285: NURSE-DRIVEN, AI-ENHANCED PALLIATIVE CARE REFERRALS IN THE ICU

Le résumé fourni par la source

Introduction: With nearly 20% of intensive care unit (ICU) patients facing end-of-life, timely palliative care (PC) is critical. Traditional electronic medical record (EMR)-based referrals often lead to delayed, disjointed care due to insufficient interdisciplinary collaboration, which can be confusing for patients and families. Our aim was to enhance early identification of ICU patients needing PC, fostering interdisciplinary collaboration—especially nursing involvement—to achieve earlier referrals and improved patient outcomes. Methods: We utilized an artificial intelligence (AI)-enhanced predictive algorithm to create a collaborative system for early ICU patient assessment. Conducted in a 30-bed ICU at a Midwest community hospital (Nov 2024 - Feb 2025), this AI-enhanced algorithm leveraged EMR and claims data to predict mortality and identify PC candidates daily. The process involved: 1) Educating nurses on PC definition and skills; 2) A dedicated palliative care nurse practitioner (NP) generating daily algorithm reports; 3) A structured workflow where the acute care NP, staff nurses, and interdisciplinary team discussed eligible patients, leading to physician approvals, often via secure messaging. The acute care NP then placed consensus-driven referrals. Results: The project resulted in 162 PC referrals and a notable increase in physician-approved consultations, as observed by ICU and PC NPs. Of these, 117 patients transitioned from active to do-not-resuscitate code status. Eligible patients identified by the algorithm and interdisciplinary team received PC consultations within 1-2 days. Key discharge diagnoses included sepsis, cardiovascular, and respiratory conditions, reflecting severe illness. Approximately 41% of patients were discharged to hospice, and a 16% death rate (predominantly sepsis) highlighted the prevalence of life-limiting conditions and the focus on improving timely end-of-life care delivery. Conclusions: By integrating AI-enhanced insights with a robust interdisciplinary approach, we improved patient identification and facilitated physician consultations. This model highlights the transformative potential of technology and collaboration in palliative care. Future research will quantify long-term benefits on patient and family well-being.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.