Barriers to Trust in the Utilization of Health-Related Social Needs Data from the Electronic Health Record

Abstract

Addressing health-related social needs (HRSN)—such as food insecurity, housing instability, transportation barriers, and financial strain—is increasingly recognized as essential to achieving equitable health outcomes. These factors play a critical role in shaping patients’ health risks, care access, and health outcomes. Consequently, understanding and addressing HRSN is vital to both clinical decision-making and population health management. The electronic health record (EHR) holds promise as a tool for the systematic collection and use of HRSN data. However, current practices are marked by inconsistencies and limited adoption of structured coding systems. Highquality, standardized data are critical for the meaningful application of HRSN information, yet multiple barriers hinder collection. These challenges are not solely technical. Patients may be reluctant to disclose sensitive social information, and clinicians may feel discomfort or uncertainty about asking, recording, or acting on such data. These attitudinal and structural obstacles introduce the potential for bias in EHRderived data, which may compromise the fairness and effectiveness of downstream applications. If unaddressed, these limitations may undermine trust in the utility and accuracy of EHR-based social factors data. This dissertation investigates the structural and attitudinal factors that influence trust in EHR-derived HRSN data, with a focus on data quality, documentation practices, and clinician perspectives. First, it evaluates the robustness of EHR-based HRSN data by comparing prevalence estimates derived from structured fields with external community benchmarks. Discrepancies highlight areas where under-documentation or misrepresentation may occur. Second, the study examines demographic and system-level characteristics associated with whether patients’ social needs are captured in the EHR, identifying patterns that may reflect disparities in documentation. Third, semi-structured interviews with clinicians provide qualitative insight into their experiences documenting HRSN, their trust in the data’s accuracy, and their perceptions of its role in clinical care. Together, these mixed-methods investigations offer a comprehensive understanding of the barriers and facilitators to trustworthy and equitable use of HRSN data within EHRs. Findings will inform future strategies to strengthen data quality, enhance clinician engagement, and ensure that EHR-derived social factors data can be reliably leveraged to support health equity and improved patient outcomes.