Background
There is a growing body of evidence that time spent sitting is an emerging health concern [
1]. Current findings have shown that sitting time is consistently associated with an increased risk of all-cause mortality [
2,
3] and numerous other negative health conditions such as obesity [
4], cardiovascular diseases [
5,
6], and type 2 diabetes mellitus [
7] as well as various other metabolic risk factors [
4,
8]. The common assumption that sufficient moderate-to-vigorous physical activity (MVPA) can compensate for prolonged sitting time must be corrected because sitting time has been found to increase the risk of various negative health outcomes independent of MVPA [
2,
9]. However, a recent review showed that the risk of premature all-cause death was attenuated but not diminished by physical activity (PA) levels and was in any case responsible for a substantial population-attributable risk fraction [
10]. Nevertheless, recent results from cross-sectional analyses suggest very little association between sitting time and cardio-metabolic biomarkers when total PA is adjusted for [
11], indicating that the interplay between sitting time and PA is not completely understood and is thus a sedentary behaviour research priority [
12]. The physiology of prolonged sitting and its relationship to health outcomes has not yet been elucidated [
13]. One suspected biological mechanism behind the adverse health effects is the following physiological response: through the absence of large muscle group contractions compared e.g., with standing, the lipoprotein lipase is suppressed that is necessary for healthy fat metabolism, and the break-down and use of glucose are reduced, which has been seen in animal models [
14]. Both mechanisms could lead to poor metabolic health with a long-term risk of different non-communicable diseases.
Sitting time is ubiquitous for most adults in developed countries and is most prevalent in three domains: in the workplace, during transport and during leisure time. The global tertiarisation of occupations as well as significant alterations in workplace environments and work practices have occurred in recent decades, largely driven by technological innovations such as computers and other labour-saving devices, resulting in sedentary work lives for many [
15,
16]. Research from Australia showed that 77% of working hours were spent sitting [
17]. Recent literature suggests harmful health consequences for prolonged sitting in workplace environments [
16,
18‐
20], and intervention studies designed to reduce this behaviour achieved the first results that support the assumption that interrupting prolonged sitting time during work is associated with reducing health risks [
21‐
24].
However, the correlates of prolonged sitting time in the workplace setting are not well understood [
25]. Overall, studies that examine correlates of sedentary behaviour are in early stages and, in most cases, are limited to overall sitting time [
26‐
28], neglecting specific domains; others often focus on TV viewing [
29,
30] or leisure-related sitting time [
8,
31,
32]. Hence, the need persists for future research that identifies the individual, social and ecological correlates of sitting time in specific domains [
33]. In addition, a gender-specific perspective is warranted based on previous findings that showed distinct gender differences concerning overall sitting time [
26] and domain-specific sitting time [
32] as well as work-related sitting time [
34]. Therefore, the socio-demographic correlates of workplace sitting are of interest for identifying the target groups with the highest need for effective workplace interventions. Separately, individual factors such as habits or attitudes towards prolonged sitting could potentially be associated with individual behavioural choices regarding reducing and reorganising sitting time during work and might be important correlates in future interventions. The first studies that concern leisure behaviour indicate that sedentary behaviours may be intentional and planned from a primary attitude base [
35]. Furthermore, it is important to learn more about the behavioural correlates of work-related sitting time. It is, for example, of interest whether MVPA behaviour in the different domains and sitting behaviour in contexts other than work are associated with sitting time during work to support healthy lifestyles, even when compensation for the prolonged sitting time during work cannot be achieved [
2].
Therefore, the aims of this study were to examine the gender-specific prevalence of work-related sitting time and to examine gender-specific associations between socio-demographic (i.e., age, education level, income level), behavioural (i.e., work-related PA, travel-related PA, leisure-related PA as well as sitting time during transport, during TV watching, during leisure computer use and during leisure time) as well as cognitive correlates (i.e., health-related beliefs about sitting time) and sitting time in the workplace.
Results
Participants reported a median sitting time during work of 2 hours (120 minutes), which accounted for 31.0 ± 24.9% of overall sitting time with no difference in genders. The age group 30–45 years had the longest sitting times at work among both men and women. Overall, the median work-related sitting time increased from 1 hour for 10 years of education (22.5 ± 23.5%) to 5 hours (45.2 ± 22.4%) for participants with university degrees. Concerning the cognitive variables, we found one gender-specific difference: women who agreed with the belief ‘When I sit for hours, I feel uncomfortable’ (see Table
1) had longer work-related sitting times.
