Concurrent and predictive validity analyses.

Aim
Assess whether school-based teacher-led screening is effective at identifying children with motor difficulties.
Methods
Teachers tested 217 children aged between 5 and 11 years old, after a one hour training session, using a freely available tool (FUNMOVES). Four classes (n = 91) were scored by both researchers and teachers to evaluate inter-rater reliability. Researchers assessed 22 children using the Movement Assessment Battery for Children (MABC-2; considered to be the ‘gold standard’ in Europe for use as part of the diagnostic process for Developmental Coordination Disorder) to assess concurrent and predictive validity.
Results
Inter-rater reliability for all individual activities within FUNMOVES ranged from 0.85–0.97 (unweighted Kappa; with 95%CI ranging from 0.77–1). For total score this was lower (κ = 0.76, 95%CI = 0.68–0.84), however when incorporating linear weighting, this improved (κ = 0.94, 95%CI = 0.89–0.99). When evaluating FUNMOVES total score against the MABC-2 total score, the specificity (1, 95%CI = 0.63–1) and positive predictive value (1; 95%CI = 0.68–1) of FUNMOVES were high, whereas sensitivity (0.57, 95%CI = 0.29–0.82) and negative predictive values (0.57, 95%CI = 0.42–0.71) were moderate. Evaluating only MABC-2 subscales which are directly related to fundamental movement skills (Aiming & Catching, and Balance) improved these values to 0.89 (95%CI = 0.52–1) and 0.93 (95%CI = 0.67–0.99) respectively.
Interpretation
Teacher-led screening of fundamental movement skills (via FUNMOVES) is an effective method of identifying children with motor difficulties. Such universal screening in schools has the potential to identify movement difficulties and enable earlier intervention than the current norm.