Long-stay Patients in the Pediatric Intensive Care Unit

Background: Long-stay patients (LSP) in pediatric intensive care unit (PICU) are associated with high healthcare resource utilization.

Objective: To identify factors associated with long hospitalization in PICU.
Methods: Retrospective cohort study of all pediatric patients admitted in a PICU of Northern Portugal in 2017 and 2018. LSP were defined as patients having a length-of-stay greater than the 90^th percentile (≥ 14 days).

Results: A total of 626 patients were admitted, 10.7% (n= 67) were LSP, with a mean age of 7.2± 6.5 years and 59.7% (n=40) were male. Using univariate binary logistic models, we determine a significant association between admission from another intrahospital department (OR=0.35, 95% confidence interval [CI] 0.21-0.59; p<0.001), emergency admission (OR=4.04, 95% CI 2.29-7.13; p<0.001), nonsurgical admission (OR=2.67, 95% CI 1.59-4.45; p<0.001), neurological disorder admitting diagnosis (OR=2.89, 95% CI 1.35-6.18; p=0.006), cardiovascular condition admitting diagnosis (OR=2.85, 95% CI 1.17-6.94; p=0.021), trauma related admission (OR=4.30, 95% CI 2.24-8.25; p<0.001) and the length-of-stay. Additionally, it was tested the association between the previously independent variables and the length-of-stay in a multivariate regression model, retaining only the significant variables. An association was determined between nonsurgical and trauma admissions with longer length-of-stay, with corresponding OR_a of 4.68, 95% CI 2.54-8.63 and 9.04, 95% CI 4.23-19.33, respectively. The model classified correctly the observations in 89.3% of the cases. Model discrimination was assessed using a ROC curve, with corresponding AUROC =0.72 (95% CI=0.65-0.78; p<0.001)

Conclusion: The clinical profile of LSP includes nonsurgical and trauma patients. A predictive model could help to identify patients with high risk of longer length-of-stay with a potential applicability for internal management regarding quality and cost-saving interventions.