Onsite versus home-office: differences in sleep patterns according to workplace

In summary, it was found that onsite workers had significantly earlier sleep timing and midpoints of sleep on weekdays compared to home-office workers, whereas there were no clinically meaningful differences regarding sleep efficiency, sleep duration, and variability in sleep timing. Importantly, also SJL did not differ between the groups. The results of the group comparison showed that the onsite group was significantly older than the home-office group. The group means differed by 4.5 years. This age difference is particularly important when considering the present results. Chronotype, as an individual difference trait, is subject to changes in expression over the course of a person’s life [32]. While young children tend to be morning types, this tendency changes strongly during adolescence [32]. Adolescents show a distinct delay in chronotype, peaking at the age of 20 years [27]. With the end of adolescence, the daytime preference shifts again and begins to advance. Roenneberg and colleagues [27] showed a continuous advancement of chronotype from the age of 20 until about 65 years. There is also a steeper regression line between the ages of 20 and 45 years than between 45 and 65 years. Most adults do not belong to an extreme chronotype and are therefore described as neither type [32]. In the sample of Randler and colleagues [32], a distribution of 21.1% evening, 71.7% neither, and 7.2% morning type was reported for the 26-year-old subjects. Among the 30-year-old subjects, 11.9% were classified as evening type, 74.8% as neither, and 13.3% as morning type. This means that the 30-year-old subjects were 9.2% less likely to be classified as evening type than the 26-year-old subjects. As the average age of the present sample groups differed by about 4.5 years (mean 25.8 years in home-office, mean 30.3 years in onsite), it could be assumed that also the chronotypes differ significantly from each other. Participants in the home-office group may therefore be more evening oriented, reflecting a higher proportion of evening types in this group than in the onsite group.

In this study two different measures of chronotype were used (MSFsc: clock time-based measure, MESSi: continuous morningness-eveningness measure). The analysis of the MSFsc and the MESSi showed no significant differences between the groups. Therefore, the expression of chronotype did not differ between the groups.

Regarding the MSFsc variable, Roenneberg and colleagues [33] described a comparable shape and width of the distribution in all age groups, with a varying mean value. Graphical comparison of the group histograms (see Supplement 1) with the expected values ([33], N ≥ 55,000) showed that the distribution in the home-office group had a similar shape. In the onsite group, two equally high peaks existed, one at 3:00 a.m. and one at 4:30 a.m., showing the shape of the distribution to be different from the reference. In the sample of Roenneberg and colleagues [33], the 26-year-old subjects showed peaks at around 4:45 a.m. compared to 4:17 a.m. for the home-office group. For the 30-year-old participants, Roenneberg and colleagues [33] reported an average MSFsc of approximately 4:30 a.m., compared to 3:51 a.m. for the onsite group. Taken together, these two discrepancies led us to consider the MSFsc of the onsite group to be unrepresentative.

In summary, a) the expected distribution of chronotype based on the age groups of the samples and b) the finding that the distribution of chronotype in the onsite group cannot be considered representative suggest that the chronotypes of the groups can generally be expected to differ. The younger group would normally have a higher prevalence of evening types/more pronounced eveningness, although this was not the case in the present samples.

Increased eveningness is associated with a higher risk of SJL. SJL has repeatedly been shown to be particularly detrimental to public health (e.g., cardiovascular outcomes, psychiatric disorders, obesity [34]). However, softer factors such as productivity at work are also significantly affected by SJL. To test whether the measured SJL in the home-office group was within the expected range, the results were compared to an earlier sample of the authors’ research group [13]. In that study, only people who worked in home-office during the COVID-19 pandemic had been surveyed. The mean age of the sample was reported to be ~ 28.6 years (SD: ~ 10.5 years), which was not different compared to the sample of the present study (p = 0.11; Supplement 2). With regard to SJL, the means of the present home-office group (0.80, SD: 0.73 h) and the previous sample (0.82; SD: 0.70 h) were also not significantly different (p = 0.87; Supplement 2). Furthermore, the chronotype within the two groups measured by MSFsc did not differ (p = 0.21; Supplement 2). It can therefore be assumed that the SJL in the present sample is representative for this age group working from home. In another sample of the same age who were not working from home, the SJL was almost 1 h higher (1.78 h) compared to the present sample with 0.8 h [35]. It could be postulated that the implementation of home-office might potentially be beneficial for young people to reduce SJL. The onsite group had an average SJL of 1.03 h, which was also compared to other samples in the current literature. As no age-matching sample could be found in the literature, the present results were compared to slightly younger and older samples. Both showed a meaningful but equal age difference compared to the present sample. Interestingly, the SJL (1.4 ± 1.0, [36]; 1.33 ± 0.88 h, [37]) of the two studies did not significantly differ (p = 0.42, t = 0.78, df = 727, Standard Error of Difference (SED) = 0.09). Therefore, these results were used as representative reference values for SJL in the current sample. Even the value closer to the current study [37] differed significantly from the SJL reported herein (p = 0.04, t = 2.07, df = 637, SED = 0.15; Supplement 2). This led to the conclusion that the SJL of the onsite group cannot be considered representative, as it was significantly below the expected range.

Although the SJL of the present two study groups were not significantly different, there was a tendency towards lower SJL in the home-office group. This tendency was reinforced by an overrepresentation of low SJL in the onsite group, while the home-office group showed more similar values to age-comparable samples. Therefore, the current results suggest that working from home might help to reduce SJL.

The present study differs from the authors’ previous studies which examined only home-office participants and not two samples in different working environments. As spatial flexibility, in addition to temporal flexibility [14], may play a role in the increasing irregularity of students’ sleep schedules, a possible relationship was investigated here.

The sleep timing of the two groups on workdays differed significantly. The home-office group went to bed later and also got up later. However, both groups had similar fluctuations regarding sleep timing variability and no difference in overall nightly sleep duration. Thus, home-office had no negative influence on sleep variability and duration in this sample. There was also no influence of spatial flexibility/working environment on the loss of sleep duration/regularity observed. Instead, data support the hypothesis that the ability to work with unrestricted temporal flexibility is the main influencing factor of the increased sleep timing regularity in students, as discussed in one of the authors’ previous studies [14] by excluding another possible influencing factor.

Sleep efficiency did not differ between the two groups. However, a difference within the groups could be detected. In the onsite group, sleep efficiency was higher on weekdays than at weekends. The opposite was true for the home-office group, where sleep efficiency was higher at weekends. The reason for this could be explained by the way the sleep efficiency variable is calculated. On weekdays, participants in the onsite group could have either gotten up immediately after waking up (e.g., because of the increased pressure to get up on time in order to commute to work compared to the weekend) or gone straight to sleep after going to bed (e.g., because of the increased exhaustion from onsite work with probably fewer breaks compared to the weekend). This would explain the high ratio of total sleep time to time in bed, which corresponds to a higher sleep efficiency. At the weekend, on the other hand, they may have stayed in bed even longer although they were already awake, or it took longer for them to fall asleep, resulting in lower sleep efficiency. In contrast, the opposite was true for the home-office group, with a higher sleep efficiency at weekends compared to weekdays. Although statistically significant, this effect was small and not clinically meaningful.

留言 (0)

沒有登入
gif