A Review of Tools to Screen for Social Determinants of Health in the United States: A Practice Brief

Abstract

Increasingly, health care institutions are called on to address social determinants of health (SDH) given the connection to morbidity and mortality across populations. However, widespread implementation of screening for health-related social conditions (HRSC) is lagging. It is estimated that half of patients who have such needs may be missed by failure to screen routinely. Health care institutions face gaps in information related to screening tools. A review of tools that screen for HRSC at the individual level is needed to share readily available and applicable tools for integration in various settings, to communicate how tools are typically administered, and to assess whether tools capture domains corresponding with Institute of Medicine (IOM) core measures. To address these gaps, an unsystematic rapid review of the literature was conducted. In addition to peer-reviewed literature, Google, PubMed, and CINAHL databases and grey literature were searched with a focus on lead agencies or scholars in the field. English language publications from 2008–2018 with content related to SDH or health-related social condition screening tools were included. Nine commonly applied tools were selected and reviewed for content, setting, and method of administration. Fifteen common domains were identified and assessed for alignment with IOM recommendations and correspondence with the construct of social stability. This review consolidates essential information about HRSC screening tools in 1 place and provides practice, policy, and research recommendations to improve HRSC screening. This review is a practice brief that can help health care institutions and clinicians implement screening and interventions related to HRSC.

Introduction

In the 10 years since the World Health Organization (WHO) released its landmark publication on social determinants of health (SDH),¹ researchers, policy makers, and health care institutions have adopted the position that the major health problems of our time must be confronted by addressing the underlying “causes of the causes.”^2,3 Assessing social conditions as key components of health is becoming more integrated into health care settings, accelerated by trends such as: (1) growing acknowledgement of the relevance of SDH in educational content for health professionals^4,5; (2) established evidence being translated into practical application with a recent proliferation of tools to screen for SDH^6,7; (3) SDH indicators are increasingly available in electronic health record (EHR) platforms with greater and lesser adoption across institutions^8,9; (4) their presence also is related to requirements of health care institutions to focus on population health through community benefits programs, among others.¹⁰

Although tools that screen for health-related social conditions (HSRC) have been developed, there is considerable variation in these measures and in the ability to incorporate screening into existing care processes. Both EHR software vendors and national experts on SDH identify lack of standardization of definitions and methods for capturing and reporting SDH in EHRs as barriers to implementation^8,9; thus, currently, critical data related to health and wellness are often missed.

This review addresses gaps in information related to HRSC screening by identifying commonly applied (or proposed) tools for US clinical practice settings, by characterizing where (in what setting) and who completes the tools, and by quantifying the domains consistently assessed in the tools. Further, this brief analyzes whether tools capture domains corresponding with core measures recommended by the Institute of Medicine (IOM¹¹; now the National Academies of Medicine), and whether these domains align with key tenets of the construct of social stability (SS).^12,13 Ultimately, this review provides practice, policy, and research recommendations to enhance the quality of existing tools and facilitate implementation.

IOM-Recommended Core Social Measures

The Committee on the Recommended Social and Behavioral Domains and Measures for Electronic Health Records, the Board on Population Health and Public Health Practice, and the Institute of Medicine were tasked to identify domains and measures that capture SDH for use in EHRs.¹¹ The IOM prioritized domains based on metrics that were already collected, ease of accessing information, and high level of evidence or import. For purposes of comparison, this review assesses IOM-recommended social domains only (not behavioral).

Social stability

Many constructs (eg, SDH, social vulnerability) are associated with the assessment of social and economic conditions affecting health. SS, often missing from the modern lexicon, has been used in clinical assessments and research for decades.^14,15 SS describes the structure and routine in one's life that allows for “maintain[ing] connections with social resources and societal expectations.”^12(p3) This stability comes from “steady social circumstances [over time] within a defined range of domains (eg, housing, employment, social ties, sufficient income, and lack of imprisonment).”^12(p3)

Absent in most HRSC assessments is an overarching framework of how domains are linked at an upstream level; for example, that multiple domains may have to be addressed or met for one to achieve (the broader construct of) stability. Thus, key tenets give the SS construct some advantage over others used to operationalize SDH at the individual level. First, SS recognizes that domains are interconnected and requires the presence of more than one to create stability. Second, SS stresses the importance of a using a defined period of time for measurement of the duration of stability in each domain. Lastly, SS considers quantity and quality (patterns of co-occurrence) of stability. Scoring domain responses can help health care teams quantify needs and assets and prioritize interventions. Noting apparent co-occurrence of domains can facilitate identifying HRSC patterns (or perhaps, profiles) for insights on interrelated social conditions and potential health/behavior risks allowing for more appropriate interventions.^12,13

