Clinical Interventions Addressing Nonmedical Health Determinants in Medicaid Managed Care
Laura M. Gottlieb, MD, MPH; Kim Garcia, MPH; Holly Wing, MA; and Rishi Manchanda, MD, MPH
The World Health Organization’s (WHO) Commission on the Social Determinants of Health has defined social determinants of health (SDH) as “the conditions in which people are born, grow, live, work, and age. These circumstances are shaped by the distribution of money, power, and resources at global, national, and local levels.”1
These conditions include social, economic, and environmental factors, such as income, education, housing, employment, transportation, and the physical layout of neighborhoods.
There is increasing evidence that SDH play a major role in the onset and progression of disease. Addressing SDH may help achieve the healthcare “Triple Aim”—improved healthcare quality for individuals and populations, and decreased healthcare costs.2
Despite this, much of the field of social epidemiology—including that related to healthcare disparities—focuses on documenting the effects of SDH rather than on ways to change or intervene on exposures.3
Existing social interventions integrated into clinical settings—which range from interventions offering housing to homeless patients to on-site food pantries and legal services clinics—can expand the traditional bounds of healthcare to address “upstream” determinants of health, including social, behavioral, and environmental conditions.4-6
Although the field of interventions related to nonmedical health determinants is growing in clinical settings around the country,7
little is known about the operational and design characteristics that define them—particularly in the context of differing payer and provider environments. Understanding how SDH interventions are designed, implemented, funded, and scaled within distinct payer environments is key to translating the growing interest around the role and replication of SDH interventions in healthcare settings into a substantive, actionable strategy.
Based on distinct membership characteristics and financial incentives defining Medicaid managed care organizations (MMCOs), these payers may be particularly well-suited to support integrated SDH approaches.8
The populations served by MMCOs are disproportionately affected by poverty and associated material deprivation, including food and housing insecurity, poor habitability, unsafe drinking water, social exclusion, low education levels, and unemployment. These conditions are known to reduce opportunities, limit choices, and threaten health.9
Additionally, over the last 15 years, there has been a nationwide increase in patients enrolled in MMCOs,10
which already enroll about half of all Medicaid beneficiaries.11
This combination of population needs and the shift toward risk-based care together encourage upstream intervention and prevention as one potential way to limit costly healthcare utilization.
Despite the apparent alignment of these structural characteristics with low-income members’ unmet social needs, there are multiple challenges limiting MMCOs from expanding social services.12
New prevention services are not easily incorporated into MMCO-state capitation agreements, so MMCOs have to cover any additional benefits out of administrative or community benefit dollars. Coding practices and other administrative requirements for MMCOs can also make it difficult to adopt new prevention services. Furthermore, any financial return related to social service investments may take many years to realize, which can decrease the financial feasibility of adoption. Finally, MMCO care delivery models, financing contracts, and organizational structures (which may span several states) can make community collaborations and public partnerships—often critical to a comprehensive approach to social service delivery—more challenging.
A Systematic Review of Clinical Interventions Addressing Nonmedical Health Determinants
Reviews completed to-date on clinical SDH interventions have offered an important glimpse into the range of potential interventions and funding mechanisms in this emerging field, but have not answered key questions about the implementation of these interventions, their financial drivers, and other characteristics specific to the context of MMCOs.7,13,14
We conducted a systematic review to identify published literature on clinical SDH interventions supported by MMCOs, the design and integration of these programs into healthcare delivery systems, and the determinants addressed and the target populations served by these programs. Our aim was to inform health policy decision makers around incentivizing these programs for broader dissemination.
Based on previous research in the area of clinical SDH interventions, models of organizational readiness for change,15
and an assessment of existing MMCO structural incentives and barriers to incorporating these interventions into routine activities, we developed a theoretical model outlining the range of ways that MMCOs might engage in activities related to SDH (Figure 1
). We hypothesized that most MMCOs would be making low-SDH investments, given the relatively recent increase in SDH research.
