Abstract:
BACKGROUND: Preschool children in low resource settings are at higher risk of missing developmental potential due to the lack of standardized and validated methods for the timely detection of children with developmental delays or neurodevelopmental disorders. The preschool teacher is a non-specialist resourceful link within the community to detect and offer interventions early. This paper discusses the preliminary iteration of designing and testing the psychometric properties of a developmental assessment for children aged 24 to 60 months in Sri Lanka. This assessment is designed to be conducted by preschool teachers in their preschool setting. METHODS: Three processes followed: 1. Designing and development of the Ragama Early Assessment for Children (REACh) complete preschool developmental assessment and a tool kit 2. Testing and training teachers on conducting the REACh assessment 3. Preliminary assessment of the psychometric properties including content validity, internal consistency, interrater reliability and concurrent validity. RESULTS: A literature search identified 11 assessments and 542 items representing cognitive, social-emotional and adaptive, language and motor domains. Content validity was assessed to select and adapt items. A complete assessment tool was designed to be administered in four settings within the preschool. This was further improved during pre and pilot testing and teacher training. Cronbach's alpha measuring internal consistency was > 0.70 for cognitive, language, social-emotional and adaptive domains across all three age groups in 1809 children. Interrater reliability was > 65% for age groups 36-47 and 47- 60 months. Concurrent validity using a clinical gold standard demonstrated sensitivity of more than 0.75 for all age groups with variable specificities (24-35 months: 0.71, 36- 47 months: 0.43 and 48-60 months: 0.67) assessed in 75 children. CONCLUSIONS: This culturally and linguistically adapted tool was tested nationally in Sri Lanka. The inte-rrater reliability between teachers and research assistants was higher than 65% for all domains in children more than 36 months. The preliminary iteration confirms it as an acceptable screening assessment for all age groups but with significantly lower specificity in the 36-47 month age group. Further improvement in certain domains together with intense teacher training is likely to enhance the validity and reliability of the assessment.