Design

To identify and critique commonly applied HRSC screening tools, a rapid review was conducted. The strategy applied here falls under Approach 3 from a recent WHO guide: a literature search of more than 1 database, including grey literature, with a search limit of both date and language, and with studies selected and data abstraction by 1 reviewer, with the absence of independent risk-of-bias assessment.¹⁶

Methods

This review involved searching databases, grey literature, and expert searching (ie, seeking out lead scholars or agencies in the field in addition to peer-reviewed literature).¹⁶ Search terms were “recommended screening tool” AND “social determinants of health”; and “recommended screening tool” AND “health-related social conditions” with the search limited to English language publications in the period of 2008–2018. First, a Google search was performed (the most commonly employed approach by health care providers seeking medical information),^17,18 yielding publications from the health care industry, regulatory agencies, professional organizations, and scholars. Websites of the following agencies were then searched with the same terms: Centers for Disease Control and Prevention, Agency for Healthcare Research and Quality, Centers for Medicare & Medicaid Services, National Academy of Medicine/IOM, American Academy of Pediatrics (Screening Technical Assistance & Resource Center), and Siren (Social Interventions Research & Evaluation Network) Metrics, Measures, and Instruments database. Finally, using the same terms and inclusion criteria, CINAHL and PubMed were searched (yielding 123 and 172 results, respectively). After removal of duplicates and review of titles and abstracts to exclude content not related to tools screening for HRSC, 56 publications remained. Complete documents were reviewed and excluded if tools were not applied in clinical settings (but retained if they had clinical application during a research study), or if a full set of screening questions or recommended measures were not publicly available. Nine tools were retained for critique.

Review

Abstraction

First, tools were described for the type of use (as clinical or research tools), type of setting (eg, emergency room, pediatric practice), type of person adminstering the tool (eg, self, nurse, physician, study worker), and number of questions. Next, tools were reviewed for domains included and domains were characterized and quantified.

Critique

Tool domains were assessed for their correspondence with IOM and SS domains. IOM social domains include: race/ethnicity, educational attainment, financial resource strain, social isolation (any assessment of social connection was accepted as corresponding with this domain), intimate partner violence (IPV) (any assessment of exposure to violence/lack of safety was accepted), and, at neighborhood/community level, median income of neighborhood based on patient's geocoded address (any assessment of neighborhood/multilevel indicator was accepted).

SS domains include: housing status (any assessment of being housed or homelessness was accepted as corresponding with this domain), residential stability (any assessment of transience/history of moving was accepted), employment (any assessment of employment was accepted), income (≥$10,000/year is SS definition; any assessment of income was accepted), legal (no incarceration in past year is SS definition; any assessment of legal issues was accepted), and relationship stability (with a main partner for ≥12 months is SS definition, any assessment of relationship status/social connection was accepted).¹³

Tools' domains also were reviewed for correspondence with key tenets of SS including: (1) evidence of stability in more than 1 domain (if tool assessed more than 1 domain, tenet was met); (2) length of time (duration) of one's SS status (any assessment of duration of a social condition was accepted); and (3) assessing the accumulation or combinations of stability (any scoring or consideration of combinations of responses in screening results met this tenet).

Results

The 9 HRSC tools reviewed are summarized in Table 1.

Table 1.

Summary of Screening Tools, Where They Are Administered, Who/How Administered

Screening tool titles, authors	Setting (type of application)	Who/how administered/number of questions
Iscreen20	Pediatric emergency department; applied in research to test for practice (research)	Self-administered or face-to-face with research assistant;
		Computer-based; 23 questions (each with 2 follow-up questions)
WE CARE23	Community health centers well-child visits (research)	Self -administered (parent of patient);
		Paper-based; 6 questions
HealthBegins22	Various clinical settings	Student, health care staff or provider;
	(clinical practice)	Paper; 15 main questions, 14 optional/additional questions
Health Leads19	Emergency departments and primary care	Self/patient, student volunteers;
	(clinical practice)	Paper (available in Spanish); 9 questions plus available additional questions per domain
PRAPARE Protocol for Responding to and Assessing Patient Assets, Risks, and Experiences Version 1.0 © 6	Various clinical settings	Staff/health care provider*;
	(clinical practice)	Computer (available in multiple languages); 12 questions plus demographics
WellRX35	Primary/family medicine clinics;	Self (patient) or medical assistant;
	(research)	Paper; 11 questions
The Accountable Health Communities (AHC) Screening Tool7 (Centers for Medicare & Medicaid Services, 2017: Billioux, Verlander, Anthony, & Alley, 2017)	Various clinical settings;	Staff/health care provider/self;
	(clinical practice)	Paper- and computer-based; 16 HRSC questions, 26 total questions
HelpStep21	Adolescent, young adult clinic; various settings	Self;
	(research and clinical practice)	Computer/web-based; 12 main questions/domains with follow-up questions
AAFP Social Needs Screening Tool; The EveryONE Project24	Family or pediatric practice;	Health care provider or patient;
	(clinical practice)	Paper (short or long form); 14 questions (long form)

^*Health care provider = physician or nurse practitioner.