We conducted a literature review examining how commercial and Medicaid-only MMCOs invest in interventions that address SDH for their patients or their network clinics’ patients. To be included in our review, an intervention or program had to meet our inclusion criteria: a) based in the United States, b) financially supported by an MMCO, c) address at least 1 social determinant of health (housing, employment, food, education, safety, legal services, or transportation), and d) be integrated into the healthcare services delivery system. The definition of SDH varies in different contexts, so we focused our search on SDH not typically addressed within the current healthcare delivery system. As a result, we excluded papers describing interventions related to health behaviors, including behaviors like tobacco use and physical activity. We also excluded papers describing interventions exclusively related to healthcare access. Clinical integration meant that a patient’s social need was identified within the clinical setting and then either referred to an external intervention program or to a social intervention conducted in the clinical setting.
We developed an electronic search strategy to scan for references in the following databases: PubMed, Scopus, and Business Source. We limited our search to publications dated 2000 to 2014 and we combined search terms using “AND” to capture at least 1 term from each of 3 major categories: SDH, healthcare settings, and intervention studies (see eAppendix for more detail [eAppendices
available at www.ajmc.com
]). Funding source was added as an element of the detailed data extraction process on references meeting other review criteria. References were also collected from national experts in the field, including the study’s national advisory group.
Our review methods included a hierarchical exclusion process. Titles and abstracts of references collected from the electronic search strategy and the national study advisory group were assessed initially based on whether or not they described an SDH program or other intervention. Where the title and abstract were insufficient to deduce if they referred to an intervention, the full-text article was reviewed. Those references that referred to SDH but did not describe an intervention were excluded; in other words, articles were excluded if they only described theories of SDH or studies on risk factors and disparities without describing any specific intervention or program addressing those risk factors or disparities.
Remaining references were excluded if they did not describe an intervention or program addressing at least 1 SDH. For example, references were excluded if interventions or programs focused exclusively on medical care and services, such as treatments, immunizations, or health behaviors; described quality improvement interventions without an SDH component, such as guidelines intended to improve clinical care coordination that did not address social needs; or exclusively described healthcare access interventions or outreach programs, such as mobile health services or Medicaid enrollment programs.
References describing an intervention addressing SDH underwent review of the full-text article to determine the degree of clinical integration and financial support from an MMCO. Those interventions or programs without any description of clinical integration or MMCO support were excluded from the final data set. For instance, interventions and programs that were community-based and had no integration with clinic-related work flows related to screening, intervention, referral, or tracking, were excluded. Two investigators reviewed each article included to determine if it met all inclusion criteria. A third investigator reviewed any articles where the reviewing investigators were in disagreement; in these cases, final decisions were made about inclusion after discussion between all reviewers. Interventions described in more than 1 reference were only counted once.
For those interventions and programs meeting inclusion criteria, we collected a detailed set of program data, including intervention name, organization, name/state of MMCO, HHS region, clinic setting, program start date, description of the intervention or program, target population, target SDH, level of intervention/prevention (primary, secondary, tertiary), model of intervention (social and/or medical approaches), level of clinical integration (extent to which clinical providers identify/address social need), study citation, study design, study findings, and an assessment of evidence quality based on the Community Guide to Preventive Services evidence rating guide (high, medium, low quality).
The electronic search strategy yielded 3975 unique references from PubMed, Scopus, and Business Source (Figure 2
). References from our national experts included, but were not limited to, reports from the Association for Community Affiliated Health Plans, Alliance of Community Health Plans, Medicaid Health Plans of America, Manatt Health Solutions, and the Institute for Alternative Futures.