Table 1
Work-related sitting time during weekdays, for all and stratified by gender (* = p < .05)
All
| 120 (28; 300) (n = 1515) | 120 (30; 300) (n = 747) | 150 (15; 300) (n = 768) | .66 |
Age
|
18–29 years | 120 (30; 360) (n = 225) | 120 (30; 270) (n = 142) | 240 (0; 372) (n = 83) | .10 |
30–45 years | 180 (30; 360) (n = 573) | 179 (30; 390) (n = 249) | 180 (25; 300) (n = 324) | .44 |
46–65 years | 120 (7; 270) (n = 717) | 120 (3; 300) (n = 356) | 120 (10; 240) (n = 361) | .85 |
Education level
|
No graduation | 0 (0; 19) (n = 6) | 150 (n = 1) | 0 (0; 0) (n = 5) | .03* |
10 years | 60 (0; 240) (n = 416) | 30 (0; 120) (n = 209) | 60 (0; 240) (n = 207) | .06 |
12 years | 120 (30; 270) (n = 584) | 120 (30; 240) (n = 260) | 120 (12; 300) (n = 323) | .73 |
13 years | 180 (30; 360) (n = 218) | 180 (20; 360) (n = 115) | 180 (59; 360) (n = 104) | .18 |
University degree | 300 (120; 420) (n = 279) | 300 (123; 420) (n = 150) | 240 (113; 420) (n = 129) | .49 |
Income groups Household net income/month |
<1500€ | 60 (0; 198) (n = 317) | 60 (0; 180) (n = 130) | 59 (0; 240) (n = 188) | .60 |
1500–2499€ | 120 (16; 300) (n = 518) | 120 (15; 300) (n = 260) | 179 (22; 300) (n = 258) | .20 |
>2.500€ | 180 (60; 360) (n = 531) | 180 (60; 360) (n = 292) | 180 (60; 360) (n = 239) | .86 |
Beliefs: ‘Sitting for long periods does not matter to me’ |
Strongly disagree | 60 (0; 240) (n = 348) | 60 (11; 180) (n = 131) | 90 (0; 300) (n = 217) | .20 |
Disagree | 180 (30; 360) (n = 314) | 179 (24; 300) (n = 162) | 180 (60; 360) (n = 151) | .16 |
Undecided | 180 (15; 300) (n = 352) | 180 (30; 331) (n = 172) | 180 (5; 300) (n = 181) | .63 |
Agree | 180 (30; 339) (n = 247) | 120 (30; 300) (n = 155) | 225 (30; 360) (n = 92) | .30 |
Strongly agree | 120 (5; 300) (n = 255) | 120 (29; 365) (n = 126) | 120 (5; 240) (n = 129) | .21 |
Beliefs: ‘When I sit for hours, I feel uncomfortable’ |
Strongly agree | 120 (5; 240) (n = 641) | 120 (30; 240) (n = 256) | 94 (0; 240) (n = 385) | .34 |
Agree | 180 (30; 360) (n = 343) | 180 (15; 360) (n = 187) | 240 (60; 383) (n = 157) |
.01*
|
Undecided | 180 (51; 300) (n = 234) | 180 (30; 300) (n = 129) | 148 (60; 300) (n = 105) | .78 |
Disagree | 60 (0; 300) (n = 158) | 60 (0; 360) (n = 93) | 98 (15; 240) (n = 64) | .52 |
Strongly disagree | 120 (5; 299) (n = 133) | 76 (5; 384) (n = 81) | 180 (5; 240) (n = 53) | .81 |
Bivariate correlations of the sitting- and PA-related behavioural variables were calculated (data not shown), and most correlations were small. The highest correlations were observed between work-related sitting time and work-related PA MET-minutes/day-1 with r = -.422 (p < .001) in men and r = -.321 (p < .001) in women.
Multiple linear regression analyses in men showed that model 1 with the PA-related behavioural correlates (18%) and model 3 with the socio-demographic correlates (13%) explained most of the variance (R
2) in work-related sitting times (see Table
2). Concerning PA behaviour, we found that sitting time during work was negatively correlated with work-related PA (β = -.43), suggesting increased sitting durations with less PA during work. Leisure-time physical activity was not a significant correlate. ‘Education’ (β = .29) and ‘income’ (β = .17) were both positively associated with ‘sitting time during work’, that is, increased education and income levels were positively associated with longer sitting times. Model 2 showed that TV-related sitting time (β = -.16) was negatively associated with work-related sitting time. In model 4, the belief ‘Sitting for long periods does not matter to me’ (recoded) (β = .10) was positively correlated with work-related sitting time, reflecting more positive attitudes towards sitting with increasing sitting durations.
