Intelligent Social Health Tool (INSIGHT) For Emergency Department Needs Screening

Document Type

Conference Proceeding

Publication Date

10-2026

Publication Title

Annals of Emergency Medicine

Abstract

Study Objectives: While non-clinical social factors are known to drive patient outcomes, capturing and acting upon these factors remain difficult in the emergency department (ED), where high-acuity care must be delivered in a rapid, time-dependent manner. Standardized screening tools are often inadequate, while critical social data are frequently buried within unstructured clinical notes. We developed the Intelligent Social Health Tool (INSIGHT), a secure web application designed to extract and aggregate information from both structured and unstructured fields within the Electronic Health Record (EHR) at the point-of-care. The primary objective of this study was to establish technical feasibility, workflow integration, and operational utility of using INSIGHT to identify needs among ED patients. Methods: We conducted a pilot study between 5/18/2025 and 7/29/2025 at a large tertiary care academic emergency department. Two board-certified emergency medicine physicians with >15 years of clinical experience served as pilot testers, utilizing the INSIGHT application to retrieve and process both structured and unstructured notes. The system architecture used the Fast Healthcare Interoperability Resources (FHIR) standard to extract a user-defined number (up to 1000) of documents including ED provider, nursing, progress, social work, and consult notes. Then, the system used a rule-based Natural Language Processing (NLP) algorithm to detect social needs and delivered the results to the user. Typically, the extraction and processing of 200 clinical notes took less than 20 seconds. The number of identified needs and appropriate wraparound services were examined after completion of the pilot. Retrospective analysis was performed to assess alignment between tool-identified needs and orders in the EHR to evaluate potential gaps in care. Results: A total of 1,557 INSIGHT searches were conducted, evaluating 1,423 unique patient cases. Of the 1,423 cases reviewed, 1,267 patients (89%) had no documented social needs. Nicotine dependence was the most frequent finding (286 cases), followed by housing instability, transportation barriers, and alcohol misuse. The physician evaluators determined that 228 patients required social work referral, 41 for substance use disorder services, and 27 for referral to a low-cost clinic for uninsured patients. A manual audit of 180 cases flagged for potential social work referral revealed that 48 patients (27%) had no recent corresponding social work order in the EHR. Conclusion: The INSIGHT application demonstrates technical feasibility as a decision support tool capable of extracting social factors from unstructured clinical documentation during an ED encounter. The pilot study revealed patients with documented needs that did not have a social work order for that encounter. This highlights a critical care gap and underscores the tool’s potential impact on clinical practice. The tool has the potential to improve referral processes and better align vulnerable patients with available wraparound support services. Future integration directly into EMR workflows will allow for further evaluation of the tool’s impact on clinical practice and downstream healthcare utilization. Image 1. User interface for the INSIGHT application. The end-user is required to input the patient’s MRN and select the number of clinical notes for processing, ranging from 100 to 1000. The tool retrieves the patient’s age, sex and race along with the number of ED and inpatient admissions during the prior 12 months. The identified social factors are listed with the date, note source, and phrases. Lastly, the tool provides options for the clinician to select the appropriate referral pathway for the patient.

Volume

88

Issue

4 Suppl 1

First Page

S314

Comments

American College of Emergency Physicians (ACEP) Research Forum, October 5-7, 2026, Chicago, IL

Last Page

S315

DOI

10.1016/j.annemergmed.2026.08.718

Link to Full Text

Share

COinS