Pediatrics

Pediatric pressure-injury prediction models show discrimination but poor validation and applicability

This systematic review found 12 pediatric pressure-injury (PI) models from nine studies with AUCs ranging 0.612–0.978. Most models lacked calibration reporting, external validation, and were rated high risk of bias by PROBAST and PROBAST+AI. Common methodological flaws included low events-per-variable, poor missing-data handling, and single-center derivation. Only two models assessed clinical utility and just one demonstrated net benefit across realistic thresholds. 1

Adoption of a pediatric ACC model associated with fewer complications and ICU back-transfers

A single-center QI study of a 26-bed ACC unit showed mean complications fell from 23.6% to 16.0% after implementation. ICU back-transfers decreased from 11.4% to 6.9%, and mean patient discharge time modestly improved. Length of stay and 7-day unplanned readmissions were unchanged, suggesting no large unintended harms. Family experience scores improved after the care model change that included cardiology-led attendings and multidisciplinary rounding. 2

Protocol: systematic review and meta-analysis of caffeine for apnea in hospitalized children (non-premature)

This protocol plans a systematic review of caffeine versus placebo or usual care for apnea in children aged 0–17 years, excluding apnea of prematurity. Critical outcomes include duration of respiratory support, ventilator-free days, time to apnea cessation, and serious adverse events. The authors will include randomized and observational studies, assess risk of bias, and use GRADE for evidence certainty. Results are pending and therefore caffeine for non-premature pediatric apnea remains not practice-changing yet. 3