As noted in the introduction, a variety of forces are aligning to create a demand for social and behavioral information in electronic health records (EHRs). These forces include the growing accumulation of evidence that social and behavioral factors play a major role in the onset and course of disease, morbidity, and mortality (McGinnis and Foege, 1993; Woolf and Braveman, 2011; Woolf et al., 2007); concerns about the costs of health care and its effects on the U.S. economy (IOM, 2012; KFF, 2012); and current and anticipated health care costs associated with the treatment of chronic conditions, such as diabetes (KFF, 2012; Thorpe et al., 2004, 2010).
There are several indications of growing interest in addressing social and behavioral determinants of health and including relevant data in EHRs. In turn, individual health systems are modifying their own EHRs to incorporate such information. Innovative programs are screening for social hardships of patients in order to provide services to address these needs as part of the clinic visit (e.g., organizations such as Health Leads, described in Chapter 6, and medical–legal partnerships). Importantly, the Office of the National Coordinator for Health Information Technology (ONC) has expressed interest in considering the inclusion of more social and behavioral domains in the EHR as a possible Stage 3 Meaningful Use requirement. All of these efforts will be facilitated by a standard set of measures to assess the most relevant domains.
The measures that merit inclusion in all EHRs (as well as which are appropriate for specific populations) are those that will enable more effective treatment of individual patients in health care settings; more effective population management for health care systems and for public health agen-
cies; and discovering the pathways that link social and behavioral factors to biological functioning, disease processes, and mortality that may inform new treatments and interventions. For this to occur, data should be accurate and useful to patients, the clinical team, systems, and researchers. Thus, in making its judgments about what to recommend for inclusion in EHRs, the committee leaned heavily on the current evidence of the link between each domain and health outcomes; the availability of a standard reliable and valid measure of the domain; the clinical usefulness of the measure; the feasibility; and the lack of sensitive information, causing patient discomfort of capturing the measure in the clinical workflow.
The committee recognizes that criteria used to evaluate measures of domains may vary in their nature and emphasis depending on the purpose for which they are being used. There is rarely a true “gold standard” for evaluating whether a given measure of a social and behavioral determinant is fully and accurately capturing the intended concept. Psychometric testing can provide evidence of aspects such as internal consistency, stability over time, convergent and discriminant validity, and the extent to which the items cover the full range of the construct. Meeting these criteria often necessitates multiple items, creating a trade-off between validity of the measure and its ease of administration. A longer, well-validated instrument may become the gold standard against which shorter instruments are tested.
Many of the measures the committee reviewed were not developed for use in clinical settings but for research on a conceptual domain in relation to health—and to other aspects of life. Researchers often develop variants of methods and measures to capture the conceptual domain in relation to the specific outcome they are studying. Finding that multiple measures of the same concept show significant associations with a single outcome or related outcomes provides evidence of the “robustness” of the association. Ironically, however, that benefit conflicts with the need for a standard assessment.
The question of clinical use of these measures is relatively new. While all of the measures the committee evaluated have strong links to health, the findings frequently emerged out of population-based surveys with self-reported health outcomes. There is a much smaller evidence base of the implications for clinical care, or of the effectiveness of actions, or interventions to modify the underlying state or to modify treatments by taking into account the patient’s social context and psychosocial risks. This is an important emerging area of research.
The committee’s findings and recommendations necessarily reflect the current status of knowledge about the social and behavioral determinants
Several domains showed a strong association to health and would be useful, but the committee could not identify adequate measures or related interventions for them. Many of the measures identified in Figure 5-1 that the committee did not recommend for inclusion in EHR are measures that would benefit from further development. Once suitable and standard measures are developed, these domains and their measures can be considered for inclusion in EHRs. The committee noted two examples of measures that would benefit from further development for EHR inclusion and are not mentioned in Figure 5-1: exposure to violence (broadly), and occupation. Violence includes a wide range of abusive behaviors and affects men, women, and children as victims and perpetrators. Although violence is an important domain related to health, the committee identified a research gap in the development and validation of measurement tools. For example, in reviewing the evidence related to violence against elders, the U.S. Preventive Services Task Force (USPSTF) found insufficient evidence to recommend routine assessment of elders for exposure to violence because of the uncertainty of the benefits and harms of doing so (USPSTF, 2013). Children are also a vulnerable population in relation to abuse, but despite great interest and concern about children’s exposure to violence, validated measurement
tools are lacking. Child abuse is often visually identified by the clinical team or outside of the health care setting, and legal systems are in place to protect the child. The existence of reporting requirements adds complexity to obtaining measures of children’s exposure to violence.
