View Abstract

A4575 - Prospective Validation of the End-of-Life Scoring-System (Ending-s) to Identify Patients at Risk for Death in the ICU
Author Block: T. Amass1, G. Villa2, R. McFadden3, T. Walsh4, A. Palmisciano4, M. Yeow3, R. De Gaudio2, S. OMahony3, J. Curtis5, M. M. Levy6; 1Pulmonary Critical Care, Brown University/Rhode Island Hospital, Providence, RI, United States, 2Anesthesia/Critical Care, Azienda Ospedaliero Universitaria Careggi, Florence, Italy, 3Palliative Care, Rush Medical Center, Chicago, IL, United States, 4Pulmonary Critical Care Research Division, Brown University/Rhode Island Hospital, Providence, RI, United States, 5Univ of Washington Harborview Med Ctr, Seattle, WA, United States, 6Pulmonary Critical Care, Rhode Island Hosp, Providence, RI, United States.
Rationale- ICU physicians need to be confident with end-of-life issues in order to provide appropriate care to dying patients and their family1. Decision making on these issues can be cumbersome in the acute setting and a scoring system able to help identify the dying patient for whom life-sustaining treatments may no longer be beneficial may support this process1-3. Previously defined scoring systems use laboratory data that are not necessarily available each day, which limits their applicability. The “END-of-Life ScorING-System” (ENDING-S) is a recently developed scoring system that incorporates clinical data such as days of mechanical ventilation, length of stay, need for vasoactive medications, presence of sepsis, and ICU length-of-stay (LOS) without requiring laboratory data. It was developed and preliminary validated with critically ill patients with an ICU LOS > 4 days in a single ICU4 to help predict which may die in the ICU. This study aimed to prospectively validate ENDING-S in multiple ICUs and compare it against the Sequential Organ Failure Assessment (SOFA) score. Methods- Adult patients with an ICU LOS >4 days admitted between September 2015 and March 2017 to Rhode Island Hospital’s medical ICU, Rush Medical Center’s medical ICU and Azienda Ospedaliero Universitaria Careggi’s medical/surgical ICU (Florence Italy) were considered for this prospective study. Patients admitted for end-of-life care, prisoners, pregnant patients and patients under 18 years old were excluded. Once enrolled, ENDING-S and SOFA scores were calculated daily and evaluated against the patient’s ICU outcome and the predictive properties of the scores evaluated through a Receiver-Operating-Characteristic (ROC) analysis. Results- 220 patients were enrolled and, among these, 21.46% died in the ICU. An ENDING-S score of 11.5, as previously defined, correctly predicted the ICU outcome in 71.4% of these patients. Sensitivity, specificity, positive and negative predictive values associated with ENDING-S cut-off of 11.5 were 68.1%, 72.3%, 60% and 89.3%, respectively. ROC-AUC for outcome prediction were 0.79 and 0.88, respectively for ENDING-S and SOFA. Conclusions- The ENDING-S score in ICU patients with a LOS greater than 4 days was able to predict a patients’ death, though not as accurately as the SOFA score. However, the ENDING-S score has the benefit of not requiring laboratory data, and has a negative predictive value of 89.3%. This scoring system can be useful to help physicians in guiding families when considering end-of-life scenarios, including engagement in palliative care, in the ICU.