INTRODUCTION

Social determinants of health (SDH) are the conditions in which individuals are born, grow, live, work, and age.1 The interplay of “community” and health status is not novel,2,3 with health-related social needs (HRSNs) such as housing, access to healthy food, and education repeatedly shown to play a fundamental role in what makes a person healthy. More recently, evidence defining political4 and structural5 drivers of health further supports the idea that determinants of health span interpersonal, community, and systemic levels. Current HRSN research also suggests that incorporating HRSN interventions within traditional healthcare (i.e., medical interventions performed in the hospital or clinic) possesses challenges.6 Moreover, the lack of social interventions is implicated as a cause of numerous health disparities, such as the prevalence of low-birth weight babies born to Black mothers7 and decreased life expectancy in individuals with a history of homelessness.8

To address HRSNs of patients, the passing of the Patient Protection and Affordable Care Act (ACA) in 20109 positioned HRSN identification and action into modern healthcare practice.10 The ACA further endorsed professionals specializing in the coordination of patients’ health and social care, community health workers for example,11 as integral members of the healthcare team.

The emphasis on HRSN prompted a growing number of studies both quantifying the impact of SDH on health status and determining the effectiveness of newly adopted strategies.10,12 Numerous studies also sought to determine the clinical impact and cost-savings of linking patients to social service systems through the provision of referrals to outside agencies. For example, evidence indicating a high prevalence of HRSNs among patients,13,14 cost-savings of CHW programs,15 and improved chronic disease management associated with support from patient navigators16 provided backing for expanding the scope of medical practice to include social interventions. Despite these findings, research describing how hospitals and clinics address HRSNs or characterize the role of community-based, social service agencies in such processes is lacking in the literature.

Minnesota possesses some of the nation’s most glaring health and social disparities; Black, Indigenous, and People of Color in Minnesota experience higher rates of incarceration, children in poverty, unemployment, and higher incidences of stroke and diabetes compared to white individuals.17 The purpose of our study was to examine how primary care clinics in the Minneapolis-St. Paul, Minnesota metropolitan area, address the HRSNs of their patients. Specifically, this paper describes the range of ways clinics address these needs after identification and initial contact with a frontline staff person is completed. Although a reasonable degree of practice variation is expected, we attempted to establish a baseline of this variability to assist future study design and program implementation. We also sought to identify the successes, challenges, and opportunities for addressing HRSNs in primary care. To accomplish these goals, we conducted qualitative interviews with staff from primary care clinics.

METHODS

Study Sample

Clinics were identified from a directory of primary care clinics compiled by the Minnesota Department of Management and Budget and were purposively sampled to include representation from Federally Qualified Health Center (FQHC), health system–associated, and University-associated clinics. Clinics were recruited to our study in two ways: an invitational email was sent to a known clinic representative by the research team or clinic leadership was contacted via phone. The sample size was not determined a priori and clinics were recruited until saturation was reached, which occurred after 12 interviews. Our final sample consisted of 21 personnel, representing 24 clinics. Clinics had between 3 and 44 providers; the sample included family medicine (n=9), pediatrics (n=2), internal medicine (n=1), or a combination of several specialties (n=12) (Table 1). Clinic interviews were performed when most convenient for the interviewees, typically during normal clinic hours. Participants did not receive financial compensation for participating. The study protocol received exempt status by the University of Minnesota Institutional Review Board (Appendix 1).

Table 1 Summary of Clinic Characteristics and Participants Interviewed
Full size table

Data Collection

Semi-structured interviews were conducted between July and December 2018. An interview guide was developed to explore factors such as clinic workforce, screening, and referral processes (Appendix 2). The interview guide was developed by the first three authors in collaboration with the senior authors, who have extensive experience with primary care research (IB) and qualitative research (CM). The interviewing team consisted of the lead authors, two MD candidates (LZ and BJ), and an undergraduate research assistant (AI). A pilot interview was conducted with a health care social worker to obtain feedback on interview content, organization, and timing. Interviews averaged 42 min (range: 27–58 min) and were conducted with at least one staff who was directly responsible for addressing HRSNs or had knowledge of the interventions employed by the clinic. All interviews were conducted by AI and at least one of the first two authors. Interviews were audio recorded and transcribed by the interview team. The first four authors, including the senior author with expertise in qualitative analysis, regularly met throughout the data collection phase to debrief, discuss initial findings, and determine when data saturation18 was reached.

