Data collection and synthesis

We identified a total of 685 articles through the PubMed query. After reviewing the titles and abstracts screening, 415 articles were included. Of these 415 articles, 324 articles included full text for qualitative synthesis. The reviewed articles were then classified according to SDoH in the EHR domains. The majority of articles focused on SDoH and health outcomes, SDoH data collection and documentation followed by NLP for SDoH, and SDoH screening tools and assessments. In the following sections, we reviewed the major themes and highlighted works for each of the five SDoH in the EHR domains (see Fig. 3, percentage see Supplement Table 3 in Supplementary Material 1).

Figure 3. PRISMA 2020 flow diagram. EHR = electronic health records; NLP = natural language processing.

SDoH screening tools

We included 29 papers (details see Supplement Material 2 – Meta Data) incorporating SDoH Screening tools into EHRs in our review. The majority of the studies utilized homegrown tools for screening SDoH, reflecting the need for tailored approaches and the limitations of existing standardized tools in certain contexts Some studies [Reference Gold, Bunce and Cowburn27,Reference Graif, Meurer and Fontana33–Reference Barton, Parke and White35] developed their own questionnaires and screening sets, reflecting a trend toward customized tools tailored to specific healthcare settings or populations. Vendor-specific tools, like the two-item screening tool [Reference Gore, DiTursi, Rambuss, Pope-Collins and Train36] integrated into Epic SDoH Wheel, were less common but still present. The screened determinants varied, but common factors included housing, food insecurity, transportation, and mental health indicators like stress and depression.

Studies targeted a diverse range of populations. For example, children were the focus in some studies [Reference Graif, Meurer and Fontana33,Reference Vasan, Kenyon and Palakshappa37], while adults were the primary subjects in other studies [Reference Bechtel, Jones, Kue and Ford34,Reference Berkowitz, Bui and Shen38]. Various healthcare settings were represented, from primary care clinics [Reference Meyer, Lerner, Phillips and Zumwalt39,Reference Nohria, Xiao and Guardado40] to emergency departments [Reference Wallace, Luther, Guo, Wang, Sisler and Wong41], as well as school-based clinics [Reference Barton, Parke and White35]. This diversity indicates the widespread recognition of SDoH’s importance across different medical environments, underscoring its growing relevance throughout the healthcare spectrum.

Active screening methods, where healthcare providers proactively administered questionnaires or interviews, were predominant (n = 28) [Reference Gold, Bunce and Cowburn27,Reference Bechtel, Jones, Kue and Ford34]. Passive methods like the analysis of EHR data [Reference Hatef, Rouhizadeh and Tia42] were less common. However, the utilization of EHR data for passive screening indicates a potential to streamline the process in the future. While many studies focused on personal health determinants (n = 19), others also assessed structural determinants like housing quality and social networks [Reference Graif, Meurer and Fontana33,Reference Nohria, Xiao and Guardado40,Reference Chukmaitov, Dahman and Garland43]. A limited number of studies (n = 8) investigated both personal and structural determinants.

The heatmap (Fig. 4) shows EHRs as the dominant data source, with surveys, interviews, and ICD codes used for specific elements. Free text fields were underutilized, suggesting an opportunity for leveraging unstructured data. Housing was studied through the most diverse methods. The heatmap highlights the importance of EHRs and the need for diverse, integrated approaches.

Figure 4. Heatmap showing the frequency of association between the top 10 social determinants of health (SDoH) elements and the top five data collection methods.