Interpersonal violence measures are an unmet need. Better evidence is available on intimate partner violence. A systematic evidence review by Nelson et al. (2004) for the USPSTF revealed that more women than men experience intimate partner violence, and most studies about screening and interventions for intimate partner violence enroll only adult women. The task force viewed screening for intimate partner violence for women of reproductive age as having moderate net benefit. The committee noted the need for more research regarding screening measures for violence for other populations, as well as the need to identify a common metric for interpersonal violence.
The second example of an important domain that is not yet feasible for inclusion in EHRs is occupation. Occupation has a number of characteristics, including employment status (e.g., working full-time, working part time, unemployed), as well as type of employment and conditions associated with the work environment (e.g., demanding physical labor, clerical work, caring for others) and prestige (e.g., unskilled labor, professional). Research has identified risks for health outcomes associated to specific jobs and informed the development of preventive measures to reduce or eliminate exposures (Sabatini et al., 2012; Ziegler et al., 2002).
Reflecting the importance of occupation to health, a number of organizations, including the American Public Health Association, the Council of State and Territorial Epidemiologists, and the American College of Occupational and Environmental Medicine, have published statements calling for the inclusion of industry and occupation in EHRs. A 2011 letter report from the Institute of Medicine observed that “occupational information could contribute to fully realizing the meaningful use of EHRs in improving individual and population health care” (IOM, 2011, p. 42). The committee agreed with the sense of the statements asserting the importance of occupation in relation to health. It not only impacts health directly, but knowledge of the characteristics and demands of their patients’ work may be relevant to clinicians in making treatment choices.
Despite its importance, the currently available measures of occupation are lengthy and complicated to code. As a result, occupation was not included in the recommended panel.
These shortcomings may be rectified in the future. The National Institute for Occupational Safety and Health (NIOSH) is currently developing and standardizing specific measures that capture a patient’s industry and occupation, including measures on work schedule, employment status, and external causes related to injury and poisoning (i.e., ICD-10 codes)
(NIOSH, 2014). Currently this coding for occupation is too time intensive to be practical for use in an EHR, but if this hurdle can be overcome, it could be added at a later time.
Some domains currently have adequate measures, but they are not freely available because of copyright protections, and they either require purchase or the measures are available only to researchers. For example, patient activation has been shown to play an important role in enabling greater patient involvement in decision making and better clinical outcomes. However, the leading measure, PAM, is proprietary and does not meet the criterion of being freely available for use in EHRs. The need for a more widely accessible instrument has been recognized in the field, and several researchers and groups are working on alternate measures. These have not yet been fully validated but should become available within the next few years.
Other domains had measures that met the criteria for inclusion in EHRs but had no agreed-upon common metric. Health literacy, stress, and food insecurity are all examples. A common metric is desirable because it provides greater comparability over time, sites, and populations; facilitating continuity over time. By relating measures of a domain to a common metric, if specific measures change and new ones develop, the common metric remains. For example, rapid advances are occurring in the development of personal devices and sensors to measure physical activity. However, regardless of the instrument used to measure physical activity, the measure can be converted into a common actionable metric (e.g., minutes of moderate or vigorous physical activity per week).
Although the existence of a common metric is desirable, its absence was not sufficient to remove a domain and its measures from committee consideration. Because several measures were included for domains lacking a common metric, future work is needed on developing such metrics.