AAFP, American Academy of Family Physicians; HRSC, health-related social conditions.

Tool, setting, administration

Of the 9 tools reviewed, 6 were regularly applied in clinical practice, and 3 were used in clinical settings as part of research studies evaluating a particular tool or the feasibility of implementing HRSC screening (see Table 1). Five of the 9 tools were applied in pediatric or family practice settings. Six tools were administered by health care providers/staff, mostly physicians, followed by medical assistants. Other tools were administered by research assistants or by self-screening. Most tools are available in paper or electronic format. Total questions per tool ranged from 6 to 23 main questions, with many supplemental questions per domain.

Domains

Many domains are consistently assessed across these tools. Fifteen common domains were identified, and 8 domains were assessed by at least 70% of tools (Figure 1). (For specific domains included per tool, see Supplementary Table S1).

FIG. 1.

Commonly assessed domains of health-related social conditions.

Some tools were more comprehensive: 5 tools included 11 of the 15 common domains (Health Leads¹⁹; Iscreen²⁰; HelpSteps²¹; PRAPARE⁶; HealthBegins²²). Safety/violence (related to individual-level violence in the home or IPV), housing, food insecurity, and employment were assessed most frequently across tools. Eight tools went beyond housing status to inquire about the state of one's shelter: the quality of/problems with one's housing or the status of gas/electric utilities. Six tools inquired about transportation and child care needs. Just over half of tools asked about legal concerns (eg, incarceration, parole/probation, or immigration status), income/source of income, and general financial hardship.

Data on social support (marital status, social connection, or social isolation indicators), demographics (age, sex/gender, race/ethnicity, sexual orientation), and health insurance status (or worry about health care costs) were collected by a minority of tools. However, information from these 3 domains may be collected elsewhere in health records.

Correspondence with IOM and SS domains

Domains were further classified as correspondings to the IOM-recommended social domains or SS domains and tenets (Table 2). None of the tools covered all IOM domains and only 1 tool addressed all SS domains.

Table 2.

Correspondence of Screening Tools and with IOM and SS Domains

IOM or social stability domain	Iscreen	WE CARE	Health Begins	Health Leads	PRAPARE	WellRX	AHC	HelpSteps	AAFP social needs
IOM-Violence	+	+	+	+	+	+	+	+	+
IOM-Education	+	+	+	+	+	+	-	+	+
IOM-Financial resource strain	+	-	+	+	-	-	-	+	+
IOM-Race/ethnicity	+	-	-	+	+	-	-	+	-
IOM-Social Support	-	-	+	+	+	-	-	+	-
IOM-Neighborhood income	-	-	-	-	-	-	-	-	-
SS*-Employment	+	+	+	+	+	+	+	+	+
SS-Housing	+	+	+	+	+	+	+	+	+
SS-Income/source	+	-	-	+	+	+	-	+	-
SS-Legal	+	-	+	+	+	-	-	+	-
SS-Social support (partner/relationship)	-	-	+	+	+	-	-	+	-
SS-Moving/transience	+	-	+	+	-	-	-	-	-

Cells with “+” signs indicate correspondence with IOM-recommended and SS domains. ^*SS domains are for the past 12 months.

AAFP, American Academy of Family Physicians; AHC, Accountable Health Communities; IOM, Institute of Medicine; PRAPARE, Protocol for Responding to and Assessing Patient Assets, Risks, and Experiences; SS, social stability.

IOM domains

Violence, education, and financial resource strain were the most consistently included domains across tools. Social support was less frequently assessed. None of the tools explicitly included a measure of neighborhood or community-/population-level variable per the recommendation of IOM. Two tools, Health Leads¹⁹ and HelpSteps,²¹ contained the most IOM recommend domains (5 out of 6).

SS domains

Five tools assessed at least 5 of 6 SS domains (Iscreen,²⁰ Health Leads,¹⁹ PRAPARE,⁶ HelpSteps,²¹ Health Begins²²). Legal issues and income were included in most but not all tools. Partner/relationship status (social support) and moving/transience were assessed less routinely.

