Related Experiment Videos

Variability in duration of stay in pediatric intensive care units: a multiinstitutional study

U E Ruttimann1, M M Pollack

  • 1Department of Pediatrics, George Washington University School of Medicine, Washington, DC, USA.

Insights

A new statistical model predicts pediatric intensive care unit (PICU) length of stay (LOS) by adjusting for patient risk factors. Organizational factors like intensivist presence and care coordination shorten LOS, while larger PICU size may increase it.

Area of Science:

  • Pediatric critical care medicine
  • Health services research
  • Biostatistics

Background:

  • Length of stay (LOS) in pediatric intensive care units (PICUs) is a key metric for resource utilization and quality assessment.
  • Predicting and understanding factors influencing LOS is crucial for effective hospital management and patient care.
  • Existing models may not adequately account for patient-specific risk factors at admission.

Purpose of the Study:

  • To develop and validate a statistical model for predicting pediatric intensive care unit (PICU) length of stay (LOS).
  • To adjust LOS predictions for patient-related risk factors identified at admission.
  • To identify institutional factors that influence LOS in PICUs.

Main Methods:

  • A prospective study involving 5415 admissions across 16 pediatric intensive care units (PICUs) selected through stratified cluster sampling.
  • Data collected included patient demographics, Pediatric Risk of Mortality (PRISM) scores, diagnoses, pre-admission care, first-day critical care modalities, and LOS.
  • Log-logistic regression analysis was employed to model LOS based on patient and institutional factors.

Main Results:

  • Patient-related predictors of LOS included PRISM score, diagnostic groups, pre-admission factors (operative status, inpatient/outpatient, prior PICU admission), and mechanical ventilation use.
  • PICU factors associated with shorter LOS (5-11%) were the presence of an intensivist, residents, and care coordination.
  • Increased ratio of PICU beds to hospital beds was linked to longer LOS; medical school affiliation and admission volume did not significantly affect LOS when adjusted for patient conditions.

Conclusions:

  • The developed statistical model effectively adjusts PICU LOS for patient-specific risk factors, facilitating comparisons of resource utilization across institutions.
  • Organizational factors promoting team-oriented care, such as intensivist presence and care coordination, are associated with reduced LOS.
  • Larger relative PICU size may inadvertently incentivize longer bed occupancy, impacting overall resource management.
Abstract

Related Concept Videos