Most of the recommended measures rely on self-report, which can be subject to error and bias. Technological advances may allow collection of more objective indicators and information on experiences that individuals may not be able to remember and report reliably. For example, rather than relying solely on a self-reported measure of sleep duration or quality,
personal devices now exist that monitor levels of sleep behavior. Sensors that record data for review and upload to the EHR if appropriate—while not without their own limitations—may eliminate or reduce the need for having to ask individuals about their behavior.
The use of EHRs is expanding at an increasing pace, and the data that are collected and stored in such records have the potential to improve clinical practice, population health management, and health research, but they may also pose challenges. As a result, the committee considered potential implications for a number of potential stakeholders if its recommendations are implemented. The implications encompass both opportunities for new or more effective actions as well as potential demands and challenges. Some of these are discussed below and summarized in Table 7-1 at the end of this section.
The ONC, through its Health Information Technology for Economic and Clinical Health Act (HITECH) program, has already made some inroads in collecting social and behavioral determinants of health that are of high clinical priority (e.g., tobacco use). The deliberations and conclusions of the committee can inform new standards for data domains that have before this time received little attention.
The core panel of measures recommended by the committee pulls in aspects that should have an important and broad effect on the need for services, the delivery of health care, and the interaction with the health care team (including electronic communication). The committee has based its recommendations on factors important to ONC, including concrete evidence about the clinical importance of each determinant, availability of standard measures of the determinant, and the expected amount of work necessary to collect the determinant. The recommendations should provide helpful guidance for further development of Meaningful Use regulations. As HITECH shifts its emphasis to improving outcomes, the social and behavioral determinants of health will be critical to selecting the right path for each patient and to making sure that benefits accrue equitably across the nation’s population.
The committee recognizes that adding anything to the HITECH program has costs and consequences. The determinants are not static; there will need to be ongoing review and maintenance of the current core set and potential additions in the future. Monitoring the process and outcomes
of implementing the addition of social and behavioral domains to EHRs and reviewing the emerging literature for new measures and domains that should be added later on will take time and resources. These demands should be viewed in the context of the potential for the addition of the committee’s core set of determinants to allow the nation to move forward in a more consistent and fair manner.
Research sponsors will have new opportunities to expand and enrich their research portfolios by supporting research using the data provided by the recommended panel. These measures assess the key determinants of health and provide new types of data in a standard form that can enable novel research. The addition of social and behavioral data would not only spur the development of new studies on social and behavioral determinants of health and of the effectiveness of health care; they can also be used in conjunction with data on genomics being linked to EHRs. The recommendations may also be useful to research sponsors working to support development and use of national data networks by informing data standards.
The flip side of these new opportunities for cutting-edge research using these determinants in EHRs is that there will be more demand for research dollars to support this work. In addition, growing awareness of the importance of having validated measures of multiple domains should also increase the need to support basic research on social and behavioral processes and how they can be modified. This work is likely to be interdisciplinary and require collaboration both among units within funding groups (e.g., across the National Institutes of Health [NIH]) and across funders (e.g., NIH, Agency for Healthcare Research and Quality [AHRQ], Patient-Centered Outcomes Research Institute [PCORI]).
The NIH has a particularly important role as the largest funder of health research. The Office of the Director of the NIH produces strategies and plans for advancing research across all of its institutes and centers. The committee sees an opportunity for better identification and use of social and behavioral determinants of health collected in EHR through the development of such a plan for this field. The Office of Behavioral and Social Science Research (OBSSR) could be the driving center, ensuring input across the many NIH institutes and centers. OBSSR could also be the visionary, informing data standards for the national data networks in the field of social and behavioral determinants of health. With the availability of new types of data collected, novel avenues for research will arise. While funding streams will need to be identified, the research garnered will likely hold tremendous benefits.
Organizations such as AHRQ, the National Quality Forum, the National Committee for Quality Assurance, and the Joint Commission will benefit from access to additional information that can enable them to better monitor the progress of health care providers and systems in providing care that improves patient outcomes and reduces disparities. The added measures provide relevant information on patient outcomes and provide greater explanations of contributors to disparities in access to care (e.g., how education or financial resource strain alters access to or use of health services). Having access to these data may also provide additional parameters upon which to make risk adjustments.
