1351: Developing an Ehr-Embedded Digital Safety Net to Reduce Acute Care Critical Deterioration Events

Introduction: Children on acute care floors face risk of critical deterioration, including cardiac arrest, urgent ICU transfer, and death. These events add approximately 100, 000 of costs per admission, prolong ICU stays, and increase in-hospital mortality. We hypothesized that the deployment of an EHR-embedded deterioration prediction model would reduce critical deterioration in pediatric acute care at our quaternary children’s hospital. Methods: Between 2017 and 2018, acute care units adopted a two-level sepsis screening process: a rule-based EHR trigger, followed by a bedside huddle and sepsis order set activation. Building on this, a deterioration prediction score was developed using 2017 admissions data. The random forest model incorporated structured clinical variables and intuitive nurse assessments to predict a composite outcome of life-threatening deterioration events, including cardiac arrest, urgent ICU transfer, or death. The model was embedded into the EHR, with iterative Plan-Do-Study-Act cycles used to refine both predictive accuracy and workflow integration over time. An interrupted time series analysis evaluated the sequential interventions impact on deterioration rates, rapid response activations, and mortality. AI tools assisted with code editing but not content creation or analysis. Results: The 313-bed hospital (181 acute care beds) recorded 143, 808 acute care admissions and 318, 678 patient days from Jan 2018–Dec 2024. Implementation of sepsis and shock screening was associated with a decline in critical deterioration from 2. 5 per 10, 000 admissions in 2018 to 0. 2 per 10, 000 between 2019–2022 (P < 0. 001). Mortality among patients transferred to the ICU fell 31%. Following deterioration model integration in Sept 2022, the rapid response activation rate fell by 0. 51 events per 10, 000 patient days each month post-implementation (P < 0. 001). Overall, the rate of critical deterioration decreased from 150. 2 per 10, 000 patient days in 2022 to 110. 4 per 10, 000 patient days in 2024. Conclusions: The development and deployment of a digital safety net was linked to a significant reduction in critical deterioration among acute care pediatric patients at our children’s hospital. Continued refinement of the model may further support clinicians in early recognition and timely resuscitation.