Ships Passing in the Night: Electronic Screening of Social Determinants of Health and Provider-Patient Interactions

• In Event: AI, Novel Technologies, and Health Equity

Abstract

Past scholarship has emphasized that achieving equitable health outcomes requires integrating social determinants of health (SDOH) into routine medical care. Yet despite decades of significant investments into electronic health record (EHR) screening of patients for SDOH, its integration into clinical practice remains spotty: Screening rates are inconsistent, referrals often stall, and many patients’ social needs go unaddressed. Why has the translation of SDOH into clinical practice proven so difficult? Existing literature has identified the many organizational, professional, and relational challenges that hinder the effective integration of EHR systems to identify and address SDOH in clinical settings. Our study builds on this literature and examines whether EHR data collected through these processes (what we label ex-ante information) actually triggers in-situ conversations about SDOH during clinical encounters. We conducted an ethnographic study, observing 74 medical consultations in a primary care setting. Our analysis revealed a systematic decoupling between data in EHR and clinician-patient interactions: most conversations about SDOH occurred when needs were not flagged ex-ante, whereas the majority of patients with pre-existing SDOH flags in their EHR had no discussion about them during their visit. We then analyzed the circumstances under which such discussions occurred, identifying the timing of visits (particularly lagged ones) and select co-occurring medical conditions (feet issues) as predictors of these conversations. Overall, our analysis advances discussions on the complex relationship between clinicians, patients and information systems by showing that medical doctors rely on alternative cues to integrate social needs into their medical practice. More broadly, our findings speak to the challenges of integrating tacit knowledge into expert-systems in an era aiming for increased human and expert-system complementarity.

Authors

• Valerio Iannucci, Boston University

• Luisa Fernanda Delgado Mejia, Boston University

• Michel Anteby, Boston University Questrom School of Business