Three challenges arise from the inclusion of these determinants of health into the quality assessment process. First, the development of clear EHR specifications for social and behavioral quality-related measures and subsequently obtaining those measures will add burdens both to the standards organizations and to already overburdened institutional information technology services. Second, inclusion will require expanding notions of what constitutes quality and how more enduring characteristics of the individual and elements exogenous to the health care system can, could, and should be incorporated into the quality assessment process. Finally, there may be challenges to incorporating social and behavioral indicators as quality indicators if these are believed by the clinical care professionals and systems to be immutable or considered out of scope of their purview. The emergence of the concept of the accountable care organization (ACO) is laying the groundwork for broader consideration of quality; however, this shift has challenges of its own and may take several years to achieve fulfillment.
The recommended measures will provide options for the Center for Medicare & Medicaid Innovation (CMMI) to expanded practice-based data and use more comprehensive information about patients and populations, especially in those areas most likely to create challenges in implementation. Social and behavioral determinants of health affect what care is needed, how to best deliver it, and how to measure success. CMMI programs can benefit from the information both at the population level—for example, to understand regional differences in uptake for different CMMI models—and at the individual level—for example, to tailor optimal treatment to the patient. This will, however, require an explicit effort to include the social and behavioral determinants of health into the CMMI programs, including demonstrating projects and evaluation.
EHR vendors and product developers are currently responding to Meaningful Use objectives with the goal of achieving interoperability between their products. The current report should help inform those activities, especially if ONC, the Centers for Medicare & Medicaid Services, and providers embrace this report’s recommendations. The emphasis on standard measures will facilitate definitions of specific fields that product developers and vendors need to create to implement the addition of social and behavioral domains when such data are desired. The committee’s recommendations provide guidance on a standard set of domains and measures that may be useful in designing population health product offerings. These offerings will require patient self-report data capture capabilities that coordinate with clinical workflows and support new ways of analyzing and visualizing social and behavioral data to aid in clinical decisions and provider population health goals. New EHR functionality will also be needed to geocode patient addresses and link to external data sources.
Health care systems and ACOs are generally charged with providing care and services for individuals and populations of individuals who enroll for care from their organizations. Currently, ACOs and health care systems maintain few standard data on social and behavioral determinants of health to help with their overall management of those they serve. The recommended set of social and behavioral domain measures will provide standard data for managing individual and population health and better risk adjustment for quality assessment and payment adjustment. Additionally, under the community benefit provisions in the Patient Protection and Affordable Care Act,1 nonprofit hospitals are required to conduct needs assessments and to document the benefit they have provided to their community. Having more robust social and behavioral determinants of health in their EHRs may contribute to each hospital’s assessment, in turn aiding their ability to allocate their resources to address identified needs.
Some domain measures that were not included in the parsimonious panel recommended for inclusion in all EHRs could provide additional useful information. They should be evaluated by these organizations depending on the populations they serve and the types of services they are providing (e.g., sexual identity). The collection and analysis of such measures could inform modifications and later decisions about inclusion in a wider array of EHRs.
1 Public Law 111-148.
Health care systems and ACOs will arguably have the greatest burden as a result of these recommendations. Challenges on small practices will likely be greater than on larger health care systems. In addition to changing their clinical information systems to capture, store, and report data, they will have to adjust their workflow to collect and act on the new data. As detailed in Chapter 6, every time a new domain is added to the EHR, systems and workflows need to change to adequately capture the data elements. Moreover, availability of data on these domains will identify areas of need. Systems that are responsible for all the care needs and are paid on a capitated or total population system (not fee for service) may be most motivated to address these needs because doing so may reduce demand for other services. Some programs have demonstrated the usefulness of targeted approaches, but most of these have never been taken to scale. This represents a challenge that all ACOs and systems will need to face.
