Exploring the Relationship Between Behavioral Health Early Prediction Model and Patient Outcomes in the Emergency Department
Abstract/Description/Artist Statement
Problem: Emergency departments (EDs) are increasingly utilized for mental health (MH) care, often resulting in overcrowding and delayed access to specialized services. A significant proportion of MH patients are discharged without a psychiatric consult, raising concerns about safety and readmission rates. This trend places additional strain on already limited behavioral health resources and contributes to care delays for patients in acute crisis. Moreover, the lack of standardized discharge criteria for MH patients increases variability in care and may compromise patient outcomes. Purpose: This study evaluates the effectiveness of the Behavioral Health Early Discharge Prediction Model (BHEDPM) in identifying MH patients suitable for early ED discharge without a mental health provider consult. It also examines patient outcomes, including readmission, hospitalization, and mortality within 30 days. Research Questions: RQ1. Is there a difference in the proportion of early ED discharges recommended by the BHEDPM protocol compared to the SDM protocol used by ED providers? RQ2. Is there a difference in demographic characteristics between patients recommended for early ED discharge by BHEDPM compared to the SDM? RQ3. Is there a difference in ED readmissions, hospital admissions (once discharged from the ED with a follow-up hospital admission) or deaths within 30 days between patients discharged by the BHEDPM compared to SDM? RQ4. Is there a difference in the proportion of MH patients who follow up with their MH provider within 7 days of ED discharge between BHEDPM and SDM? RQ5. Which variables (personal characteristics, MH diagnosis, discharge model) influence ED readmission or deaths within 30 days of discharge? Methods: This causal-comparative study uses retrospective data from four EDs within a large integrated health system. Participants include adults (≥18 years) discharged with MH diagnoses between January–April 2024 (pre-intervention) and January–April 2025 (post-intervention). Data will be extracted from electronic health records and analyzed using descriptive statistics, t-tests, chi-square tests, and logistic regression. Outcomes: The study will assess whether BHEDPM improves discharge safety and reduces unnecessary psychiatric consults. Anticipated outcomes include comparable or improved 30-day readmission and mortality rates in the BHEDPM group. Significance: Findings may support the integration of AI tools in ED workflows to enhance discharge planning, optimize resource use, and improve continuity of care for MH patients. This research has implications for nursing practice, health policy, and future AI-driven interventions in behavioral health.
Faculty Advisor/Mentor
Kathie Zimbro, PhD, RN
Faculty Advisor/Mentor Email
kzimbro@odu.edu
Faculty Advisor/Mentor Department
Graduate Nursing
College/School Affiliation
Ellmer School of Nursing
Student Level Group
Graduate/Professional
Presentation Type
Poster
Exploring the Relationship Between Behavioral Health Early Prediction Model and Patient Outcomes in the Emergency Department
Problem: Emergency departments (EDs) are increasingly utilized for mental health (MH) care, often resulting in overcrowding and delayed access to specialized services. A significant proportion of MH patients are discharged without a psychiatric consult, raising concerns about safety and readmission rates. This trend places additional strain on already limited behavioral health resources and contributes to care delays for patients in acute crisis. Moreover, the lack of standardized discharge criteria for MH patients increases variability in care and may compromise patient outcomes. Purpose: This study evaluates the effectiveness of the Behavioral Health Early Discharge Prediction Model (BHEDPM) in identifying MH patients suitable for early ED discharge without a mental health provider consult. It also examines patient outcomes, including readmission, hospitalization, and mortality within 30 days. Research Questions: RQ1. Is there a difference in the proportion of early ED discharges recommended by the BHEDPM protocol compared to the SDM protocol used by ED providers? RQ2. Is there a difference in demographic characteristics between patients recommended for early ED discharge by BHEDPM compared to the SDM? RQ3. Is there a difference in ED readmissions, hospital admissions (once discharged from the ED with a follow-up hospital admission) or deaths within 30 days between patients discharged by the BHEDPM compared to SDM? RQ4. Is there a difference in the proportion of MH patients who follow up with their MH provider within 7 days of ED discharge between BHEDPM and SDM? RQ5. Which variables (personal characteristics, MH diagnosis, discharge model) influence ED readmission or deaths within 30 days of discharge? Methods: This causal-comparative study uses retrospective data from four EDs within a large integrated health system. Participants include adults (≥18 years) discharged with MH diagnoses between January–April 2024 (pre-intervention) and January–April 2025 (post-intervention). Data will be extracted from electronic health records and analyzed using descriptive statistics, t-tests, chi-square tests, and logistic regression. Outcomes: The study will assess whether BHEDPM improves discharge safety and reduces unnecessary psychiatric consults. Anticipated outcomes include comparable or improved 30-day readmission and mortality rates in the BHEDPM group. Significance: Findings may support the integration of AI tools in ED workflows to enhance discharge planning, optimize resource use, and improve continuity of care for MH patients. This research has implications for nursing practice, health policy, and future AI-driven interventions in behavioral health.