Identifying Patients at Risk for Unplanned Extubation in a Pediatric ICU

Abstract

Background:

Unplanned extubation (UE) poses a substantial risk, potentially resulting in clinical instability, cardiac arrest, and need for re-intubation. Our objective was to develop a UE risk assessment scoring (UERAS) tool to identify contributing factors, facilitate mitigation strategies, and decrease UE. We hypothesized that our UERAS would demonstrate high reliability and UERAS implementation would be associated with UE reduction.

Methods:

In a 75-bed pediatric ICU, we developed the UERAS with the range score at 0–18. We predetermined UERAS ≥ 6 as high risk. For inter-rater reliability (IRR), the UERAS was piloted in a 24-bed section of the PICU, where 2 respiratory therapists (RTs) independently evaluated patients undergoing invasive mechanical ventilation. Upon demonstrating high IRR, UERAS was expanded across the 75 beds. Standardized mitigation actions for high-risk patients were developed including immediate clinician alerts via electronic health record chat function and interdisciplinary bedside huddles within 1 h.

Results:

During March to April 2023, a sample of 50 paired observations by 16 RTs showed a Pearson’s correlation coefficient of 0.997 (P < .001). During March 2023 to May 2024, 3,206 subsequent assessments in 476 subjects (median age 5 years [interquartile range (IQR) 1–13]) were reported with the UERAS score median 2 (IQR 0–4). Three hundred twenty-eight assessments (10.2%) met high-risk criteria (UERAS ≥ 6), including 302 assessments that scored at risk for sedation concerns. One hundred twenty-one high-risk subjects (36.8%) experienced an increased sedative as part of mitigation actions. Compared to the preimplementation phase (July 2022 to March 2023), UERAS implementation phase (July 2023 to March 2024) had significantly lower UE events: 0.44/100 ventilator days (13 UEs) versus 0.14/100 ventilator days (6 UEs), absolute difference 0.28/100 UE events, 95% CI 0.003–0.563, P = .049. However, the UERAS tool demonstrated limited efficacy in predicting UE. Although many high-risk scores did not correspond to events, some UEs occurred in subjects not identified as high risk.

Conclusions:

The UERAS scores showed high IRR. Implementation of the UERAS with mitigation actions was associated with a reduction in UEs.

Graphical Abstract

This is a visual representation of the abstract.