Video Analysis of Parent-Child Interactions in Behavioral Sleep Disorders: Development of a Scoring Algorithm

Introduction: Behavioral sleep disorders, including chronic insomnia (CI), are generally assessed by subjective parent interview. However, evidence suggests that parental report of children's overnight behaviors is unreliable, perhaps due to recall bias or confusion due to sleep deprivation. Video technology has been used clinically to capture complex behavioral disorders in children during the day. However, there is no standardized means of analyzing child and parent behavior at bedtime or during the night. We aimed to create an algorithm for this purpose. Methods: Child brain tumor survivors (a population previously shown to have a high prevalence of CI) were screened for difficulties initiating and maintaining sleep using sub-scales from the Sleep Disturbance Scale for Children. Those who screened positive (n = 3) then completed a detailed parent interview to confirm a clinical diagnosis of CI. One night of home video footage was obtained from initial settling period to morning waking (SOMNOmedics camera). Footage was imported into BORIS^© software and a coding system for parent and child behavior was developed over multiple iterations until agreeable inter-rater reliability (>70%) was achieved between two independent coders. Results: The final coding categories were: 1) time domains, 2) physical environment, 3) child global status, 4) location, 5) activity, and 6) physical interaction. This achieved 74% inter-reliability in its last iteration. Discussion: A statistically acceptable behavior scoring algorithm was achieved. With further development, this tool could be applied clinically to investigate behavioral insomnia and in research to provide more objective outcome measurement.