Feasibility of community‐based remote, biometric data collection in persons with Alzheimer’s disease and behavioral symptoms

Abstract

AbstractBackgroundAging at home with Alzheimer’s disease (AD) is progressively difficult due to cognitive dysfunction interfering with persons’ functional activity and behavioral regulation. Community‐based collection of biometric data via wearable devices provide an innovative approach as a measure of behavioral symptomology. This study analyzes feasibility of remote, biometric data collection during a non‐pharmacological, home‐based randomized controlled trial of older adults with AD.MethodsIndividuals, aged 65 or older, with diagnosis of AD as primary dementia type and behavioral symptoms who resided in the community were included in this prospective study. ActiGraphs were mailed to care partners who implemented device use and management. Feasibility was assessed via accepting and wearing of the device by the person with AD through care partner facilitation at baseline and post‐intervention (continuous wear for 6 days at both time points). Baseline actigraphy data (mean motor activity; MMA) were assessed to determine utility of acceptance/wearing the device as means to collect biometric data. Secondary analyses assessed Pearson correlation of demographic information including cognition measured by the Montreal Cognitive Assessment and the Neuropsychiatric Inventory‐Questionnaire (NPI‐Q).ResultsTwenty‐one participants with AD engaged in the study. Eighteen participants completed baseline data collection (three caregivers refused the device on behalf of person with AD). Fourteen participants or 66.7% (95% CI: 46.6% to 86.8%) also completed post‐intervention data collection (one dropout in mid‐study and three care partners refused the device at second time point). Baseline MMA data were measured: 24‐hour MMA = 608,913; daytime MMA = 253,534; evening MMA = 308,113; nighttime MMA = 47,079. Age (x̄ = 77.8±1.7; r = ‐.56, p‐value = 0.038) was negatively correlated and cognitive impairment (x̄ = 11.13±2.3; r = .789, p‐value = 0.002) was positively correlated with 24‐hour MMA. NPI‐Q total and emotional items (agitation, anxiety, liability) were not correlated in this sample.ConclusionCommunity‐based biometric data collection for older adults with AD and behavioral symptoms is feasible. Behavioral disturbance and care burden likely lowered acceptability of the device. Future power calculations anticipating 66% compliance maybe sufficient for primary outcome analyses in community‐based data collection. Additional studies are needed to validate wearable devices as a primary outcome measure for behavioral symptoms in community‐based participants as this study was not sufficiently powered for such analyses.