Clinical care providers and teams and administrative staff will have access to a more comprehensive picture of the patient state. Standard collection of social and behavioral domain measures along with interoperability of records could allow for different health providers to screen and access patient data and eliminate the need for burdening the patient and clinical team with redundant questions and entry of data. Access to such information could enable providers to engage more effectively with patients in shared decision making with the clinical team about treatment options, prevention, and care. As with health systems, providers will need to adjust their workflow to collect data on social and behavioral determinants of health and modify their clinical information technology systems accordingly to incorporate collection, review, and action on the data. Health professionals will likely need additional education about social and behavioral factors and interventions. Interprofessional education and other training opportunities to create links between public health, medical, nursing, and other health professional education will likely aid efforts to advance the collection and use of social and behavioral information.
The committee recognizes that effective population health management adds burden in the short term to the health system, especially to smaller practices. However, in the long term, the result will be improved patient and population health outcomes. Recognizing this, the committee recommended a parsimonious panel of domains with standard measures hoping to minimize this barrier to adoption.
Patients and their advocates will have access to new tools to help assure better care and outcomes. The committee’s recommended measures were selected as patient-centered tools and are outcome oriented whenever possible. Advances in health care require that individuals participate knowledgeably and actively in their own health care to realize the full benefit of shared decision making. Inclusion of standard data and language about social and behavioral determinants of health should help patients avoid redundant reporting and achieve more effective patient-centered care. At the same time, some of the social and behavioral questions may deal with sensitive or uncomfortable issues, and patients and advocates may have privacy concerns. Some may not understand the links between social and behavioral states and their health. Community and patient education efforts will be needed to address these concerns and prepare individuals to participate in shared decision making and interventions. Patients and patient advocacy groups should be active participants in ensuring that privacy protections are in place—particularly when data that are considered sensitive are involved.
Public health agencies have long understood the connection between social and behavioral determinants of health and the health of communities. However, there has been limited data capture of social determinants on an individual level or links with documented health outcomes. Having these determinants routinely collected as part of the EHR will enable communities to better understand how these determinants are affecting health and to develop community-wide interventions to improve population health.
Greater use of electronic health care data by public health agencies could enable better coordination of efforts and help break down the artificial walls between public health practice and clinical care. Public health programs may also be especially well qualified to address privacy concerns. They can inform the broader public about social and behavioral determinants of health and how the information on their characteristics will be used in both clinical and public health practice. While some data can be de-identified, it will be necessary for people to understand when their medical data will be used in a way that identifies them and when it will be used in ways that would not identify them. Achieving this understanding may take a robust public education campaign.
Most of our knowledge of social and behavioral determinants of health is derived from specific research projects (e.g., cross-sectional phone surveys, cohort designs). Including measures of social and behavioral determinants of health into EHRs provides unique opportunities and challenges for researchers. Integrating measures of social and behavioral determinants of health with both historical and concurrent clinical data will enable researchers to test in diverse populations important hypotheses about the strength of relationships with various indices of health, timing of their influence across the life span, and pathways that connect these domains with health, informing interventions and treatment. It would also enable the identification of social and behavioral subgroups that would benefit from specific current treatments and interventions.
Shortcomings in existing measures pose both challenges and opportunities. As mentioned earlier, some of the measures for domains are not ready to be used at present. The committee’s analysis provides some initial ideas about which aspects of given measures need better documentation. Researchers can make important contributions by developing standards on how to best collect data, establishing which measures are well suited for capturing valid, actionable information and minimizing the burden and increasing the feasibility of collecting information on social and behavioral determinants of health. This should include best practices for asking patients about sensitive issues or how to collect the data (i.e., self-report on a computer, in-person, or by a social worker). Where appropriate, research can be used to develop metrics for the domains. Because the measures need to be useful clinically, some domains require developmental studies regarding effective ways to change social and behavioral factors that lead to improved health. For example, studies on how to improve psychological assets, such as optimism, conscientiousness, and self-efficacy, are needed along with documentation on how such changes affect health and health care utilization.