Challenges and opportunities

The reviewed studies collectively highlight the challenges in standardizing SDoH screening across various contexts but also point toward the potential benefits of such screenings to patient care. The diversity in approaches reflects the complexity of addressing SDoH in clinical practice. However, it also demonstrates a concerted effort toward more comprehensive patient care. The prevalence of homegrown tools [Reference Graif, Meurer and Fontana33,Reference Barton, Parke and White35,Reference Nohria, Xiao and Guardado40,Reference Yazar, Kang, Ha, Dawson and Dawson44], indicates a trend toward customization, tailored to specific patient populations and healthcare settings. This is likely due to the unique needs and circumstances of different patient demographics. The variability in tools and approaches (e.g., the number of questions in studies [Reference Gold, Bunce and Cowburn27,Reference Okafor, Chiu and Feinn45], and the use of paper-based vs. EHR-based tools [Reference Bechtel, Jones, Kue and Ford34,Reference Meyer, Lerner, Phillips and Zumwalt39] highlight the challenges in standardizing SDoH screening. This variability could impact the comparability of data and the scalability of successful approaches. Despite the challenges, the focus on SDoH screening illustrates a shift toward more personalized patient care. Recognizing and addressing social and behavioral factors [Reference Barton, Parke and White35,Reference Herrera, Fiori, Archer-Dyer, Lounsbury and Wylie-Rrosett46] can lead to more effective healthcare interventions and better health outcomes.

SDoH data collection and documentation

In our review, we identified 76 articles (see Supplement Material 2 – Meta Data) describing SDoH data collection and documentation practices. These studies focused on engagement with populations and leveraged a variety of technologies to support collection and documentation processes.

The reviewed studies demonstrated that various technologies were employed to support the collection and documentation of SDoH data. Screening tools and questionnaires were also commonly used, often integrated directly into the EHR system. For example, Boston Medical Center’s THRIVE tool [Reference de la Vega, Losi and Sprague Martinez47] utilized a paper screening form that was entered into the EHR by medical assistants, while the University of California, San Diego [Reference Lee, Saseendrakumar and Nayak48] used a local EHR database (Epic) to collect SDoH data through ICD codes and discrete data fields. The OCHIN network of community health centers [Reference Gruß, Bunce, Davis, Dambrun, Cottrell and Gold49] developed an EHR-based screening questionnaire to assess various SDoH domains, and Wake Forest School of Medicine [Reference Palakshappa, Benefield and Furgurson50] employed a tablet-based digital health system to screen for food security, housing, and transportation needs.

In addition to EHRs and screening tools, some studies leveraged web-based platforms and applications for SDoH data collection. For instance, the COMPASS-CP study[Reference Duncan, Abbott and Rushing51] used a web-based application and iPad application to capture patient-reported outcomes related to physical, mental, and social well-being, as well as financial challenges and caregiver needs.

NLP techniques were also employed to extract SDoH information from unstructured EHR data. The OSF HealthCare System[Reference Stewart de Ramirez, Shallat, McClure, Foulger and Barenblat52] utilized the Pieces NLP system to identify SDoH factors from EHR notes, demonstrating the potential of NLP in automating the extraction of SDoH data from free-text clinical documentation.

Furthermore, studies used geocoding techniques to link patient addresses with external data sources, such as census tract socioeconomic data and community-level characteristics. The Envirome Web Service [Reference Kane, Wang and Gerkovich53], developed by Children’s Mercy Hospital, integrated census tract data with patient EHRs using real-time geocoding, enabling a more comprehensive understanding of patients’ social and environmental contexts.

Our review identified diverse methods for SDoH data collection and integration. Nine studies incorporated qualitative approaches, including interviews with patients and clinicians, community engagement initiatives, focus groups, and town hall meetings [Reference Potharaju, Fields and Cemballi54–Reference Duberstein, Brunner and Panisch62]. These methods provided rich, contextual information about SDoH factors and their impacts on health outcomes.

Technology-assisted collection methods were also prevalent, with seven studies utilizing various tools for SDoH data gathering. These included paper-based entry, iPads/tablets, patient or clinician-facing web portals, and other web-based toolkits and forms [Reference Palakshappa, Benefield and Furgurson50,Reference Duncan, Abbott and Rushing51,Reference Gold, Bunce and Cottrell63–Reference Arevian, Springgate and Jones67].