A total of 111 articles were identified that included SDH interventions. These 111 articles were screened to determine whether they described interventions integrated into healthcare delivery systems and whether they were funded by MMCOs. Although many described serving Medicaid populations, only 13 articles described Medicaid managed care–supported interventions addressing SDH integrated within a clinical setting. The final data set included 25 interventions and programs described within those 13 articles.12,16-27
Seven references described 1 primary SDH-focused intervention or program12,16-18,21,22,26
and 6 references described more than 1 intervention.19,20,23-25,27
Targeted Populations, Targeted Social Determinants, and Intervention Models
Target populations included specific demographically defined groups—primarily low-income individuals, children, or families. Other population groups included seniors, minority groups (racial, ethnic), those experiencing homelessness, and those who were broadly eligible for Medicaid. Some interventions defined target populations based on healthcare utilization patterns (eg, high-cost, high-utilizer), while others focused on patients with specific health conditions (eg, asthma, hypertension, diabetes and other dietary-related chronic health issues, HIV, multiple sclerosis, mental illness).
Seven interventions specifically targeted a single social issue, such as housing (4),12,21,23
food quality and availability (2),23
and employment (1).23
The remaining 18 interventions provided more comprehensive services to address multiple SDH. Interventions varied widely in how they addressed SDH within the clinical setting. Specific intervention components included variations on team-based approaches, including case managers, social workers, community health workers, and other nonprofessional staff integrated into clinical teams—although the degree of integration and communication with other clinical staff was rarely described. Intervention settings also differed across programs, including both training and education delivered on site, referrals to off-site programs, and home visits.
The majority of the programs identified were designed to address both social and medical needs of the patients being served. For example, asthma programs typically included elements focused on pharmacological management and specific housing risk factor reduction.21
Other case management programs were designed to facilitate pharmacy and appointment access in addition to social service linkages.17,20,27
Eleven program descriptions included some empirical evaluation indicating effects of the intervention on health system outcomes. Five program evaluations reported reductions in emergency department (ED) visits16-19
and reductions in hospital admissions.16,17,19,20
The impact of 1 home-based intervention targeting high-risk asthma patients on ED use and hospital admissions reported mixed results. Although the initial evaluation of enrollees’ pre- and post-utilization patterns showed significant reductions in hospital admissions and ED visits, a subsequent, more rigorous evaluation comparing an intervention group with an untreated control group found no overall differences in utilization patterns between the 2 groups.21
Several programs reported cost savings associated with changes in enrollee utilization patterns.16,17,19,20,22,23
Three studies described higher levels of patient satisfaction as a result of social intervention programs,18,22,23
and 1 study reported an increase in quality of care.18
The majority of program descriptions included no outcomes data or other return-on-investment information. The eAppendix includes a complete list of interventions and programs included in this review (eAppendix Table 1).
This study provides an overview of ways in which MMCOs help address patients’ nonmedical needs by using healthcare services as a venue for social needs screening and related social needs interventions. The 25 programs captured are geographically dispersed across the United States and across multiple MMCOs. The strong majority of programs identified target high healthcare–utilizing patients with specific chronic health conditions. Within target disease groups, programs frequently focus on specific racial or ethnic groups and low-income, homeless, or other specific sociodemographically defined populations. The existing literature provides no empirical data showing that MMCOs engage in universal social screening, needs assessments, or resource mobilization to address the social needs of all members. Based on our theoretical model, these findings suggest that MMCOs are making low, or at most, low to moderate investments in SDH interventions and are not yet systematically engaged in comprehensive SDH strategies to improve health or change healthcare utilization patterns of enrolled patients.
This review differs from previous reports examining SDH interventions in 2 important ways. First, it focuses exclusively on MMCOs because these organizations are relatively well-positioned and incentivized to address the social needs of their patient populations to improve healthcare outcomes and service utilization. Although we found many interventions and programs addressing SDH that serve Medicaid patients, this review focuses on the few that are financed or directly supported by MMCOs. Second, the review focuses on programs that have at least some degree of clinical integration, meaning that individual patients are being screened for social needs and connected with relevant services based on being patients in a clinical care delivery system. Understanding the degree to which SDH programs are clinically integrated is key for MMCOs, which must decide whether spending on SDH interventions will be linked to a plan’s patient care and quality improvement activities or to its community benefit activities. Previous reports describing and examining innovations in addressing SDH have either not been specific to MMCOs or have included both clinical and community-level interventions.7,9,13,23
The 25 program descriptions we identified provide little detail on key program characteristics or MMCO decision-making processes that could help establish and disseminate best practices, such as the role of internal or external financial or other drivers or barriers to undertaking these interventions; any relevant community needs assessments on which interventions are based; or the role of executive sponsors, project owners, and key stakeholders in shaping the interventions. Similarly absent are descriptions of the return-on-investment calculations required to sustain these types of programs. This lack of information on organizational decision making, “readiness” assessments, and management processes, combined with the lack of rigorous evaluation of the impacts that these types of interventions have on health outcomes or health services, limits the capacity to understand and disseminate best practices in SDH-related interventions among MMCOs.