Some of the challenges that researchers will face in tackling this research agenda simply reflect current challenges, but with added complexity with the interface with EHRs. Maintaining patient privacy and confidentiality in the process of integrating data among a number of sources is paramount. Some of the data are particularly sensitive and require special protections. Inclusion of patients into research teams and internal review boards may help to develop solutions to these issues and would be helpful to all concerned. Finally, research on social and behavioral determinants of health using EHRs will require collaborations among social scientists, informaticians, health services, and public health researchers. As with any interdisci-
plinary work, such collaboration introduces complexity and challenges, but it results in better solutions to important health problems of our country.
Payers and employers have a keen interest in the health of their populations. Many rely on health risk appraisals (HRAs) to inform their employees about social and behavioral risks and risk-reduction strategies and to guide their programs and benefit plans. Several of the measures not meeting the criteria for usefulness and readiness for EHRs may be useful in HRAs. The availability of widely accepted standard measures of the social and behavioral determinants of health should greatly enhance the value of HRAs for use in employee wellness programs by enabling them to make more accurate risk assessments and adjustments.
Payers and employers will need to address individual and special populations’ concerns about privacy, especially for sensitive domains. In addition, they may be challenged as they gain knowledge of these determinants of health in their enrolled and employed populations because the resulting awareness of these risks may increase pressures to create programs to improve health that address these and other determinants of health. Employers and payers will need to have the ability to prioritize efforts to address social and behavioral determinants of health and to offer effective programs incorporating social and behavioral determinants of health in their wellness and health promotion programs.
TABLE 7-1 Opportunities and Challenges to Integrating Social and Behavioral Determinants of Health into EHRs for Various Stakeholder Groups
|Office of the National Coordinator for Health Information Technology (ONC)||
|Research sponsors (e.g., NIH, PCORI, AHRQ)||
|Quality improvement organizations (e.g., AHRQ, NQF, NCQA, Joint Commission)||
|Center for Medicare & Medicaid Innovation (CMMI)||
|Health care systems and accountable care organizations||
|Patients and patient advocacy groups||
|Public health (Centers for Disease Control and Prevention, state, local and territorial public health departments)||
|Payers and employers||
The range and characteristics of measures available to assess the social and behavioral determinants of health are likely to expand in the coming years. The rapid pace in developing new methods for capturing internal states, health-related behaviors, and self-reports on a wide range of characteristics and experiences will undoubtedly yield new measures within a relatively short time. There is no current forum or process for evaluating results of ongoing research and reviewing recommendations of social and behavioral determinants of health that meet acceptable criteria for inclusion in EHRs.
The recommendations made by the committee in Chapter 5, set forward a coherent, parsimonious panel of social and behavioral domains and measures that should be included in all EHRs. In brief, the committee recommended use of standard measures for four domains that are already being regularly collected (race and ethnicity, tobacco use, alcohol use, and residential address), and the addition of eight additional domains (educational attainment, financial resource strain, stress, depression, physical activity, social isolation, intimate partner violence [for women of reproductive age], and neighborhood median-household income).
The committee believes that, taken together, these provide a set of psychosocial vital signs whose inclusion in the EHR will have sufficient benefit to justify the additional time and effort to collect the data, and the added demands to use the resulting information to improve care. The committee further identified three recommendations for future directions in the field of capturing recommended social and behavioral domains and measures for EHRs as follow.
Finding 7-1: Standardized data collection and measurement are critical to facilitate use and exchange of information on social and behavioral determinants of health. Most of these data elements are experienced by an individual and are thus collected by self-report. Currently, EHR vendors and product developers lack harmonized standards to capture such domains and measures.
Recommendation 7-1: The Office of the National Coordinator for Health Information Technology’s electronic health record certification process should be expanded to include appraisal of a vendor or product’s ability to acquire, store, transmit, and download self-reported data germane to the social and behavioral determinants of health.
Finding 7-2: The addition of social and behavioral data to EHRs will enable novel research. The impact of this research is likely to be greater if guided by federal prioritization activities.
Recommendation 7-2: The Office of the Director of the National Institutes of Health (NIH) should develop a plan for advancing research using social and behavioral determinants of health collected in electronic health records. The Office of Behavioral and Social Science Research should coordinate this plan, ensuring input across the many NIH institutes and centers.
