JMIR Form Res. 2022 Sep 22. doi: 10.2196/38562. Online ahead of print.
ABSTRACT
BACKGROUND: The COVID-19 pandemic has greatly boosted working from home as a way of working, which is likely to continue for most companies in the future, either in fully remote or hybrid form. To manage stress levels in employees working from home, insights into the stressors and de-stressors in a home office first need to be studied. We present an international remote study with employees working from home by making use of state-of-the-art technology, i.e. smartwatches and questionnaires through smartphones.
OBJECTIVE: Firstly, to determine stressors and de-stressors in people working from home. Secondly, to identify smartwatch measurements that could represent these stressors and de-stressors.
METHODS: Employees working from home from three regions of the world (United States, United Kingdom, and Hong Kong) were asked to wear a smartwatch continuously for seven days, fill in five questionnaires each day and two additional questionnaires before and after the measurement week. The entire study was done remotely. Univariate statistical analyses comparing variable distributions between low and high stress levels were followed by multivariate analysis using logistic regression, considering multicollinearity by using Variance Inflation Factor filtering.
RESULTS: A total of 202 people participated, with 198 participants finishing the experiment. Stressors found are other people and daily life getting in the way of work (P=.05), job intensity (P=.007), a history of burn-out (P=.03), anxiety towards the pandemic (P=.04) and environmental noise (P=.008). De-stressors found are access to sunlight (P=.02) and fresh air (P<.001) during the workday and going outdoors (P<.001), having breaks (P<.001), exercising (P<.001), and having social interactions (P<.001). The smartwatch measurements positively related to stress were number of active intensity periods (P<.001), number of highly active intensity periods (P=.04), steps (P<.001) and the standard deviation in heart rate (P<.001). In a multivariate setting only history of burn-out (P<.001) and family and daily life getting in way of work (P<.001) were positively associated with stress, while self-reports of social activities (P<.001) and going outdoors (P=.03) were negatively associated with stress. Stress prediction models based on questionnaire data had similar performance (F1=.51) compared to models based on automatic measurable data alone (F1=.47).
CONCLUSIONS: The results show that there are stressors and de-stressors when working from home that should be considered when managing stress in employees. Some of these stressors and de-stressors are (in)directly measurable with unobtrusive sensors, and prediction models based on this data show promising results for the future of automatic stress detection and management.
CLINICALTRIAL: This study is registered under Registration ID NL9378 in the Dutch Trial Register (NTR).
PMID:36265030 | DOI:10.2196/38562
