Abstract
Introduction/Objectives:
Stable housing is a well-recognized health-related social need (HRSN) with little attention in medical training. This study examined the efficacy of a resident housing curriculum on referrals to a Community Health Worker (CHW) intervention to assist patients with unmet housing needs.
Methods:
This was a retrospective cohort study conducted on Internal Medicine residents at a large urban residency program in Bronx, NY. We utilized multivariate mixed-effects logistic regression to determine whether clinician curriculum exposure increased CHW referrals among patients who self-reported housing needs in a HRSN screening tool between July 2021 and August 2024.
Results:
Nine hundred six unique patients screened positive for unmet housing needs, and 303 (33.4%) patients were referred to CHWs by eligible clinicians (n = 118). Clinician exposure to the curriculum was not a significant predictor of CHW referrals (aOR = 1.03, 95% CI = 0.69-1.54), adjusting for covariates. Patient age (aOR = 0.98, 95% CI = 0.97–0.99), Spanish language (aOR = 1.65, 95% CI = 1.09-2.51), and Medicaid coverage (aOR = 1.61, 95% CI = 1.03-2.51) were associated with likelihood of referral.
Conclusions:
Our findings demonstrate that the curriculum did not increase CHW referrals. Multimodal educational and systemic interventions that support the use of existing workflows may be needed to increase uptake of interventions to address HRSNs.
Keywords
Introduction
Housing insecurity is widely recognized as a major driver of healthcare access and outcomes. 1 Substandard housing conditions—such as overcrowding, exposure to environmental hazards, and poor sanitation—have been associated with exacerbations in chronic physical and psychiatric illnesses.2,3 Moreover, housing insecurity is linked to higher healthcare utilization, including higher rates of emergency department and hospital use compared to the general population. 4 In the United States (US), homeless individuals experience a 3.5 times higher mortality rate, underscoring the critical importance of stable housing.5,6 Stable housing is associated with reduced stress, improved chronic disease management, and better overall health, with reports showing that it can reduce the risk of chronic conditions like diabetes and hypertension by up to 21%.7,8 Recent WHO guidelines also emphasize the importance of adequate housing in preventing both communicable and non-communicable diseases, emphasizing its critical role in public health initiatives. 8
Many interventions have emerged to address homelessness and housing insecurity in the US that also demonstrate a health benefit, such as the provision of low-threshold affordable housing via the Housing First model, rental assistance programs, and housing navigation programs.9 -11 Furthermore, health systems are increasingly recognizing the need to address individually experienced unmet health-related social needs (HRSNs), such as unstable housing, through the implementation of screening and referrals interventions to identify patients within the clinical settings and link patients to essential social services.12 -14 Given the importance of housing to health and wellbeing, and the presence of available resources to address housing needs, health systems may have a particular role to play in linking patients to housing-related resources. For this purpose, growing evidence supports the use of patient navigators such as Community Health Workers (CHW), who often come from the communities they serve and enter the role through specialized training programs rather than formal licensure, and have specific expertise in linking patients to health-promoting resources.15,16
Despite the growth of HRSN screening and referrals interventions in clinical settings, there is inconsistent training on social determinants of health (SDOH) in medical education. Medical residency programs are increasingly acknowledging the importance of SDOH,17,18 but curricula remain fragmented, lack standardization, 19 and fail to provide targeted education on specific unmet HRSNs like housing. 20 Lack of satisfaction with current curriculum on health disparities was highlighted by a recent study with approximately 20% of residents reporting an unsatisfactory rating of “fair” or lower. 21 While some residency programs also provide valuable exposure to HRSNs through community partnerships or elective rotations, the absence of a core, structured curriculum precludes many trainees from developing the skills necessary to address housing-related and other health inequities.22,23
Integrating standardized training on specific HRSNs like housing demonstrates potential for improving both housing outcomes and overall health for vulnerable populations 24 ; however, previous research has shown that residency programs often fail to provide clinicians sufficient instruction on how to refer patients to CHWs for housing assistance. 25 There is a lack of training on how to integrate CHWs into clinical decision-making, which limits the potential for effective collaboration on patients’ HRSNs within the clinical team.26,27