Tenets of SS

The key tenets of SS were less consistently addressed than the SS domains.

Tenet 1: More than one domain necessary for SS. All the tools assessed more than 1 HRSC domain.

Tenet 2: Duration/Length of time. Most tools (67%) considered the duration of a HRSC domain. Table 3 describes the tools that assess duration, including the domains and the time period assessed. Four tools assessed the duration of status for more than 1 domain. Food insecurity was the most common domain for which duration was assessed.

Table 3.

Summary: Health-Related Social Conditions Tools and Assessment of Duration of a Social Condition Status

Title of tool	Number of domains duration	Duration of status
Iscreen	1 domain: move/transience	Past 6 months
WellRX	1 domain: food insecurity	Past 2 months
Health Leads	3 domains: utilities, food;	Past 12 months; next 2 months
	housing security
PRAPARE	10 domains: food insecurity; utilities; medicine/health care; clothing; child care; phone; incarceration; intimate partner violence; seasonal or migrant work; household income	Past 12 months
HealthBegins	2 domains: housing, moving	Past 1 month
AAFP Social Needs Screening	4 domains: housing, transportation, utilities, food insecurity	Past 12 months

AAFP, American Academy of Family Physicians; PRAPARE, Protocol for Responding to and Assessing Patient Assets, Risks, and Experiences.

Tenet 3: Accumulation and Type of Stability (scoring and combinations or pattern of stability). Only 1 tool, HealthBegins,²² explicitly used a scoring system (some items weighted, and score used to rank need and interventions). WE CARE²³ did not have a score tool, but the respondent was asked to rank the top 3 priorities. The American Academy of Family Physicians Social Needs Tool²⁴ scored the personal safety indicator (only) to develop a safety action plan. Combinations or patterns of social conditions were not assessed by the tools or analyzed in results. There was no discussion of interpreting results based on HRSC response profile as a potential predictor of risk or protection. Most tools allowed the patients or providers to interpret and prioritize results and need for intervention.

Discussion

Nine recommended, publicly available, and commonly applied HRSC screening tools were identified via rapid search technique. In a survey of domains, 15 were commonly included and 6 tools assessed the same 8 domains, demonstrating some conceptual uniformity in current HRSC screening. Yet, there was great variation in definitions and measurement (lack of standardization).

Exposure to violence was the domain most commonly assessed. Screening for violence may already be in place in clinical settings given the US Preventive Services Task Force (B grade) recommendation on screening for IPV, especially among women of childbearing age.²⁵ Employment, food insecurity, housing, and education were the next most commonly assessed. Components of these domains may already be routinely gathered for purposes not specifically related to HRSC screening, and therefore readily integrated into practice. For example, documenting a patient's address may provide information on homelessness, assessing occupational exposures may indicate employment, and literacy/health literacy assessments may inquire about highest level of education attained. The prevalent integration of food insecurity screening may be related to the availability of a concise, validated, and reliable measure.²⁶

Missing domains

There were some notable omissions to the most commonly assessed domains, especially in key areas of SS and IOM recommendations: legal stability, moving/transience, social support, and multilevel assessment.

Missing measurements

Many tools inquire about the current status of one's condition without regard to the duration/length of time that one has experienced that status. Not knowing if a condition is temporary or chronic limits understanding and vital contextualization. Another measurement challenge lies in the absence of scoring of tools and lack of consideration for combinations or patterns of responses. Finally, missing in all tools was the assessment of multilevel influence such as neighborhood/community factors.

Limitations

This review is limited in its scope and depth and subject to bias. Other tools (published and unpublished) may assess similar conditions but were not reviewed here. A strength of this review is that it replicated real-world search strategies by first examining the broadly available grey literature and then reviewing scientific publications. Further, the aim of the review was not a systematic assessment of the evidence for HSRC screening. Rather, it summarizes commonly applied or recommended tools that could be readily adapted to other settings. Another potential limitation is the review's focus on contrasting tool domains with domain recommendations of IOM and the construct of SS while not evaluating other merits or deficits. However, this focused critique allows valuable information and relevant suggestions to be conveyed to practitioners, researchers, and policy makers to increase the usefulness of data gathered by existing tools.

Implications for Policy and Practice

Practice recommendation 1

As calls for government agencies and institutions to guide standardization of what is required in SDH/HRSC data capture, these entities, and developers of HRSC tools or EHR software, as well as the practitioners who adopt them, should consider inclusion of certain domains in all tools, particularly legal issues, moving/transience, social support, and multilevel assessment.

