Sensitivity and specificity of General Movements Assessment for diagnostic accuracy of detecting cerebral palsy early in an Australian context.

Abstract

The aim of this study was to calculate the sensitivity and specificity of the General Movements Assessment (GMA) for estimating diagnostic accuracy in detecting cerebral palsy (CP) in an Australian context by a newly established NSW rater network. A prospective longitudinal cross-sectional study was conducted. The GMA was blind-rated from conventional video by two independent certified raters, blinded to medical history. A third rater resolved disagreements. High-risk population screening for CP using the GMA during the fidgety period (12-20 weeks) was carried out in four neonatal intensive care units and one CP service over a 30-month period (2012-2013). Participants were 259 high-risk infants. Sensitivity and specificity values were calculated with true positives defined as a confirmed diagnosis of CP from a medical doctor. Of the 259 infants assessed, 1-year follow-up data were available for 187. Of these, n = 48 had absent fidgety (high risk for CP), n = 138 had normal fidgety (low risk for CP), and n = 1 had abnormal fidgety (high risk for a neurological disorder). Of the 48 with absent fidgety movements, 39 had received a diagnosis of CP by 18 months and another 6 had an abnormal outcome. Of the n = 138 normal fidgety cases, n = 99 cases had a normal outcome, n = 38 had an abnormal outcome but not CP, and n = 1 had CP. For detecting CP, we had a sensitivity of 98% and specificity of 94%. GMA was feasible in an Australian context and accurately identified CP with a sensitivity and specificity comparable with European standards and published neuroimaging data.