An Audit of a Modified Pediatric Early Warning Scoring System in a Tertiary Pediatric Hospital

Hosp Pediatr. 2024 Dec 1;14(12):1017-1025. doi: 10.1542/hpeds.2024-007799.

Abstract

Objective: Pediatric early warning systems identify patients at risk for deterioration. We aimed to determine if a locally adapted pediatric early warning system [KK hospital early warning score (KEWS)] can distinguish patients who require transfer from general ward to high dependency unit (HDU) or PICU, from those who do not.

Methods: We conducted a retrospective case-control single-center audit. All unplanned general ward to PICU/HDU transfer from January to December 2021 were included as cases. We collected KEWS at the time of transfer and at 2, 4 and 6 hours prior to transfer. Univariate and multivariable logistic regression with imputation analysis were performed to identify variables associated with unplanned PICU/HDU transfer.

Results: There were 784 patients (196 cases and 588 controls). Median age of cases and controls were 3.3 (interquartile range: 1.0-9.7) and 3.9 (0.8-11.3) years, respectively. At the time of transfer, median (interquartile range) KEWS for cases and controls were 3 (0, 4) and 0 (0, 1), respectively (P <.001). At all timepoints, KEWS was able to distinguish between children who did and did not require transfer to HDU/PICU (At time of transfer: area under the curve = 0.80, 95% CI = 0.75-0.85, P <.001). After adjusting for age, category of admission, frequency of ordered monitoring, KEWS predicted HDU/PICU transfer with an odds ratio of 2.34 (95% CI 1.82 to 3.00, P <.001).

Conclusions: KEWS was able to distinguish patients requiring transfer from GW to HDU/PICU from those not needing a transfer.

MeSH terms

  • Case-Control Studies
  • Child
  • Child, Preschool
  • Early Warning Score*
  • Female
  • Hospitals, Pediatric*
  • Humans
  • Infant
  • Intensive Care Units, Pediatric*
  • Male
  • Medical Audit
  • Patient Transfer* / statistics & numerical data
  • Retrospective Studies
  • Tertiary Care Centers*