Finding 7-3: Advances in research in the coming years will likely provide new evidence of the usefulness and feasibility of collecting social and behavioral data beyond that which is now collected or which is recommended for addition by this committee. In addition, discoveries of interventions and treatments that address the social and behavioral determinants and their impact on health may point to the need for adding new domains and measures. There is no current process for making such judgments.
Recommendation 7-3: The Secretary of Health and Human Services should convene a task force within the next 3 years, and as needed thereafter, to review advances in the measurement of social and behavioral determinants of health and make recommendations for new standards and data elements for inclusion in electronic health records. Task force members should include representatives from the Office of the National Coordinator for Health Information Technology, the Center for Medicare & Medicaid Innovation, the Agency for Healthcare Research and Quality, the Patient-Centered Outcomes Research Institute, the National Institutes of Health, and research experts in social and behavioral science.
With the passage of the Patient Protection and Affordable Care Act of 2010 the United States has begun to expand health coverage to millions of uninsured Americans, and increased attention has been given to population health management. Currently, the limited availability of social and behavioral determinants of health in EHRs limits the capacity of health systems to address key contributors to the onset and progression of disease. Addition and standardization of a parsimonious panel of social and behavioral measures into EHRs will spur policy, system design, interoperability, and innovation to improve health outcomes and reduce health care costs.
IOM (Institute of Medicine). 2011. Incorporating occupational information in electronic health records: Letter report. Washington, DC: The National Academies Press.
IOM. 2012. For the public’s health: Investing in a healthier future. Washington, DC: The National Academies Press.
KFF (Kaiser Family Foundation). 2012. Health care costs: A primer. Key information on health care costs and their impact. Washington, DC: Kaiser Family Foundation.
McGinnis, J., and W. H. Foege. 1993. Actual causes of death in the united states. JAMA 270(18):2207–2212.
Nelson, H. D., P. Nygren, Y. McInerney, and J. Klein. 2004. Screening women and elderly adults for family and intimate partner violence: A review of the evidence for the U.S. Preventive Services Task Force. Annals of Internal Medicine 140(5):387–396.
NIOSH (National Institute of Occuptational Safety and Health). 2014. Comments from the National Institute for Occupational Safety and Health for the IOM Committee on the Recommended Social and Behavioral Domains and Measures for Electronic Health Records. Washington, DC: Institute of Medicine.
Sabatini, L., A. Barbieri, V. Lodi, and F. S. Violante. 2012. Biological monitoring of occupational exposure to antineoplastic drugs in hospital settings. Medical Lavatory 103(5): 394–401.
Thorpe, K. E., C. S. Florence, and P. Joski. 2004. Which medical conditions account for the rise in health care spending? Health Affairs (Web exclusive):437–445. http://content.healthaffairs.org/content/early/2004/08/25/hlthaff.w4.437.full.pdf?origin=publication_detail (accessed August 31, 2014).
Thorpe, K. E., L. L. Ogden, and K. Galactionova. 2010. Chronic conditions account for rise in medicare spending from 1987 to 2006. Health Affairs 29(4):718–724.
USPSTF (U.S. Preventive Services Task Force). 2013. Screening for intimate partner violence and abuse of elderly and vulnerable adults. U.S. Preventive Services Task Force recommendation statement. http://www.uspreventiveservicestaskforce.org/uspstf12/ipvelder/ipvelderfinalrs.htm (accessed July 30, 2014).
Woolf, S. H., and P. Braveman. 2011. Where health disparities begin: The role of social and economic determinants—and why current policies may make matters worse. Health Affairs 30(10):1852–1859.
Woolf, S. H., R. E. Johnson, R. L. Phillips, and M. Philipsen. 2007. Giving everyone the health of the educated: An examination of whether social change would save more lives than medical advances. American Journal of Public Health 97(4):679–683.
Ziegler, E., H. J. Mason, and P. J. Baxter. 2002. Occupational exposure to cytotoxic drugs in two uk oncology wards. Occupational and Environmental Medicine 59(9):608–612.