Data Analysis

Interviews were coded via thematic analysis19 using inductive and deductive coding. This method involved in-depth review of transcript data and creating categorical units that were used to identify themes in the coded data. First, three study members (LZ, BJ, AI) performed a review of interview transcripts, including reading transcripts, open coding, and discussing notes in team meetings. Based on this review, a preliminary codebook was developed that included deductive (reflecting the focus of the interview guide) and inductive codes. The codebook included definitions of codes with examples of when codes should/should not be used (Appendix 3). The first four authors each coded two transcripts using the preliminary codebook and noted where additions or clarifications should be made. After finalizing the codebook, the two lead authors (LZ and BJ) independently coded 50% of the transcripts using the tag and category sets feature on Microsoft Word. The two coders then compared their coding results to determine if there were inconsistencies in their application of codes. After approximately 6–7 of the total transcripts were coded, the first four authors met to reevaluate codes and add additional case examples to the codebook. The third coder (AI) reviewed each transcript to verify that codes were applied consistently. Finally, the codes were organized into themes and exemplar quotes were identified by the first two authors.

RESULTS

We identified significant variation in four areas present in the processes used by clinics to address patients’ HRSNs after identification and initial contact by frontline staff: clinic personnel involved in addressing needs, clinic referral processes, “resource” and “success” definitions, and barriers to accessing community-based supports. Each of these areas supported our first research aim to identify the range of ways clinics address these needs after identification and initial contact with a frontline staff person. Moreover, clinic referral process, “resource” and “success” definitions, and barriers to accessing supports also relate to our second research aim to identify successes, challenges, and opportunities.

Clinic Personnel Involved in Addressing Needs

After a patient has a HRSN identified, the individual is typically referred to a clinic employee whose role is to address the concern. These frontline staff are tasked with triaging patient needs and offering in-house services or referrals to outside agencies. While this process was present in most clinics, we identified significant variation in titles, education, funding, and responsibilities of those staff involved after the initial identification of a HRSN.

Staff titles were variable and included clinical coordinator, social work care coordinator, community health worker, and nurse care coordinator. Moreover, the training background of frontline clinic staff spanned several education levels: from high school diploma to certificate program to doctoral degree (e.g., “... [Name] is extremely overqualified for the position; he is a doctor in [Country] …” (I09, Program Manager)). Across clinics and health systems, staff with the same or similar titles had different levels of education; clinics did not have consistent educational standards for these staff and often prioritized skills over educational background (e.g., “... our focus was on [Resource Navigator’s] interpersonal skills and their ability to show empathy …” (I11, Program Manager)).

When asked to define frontline positions’ roles and responsibilities, respondents provided answers ranging from broad responsibilities (e.g., “I basically attend to the social needs of the patients ... anything and everything.” (I03, Social Work Care Coordinator)) to focused tasks (e.g., “... our community health workers just talk to women about family planning.” (I05, Quality Manager)). Specific responsibilities depended on numerous factors, such as common needs of the clinic’s patient population, presence of additional support staff, size of the staff’s patient panel, and requirements imparted by funders. The roles of grant-funded positions were often well-defined and included specific interventions intended for specific populations (e.g., “... [Position] specifically focus(es) on immigrants and refugees that have experienced trauma ...” (I03, Social Work Care Coordinator)).

Positions funded through the clinic’s payroll were viewed as a solution to a well-defined clinic shortcoming that was either 1) deemed to be worth investing in (e.g., “... it’s [organization] choosing to use their funds to employ positions that aren't … reimbursed for services but they think they're getting benefits in other ways ...” (I08, Social Worker Care Coordinator)) or 2) a requirement of the clinic’s designation or care delivery model, for example a patient-centered medical home or FQHC.

While respondents overwhelmingly performed daily tasks inside the clinic, some noted extension into the community or performing home-based visits (e.g., “We also do face to face, home visits ... I think [clinic] is unique in that we do [these] ...” (I07, Social Work Care Manager)). Accompanying patients to community-based referral sites and visiting community organizations receiving the highest proportion of referrals represented the primary reasons frontline staff step beyond clinic walls, although these were not endorsed as common practices.

