Assessment of Sleep Duration and Number of Awakenings Based on Ankle and Wrist Actigraphy in Medical Hospitalized Older Patients

Abstract

Background Most studies performed in the hospital assess sleep using self-reports; few rely on actigraphy. Although wrist actigraphy is commonly used for sleep assessment in field studies, in-hospital assessment may be challenging and cumbersome because other more necessary monitoring devices are often attached to patients’ upper limbs; these may affect interpretation of wrist activity data. Placement on the ankle may be a viable solution. Objective To compare total sleep time (TST) and number of awakenings (NOA) using concomitant wrist and ankle actigraphy, as well as self-reports in a sample of older adult patients hospitalized in medical units. Methods This was a prospective observational study. Objective sleep data were collected using ankle and wrist actigraphy, and subjective data using sleep diary. Repeated measures mixed model analysis was performed, adjusting for age, gender, sleep medications, symptoms severity, interaction between types of measure, and night number. Results Twenty-one older adults (65+) wore ankle and wrist actigraphy devices and subjectively estimated sleep parameters for an average of (2.15 ± 1.01) nights, with 40 nights available for analysis. TST was lower for wrist than ankle actigraphy ( F(2,87) = 7.92, p = .0007). Neither differed from self-reports. NOA differed between all types of measure (ankle, 8.58 ± 6.66; wrist, 15.49 ± 7.47; self-report, 1.81 ± 1.83; F(2,85) = 47.66, p < .001). No significant within-subject variations and no interaction between devices and repeated measures were found. Conclusions Despite differences between ankle and wrist assessments, all three methods provided consistent TST estimation within participants. Findings provide preliminary support for the use of ankle actigraphy for sleep assessment in hospital settings.