Comparison between Three-Night and Single-Night Actigraphic Sleep Assessment During Early Hospitalization

Abstract

To determine whether a single-night actigraphic sleep assessment during early hospitalization can approximate a 3-night average and characterize inpatient sleep- and rest-activity rhythm.

In the present multicenter, cross-sectional study conducted in 7 Czech hospitals, 64 inpatients underwent wrist actigraphic monitoring during the first 3 nights of hospitalization. Actual sleep time, sleep latency, mobile time, sleep efficiency, fragmentation index and average light exposure were analyzed. First-night values were compared with 3-night means using the Wilcoxon test. Individual-level agreement was evaluated using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. Non-parametric circadian rhythm analysis (NPCRA) variables were reported descriptively.

The 3-night mean sleep duration was 342 ± 83 minutes, with reduced sleep efficiency and increased fragmentation. No significant differences were observed between first-night values and 3-night averages for actual sleep time, mobile time, sleep efficiency, fragmentation index or average light exposure. Agreement was good for these parameters (ICC_single ≈ 0.71–0.75), whereas sleep latency showed lower reliability (ICC_single = 0.51). The Bland-Altman analysis demonstrated no systematic bias, with up to 6.3% of observations outside the 95% limits of agreement for most parameters and 9.4% for sleep latency. The NPCRA suggested preserved relative amplitude but fragmented rest-activity rhythm.

Single-night actigraphy provides reliable estimates for most inpatient sleep parameters and may represent a pragmatic alternative to multi-night monitoring in acute hospital settings. Sleep latency is less stable and may require repeated assessment.