Home › Paediatrics › Machine learning predicts paediatric code events

Paediatrics

Machine learning predicts paediatric code events

Validation of a Hospital-Wide Risk Stratification Model for Predicting Code Events Among Critically Ill Children
Resuscitation Plus · Added 27 July 2026

Study on: Validation of a Hospital-Wide Risk Stratification Model for Predicting Code Events Among Critically Ill Children. No abstract available for detailed assessment.

Study type

Cohort

Design

Single-centre retrospective external validation study (2020-2023)

Participants

5,254 paediatric encounters, of which 2.6% had at least one code event

Main finding

pCREST predicted code events within 12 h with c-statistic 0.80; 13-fold improvement in detection over outcome prevalence; high performance up to 18 h pre-event; transfer learning did not improve performance

Caveat

Single-centre retrospective validation; in-hospital PICU setting with low event rate (precision-recall curve 0.034)

Bottom line

Context only

Learning goal

Recognise early signs of deterioration in critically ill children to anticipate and prevent paediatric code events.

More

• Topic: Paediatrics
• Daily scan: 27 July 2026

Related papers on Paediatrics

• Tactile stimulation site in neonatal resuscitation

  Resuscitation · 27 Jul 2026 · RCT · Supports current practice

  There was no difference in 5-min SpO2, need for further resuscitation in DR, cerebral oxygenation, and in-hospital complications with the three modes of TS, sole rub, trunk rub and back rub.

• Poor field differentiation of infant upper vs lower airway disease

  Prehospital Emergency Care · 27 Jul 2026 · Cohort · Context only

  EMS clinicians documented similar primary impressions and administered many of the same treatments across both upper and lower airway groups, demonstrating minimal differentiation in prehospital management of young children.

• Synchronised ventilation faster ROSC in neonatal piglets

  Resuscitation Plus · 27 Jul 2026 · Simulation · Hypothesis-generating

  In a neonatal piglet model of asystolic cardiac arrest, CCSV resulted in a significantly faster time to ROSC compared to 3:1 C:V.

• Observation cuts CT in paediatric abdominal trauma

  Annals of Emergency Medicine · 27 Jul 2026 · Cohort · Supports current practice

  Observation with deferred CT decisionmaking was associated with less frequent CT use, particularly in patients with intermediate clinician suspicion, without an increase in missed injuries.