Use of sensing technologies to assess sleep quality and physical activity levels in nursing home residents with dementia taking nightly psychotropic drugs for sleep disturbance: a cross-sectional study.

Sleep disturbances are common in people with dementia and nightly prescribed psychotropic drugs, such as sedatives or antidepressants, can increase risks such as injury, inactivity, and behavioral symptoms. Treatment decisions currently rely on periodic, proxy-rated questionnaires that may miss important daily fluctuations in sleep and activity. We explore whether sensing technologies provide insights into distinct differences in sleep characteristics and activity levels in nursing home residents with dementia who are prescribed nightly psychotropic drugs for sleep disturbances. Forty-seven participants were recruited from four nursing homes in Bergen, Norway, and stratified according to prescribed nightly psychotropic drug use for sleep disturbance: 1) none, 2) medications with short half-lives, 3) medications with long half-lives. Garmin Vivoactive5 and Venu3, Vital Things Somnofy sleep monitor, and traditional questionnaires (Physical Self Maintenance Scale and Neuropsychiatric Inventory-Nursing Home version) were used for data collection. Digital metrics included Euclidean Norm Minus One (ENMO; day/night/24-h), Sleep Regulatory Index (SRI), Sleep Efficiency (SE), Total Sleep Time (TST), Sleep Fragmentation Index (SFI), and no presence (time out of bed). Thirty participants (73-100 years old) were included for analysis. Groups taking psychotropic medications were awake for longer periods (WASO: chi2 = 8.7, p = 0.01) and had poorer sleep regularity (SRI: chi2 = 20.6, p = 0.0001). Participants taking psychotropic drugs had less physical activity (day/night/24-h ENMO), with greatest differences between those on medications with a long half-life (day: chi2 = 9.48, p = 0.009; night: chi2 = 12.83, p = 0.002; 24-h: chi2 = 8.23, p = 0.02) and those not on nightly psychotropic medications. The digital biomarkers collected using the selected sensing technologies offered nuanced information regarding sleep behaviors and physical activity levels, providing detailed distinction between the groups. Sensing technologies may be a promising companion to the currently used proxy-rated assessment tools for sleep disturbance and physical activity levels for people with dementia residing in nursing homes.