Several studies (n = 6) made use of publicly available, external data resources to infer structural SDoH information for a given population. The most common external SDoH data sources linked to EHRs were US census and community survey data (at both patient/individual and area/neighborhood levels), administrative data/claims records, and disease registries. Commonly linked community surveys and systems include the Behavioral Risk Factor Surveillance System, the National Health and Nutrition Examination Survey, the National Health Interview Survey [Reference Melton, Manaktala, Sarkar and Chen68], the National Institutes of Health PROMIS® (Patient-Reported Outcomes Measurement Information System) [Reference Aghdaee, Gu, Sinha, Parkinson, Sharma and Cutler69], the National Survey of Children’s Health [70], the Center for Disease Control Youth Risk Behavior Surveillance System [Reference Ettinger, Landsittel and Abebe71], the Center for Medicare and Medicaid Services (CMS) Accountable Health Communities’ Health-Related Social Needs Screening Tool [Reference Trinacty, LaWall, Ashton, Taira, Seto and Sentell72], the National Center for Education Statistics, the Uniform Crime Reports, and the American Community Survey [Reference Cottrell, Hendricks and Dambrun73–Reference Udalova, Carey and Chelminski76]. Longitudinal study data included the National Longitudinal Study of Adolescent to Adult Health (Add Health) [Reference Sacco, Chen, Wang and Aseltine77]. Other administrative data sources included the Healthcare Cost & Utilization Project (HCUP) Nationwide Readmissions Database [Reference Gottlieb, Tirozzi, Manchanda, Burns and Sandel78], claims data [Reference McCormack and Madlock-Brown79], and Medicaid data warehouse [Reference Hewner, Casucci and Sullivan80]. Few studies describe use of disease-specific registries e.g. cancer registries such as SEER-CMS, SEER-Medicare, and SEER-Medicaid [Reference Dusetzina PhD, Enewold Mph PhD, Gentile PhD, Ramsey Md PhD and Halpern81]. While these sources may not directly capture SDoH information, they can provide proxy measures related to healthcare utilization patterns, access to care, and socioeconomic status. However, the use of administrative and claims data for SDoH analysis has limitations, as they may lack the granularity and specificity of data collected directly from patients or through dedicated SDoH screening tools.

Eight studies describe methods for inferring structural SDoH using geocoding of patient addresses and linking to public census tract data [Reference Kane, Wang and Gerkovich53,Reference Van Brunt66,Reference Biro, Williamson and Leggett82,Reference Comer, Grannis, Dixon, Bodenhamer and Wiehe83] These studies employed various geocoding techniques to convert patient addresses into geographic coordinates, which could then be mapped to specific census tracts or other geographic units. For example, the Envirome Web Service, developed by Children’s Mercy Hospital, used real-time geocoding to link patient addresses with census tract data [Reference Kane, Wang and Gerkovich53]. This geocoding approach enabled the integration of information related to neighborhood and community-level characteristics (e.g., SES, crime incidence, and health facility locations) [Reference Gardner, Pedersen, Campbell and McClay74,Reference Hughes, Phillips, DeVoe and Bazemore85,Reference Bazemore, Cottrell and Gold86] and neighborhood factors (e.g., poverty level, education, employment status, etc.) [Reference de la Vega, Losi and Sprague Martinez47,Reference Lee, Saseendrakumar and Nayak48,Reference Kane, Wang and Gerkovich53,Reference Melton, Manaktala, Sarkar and Chen68,Reference Luijks, van Boven, olde Hartman, Uijen, van Weel and Schers87–Reference Rousseau, Oliveira, Tierney and Khurshid89]. By linking patient locations to area-level SDoH data, these studies were able to provide a more comprehensive understanding of patients’ social and environmental contexts, even when individual-level SDoH data were not available in the EHR.

