Validating Movement-Based Sleep Assessment in a Pediatric Sample with Developmental Disabilities and Associated Motor Impairment and Intellectual Disability: A dissertation submitted to the faculty of the University of Minnesota

Individuals with developmental disabilities (DD) have increased rates of sleep
disorders compared to their typically-developing peers. Poor sleep can have serious
negative consequences for the individual and their family’s health and quality of life,
underscoring the need for appropriate sleep assessment and intervention. Objective
measures of sleep offer advantages over commonly-used proxy-report sleep measures,
such as the ability to directly measure sleep as it is happening and not requiring recall.
Polysomnography (PSG) is the current gold-standard, but it has a high resource and
personnel burden, limiting its availability and utility for the number of children with DD
who need sleep assessment. Actigraphy (ACT) is an alternative objective, non-invasive,
cost-effective tool that measures movement via a watch-like device to score sleep and
wake. ACT has been extensively validated against PSG in typically-developing adults
and children, but there is limited validation in DDs. The purpose of this study was to
conduct an initial validation of ACT in the largest sample of children with DD and
associated motor impairment and intellectual disability to date. Fifty-one participants
wore an ACT device while undergoing standard-of-care PSG. Results indicated that the
low movement threshold setting of ACT analyses (i.e. less movement required to score
wake) had the best specificity without sacrificing excellent sensitivity, compared to the
medium and high thresholds. Overall, ACT had high ability to detect true sleep (0.93),
but weak ability to detect true wake (0.45). This pattern is consistent with the broad ACT
validation literature. However, these data show wide variability in agreement, indicating
ACT shows high agreement for some but poor agreement for others. Summary-level
comparisons and Bland-Altman plots showed ACT slightly over estimated onset latency
and sleep efficiency, over estimated total sleep time, but underestimated wake after sleep onset. There was wide variability in individual differences between PSG and ACT. There may be relationships with motor impairment and age, but future work with a larger sample is needed to elucidate these potential relations. This work is necessary to provide guidance on ACT’s use in DD populations with associated motor impairment and
intellectual disability