Development of an electronic Poor Outcome Screening (ePOS) Score to identify critically ill patients with potential palliative care needs

Nora Luethi; Sarah D Wermelinger; Alan G Haynes; Marie Roumet; Maud Maessen; Barbara Affolter; Martin Müller; Joerg C Schefold; Steffen Eychmueller; Luca Cioccari

doi:10.1016/j.jcrc.2022.154007

Development of an electronic Poor Outcome Screening (ePOS) Score to identify critically ill patients with potential palliative care needs

J Crit Care. 2022 Jun:69:154007. doi: 10.1016/j.jcrc.2022.154007. Epub 2022 Feb 16.

Authors

Affiliations

¹ Department of Radiation Oncology, University Center for Palliative Care, Inselspital, Bern University Hospital, University of Bern, Switzerland; Department of Emergency Medicine, Inselspital, Bern University Hospital, University of Bern, Switzerland. Electronic address: nora.luethi@med.unibe.ch.
² Department of Intensive Care Medicine, Inselspital, Bern University Hospital, University of Bern, Switzerland.
³ CTU Bern, University of Bern, Switzerland.
⁴ Department of Radiation Oncology, University Center for Palliative Care, Inselspital, Bern University Hospital, University of Bern, Switzerland; Institute for Social and Preventive Medicine, University of Bern, Switzerland.
⁵ Department of Radiation Oncology, University Center for Palliative Care, Inselspital, Bern University Hospital, University of Bern, Switzerland.
⁶ Department of Emergency Medicine, Inselspital, Bern University Hospital, University of Bern, Switzerland.
⁷ Department of Intensive Care Medicine, Inselspital, Bern University Hospital, University of Bern, Switzerland; Department of Intensive Care Medicine, Kantonsspital Aarau, Aarau, Switzerland.

PMID: 35183039
DOI: 10.1016/j.jcrc.2022.154007

Abstract

Purpose: To develop and validate an electronic poor outcome screening (ePOS) score to identify critically ill patients with potentially unmet palliative care (PC) needs at 48 hours after ICU admission.

Materials and methods: Retrospective single-centre cohort study of 1'772 critically ill adult patients admitted to a tertiary academic ICU in Switzerland between 2017 and 2018. We used data available from electronic health records (EHR) in the first 48 hours and least absolute shrinkage and selection operator (LASSO) logistic regression to develop a prediction model and generate a score to predict the risk of all cause 6-month mortality.

Results: Within 6 months of the ICU admission, 598 patients (33.7%) had died. At a cut-off of 20 points, the ePOS score (range 0-46 points) had a sensitivity of 0.81 (95% CI 0.78 to 0.84) and a specificity of 0.51 (0.48 to 0.54) for predicting 6-month mortality and showed good discriminatory performance (AUROC 0.72, 0.67 to 0.77).

Conclusions: The ePOS score can easily be implemented in EHR and can be used for automated screening and stratification of ICU patients, pinpointing those in whom a comprehensive PC assessment should be performed. However, it should not replace clinical judgement.

Keywords: Intensive care; end-of-life care; mortality; outcome; palliative care; prediction scores.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Adult
Cohort Studies
Critical Illness*
Electronics
Hospital Mortality
Humans
Intensive Care Units*
Palliative Care
Retrospective Studies