External data provided various socioeconomic factors (income, education, employment, poverty level, air quality), neighborhood variables (segregation, safety, and walkability), and health behaviors (diet, exercise, and smoking) [Reference de la Vega, Losi and Sprague Martinez47–Reference Gruß, Bunce, Davis, Dambrun, Cottrell and Gold49,Reference Stewart de Ramirez, Shallat, McClure, Foulger and Barenblat52,Reference Potharaju, Fields and Cemballi54,Reference McCormack and Madlock-Brown79,Reference Wang, Pantell, Gottlieb and Adler-Milstein90–Reference Freij, Dullabh, Lewis, Smith, Hovey and Dhopeshwarkar97]. These complemented and expanded the individual-level SDoH data (food/housing security, transportation, interpersonal violence, etc.) captured directly in EHRs [Reference Tenenbaum, Christian and Cornish98–Reference Ofili, Schanberg and Hutchinson100]. A small subset of studies (n = 3) aimed to integrate EHR, genomic, and public health data to examine the intersection of lifestyle, genetics, and environmental influences [Reference Lee, Saseendrakumar and Nayak48,Reference Kyriazis, Autexier and Brondino101,Reference Khatib, Li, Glowacki and Siddiqi102].

Challenges and opportunities

Although these works highlight the potential for study of personal and structural SDoH, there is considerable effort for systematically collecting, linking, and analyzing SDoH data from external sources together with EHR data at the community, state, and national levels [Reference Smith, Gigot and Harburn103,Reference Cook, Espinoza and Weiskopf104]. The adoption of common data models to improve standardization and interoperability of collected SDoH data remains limited [Reference Lindemann, Chen, Wang, Skube and Melton105]. Moreover, there is a scarcity of research demonstrating how this integrated information could be leveraged to connect individuals with identified SDoH risk factors to appropriate social programs.

NLP in SDoH

In our review, we identified 36 articles (details see Supplement Material 2 – Meta Data). describing NLP methods for powering SDoH studies. Many SDoH elements are captured in clinical free-text notes, such as progress notes, discharge summaries, and social work assessments. These unstructured data often contain rich, contextual information about patients’ social circumstances that may not be fully captured in structured fields. For example, free text might include detailed descriptions of a patient’s living situation, family dynamics, or barriers to accessing care. In contrast, structured data elements typically consist of predefined fields or checkboxes in the EHR, such as standardized screening questionnaires or ICD-10 Z-codes for social factors [Reference Dorr, Quiñones, King, Wei, White and Bejan106–Reference Dorr, Bejan, Pizzimenti, Singh, Storer and Quinones108]. The use of NLP techniques is crucial for extracting and analyzing SDoH information from these unstructured sources, as it allows researchers to access a wealth of data that might otherwise remain untapped. NLP can identify mentions of social factors, assess their relevance, and even determine the severity or impact of these factors on the patient’s health.

