Feasibility study of wearable devices in predicting postpartum depressive symptoms during the puerperium

Abstract

Background: With the development of technology, wearable devices have gradually become a popular remote health monitoring tool, which have been demonstrated to be a simple and feasible way to record physiological data. Postpartum depression (PPD) is a serious health problem that cannot be ignored, and sleep, physical activity, and body temperature are reported to be associated with the development of PPD in some research. However, in most studies, such kind of data are always limited to self-reported information. Some studies have demonstrated some discrepancies between subjective and objective data. So far, research investigating the relationship between wearable-based objective data and PPD symptoms in China is rare. The prediction power of wearable-based sleep, physical activity, and body temperature on PPD symptoms during the puerperium needs to be further explored. Objective: The purpose of this observational study was to explore the feasibility of sleep, steps and body temperature derived from wearable devices in predicting PPD symptoms among puerperium women. Methods: From September 2022 to December 2022, 84 pregnant women who met the inclusion criteria were selected from the Fetal Heart Rate Monitoring room of the Obstetrics and Gynecology Hospital of Fudan University in Shanghai. Participants were followed up for 42 days postpartum. The Pittsburgh Sleep Quality Index Scale (PSQI) was used to assess subjective sleep quality during this period. Objective sleep quality, physical activity, and body temperature of participants were recorded with smart watches. The Edinburgh Postpartum Depression Scale (EPDS) was used to assess their mental health at the end of the puerperium. After the study, a questionnaire about their experience of wearing smart watches during the puerperium was completed. The correlation and agreement of subjective and objective sleep were analysed by Spearman correlation analysis and Bland-Altman plot. A binary logistics regression model was constructed to predict the occurrence of PPD symptoms. Findings: When investigating the acceptability, in terms of the influencing factors of women’s wearing smart watches, being busy taking care of babies, forgetting or did not want to synchronize data, and inconvenient breastfeeding were the top three reasons. In terms of participants’ experience about wearing the smart watch during the puerperium, 27 (40.3%) reported that they needed assistance during the process, and sleep monitoring, time checking, and heart rate monitoring were the top three functions of interest to postpartum women. Objective sleep efficiency (SE), total sleep time (TST), light sleep, deep sleep, rapid eye movement, wake after sleep onset (WASO) were correlated with subjective sleep duration, habitual sleep efficiency, and global PSQI score with r ranging from 0.253 to 0.504. The Bland-Altman plot showed that although the mean difference between subjective and objective TST was only 0.05 h, the limits of agreement ranged from -2.21 to 2.1 h. Although the mean difference between subjective and objective SE was only 0.06, the limits of agreement ranged from -0.31 to 0.19. Participants with PPD symptoms history and each increase in global PSQI scores were associated with 23.602-fold and 1.431-fold increased odds of having PPD symptoms during the puerperium respectively. Conclusions: The occurrence of PPD symptoms during the puerperium is related to subjective sleep quality rather than objective sleep measures. Strategies to boost subjective sleep quality should be considered during postnatal health care to improve women’s mental health status.