Delayed babbling at 10 months : observation, detection and a two-year follow-up

Abstract

Babbling variables that can predict later expressive language, such as presence and onset of canonical babbling (CB), presence of anterior stops and consonant inventory, are suitable targets for a CB screening and the Swedish Child Health Services (CHS) is a suitable framework for such a screening. The overall aim of the thesis was to find means for identification and valid care for children with delayed CB and foundations for future intervention. The specific aims were to investigate babbling observation as method for assessment of babbling, the predictive validity of a CB screening at 10 months, follow-up assessments of children who failed the CB screening and parent’s use of a responsive approach in interaction with their 10-month-old children who either lacked or used CB.

The thesis consists of five studies, including 1219 participants at 10 months recruited for the project and material for another 149 children aged 9 to 21 months collected for other studies. Children without and with known medical conditions, such as hearing loss, cleft palate, and neurodevelopmental disorders, were included. In study I and II the validity and accuracy of observation as a method for identifying CB was investigated. In study III, the predictive validity of a CB screening was evaluated by comparing the result of a CB screening at 10 months with the result of the language screening in use within the CHS at 2.5/3 years. In study IV, babbling, expressive language, cognitive, and motor abilities in children who failed the CB screening at 10 months were followed-up at 12, 18 and 36 months by babbling observations, parent questionnaires (SECDI) and clinical assessments (SVANTE, Bayley). In study V, parent’s responses following two different child utterance types in a clinical group of children were examined, to find a starting point for an intervention for children with delayed CB.

The results showed that babbling observation during interaction with a parent was a valid and reliable method to identify children who used/did not use CB and other predictive consonant variables. Observation of the variables could, to a clinically significant level, separate children in a clinical and a non-clinical group. The CB screening at 10 months had high specificity and negative predictive value. However, the sensitivity was low. For some children who failed the CB screening, expressive language was delayed at 18 months and the delay continued up to 36 months. For one out of five children, the delay in CB was an early sign of a more comprehensive developmental disability. Regarding parent’s response types, parents to children with delayed babbling used significantly more acknowledgements following CB than following other vocalizations, which can help children stay in the interaction but does not provide word-shaping models.

The CB screening at 10 months, using valid questions to parents in the first step and valid, reliable, and easy-administered observations of CB variables in the second step, could dismiss most of the children who passed the language screening at 2.5/3 years and identify some of the children who failed the language screening at that age. Although there is a gain in valid and specific information on the babbling of the identified children and benefits of early intervention, the CB screening as a new universal screening within the CHS is problematic. Instead, including the specific CB questions in the current developmental surveillance at the 10-month-visit is suggested, and if needed they can be followed by a specialist assessment.