Several studies focused on lexicon development using methods such as lexical associations, word embeddings, term similarity, and query expansion. Lexicons and regular expressions have been demonstrated to extract SDoH and psychosocial risk factors [Reference Navathe, Zhong and Lei18,Reference Conway, Keyhani and Christensen109–Reference Morrow, Zamora-Resendiz and Beckham112], learn distinct social risk factors by mapping them to standard vocabularies and code sets including ICD-9/10, ICD Z codes, Unified Medical Language System, and SNOMED-CT. Most articles (n = 15) describe rule-based approaches using regular expressions and/or hybrid machine learning methods leveraging platforms. Five articles highlighted well-known rule-based toolkits and platforms adapted with lexicons and regular expressions for SDoH extraction including Moonstone, Easy Clinical Information Extraction System, Medical Text Extraction, Reasoning and Mapping System, Queriable Patient Inference Dossier, and Clinical Event Recognizer [Reference Navathe, Zhong and Lei18,Reference Conway, Keyhani and Christensen109–Reference Morrow, Zamora-Resendiz and Beckham112]. Other articles (n = 7) describe rule-based systems paired with traditional machine learning approaches i.e., an ensemble, particularly using NLP systems such as General Architecture for Text Engineering, Clinical Language Annotation, Modeling, and Processing Toolkit, Extract SDOH from EHRs, Yale clinical Text Analysis and Knowledge Extraction System, Relative Housing Stability in Electronic Documentation, and toolkits such as spaCy and medspaCy in conjunction with conditional random fields and support vector machines (SVM) [Reference Chilman, Song and Roberts113–Reference Wang, Long and McGinnis116]. In contrast, several investigators have leveraged open-source NLP toolkits like spaCy and medspaCy without supervised learners to extract SDoH variables [Reference Hatef, Rouhizadeh and Nau117–Reference Chapman, Jones and Kelley119]. Other studies (n = 19) have solely leveraged traditional supervised and unsupervised learning techniques, SVM, logistic regression (LR), Naïve Bayes, Adaboost, Random Forest, XGBoost, Bio-ClinicalBERT, Latent Dirichlet Allocation, and bidirectional long short-term memory [Reference Vaswani, Shazeer and Parmar125] to extract and standardize social and behavioral determinants of health (SBDoH), e.g., alcohol abuse, drug use, sexual orientation, homelessness, substance use, sexual history, HIV status, drug use, housing status, transportation needs, housing insecurity, food insecurity, financial insecurity, employment/income insecurity, insurance insecurity, and poor social support. In more recent years, nine studies have focused on the training and tuning of deep learning approaches, primarily transformer-based [Reference Yu, Yang, Guo, Bian and Wu126–Reference Han, Zhang and Shi135] approaches i.e., Bidirectional Encoder Representations from Transformers (BERT), RoBERTa, BioClinical-BERT models for extracting SBDoHs including relationship status, social status, family history, employment status, race/ethnicity, gender, social history, sexual orientation, diet, alcohol, smoking housing insecurity, unemployment, social isolation, and illicit drug use—from clinical notes, PubMed, among other specialized texts, e.g., LitCOVID [Reference Yu, Yang, Guo, Bian and Wu126–Reference Han, Zhang and Shi135]. The frequency of papers for SDoH extraction NLP algorithms within EHR systems, highlighting the combinations and intersections of utilized methodologies can be found in Fig. 5.

Figure 5. UpSet Plot of natural language processing for social determinants of health (SDoH) algorithms ∼electronic health record (EHR) integration. *Outlining the distribution of papers in which each approach/method individually and in combination was described in the study. *Supervised machine learning includes traditional machine learning methods (naive bayes, support vector machine, logistic regression, random forest, etc), excluding neural networks and pretrained approaches.

Challenges and opportunities

Although these works highlight the potential for extracting SDoH from texts, several challenges remain. Few studies focused on lexicon development, make use of standard terminologies for encoding SDoH data, and explore deep extraction and representation of SDoH attributes and relationships. Also, many studies focus on extraction and encoding SDoH data from a single site and fail to assess the portability of methods to new textual data sources beyond clinical notes and PubMed articles such as digital technologies and chatbots. The introduction of shared datasets like the Social History Annotated Corpus is an important step toward demonstrating generalizability of NLP-powered, SDoH extraction systems. Emerging generative models may also improve upon the state-of-the-art demonstrated by common shared task datasets.

