Background and Objectives: long-stay patients (LSPs) in PICU represent a unique group of patients in terms of morbidity, mortality and healthcare resource consumption. We described the epidemiology, clinical characteristics and outcome of LSPs and determined predictive factors identifying LSPs at admission.
Methods: retrospective case-control study reviewing all children admitted to "Soroka" University Medical Center PICU during 2013. Patients were identified as LSP if hospitalized at PICU for >6 days.
Results: LSPs represented 21.8% of the PICU population and had a higher mortality rate (9.3%) than the average PICU rate of 5.7%. Multivariate regression analysis determined which clinical characteristics available within the first 24 hours of hospitalization were associated with LSPs and to create a predictive algorithm. We found that a diagnosis of respiratory disease at admission is predictive of long stay (OR 5.13, p<0.001); higher serum levels of albumin were found to be associated with a reduced risk for prolonged hospitalization (OR 0.45 p=0.01). LSPs were younger and have been admitted to the PICU previously. During hospitalization, LSPs utilized more hospital resources than non-LSP: invasive mechanical ventilation (61.5% vs. 19.6% p<0.001), central catheter insertion (51.5% vs. 12.3%, p<0.001), blood products (36.9% vs. 9.7% p<0.001), antibiotics (93.8% vs. 51.4% p<0.001) and vasoactive treatment (26.4% vs. 9.2%).
Conclusions: LSPs have higher mortality rates and consume more resources than non-LSPs. LSPs were younger and had previous PICU admissions. The predictive analysis revealed a significant correlation between length of hospitalization and serum hypoalbuminemia at admission in patients with respiratory failure.
