283: PREDICTING ETHICS CONSULTATIONS IN PEDIATRIC PATIENTS SUPPORTED ON ECMO

Abstract

Introduction: No publications quantify or characterize patient-level ethical issues in pediatric extracorporeal membrane oxygenation (ECMO). We sought to examine characteristics associated with referral to a hospital ethics consult (EC) service for pediatric ECMO patients. Methods: This is a single center retrospective study in a quaternary pediatric hospital. Local ECMO and ethics databases and records were cross referenced to include all patients on ECMO from 2012 through 2021 and identify those with EC. We compared demographic and clinical variables according to EC. To define optimal cutoffs for EC for run duration, length of stay (LOS) and sum of procedures and complications, we fit receiver operating characteristics (ROC) curves. To identify independent predictors of EC in ECMO patients we used forward stepwise selection including 7 a priori defined covariates in a multivariable logistic regression model (age, ECMO location, ECPR and multiple runs as well as the three optimal cutoff variables). Results: Of 601 ECMO patients, 27 received EC (4.5%). EC was associated with multiple ECMO runs, more complications/procedures, longer LOS and ECMO duration, cardiac service cannulation location, decannulation outcomes (bypass or recannulation) and higher in-hospital mortality (p< 0.05). Cutoffs for EC were LOS >52 days, run duration >160hrs and >6 procedures and complications. Independent associations with EC included LOS>52 days (adjusted odds ratio (aOR) 7.9, 95% CI [3.0, 21.1]), run >160hrs (aOR 3.3, 95%CI [1.1, 9.7]), procedure and complication sum >6 (aOR 2.5, 95%CI [1.0, 6.0]) and age (adjusted coefficient 1.3, 95%CI [1.1, 1.5]). The model showed good discrimination (area under the ROC curve 0.88), fit (Hosmer Lemeshow test p = 0.723) and no collinearity (VIF < 1.4). Conclusions: We observed a low EC rate compared to adult studies. Older age, longer LOS, prolonged runs and multiple procedures or complications were predictors of EC. These data may guide earlier identification of patients at risk for significant ethical issues and inform ethically supportable frameworks around this invasive care. Future work should include model validation and quality improvement projects to improve EC utilization and evaluate the influence on patient experience and quality of care.