Early high risk of cerebral palsy classification is predictive of cerebral palsy at 2 years: an implementation cohort study

To determine the predictive accuracy of an early high risk of cerebral palsy (CP) classification for CP diagnosed by 2 years' corrected age within an implementation study of international clinical CP guidelines. Implementation cohort study. Eleven Australian neonatal intensive care units. 453 infants born 2019-21 <28 weeks' gestation, or ≥28 weeks with other newborn-detectable risk factors for CP. Implementation included providing professional development for clinicians, technology (smartphone app) and health network peer support. Infants were classified as high risk of CP if they had abnormal findings on at least two of the following three assessments: neonatal neuroimaging, General Movements Assessment at 3-4 months or Hammersmith Infant Neurological Examination. Baseline perinatal data and 2-year outcome data were collected from medical record review. Any parent-report of CP at the 2-year interview was confirmed by medical records and/or a paediatrician's report. We calculated predictive values for high risk of CP classification for confirmed CP at 2 years. We obtained 2-year outcomes from 425 infants (95%). High risk of CP was classified in 105 (25%) of these infants at a mean age of 3.5 months (SD 2.5). This classification demonstrated 91% sensitivity (95% CI 82% to 96%), 90% specificity (95% CI 86% to 93%) and 90% accuracy (95% CI 87% to 93%) for predicting CP, with a mean age of diagnosis of 10.8 months (SD 6.3). Being classified as high risk of CP using a combination of neuroimaging, General Movements Assessment and/or Hammersmith Infant Neurological Examination can predict CP by 2 years of age with high accuracy.