Given the need to effectively address HRSNs in clinical settings, and the lack of effective education on this topic in medical residency, the aim of this study was to examine the efficacy of a didactic housing curriculum intervention on clinician referral to a HRSN CHW-led intervention. By understanding the impact of focused didactic interventions for connection to HRSN resources, we hope to elucidate pathways to empower future healthcare clinicians to address housing-related needs and promote health equity. 28
Methods
From September 2022 to August 2024, a multi-part longitudinal curriculum focused on the impact of housing and health was delivered to residents in all 3 years of training in the Internal Medicine residency program of a large urban academic medical center. Housing was chosen as a focal HRSN because of its clear impact on health, and the existence of a robust local service infrastructure to address housing. Curriculum components consisted of 60- to 90-min didactic lectures on the impact of housing on health, HRSN screening and referral workflow and use of EHR-supported CHW electronic referral order, as well case-base sessions discussing complex housing cases (see Table 1 for full description of curricular components). In order to identify the impact of different elements of the curriculum, and due to the ongoing nature of the curriculum with certain components not completed at the time of this study, this evaluation focuses on the first didactic component (“Introduction to Housing and Health” in Table 1), which highlights the impact of housing on health, the importance of addressing this social need as providers, and introduces the CHW intervention available to patients in the outpatient clinical setting.
Overview of Housing and Health Residency Curriculum Components.
In 2017, the health system piloted the Community Linkage to Care (CLC) program, which introduced standardized HRSN screening and CHW referral support. 12 The 10-item HRSN screener was adapted from the validated Health Leads screening toolkit, integrated into the health system’s electronic health record (EHR) in April 2018, and made available to all clinical practices. As part of the CLC program, CHWs were initially recruited, trained, and managed by a local community-based organization. The Community Health Worker Institute (CHWI) at the health system in question was deployed in June 2022 as an internal workforce to optimize the integration of CHWs into clinical care, improve patient access to healthcare, and provide social service navigation for patients with HRSNs. 15
Patients are screened for HRSNs during routine clinical encounters. If a patient screens positive for at least 1 HRSN, clinicians review the results of the screener with the patient and their family and ask whether they want help in addressing their HRSN(s). If the patient wants help, clinicians refer the patient to the CHW through an electronic referral order sent via the EHR. If a CHW is not available at the clinical practice, clinicians may directly refer the patient to social services using the EHR-supported social service directory. Clinicians are required to complete the electronic referral orders through the EHR, and CHWs are instructed to request orders if they receive an in-person warm handoff, or transfer of care.
The study was limited to patients (1) who self-reported a housing insecurity need, housing quality need, or both in a documented HRSN screener (2) between July 1, 2021 and August 31, 2024, (3) with an Internal Medicine clinician who received the curriculum intervention. Patients were excluded if their HRSN screener was not documented in an outpatient Internal Medicine practice.
The primary outcome was defined as whether a patient was referred to a CHW via electronic referral order by their clinician (0: not referred, 1: referred). The primary predictor was defined as whether the patient’s encounter/order was completed before or after clinician exposure to the housing curriculum intervention (0: before clinician exposed to curriculum, 1: after clinician exposed to curriculum). The predictor is based on the date of referral order for patients referred to a CHW and the date of screening encounter for patients not referred. This accounts for potential delays between documentation of the screener and electronic referral order; however, it is recommended that clinicians complete referrals on the date of screening. Socio-demographic and encounter covariates utilized in the analysis included resident year (Year 1, Year 2, and Year 3), age, sex (Male or Female), race/ethnicity (White, Spanish/Hispanic/Latino, Black/African American, Asian/Native Hawaiian/Other Pacific Islander, Other, or Declined to Report), language (English, Spanish, or Other), and primary insurance (Commercial, Medicaid, Medicare, or Other).
