Uso de smartwatch e da plataforma sênior saúde móvel para avaliação da qualidade do sono em idosos comunitários: um estudo comparativo entre métodos

Abstract

Introdution: it is estimated that approximately 50% of the elderly population has insufficient sleep, impairment due to occasional interruptions, lack of time spent on restorative adjustments and reduced total sleep time, often correlated with chronic degenerative diseases, physical-functional and cognitive declines. Considering that sleep quality is an important indicator of health status, several tools have been used to assess it, including wearable devices such as smart watches with integrated remote monitoring platforms, which allow continuous longitudinal data collection sleep goals. Objective: to examine the correlation and agreement between sleep quality parameters of community-dwelling elderly people measured by subjective (Pittsburgh Sleep Quality Index-PSQI) and objective (Fitbit smartwatch and Senior Saúde Móvel platform) methods. Methodology: this is an observational research, descriptive-correlational and analytical, with a longitudinal design, with a quantitative approach. The sample was non-probabilistic, consisting of 60 individuals aged between 60 and 90 years. Participants were monitored remotely for a period of 7 consecutive days. To evaluate the correlation and agreement between objective and subjective measurements (continuous variables), Spearman's Correlation Coefficient, Bland-Altman graphic analysis and Student's t-test or Wilcoxon test were applied. Comparison between sleep quality classifications was established using the Kappa test. Results: significant correlations were found between the mean total sleep time from the PSQI and the mean total sleep time from the smartwatch (p=0.033) and between the mean sleep onset latency estimated by both methods (p=0.018). No significant correlations were identified between sleep efficiency measures. The Bland-Altman plot revealed that the smartwatch tended to slightly underestimate the PSQI parameters total sleep time (-12.3 ± 1min) and sleep latency (-5.97 ± 23min), but the differences between the measurements were within of confidence limits, demonstrating substantial agreement between the methods. The results were confirmed by Student's t-test or Wilcoxon test with p>.05, showing that the means were not significantly different. However, the mean differences in sleep efficiency estimated by Fitbit were considered higher than those of the PSQI and statistically significant (15 ± 17.17%; p<.05). Regarding comparisons between global sleep quality classifications, the results obtained from the Kappa test demonstrated good agreement between the two forms of classification tested in comparison to the PSQI, however, the set of measures derived from the PSQI and Fitbit proved to be slightly superior for tracking poor sleep quality to classification based solely on objective data estimated by Fitbit. Conclusion: in a sample of community-dwelling elderly people with good cognitive capacity and without serious sleep disorders, Fitbit parameters present acceptable mean differences and indicate significant correlations with the PSQI, the gold standard method for screening poor sleep quality. Such findings suggest that the use of Fitibit smartwatches with an integrated monitoring platform to track possible poor sleep quality may be appropriate. Furthermore, the adopted platform could ultimately be an effective tool for long-term home monitoring of sleep-wake behavior. However, more validation studies are needed.