• Nazleen Bharmal, Alex Rennick, Amy Shideler, Madeleine Blazel, Robert Jones, Chi' Wilson, and Elizabeth R Pfoh.
• Community Health & Partnerships, Cleveland Clinic Community Care, Community Health, Cleveland, OH, USA. bharman2@ccf.org.
• J Gen Intern Med. 2023 Sep 1; 38 (12): 269527022695-2702.

BackgroundHealth systems are screening patients for health-related social needs (HRSN) but the optimal approach is unknown.ObjectiveTo describe the variation in responding to an HRSN questionnaire delivered via patient portal, and whether referral to and resources provided by social workers differed by response status.DesignRetrospective observational study.ParticipantsPrimary care patients with a visit between June 2020 and January 2022.InterventionHRSN questionnaire MAIN MEASURES: We identified each patient's index visit (e.g., date of their first questionnaire response for responders or their first visit within the study period for non-responders). Through the EHR, we identified patients' demographic characteristics. We linked the area deprivation index (ADI) to each patient and grouped patients into quintiles. We used multilevel logistic regressions to identify characteristics associated with responding to the questionnaire and, for responders, reporting a need. We also determined if responder status was associated with receiving a social worker referral or receiving a resource. We included patient demographics and ADI quintile as fixed variables and practice site as a random variable.Key ResultsOur study included 386,997 patients, of which 51% completed at least one HRSN questionnaire question. Patients with Medicaid insurance (AOR: 0.62, 95%CI: 0.61, 0.64) and those who lived in higher ADI neighborhoods had lower adjusted odds of responding (AOR: 0.76, 95% CI: 0.75, 0.78 comparing quintile 5 to quintile 1). Of responders, having Medicaid insurance (versus private) increased the adjusted odds of reporting each of the HRSN needs by two- to eightfold (p < 0.01). Patients who completed a questionnaire (versus non-responders) had similar adjusted odds of receiving a referral (AOR: 0.91, 95% CI: 0.80, 1.02) and receiving a resource from a SW (AOR: AOR: 1.18, 95%CI: 0.79, 1.77).ConclusionHRSN questionnaire responses may not accurately represent the needs of patients, especially when delivered solely via patient portal.© 2023. The Author(s), under exclusive licence to Society of General Internal Medicine.

13 keywords...

hide…