Preprint · Mar 12, 2026
A one-year observational study using digital biomarkers from sensing technologies to assess changes in physical activity levels and sleep quality in nursing home residents with dementia. (Preprint)
By Boyle LD, Pătraşcu M, Husebo BS et al.
BACKGROUND Proxy-rated questionnaires remain the standard for assessment of activity and sleep for people with dementia living in nursing homes. Sensing technologies, such as wearables, can generate continuous data that provides quantitative insights into daily activities and behavioral and psychological symptoms, such as sleep disturbance. This study explores the utility of sensing technologies in the detection of changes in physical activity levels and sleep behaviors over time. OBJECTIVE This study aims to explore the long-term capabilities of multi-modal sensing technologies for assessing physical activity levels and sleep quality using selected digital biomarkers for nursing home residents with dementia. Objectives were for observation to be aligned with real-world conditions in which such sensing technologies would be applied within a nursing home environment, and to assess whether distinct differences in selected digital biomarkers can be observed accurately and reliably longitudinally. METHODS This study included eleven participants (79–93 years) recruited from two dementia care units in Norway. A smartwatch (Garmin Vivoactive5/Garmin Venu3) and radar-based system (Vital Things, Somnofy) were used to collect 7-days and 6-nights of data on physical activity levels and sleep quality at baseline, 6 months, and 1-year. The Personal Self-Maintenance Score and Neuropsychiatric Inventory–Nursing Home Version (nighttime behaviors section K), were also administered. Digital biomarkers included Euclidean Norm Minus One (ENMO), Sleep Efficiency (SE), Wake After Sleep Onset (WASO), Sleep Regulatory Index (SRI), Sleep Fragmentation Index (SFI), Total Sleep Time (TST), and time out of bed (no presence). RESULTS Nine total participants were included in the final analysis. Differences were found in Nighttime ENMO (P= .01) and in four of the sleep biomarkers: TST (P=.02), SE (P= .02), WASO (P= .01), and SRI (P=.01). Long-term reliability of the group ENMO was poor (ICC: 0.00-0.02), however, between individuals at each timepoint was moderate - strong (0.58-0.79). Adherence and acceptability of the technologies was high (88-96%), and application of the devices was well tolerated by the participants with no adverse events. CONCLUSIONS The use of sensing technologies could enable more objective, data-driven future care models for people with dementia residing in nursing homes, however, the results emphasized in this study require the recommendation for cautious, well-designed use of digital biomarkers for clinical decision-making. CLINICALTRIAL The DIPH.DEM study was approved by the Regional Committee for Medical and Health Research Ethics (REK) in Norway in October 2023: approval number 634938.