Predictive validity of verbal and non-verbal communication and mother-child turn-taking at 12 months on language outcomes at 24 and 36 months in a cohort of infants experiencing adversity: a preliminary study

Background: Parent-reported measures of early communication have limitations for use with infants experiencing adversity. Observational measures of early non-verbal and verbal communicative behaviours and mother-child turn-taking may provide a complementary method of capturing early communication skills for these children. Aims: To explore the predictive validity of verbal and non-verbal behaviours and mother-child conversational turn-taking (fluency and connectedness) at child age 12 months in relation to language measures at 24 and 36 months in a cohort of infants experiencing adversity. Methods & Procedures: Pregnant women experiencing adversity were recruited from maternity hospitals in Australia. At 12 months, 190 infants were videoed during mother-child free-play. Verbal and non-verbal communicative behaviours and fluency and connectedness were measured from the 12-month videos. Predictive validity of 12-month behaviours was calculated in relation to mean length of utterance and number of unique words at 24 months and Clinical Evaluation of Language Fundamentals Preschool-Second Edition (CELF-P2) Core Language scores at 36 months. Outcomes & Results: All 12-month behaviours had adequate specificity but poor sensitivity when compared with other predictive validity studies using published early language measures. However, in adjusted regression models, fluency and connectedness and verbal behaviours at 12 months predicted unique words at 24 months. Fluency and connectedness also predicted CELF-P2 scores at 36 months. Conclusions & Implications: Findings reconfirm the difficulty in early identification of children at risk of later language difficulties. All 12-month measures were more accurate at identifying those children who will have better language than those children who will not. As fluency and connectedness was the only measure to predict 24- and 36-month language in adjusted regression models, it may be an important factor to consider when measuring early language skills for infants experiencing adversity. Future research could combine observational measures of early communication and fluency and connectedness with other predictors of language to try to increase prediction accuracy.