on race and ethnicity; consequently, investigators have had to develop algorithms for interpolating race data. Bird speculated that these beginning analytical efforts may be enhanced by the new NIH Institute for Minority Health and Health Disparities.
A participant asserted that NIH has done a better job of including more women in its sponsored trials and surveys, but NIH is not necessarily doing subgroup analysis on women. Other federal agencies are doing less. For example, the U.S. Food and Drug Administration (FDA) is not required to do subgroups analysis on women (as NIH is), and that agency seems reluctant to push subgroup analysis on studies conducted by industry. The participant posited that the lack of analyses by gender affects the validity of those studies, and FDA is therefore making decisions about whether to approve a drug or device on the basis of inadequate studies. For example, the questioner suggested that cardiac devices are not being studied for their effects on women.
Bird responded that it is important for federal agencies to set requirements for analyses by gender both for grant reports and published journal articles. The requirement for gender analysis is gaining ground in Europe, she noted, where it is becoming the standard. Other countries make an effort, when deciding about health and health care funding, to give women “a seat at the table.” She stated that the NIH Office of Research on Women’s Health is able to promote this approach, but the organization is small and has limited ability to implement the widespread changes in approach throughout NIH.
A participant suggested the creation of a repository for federally funded research in order to create a body of knowledge on women’s health issues. The repository might be able to support investigators to readily research knowledge about women’s issues.
Another participant returned to the role of FDA, noting that its Office of Women’s Health published a report in 2013, which indicated that while
there are some areas where progress still needs to be made, in the majority of the reports the gender analyses are being done. The FDA is now trying to find ways that advance the knowledge and address women’s issues in the current climate. Bird suggested that a step in the direction of subgroup analysis would be for the FDA to require that one of the publicly available pieces of information be whether the particular medication works in women.
Another participant pointed out that surveillance data from the U.S. Department of Health and Human Services do not adequately present data disaggregated by gender. Surveillance data, which include thousands and thousands of interviews or records review, should easily be able to provide information by gender.
Bird responded that agencies claim that providing data by gender cuts the data too thin for reliable analyses. The counterargument is that women are the majority of the population and that it is possible to draw more conclusions from the available data than is now being done. One approach is to isolate one area for focus. The reason for the success in treating breast cancer is that the women’s health movement focused on that disease, she said. Today, the money that goes to research on breast cancer research is three times more than the money that goes to all research on heart disease.
A participant suggested that it would be useful to engage editors and reviewers of research-based papers in focusing on the influence of gender. In response, Bird reported that one publisher, Elsevier, is including this emphasis in its new editing and review system. Nancy Adler (University of California, San Francisco) reported that there was an Institute of Medicine workshop that focused on this issue, bringing together the editors to seek a solution (Institute of Medicine, 2012). She suggested that this is an important area for research and that there are always new opportunities since journal editors change over time.
This page intentionally left blank.