Clinic Referral Processes

After identification of a need and consultation with frontline staff, referral to outside agencies represented the primary way in which clinics addressed patients’ HRSNs. Regarding the provision of a referral, three overarching components of a clinic’s process were identified: knowledge and assessment of community agencies/organizations, post-referral tracking, and referral data collection.

Knowledge and Assessment of Community Agencies/Organizations

Frontline staff noted that being able to discuss referral site location, eligibility requirements, and patient experiences at a given site was integral to providing effective referrals. Organizations addressing housing, food insecurity, and legal status were commonly cited referral sites. Knowing which organization was appropriate for a patient depended on numerous factors, including staff knowledge, previous professional experience, success at a given site, or formal assessment of a referral site.

Assessment of community agencies, organizations, or other referral sites was performed in a variety of ways. Several respondents indicated that calling a new or unknown organization was the most effective way to assess a potential site (e.g., “... you have to really learn about that resource ... ‘Who are you serving, what’s eligible …’” (I12, Medical Assistant)).

In-person visits to community-based referral sites were a mainstay for only a few clinics. Numerous other respondents utilized “… connecting with other social workers or the people doing the work.” (I03, Social Work Care Coordinator) to learn about community supports.

Information gathered from the above-described methods was compiled in several ways. While a number of clinics kept “... lots of books, lots of resource pages ...” (I08, Social Worker Care Coordinator) that were updated periodically, others utilized collections of pamphlets or brochures.

Despite possessing quality information, a shortcoming of brochures, pamphlets, and lists of phone numbers was the need to update agency contact information, eligibility requirements, and services provided (e.g., “... there’s tons of [organizations] that I don’t know about ... or the requirements change or agencies close …” (I04, Social Worker)).

Few clinics opted to purchase HRSN databases, such as NowPow, in which a patient’s needs are inputted into the database and a print-out of community services is generated based on several factors, including zip code. For clinics utilizing databases, this print-out is then given to patients to address their needs (e.g., “... [NowPow] just printed out [nutrition and family support information] and the patients have to call ...” (I10, Social Work Care Coordinator)). Compared to pamphlets, a commonly cited benefit of using a database was that agency information is frequently updated. Other community level databases are available for specific needs, such as youth housing options; however, these are uncommonly used.

Post-Referral Tracking

Referrals intended to address an identified HRSN were most commonly in the form of agency contact information. Few respondents indicated that patient or referral information was communicated with the referral agency. Discussions with patients concerning expectations and next steps post-referral were performed by some respondents; however, “... sometimes we’re just handing them a piece of paper ...” (I08, Social Worker Care Coordinator).

Frontline respondents’ case loads ranged from roughly ten to “... about 5500 patients.” (I06, Nursing Supervisor). Follow-up timing after the provision of a referral varied significantly: from “... every 7-10 days they'll get a call ...” (I11, Program Manager) to “... contact at least once in six months.” (I04, Community Health Worker). Follow-up protocols often depended on patient panel requirements, grant requirements, the specific need, the patient’s level of engagement, or how well the patient was doing clinically.

Respondents described tracking referrals as both a challenge and opportunity to address HRSNs more effectively. Many clinics noted tracking “... how long they’ve met with the patient and what resources they gave …” (I10, Care Coordinator Supervisor); however, only one clinic staff indicated that referral sites were contacted, albeit uncommonly, to assess whether the patient utilized the referral. Most clinics relied on patients returning to the clinic to determine the status of a referral (e.g., “... I don’t know how you would even track ... besides contacting the patients …” (I12, Director of Behavioral Health)).

Documentation of referrals provided to patients was performed several ways. Some clinics utilized areas within the electronic health record (EHR) to document patient needs and referrals provided. Others have not incorporated HRSN-specific documentation within their EHR and relied on other methods (e.g., “... we've just tracked [referrals] on a piece of paper with sticky-notes connected to it.” (I05, Quality Manager)).

