Poster

  • P317

Performance and accuracy of an automated electronic screening tool in a mixed surgical ICU collective

Abstract

We previously introduced a simple, computer-assisted, prospective tool designed to address the challenges in identifying potential organ donors in the ICU. This tool screens daily for ascertained scores (RASS and pupillary reflex) in the electronic patient data management system and notifies the recipient on a 12-hour-basis of potential patients. This study aims to evaluate the tool's effectiveness over an extended period, highlighting its potential benefits.

The software was implemented in a mixed surgical ICU (traumatic, abdominothoracal, vascular and neurosurgery) from January 2020 to December 2022 (24 months). Its performance was compared with an existing retrospective system that identifies patients who died with an ICD-10-diagnosis of primary or secondary brain damage.

The tool generated 1225 notifications for 240 different patients (median of 3; mean of 5.1, SD 6.0). Of these patients, 109 (45.4 %) died in the hospital while 34 (14.2 %) were confirmed to be brain dead in the course of the ICU stay and 16 (6.6 %) proceeded to organ donation. The established application identified 102 patients, who died with the diagnosis of primary or secondary brain damage, 19 of whom were not detected by the prospective tool. Reasons for the inability to identify those patients were very short ICU stay, cardiac arrest without return of spontaneous circulation during CPR or documentation inaccuracies. The vast majority of those undetected cases did not meet criteria for potential brain death.

In this diverse surgical patient cohort, the assistive tool demonstrated potential in increasing awareness for potential organ donation. It proved reasonably accurate and effective for daily use.