SDoH and health outcomes

In our review, we identified 164 articles (details see Supplement Material 2 – Meta Data). describing SDoH and health outcomes. SDoH and health outcome studies examined a wide variety of health-related events and outcomes in relation to SDoH factors. A predominant focus was on infectious disease outcomes, with 16 studies examining drivers of COVID-19 hospitalization, mortality, treatment disparities, and differences in positivity rates across social groups [Reference Holbert, Andersen, Stone, Pipkin, Turcotte and Patton140–Reference Sills, Hall and Cutler149]. Another major category included healthcare utilization metrics like preventable hospital readmissions (n = 11) [Reference Holbert, Andersen, Stone, Pipkin, Turcotte and Patton140–Reference Sills, Hall and Cutler149], ED reliance ( n = 16) [Reference Luo, Tong and Crotty173], and telehealth adoption [Reference Schwartz, Kolak and Pollak174–Reference Ares-Blanco, Polentinos-Castro, Rodríguez-Cabrera, Gullón, Franco and Del Cura-González184]. Beyond infectious outcomes and healthcare utilization, studies also assessed chronic disease control across conditions like diabetes (n = 11) [Reference Schwartz, Kolak and Pollak174–Reference Ares-Blanco, Polentinos-Castro, Rodríguez-Cabrera, Gullón, Franco and Del Cura-González184], hypertension [Reference DuBay, Su and Morinelli187,Reference Ghazi, Oakes, MacLehose, Luepker, Osypuk and Drawz188], kidney disease [Reference Tomayko, Flood, Tandias and Hanrahan150,Reference Cereijo, Gullón and Del Cura179,Reference Lange, Kompaniyets and Freedman189–Reference Boone-Heinonen, Tillotson and O’Malley193], and obesity (n = 7) [Reference Hong, Nguyen and Lee171,Reference Nemirovsky, Klose and Wynne194–Reference Sathe, Accordino, DeStephano, Shah, Wright and Hershman203], along with risk factors like elevated blood pressure and cardiovascular events. Some studies focused on cancer (n = 11) screening, diagnoses, treatment disparities, and survival outcomes [Reference Tomayko, Flood, Tandias and Hanrahan150–Reference Avalos, Nance and Zhu155], while others addressed mental health (n = 6) indicators [Reference Hayes-Larson, Ikesu and Fong204] ranging from dementia incidence [Reference Alemi, Avramovic, Renshaw, Kanchi and Schwartz205,Reference Blosnich, Montgomery and Dichter206] to suicide (n = 2) risk factors [Reference Liu, Hung and Liang207,Reference Freeman, Stanhope, Wichmann, Jamieson and Boulet208]. Additional outcomes evaluated included maternal morbidity (n = 2) [Reference Lê-Scherban, Ballester and Castro177,Reference Cottrell, O’Malley and Dambrun209] and pediatric health metrics, ranging from vaccine completion rates to epilepsy-related consequences.

The most common quality measures reported were standardized condition control thresholds like HbA1c levels for diabetes control [Reference Chakravarthy, Stallings and Velez Edwards195], blood pressure (n = 5) levels for hypertension control, and established cancer staging guidelines [Reference Alemi, Avramovic, Renshaw, Kanchi and Schwartz205]. Some studies used validated risk prediction models for outcomes like hospital readmissions (n = 5), suicide risk [Reference Kind and Buckingham210,Reference Ryan Powell, Sheehy and Kind211], or mortality (n = 6). Beyond clinical indicators, several studies incorporated validated SDoH indexes like the Area Deprivation Index [212], Social Deprivation Index [213], and CDC’s Social Vulnerability Index [Reference Faison, Moon and Buckman163,Reference Nieuwenhuijse, Struijs, Sutch, Numans and Vos183,Reference Ares-Blanco, Polentinos-Castro, Rodríguez-Cabrera, Gullón, Franco and Del Cura-González184,Reference Otero, Scott, Martin and Huesing196,Reference Sathe, Accordino, DeStephano, Shah, Wright and Hershman203,Reference Gordon, Banegas and Tucker-Seeley214–Reference Kapoor, Lynch and Lacson221]. In terms of analysis approaches, common methods included multivariate regression models like LR (n = 13) [Reference Johnson, Zhu and Garrard142,Reference Adie, Kats and Tlimat200,Reference Ahmad, Kong and Turner216] and Cox proportional hazards models (n = 3) [Reference Thompson, Sharma and Smith219] to assess adjusted outcome associations with SDoH factors. Other advanced techniques included machine learning algorithms [Reference Cobert, Lantos and Janko222], geospatial analysis for clustering [Reference Schlauch, Read, Koning, Neveux and Grzymski154], and phenome-wide association studies [Reference Egede, Walker and Linde223,Reference Eakin, Singleterry, Wang, Brown, Saynina and Walker224].

