OBJECTIVES: Gold-standard overnight polysomnography does not reliably capture highly variable sleep patterns across the 24-hour day that are common with dementia and often problematic for carers. We evaluated the reliability of automatically scored actigraphy data as an alternative. METHODS: Actigraphy recordings were analysed from 15 community-dwelling people with dementia (135 days total) and 14 of their family carers (124 days total). Manual scoring used participant sleep diaries to identify sleep periods. Automated scoring used the manufacturer's algorithm to score entire records. RESULTS: For people with dementia, automated scoring identified more sleep fragmentation at night and increased sleep during the day, with comparable sensitivity but lower specificity than for carers. CONCLUSIONS: Automated scoring offers reasonable agreement with manual scoring and may better describe the fragmented nature of dementia-related sleep, which can be challenging to record accurately in a sleep diary. Automated scoring reduces participant burden and could improve research and treatment protocols.
OBJECTIVES: Gold-standard overnight polysomnography does not reliably capture highly variable sleep patterns across the 24-hour day that are common with dementia and often problematic for carers. We evaluated the reliability of automatically scored actigraphy data as an alternative. METHODS: Actigraphy recordings were analysed from 15 community-dwelling people with dementia (135 days total) and 14 of their family carers (124 days total). Manual scoring used participant sleep diaries to identify sleep periods. Automated scoring used the manufacturer's algorithm to score entire records. RESULTS: For people with dementia, automated scoring identified more sleep fragmentation at night and increased sleep during the day, with comparable sensitivity but lower specificity than for carers. CONCLUSIONS: Automated scoring offers reasonable agreement with manual scoring and may better describe the fragmented nature of dementia-related sleep, which can be challenging to record accurately in a sleep diary. Automated scoring reduces participant burden and could improve research and treatment protocols.