Predictive modeling for social needs in emergency department settings

Unmet social needs create immediate risks to health, increase utilization, wait times and costs, and contribute to provider burnout. Due to the high prevalence of unmet social needs such as housing or income insecurity among patients, the emergency department (ED) is an opportune setting for intervention. Problematically, social

needs frequently go unscreened and unaddressed in EDs. Basic workflow issues and time constraints inhibit screening. Patients may decline screening or avoid questions they deem stigmatizing. While numerous ques- tionnaires exist to screen for a broad number of social needs, their reliability and validity are unknown. Predictive

modeling combined with clinical decision support (CDS) could overcome the above challenges that limit screen- ing and perpetuate ED patients' unmet social needs. Our long-term goal is to support effective care by enabling provider access to clinical and social context information. The objective of this proposal is to implement and

evaluate a CDS tool that identifies ED patients needing a referral to the social providers best equipped to address social needs. Our central hypothesis is that the purpose of screening is to inform referrals to appropriate services and that, in the context of social needs, social workers, dietitians, and behavioral health counselors are the

professionals best suited to meet patients' needs. Leveraging a proven technological infrastructure and collabo- ration with an urban, safety-net ED, this project will accomplish three aims. Aim 1, Compare the effectiveness of predictive modeling vs. questionnaire-based screening in identifying patients in need of social and behavioral

services, compares the performance of predictive modeling against questionnaire-based screening. Predictive modeling will leverage a unique combination electronic health record, health information exchange, social ser- vice organization, and public health data sources. Aim 2, Identify ED providers' and patients' perceptions of

screening for unmet social needs using predictive modeling and questionnaire-based screening, utilizes qualita- tive methods grounded in implementation and patient-centered innovation theoretical frameworks to understand ED patient, provider, and staff perceptions of enablers and barriers to screening. Aim 3, Quantify the impact of

real-time screening for social needs on subsequent utilization, will implement and evaluate a CDS intervention (using the best performing approach from Aim 1 and guided by the findings of Aim 3) that facilitates appropriate referrals to social and behavioral providers in a pre-post with comparison group longitudinal design. Outcomes

of interest are reduced ED revisits, increase follow-up visits with primary care providers. The proposed research is significant because it directly compares two approaches to addressing the widespread problem of unmet social needs. This proposal is innovative by applying predictive modeling with personal, social service, and clinical

context data, and by shifting social screening research to the ED. By working with an urban safety-net hospital, this research addresses the priority populations of socioeconomically disadvantaged and minority populations who are disproportionality burdened by unmet social needs.