SDoH interventions

A total of 19 papers (details see Supplement Material 2 – Meta Data) collected supplementary SDoH data to support population health intervention initiatives targeting hospitals/clinics (n = 10) or communities (including primary care, n = 9) at the meso (institution) level. Two articles discussed policy potential and proposed policy reform at the macro (system) level [Reference Sokan, Stryckman and Liang225]. The majority of the selected research (n = 16) focused on implementing a social and healthcare-supportive program to address the social needs of the target population. Interventions were implemented in various settings for hospital-based initiatives, including posthospital discharge [Reference Tung, Abramsohn and Boyd226], the emergency department [Reference McCrae, Robinson, Spain, Byers and Axelrod227–Reference Embick, Maeng and Juskiewicz231], and clinics specializing in different medical disciplines [Reference Sandel, Hansen and Kahn232–Reference Berkowitz, Hulberg and Placzek235]. On the other hand, community-based initiatives concentrated mainly on integrating interventions into primary care services [Reference Henize, Beck, Klein, Adams and Kahn236,Reference Gyamfi, Cooper and Barber237].

The social and healthcare supportive programs included a range of initiatives designed toward improving community health. These initiatives encompassed the introduction of new healthcare programs [Reference Liu, Cruz, Gamcsik and Harris229,Reference Henize, Beck, Klein, Adams and Kahn236,Reference Emmert-Aronson, Grill, Trivedi, Markle and Chen238], health education and coaching [Reference Sandel, Hansen and Kahn232], the strengthening of medical-legal partnerships [Reference Krist, O’Loughlin and Woolf233,Reference Berkowitz, Hulberg and Placzek235,Reference Taber, Williams and Huang239], the enhancement of integrated care planning [Reference Tung, Abramsohn and Boyd226,Reference Battaglia, Freund and Haas240,Reference Loo, Anderson and Lin241] and the improvement of patient navigation [Reference McCrae, Robinson, Spain, Byers and Axelrod227,Reference Henize, Beck, Klein, Adams and Kahn236,Reference Emmert-Aronson, Grill, Trivedi, Markle and Chen238]. Only a few papers (n = 3) have examined the potential of incorporating the family or social support element into their intervention design [Reference Sokan, Stryckman and Liang225]. Meanwhile, four studies investigated the potential for enhancing resource allocation through surveying outcomes. The improvement objectives encompassed the allocation of staff and equipment [Reference Messmer, Brochier, Joseph, Tripodis and Garg228], the enhancement of patient navigation [Reference Mathias, Mahali and Agarwal230,Reference Zuckerman, Mourani and Smith242], and the transformation of health service practices [Reference Gyamfi, Cooper and Barber237].

Our review identified a range of target populations receiving SDoH interventions. Several interventions focused on specific demographic groups, including racial/ethnic minorities (e.g., Blacks for hypertension control) [Reference McCrae, Robinson, Spain, Byers and Axelrod227,Reference Messmer, Brochier, Joseph, Tripodis and Garg228], specific age groups (both pediatric and adult populations) [Reference Embick, Maeng and Juskiewicz231], and women’s health [Reference Sokan, Stryckman and Liang225]. Health condition-specific interventions were also common, targeting chronic diseases such as heart failure [Reference Gyamfi, Cooper and Barber237], COPD [Reference Krist, O’Loughlin and Woolf233], diabetes [Reference Mathias, Mahali and Agarwal230], and hypertension [Reference Embick, Maeng and Juskiewicz231,Reference Berkowitz, Hulberg and Placzek235], as well as mental health conditions [Reference Taber, Williams and Huang239] and specific diseases like lupus [Reference Hsu, Hertel and Johnson234,Reference Taber, Williams and Huang239,Reference Loo, Anderson and Lin241,Reference Zuckerman, Mourani and Smith242]. Many studies focused on vulnerable or underserved populations.

Health outcomes, such as improvements in health metrics, reductions in disease incidence, changes in vital signs, and quality of life, are commonly used as measures to determine the feasibility of initiating an intervention (n = 10). Several studies have also assessed social SDoH in relation to patient satisfaction and acceptability [Reference Alcaraz, Wiedt, Daniels, Yabroff, Guerra and Wender243–Reference Dagenais, Russo, Madsen, Webster and Becnel245]. Since most programs were new efforts, it was not possible to determine the effectiveness of the intervention in the short term, nor could the potential generalizability be assessed.