Post-Referral Data Collection

Respondents indicated numerous outcome measures related to their clinics’ HRSN-specific interventions. Data collection often stemmed from clinics needing “... to prove this is valuable to our patients and … has further impact on [patients’] health ...” (I03, Clinic Supervisor) in order to secure future funding or satisfy grant requirements.

Healthcare utilization measures, including emergency room visits and hospital readmissions, were the most commonly cited data collected by clinics. A limited number of clinics looked at social needs–related outcomes, such as the number enrolled to receive county benefits or housing rates. Respondents also noted that their clinics viewed HRSNs as drivers for poor clinical outcomes (e.g., “… we want to get people connected to follow-up care so they are not being readmitted.” (I04, Social Worker)), and were most interested in measuring improvements in patients’ clinical status.

When asked about collecting data or performing research on their work with patients, respondents described a host of significant challenges. For some, the power behind a patient’s story or resilience cannot be represented within the scope of quantitative outcome reporting:

... we are asked to compartmentalize these outcomes and report on them and figure out how to make it standardized; it is extremely irritating, especially [to] folks on the ground providing the care. They intrinsically know that there is no measure ... (I09, Program Manager)

For others, there are current gaps in what data is commonly measured, such as “... patient and provider satisfaction.” (Interview 03, Clinic Supervisor).

“Resource” and “Success” Definitions

Respondents classified a “resource” in one of three ways: pertinent information about a community-based good or service, provision of a referral to an outside agency, or a tangible good or service provided to a patient. Use of “resource” by respondents commonly included more than one of the abovementioned classifications throughout the interview.

Similarly, defining a successful patient interaction or intervention yielded varying responses. “Success” definitions included a patient returning to the clinic, providing information specific to the patient’s HRSNs, the patient reaching a referral organization, or a patient receiving a tangible good or service (e.g., “... it was successful because I gave them the resources ... for [them] it was not successful because [the patient] was not leaving with a place to stay.” (I10, Social Work Care Coordinator)). Examples of a tangible good or service included enrolling in county benefits and accessing childcare.

While each respondent provided a definition of “success,” definitions were often formed at the time of the interview (e.g., “(Laughter) That’s a wonderful question.” (I03, Clinic Supervisor)) and had not been established by the clinic itself.

Barriers in Accessing Supports

Respondents described minor variation in barriers to accessing supports after initial identification of a HRSN; such barriers were generally classified as patient barriers and clinic barriers. Patient barriers included health and social system literacy, access to reliable transportation, other social stressors (e.g., “... financial [concern] then ties to housing ... lots of mental health [concerns] which affects all of those pieces as well.” (I10, Social Work Care Coordinator)), and meeting eligibility for services. For example, an individual interested in being connected with organizations that address housing may be limited by their insurance status, county of residence, or legal status.

Respondents, including frontline and management staff, expressed certain barriers they face in addressing patients’ HRSNs. Limited patient follow-up was commonly cited as a barrier; respondents stated that patients may present to the clinic and have their HRSNs identified; however, this may be the only interaction “… because they may not come back. “ (I02, Social Services Supervisor).

Limited communication between clinics and community-based agencies was noted to be a barrier by frontline respondents. Without consistent interaction with referral sites, respondents indicated that they were unable to provide patients with reasonable expectations. Moreover, the fact that “... a lot of [agencies] are no longer in business or the non-profits are gone …” (I12, Medical Assistant)) was a significant barrier for staff when determining an appropriate referral site for a patient.

DISCUSSION

In Minnesota and across the nation, health systems are employing interventions, including referrals to outside agencies, to combat health-related social inequities. This study explored the range in ways Minneapolis-St. Paul primary care clinics address patients’ HRSNs after identification and initial contact with frontline staff. Our results, which include input from frontline and management staff across the spectrum of clinic designations, highlight the variability in approaches and define challenges in translating published interventions to real-world practice. These findings also address notable gaps in the Donabedian framework, including patient-centeredness, information technology, and viewing healthcare as a system.20