Legal

The United States has, by far, the largest proportion of residents incarcerated in the world.²⁷ Further, the detention of individuals with unclear immigration status is increasingly common and destabilizing for individuals and families (467,000 people were detained at the United States–Mexico border in 2018).²⁸ Including legal concerns in HRSC is important, especially as minority and immigrant populations, persons living in poverty, or persons with mental health conditions are disproportionately detained.²⁹

Moving/transience

In multiple studies, behaviors related to HIV acquisition or problematic drug use,^13,30 and recently, adverse birth outcomes,³¹ were associated with having moved or transience in the past year; thus, it may be important to assess moving as a contributing destabilizing influence to one's health.

Social support

Per WHO, “…greater support from families, friends and communities is linked to better health.”³² Importantly, whether and how social support is assessed may reveal sources of resilience or of stress and there are multiple valid and reliable tools for health care institutions to consider when assessing this concept.³³

Multilevel assessment

Individual-level social conditions stem from the social/community environment. Interventions at the individual level may not have limited impact in the long run. For example, geocoding addresses as part of screening, as recommended by IOM, can create awareness of geographic communities for prioritized interventions, particularly relevant for health systems with community benefit districts.¹¹ In fact, EHR vendors “described a common request from clients to use secondary SDH survey data to identify “hot spots” or areas of high social need in the communities they serve.”^8(p11)

Practice recommendation 2

Health care institutions/clinicians who adopt screening should assess duration of HRSC status. The element of time was included in most but not all tools. With a simple addition, HRSC tools can be modified to assess duration of a condition to better inform service to clients and for clinical and service utilization purposes.

Practice recommendation 3

Authors of tools should provide guidance to quantify assessments, or health care institutions/clinicians should score responses and document combinations or patterns of HSRC needs. Most reviewed tools did not contain scoring or recommendations for interpretation of results at the individual level nor representation of captured data at the aggregate level. There is evidence that meeting 3 or more domains of stability reduces likelihood of experiencing exposure to health risks.^13,34 Tools that do not quantify or qualify HRSC needs hinder providers' ability to prioritize or develop appropriate interventions, and perhaps to prevent health risks. Further, noting apparent co-occurrence of domains allows for documenting HRSC patterns and interrelated needs and risks. The complexity of HRSC clearly implicates the need for interprofessional care teams to address multiple domains. Finally, standardization of interpreting and reporting SDH/HRSC data is another common request among clients of EHR vendors.⁸

Practice recommendation 4

Bottom line

More health care institutions/practitioners should systematically assess HSRC. The tools reviewed were applied primarily in pediatric clinical settings, yet HRSC screening should be applied universally in clinical settings. Half or more patients who have HRSC needs may be missed by not screening routinely.^35,36 As Dr. Alicia Cohen noted, “We recognize that unmet social needs have critical implications for health, and however, our ‘eyeball’ test is incredibly poor.”^37(col1) If we fail to screen, we miss opportunities to intervene. However, successful screening leads to successful referrals and needs being met.^23,35,36 “When provided with services to address health-related social problems, the majority chose to receive help; nearly half successfully addressing their priority concern.”³⁸ More institutions serving different populations should begin implementing screening for HRSC, but specific populations may need tailored questions. This process will be greatly facilitated with the adoption of the increasingly available packages of SDH tools within EHR systems. Already, clients of major EHR vendors are interested in SDH/HRSC data to target specific populations such as persons living with diabetes or chronic pain.⁸

Policy and research recommendations

Gathering baseline data to contrast with results post implementation of HRSC screening will be needed for evidence of the value added from use of the tools (eg, determining the proportion of new needs detected and new interventions done). As more health systems are asked to implement HRSC tools or integrate these indicators as part of EHR systems, implementation science research studies are needed to establish best practices for integration of new practices.

Next, evaluating the uptake or utilization of referrals/services and the outcomes of these interventions will be important to build sustained support for health care system policy related to screening and addressing HRSC. Clinical trials reporting outcome information demonstrated successful referrals and uptake of services.^23,38 However, establishing standard guidance from funding or regulatory agencies on how to measure outcomes of screening and interventions (especially related to that complicated and distal outcome, improved health) will be important. These data can then be leveraged in advocacy for broader policy change at the health care institution, community, or other structural level to alter SDH and thus improve the health of populations.

Conclusion

Routine assessment of social conditions is needed given their link to health. Further, health care agencies soon may be required to do so. This brief can serve as a focused resource of identified and recommended HRSC tools, a critique of tool domains and their alignment with specific recommendations, and a source of clear suggestions for improving existing tools and their implementation into practice, as well as next steps for research and policy.

Author Disclosure Statement

The authors declare that there are no conflicts of interest.

Funding Information

No funding was received for this article.

Supplementary Material

Supplementary Table S1