Data on HRSN screening and CHW electronic referral orders were extracted from the EHR using Microsoft SQL Server, version 18, to query data from the Epic EHR Data Warehouse. Additional data on linkage to CHWs were collected and managed using Research Electronic Data Capture (REDCap) tools. REDCap is a secure, web-based software platform designed to support data capture for research studies.29,30
We utilized a multivariate mixed-effects logistic regression model to estimate the association. This model included fixed effects for the primary predictor, EHR referrals to CHW, and covariates and random effects for patient and clinician identifiers. We presented estimates using adjusted odds ratios and 95% confidence limits using RStudio (RStudio: Integrated Development for R. R Studio, PBC, Boston, MA). This study was approved by the study team’s Institutional Review Board (2017-8434).
Results
Between July 2021 and August 2024, 906 unique patients self-reported at least 1 housing-related need in the HRSN screener during 953 total clinical encounters. Approximately 61.9% (n = 590) encounters occurred after clinician exposure to the housing curriculum. Of the unique patients with housing needs, 303 (33.4%) were referred to a CHW through the EHR, resulting in a total of 307 referrals. There were 118 clinicians exposed to the housing curriculum, of which 114 completed at least 1 screen and 79 completed at least 1 EHR referral. Of the 906 patients with housing-related needs, there were 122 patients referred to a CHW, as documented in the CHW REDCap database, without an EHR referral from an eligible clinician. These referrals were excluded from the analysis.
Patients who screened positive for housing-related needs in our sample were primarily female (57.0%), English-speaking (73.2), and indicated Medicaid (58.9) as their primary insurance. The majority identified as Spanish/Hispanic/Latino (51.7) or Black/African American (34.7). Socio-demographic characteristics of patients referred to a CHW via the EHR were comparable to those screened positive (Table 2).
Descriptive Characteristics a of Unique Patients Screened for Health-Related Social Needs (HRSNs) Who Self-reported Housing Quality or Security Needs and Were Referred to a Community Health Worker (CHW), July 2021 to August 2024.
Descriptive socio-demographic characteristics reported according to index/first screen per unique patient in study period.
In the mixed-effects logistic regression model (Model 1), clinician exposure to the housing curriculum was not associated with EHR-supported referral to a CHW (aOR = 1.03, 95% CI = 0.69-1.54, P = .89). Patient age, preferred language, and primary insurance were associated with likelihood of referral. Spanish-speaking patients (aOR = 1.65, 95% CI = 1.09-2.51, P = .019) and patients insured through Medicaid (aOR = 1.61, 95% CI = 1.03-2.51, P = .036) demonstrated higher odds of EHR referral to a CHW. Meanwhile, older patients had lower odds of EHR referral (aOR = 0.98, 95% CI = 0.97–0.99, P = .0065) with each additional year of age.
Adjusted Mixed Effects Regression Model of Patients with Housing Quality or Security Needs, July 2021 to August 2024 (n = 906).
Discussion
Our study evaluated a single didactic component of a longitudinal curriculum focused on the intersections of housing and health for Internal Medicine residents to assess changes in clinician referrals to integrated CHW programs. There was no evidence of difference in clinician referrals to CHWs, using EHR-supported tools, after receipt of the didactic lecture in this study. Patients demonstrated higher odds of being referred, however, if they were Spanish-speaking or insured through Medicaid and lower odds if they were older. Based on these findings, a didactic intervention alone may not be sufficient to increase engagement by medical trainees in a HRSN referral intervention.
