Predicting time to hospital discharge for extremely preterm infants.

Area of Science:

• Neonatal Medicine
• Clinical Epidemiology
• Health Services Research

Background:

• Decreasing mortality in extremely preterm infants (<27 weeks' gestational age) raises concerns about healthcare resource utilization.
• Accurate prediction of hospital discharge timing is crucial for resource planning, family support, and quality improvement initiatives.

Purpose of the Study:

• To develop, validate, and compare predictive models for hospital discharge time in extremely preterm infants.
• Models were based on time-dependent covariates and the presence of key risk factors.

Main Methods:

• Retrospective analysis of 2254 infants born <27 weeks' gestational age (July 2002-December 2005).
• Modeled time to discharge (postmenstrual age) using linear and logistic regression with time-dependent covariates.
• Evaluated models incorporating perinatal factors only, perinatal + early-neonatal factors, and perinatal + early-neonatal + later factors.
• Assessed simplified models using 5 key risk factors for early/late discharge prediction.

Main Results:

• Prediction of exact postmenstrual age at discharge was poor.
• Models including later clinical characteristics demonstrated improved prediction of early or late discharge (AUC: 0.76-0.83).
• Simplified key-risk-factor models showed comparable predictive performance (AUC: 0.75-0.77) to comprehensive models.

Conclusions:

• Predicting discharge solely on perinatal factors is inadequate.
• Incorporating later-occurring morbidities significantly improves discharge prediction accuracy.
• Clinically applicable strategies using a few key risk factors offer comparable prediction to complex models.