Our review revealed several reports of case management programs for high-risk patient populations that include both social and medical components. This blended approach of social and medical case management may be an important target for scaling nonmedical health interventions within MMCOs. A recent report from John Snow, Inc (JSI) suggests that these services elicit a better response from clients if initiated at the provider level rather than at the payer level, which could incentivize MMCOs to fund provider-delivered programs.13
Both the JSI report and a related Commonwealth Fund issue brief authored by the Center for Health Care Strategies further suggest that the absence of assured flexible use of Medicaid managed care capitation rates may prevent MMCOs from transitioning from traditional case management of medical services to case management that includes behavioral and social needs coordination.13,14
Ensuring flexible funding for managed care capitation rates could improve MMCO case management programs that address a combination of patient medical and social needs. This funding may be accessible via the Affordable Care Act’s Health Homes program, which does require both comprehensive care management and increased referrals to community and social support services28
; the Health Homes program already supports intensive case management activities in 19 states,28
though only 1 was identified in this systematic review.29
References meeting our inclusion criteria show wide variation in associated key words and terms, which made it difficult to develop a comprehensive, practical electronic search strategy. The WHO definition of SDH is very broad, and the process of translating that broad concept into meaningful search terms that capture specific social determinants interventions is complex. Furthermore, some MMCOs may choose not to publish information regarding successful programs in order to maintain advantage in a competitive marketplace. To minimize the challenges inherent to this search strategy, we supplemented the electronic search using references provided by national experts. Nevertheless, there could be programs that meet our inclusion criteria that were not identified by these methods.
Our methods included a hierarchical exclusion process: references were initially assessed based on whether or not they described an SDH program or other intervention. Those that referred to SDH but did not describe an intervention were excluded. Remaining articles were then reviewed to determine whether there was some degree of clinical integration for the intervention. In the 25 programs captured, there was a considerable range in the extent to which clinical integration was described. For example, one program simply said that a social worker was added to care teams (article not included), while another program more clearly described the integration, including information about how the social worker addressed patients’ social needs (article included). It is possible that some programs were clinically integrated, but the integration was insufficiently described in the reference article to justify inclusion. Future efforts should supplement available information via key informant interviews, organizations’ annual reports, or other data sources. The lack of existing information may negatively influence dissemination or quality improvement efforts.
Additionally, many references in the original search did not include a description of the funding mechanism for the program. In other cases, funding was from a source other than Medicaid managed care entities. Interventions that met the other inclusion criteria but were not clearly funded by MMCOs are listed in eAppendix Table 2. To improve scaling and dissemination efforts, journals may consider requiring reporting of funding mechanisms.
This review is an important first step toward understanding how MMCOs are making investments in clinical nonmedical health determinants. Although risk-based capitated payment systems serving low-income populations provide incentives for incorporating models to address SDH, real and perceived local, state, and federal barriers can dis-incentivize adoption. In the context of increasing federal and state funding experimentation supporting Medicaid investments in SDH, MMCOs hoping to invest in these interventions will require detailed implementation, operations, scaling, and sustainability descriptions from other programs that have begun to make these investments.
The authors wish to thank Stephanie Chernitskiy for her editing and assistance with the paper’s figures and Beena Patel for her assistance in the literature review. Support for this study was provided by the Commonwealth Fund.