Didactic interventions may be sufficient in recognizing the importance of addressing HRSNs; however, CHW-based referral interventions require trainees to work within a multidisciplinary team, understand systems-level resources, and set expectations with patients, which are complex skills that may require more intensive training. Further training on practice-specific workflows, engagement with multidisciplinary teams, and ongoing administrative support may be needed to increase medical trainee use of CHW referral interventions. As postgraduate medical education continues to prioritize education around SDOH and HRSNs, more attention should be paid to how education on these topics may differ from traditional clinical topics. Beyond medical training, this is consistent with qualitative evaluations of health system HRSN screening and referral programs, which identify clinician buy-in as a common obstacle to effective implementation. This study reinforces that outlining the significance of an issue alone is likely insufficient in encouraging clinician uptake of these programs. Residency remains an opportune time to develop these skills for new clinicians.
Furthermore, our study may underscore the importance of system-level changes in the clinical workflow of facilitating access to social resources. In other words, while a well-thought-out approach to medical education is necessary, systematic changes towards how clinical teams manage HRSNs may be equally, if not more, crucial to increase CHW referral rates. More care should be focused on systematic barriers (eg, integration of screening tools into clinical workflows or time constraints due to responsibility overflow of the providers) so that the impact of educational efforts is able to be fully actualized. Future exploration of the impact of infrastructure limitations on CHW referral interventions may be warranted.
Our study did find that patient characteristics, such as Spanish language preference and Medicaid as primary insurance, were positive predictors of referral to CHWs. These are consistent with prior observations from the health system, and possibly suggest that traditional barriers to care (ie, non-English speaking, public insurance) can be effectively addressed with appropriate investment and systemic supports targeted to these vulnerable populations. 31
There are several limitations that affect this study. Practice-specific barriers and facilitators affected the rates of referral to CHWs at each practice, including the presence of clinician champions, availability of CHWs on-site, gaps in full-capacity team of CHWs, lack of CHW appointment EHR scheduling, among others. Another significant limitation was that several patients (n = 122) were contacted by a CHW without a clinician EHR referral, which suggests that patients are being referred to CHWs outside the recommended workflow and may underestimate the effect of the final model. The CHW REDCap database does not collect information on the referring clinician or the mechanism of referral for these patients, but possible mechanisms include clinicians completing in-person warm handoff without EHR referral, patients self-referring to CHWs after learning about CHW services from ancillary staff (eg, front desk team, nursing team), other patients, or community members, or visible materials in the practice. If clinicians are referring patients outside the EHR, this may suggest that didactic interventions increase clinician awareness of unmet HRSNs and prompt further conversations with their patients, even if they are not adhering to the recommended workflow. Further exploration of provider interactions with these patients is needed. There were also limitations to the didactic intervention with the curriculum components delivered at different times in the year, which may have impacted the trainee’s pre-existing knowledge and use of the HRSN referral program.
Conclusions
Integrated CHW-led interventions demonstrate immense potential to address housing needs in the clinical setting but, clinician use of these interventions appears limited. Dedicated didactic interventions during medical training can support the effective use of CHW interventions to address unmet HRSNs, however, didactic interventions need to shift from focusing solely on increasing awareness of HRSNs and provide practice-specific directed instruction on the workflows and practical utilization of referral interventions. Understanding how patients access CHWs both through and outside established workflows can better optimize HRSN screening and referral interventions to increase clinician engagement and patient access. Further investigation may involve textual analysis of clinician communication about referrals in clinician notes or qualitative interviews with clinicians, patients, and CHWs to clarify how patients are being connected with CHWs.
Footnotes
Acknowledgements
The authors would like to acknowledge the program staff at the Community Health Worker Institute (CHWI) at Montefiore Einstein for always going above and beyond to connect our patients to essential social services.
Ethical Considerations
All research was approved by the Albert Einstein College of Medicine Institutional Review Board (2017-8421; 2017-8434).
Consent to Participate
The Albert Einstein College of Medicine Institutional Review Board granted our study a waiver of informed consent as this was a retrospective cohort study including data routinely collected by the health system.
Consent for Publication
Not applicable.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project described was supported by the National Center for Advancing Translational Sciences (NCATS), National Institutes of Health, through CTSA award number (UM1TR004400). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The project described was also supposed by the American Medical Association (AMA) Reimagining Residency Grant.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.