Validation of the Jefferson Health-Related Social Needs Screener

Sara Beachy^1,2*Kristin Rising^2,3,4Richard W Hass⁴Terry Hyslop⁵Bella Muti⁶Mackenzie Kemp²Rhea Powell⁷Cara Martino⁸Baligh Yehia^6,8Joe Cacchione⁸Patricia Henwood^3,4,8

• ¹Department of Psychiatry, Sidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, United States
• ²Center for Connected Care, Thomas Jefferson University, Philadelphia, PA, United States
• ³Department of Emergency Medicine, Sidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, PA, United States
• ⁴Thomas Jefferson University College of Population Health, Philadelphia, United States
• ⁵Thomas Jefferson University Department of Pharmacology Physiology & Cancer Biology, Philadelphia, United States
• ⁶Thomas Jefferson University Sidney Kimmel Medical College, Philadelphia, United States
• ⁷Department of Medicine, Sidney Kimmel Medical College, Thomas Jefferson University, Philadelphia, PA, United States
• ⁸Jefferson Health - Northeast, Philadelphia, United States

Introduction: Screening for health-related social needs (HRSN) is a growing national health priority. While multiple HRSN screening tools currently exist, none to our knowledge have been evaluated using robust statistical analyses. The goal of this work is to provide results from a validation study of the Jefferson HRSN screener conducted across inpatient and outpatient settings. Methods: This retrospective cross-sectional psychometric study included HRSN assessments conducted across inpatient and outpatient settings with adult patients from March 2023 to May 2024. The study was conducted across a 17-hospital academic health system serving a diverse community in a 9-county area crossing two states. Participants answered the HRSN screener, which includes eight questions across seven HRSN domains (financial, food, housing, utilities, transportation, violence/safety, and social connection) and two follow up questions, as part of standard healthcare encounter procedures. The measure was assessed with item response theory and a two-parameter logistic model. A follow-up analysis using Latent Class Analysis (LCA) was used to assess whether HRSN items and demographic variables could be used to identify people with higher levels of social vulnerability index (SVI). Higher SVI indicates higher levels of needs based on community and neighborhood related factors. Results: The final sample included data from 302,929 adults. Patients were relatively evenly distributed across ages (< 45 years, 32%; 45-64 years, 32%; 65-84 years, 30%; 85+, 4%). Most patients were Non-Hispanic (87%), White (66%), and female (59%). A third of patients were in the medium-high (18%) and high (15%) SVI areas. Positive responses across questions ranged from 0.90%-5.90%. Slopes ranged between 1.67-3.77, and difficulty parameters ranged between 2.20-3.31, indicating that the items can detect a high level of need. LCA results suggested that the eight HRSN items combined with basic demographic variables could help identify people with higher HRSN. Discussion: The Jefferson HRSN screener provides a valid approach for HRSN screening across healthcare settings. The eight screening questions, combined with additional questions to evaluate the patient's desire for help and urgency, can be used to identify patients needing additional resources to address fundamental social needs potentially contributing to health disparities.