Variation among clinics’ approaches to addressing HRSNs was present in most processes of care, including titles and roles of staff, referral procedures, patient follow-up, and barriers to accessing support. Such variability in practices raises questions concerning the role of standardization of HRSN-specific care within medicine. Several frontline professionals, such as CHWs21,22 and public health nurses,23 attempted to define scopes and standards of practice. The goals of defining such scopes and standards are numerous and often include advocacy for a profession that is new or similar to other existing positions. However, our findings revealed potential problems with standardization of frontline staff who address patients’ social needs. The range in titles and responsibilities noted by respondents in this study are dependent on several health system– and clinic-level factors, such as the availability of in-house clinic services or support staff, funding mechanism, and needs of the patient population served. Moreover, we identified significant differences in titles of positions between clinics and responsibilities of positions with similar titles. While there are benefits of defining profession-specific scopes or standards of practice, the importance of tailoring HRSN interventions should be carefully balanced with any standardization measures.

The results of this study also highlight challenges with data collection and outcome reporting. Many respondents indicated that post-intervention or post-referral data collection was used to document value for both patients and clinics. The discrepancy between what is important to a patient versus a health system has been previously discussed24 and is compounded by the perspective that social work research is better suited for reporting experiential or values-centered findings than quantitative analyses.25,26,27 Descriptions of HRSNs interventions, most commonly those utilizing referrals to outside agencies, are abundant and suggest promising results.28,29,30,31,32 The inclusion of basic components of these interventions, however, is often missing. To more effectively analyze pertinent research and aid the translation to widespread practice, we encourage researchers to place greater emphasis on defining the processes used in future studies, such as the following: defining “resource” as it pertains to provision of information or referral, identifying all members of the clinical team involved in addressing HRSNs, and commenting on the documentation strategies used. In addition to more complete reporting, we further suggest primary care clinics identify the processes they employ to address the spectrum of patients’ needs; for example, “How do staff communicate with patients about HRSNs, what education is available to staff about the communities they serve, and is there any information exchange with outside agencies?” Answering these questions will identify strengths, discover opportunities, and lead to better care for vulnerable patients.

The referral loop refers to the process in which a referral to an outside agency is provided to address an identified need.33,34,35 After reaching the referral organization and receiving the appropriate service, the referral loop is considered closed when the organization that met the patient’s need communicates this information back to the referring organization. Numerous studies exist in the literature that describes this process.36,37,38 In these studies, commonly cited challenges with “closing” the referral loop include health system documentation of HRSNs,39 communication with community-based referral agencies,6 and capacity of outside agencies to handle referrals.40

Our findings suggest that many clinics appreciate the importance of closing the referral loop. However, most are unable to do so because of limited resources or limited data tracking beyond grant-required reporting. Moreover, absent communication streams with community-based agencies represent a significant barrier to determining the success of a referral and to providing the patient with reasonable expectations for the referral process. Communication strategies or collaboration with outside agencies are also rarely discussed in the literature. To our knowledge, no study evaluating referrals as a means of addressing HRSNs reports on other professionals in the community the patient may be working with — such as a county social worker or shelter-based care coordinator — although high cross-sector care utilization among patients has been identified and poses an intriguing basis for future investigation.41 Current studies, despite detailing “successful” interventions,28,29,30,31,32,42 do not accurately reflect real-world practice as we describe in this study and possess incomplete recognition of a patient’s engagement with additional community organizations. Therefore, broad adoption of the reported interventions in the literature remains limited.

This study was strengthened by the rigorous coding process involving multiple team members who independently coded and reviewed data. The strength of the sample was the wide range of primary care clinics represented. The primary limitation of the study was that the sample only represented clinics in the Twin Cities metropolitan area and may not capture the range of strategies used to address HRSNs in rural or other metropolitan areas. An additional limitation is that we only interviewed clinics that were able to identify someone who addresses HRSNs; we may not have captured strategies used by primary care providers themselves, independent of other clinic personnel. Finally, this paper only reports on post-identification interventions used by clinics and does not describe other aspects of identifying and meeting patients’ HRSNs.

There are certain barriers when performing HRSN-specific research; particularly in trying to capture the impact of “community” in interventions, determine effective implementation strategies, and in performing rigorous randomized control trials.43,44,45,46 Future research should aim to characterize relationships between health systems and social service agencies, describe the full range of social supports utilized by patients, and assess the impact of clinic-based HRSN interventions on patient health status using both quantitative and qualitative methods.