Table 2
Results from the multiple linear regressions on the contributions of the socio-demographic, behavioural and cognitive correlates to the dependant variable “sitting time during work” for men
Work-related PA MET minutes/day-1
| -.306 | .024 | -.425 | ≤.001*** | | | | | | | | | | | | |
Transport-related PA-1
| .069 | .094 | .025 | .461 | | | | | | | | | | | | |
MET-minutes/day | | | | | | | | | | | | | | | | |
Leisure-related PA | -.120 | .127 | -.032 | .346 | | | | | | | | | | | | |
MET-minutes/day-1
| | | | | | | | | | | | | | | | |
Sitting time transport | | | | | .035 | .106 | .028 | .768 | | | | | | | | |
Sitting time TV | | | | | -.382 | .090 | -.156 | <.001*** | | | | | | | | |
Sitting time computer | | | | | .166 | .112 | .058 | .137 | | | | | | | | |
Sitting time leisure | | | | | -.113 | .095 | -.046 | .234 | | | | | | | | |
Age | | | | | | | | | .582 | .516 | -.041 | .260 | | | | |
Education level | | | | | | | | | 35.348 | 4.430 | .294 | ≤.001*** | | | | |
Income level | | | | | | | | | 37.777 | 8.106 | .170 | ≤.001*** | | | | |
‘Sitting for long periods does not matter to me’ b
| | | | | | | | | | | | | 12.801 | 4.726 | .104 | .007** |
‘When I sit for hours, I feel uncomfortable’ a
| | | | | | | | | | | | | -2.527 | 4.710 | -.021 | .592 |
R
2
| 0.179 | 0.009 | .130 | .010 |
Among women, multiple linear regression also showed the highest explained variance in models 1 (11%) and 3 (10%) (see Table
3). The PA-behavioural correlates showed that work-related (β = -.32) and transport-related PA (β = -.07) were both negatively associated with work-related sitting time; sitting time during work increased with decreased PA during transport and work. Leisure-related PA was also not associated with work-related sitting time. Concerning socio-demographics, age (β = -.14) was negatively correlated with work-related sitting time, showing longer work-related sitting times in younger women. ‘Education’ (β = .21) and ‘income’ (β = .13) were both positively associated with ‘sitting time during work’, demonstrating that education and income levels increased with increased sitting time during work. Model 2, with sitting-related behaviour correlates, found negative associations with TV-related sitting time (β = -.16), that is, decreased TV sitting time with increased sitting time during work. For the cognitive variables, we found no associations.
Table 3
Results from the multiple linear regressions on the contributions of the socio-demographic, behavioural and cognitive correlates to the dependant variable “sitting time during work” for women
Work-related PA MET-minutes/day-1
| -.330 | .035 | -.324 | ≤.001*** | | | | | | | | | | | | |
Transport-related PA | -.163 | .077 | -.073 | .034* | | | | | | | | | | | | |
MET-minutes/day-1
| | | | | | | | | | | | | | | | |
Leisure-related PA | .112 | .135 | .028 | .410 | | | | | | | | | | | | |
MET-minutes/day-1
| | | | | | | | | | | | | | | | |
Sitting time transport | | | | | .103 | .135 | .028 | .443 | | | | | | | | |
Sitting time TV | | | | | -.382 | .090 | -.156 | ≤.001*** | | | | | | | | |
Sitting time computer | | | | | -.052 | .093 | -.020 | .579 | | | | | | | | |
Sitting time leisure | | | | | -.161 | .093 | -.063 | .083 | | | | | | | | |
Age | | | | | | | | | -2.254 | .610 | -.138 | ≤.001*** | | | | |
Education level | | | | | | | | | 33.019 | 6.279 | .205 | ≤.001*** | | | | |
Income level | | | | | | | | | 27.690 | 8.297 | .128 | .001** | | | | |
‘Sitting for long periods does not matter to me’ b
| | | | | | | | | | | | | -0.216 | 4.368 | -.002 | .961 |
‘When I sit for hours, I feel uncomfortable’ a
| | | | | | | | | | | | | 3.201 | 4.900 | .024 | .514 |
R
2
| .109 | .032 | .100 | .001 |
Conclusion
The present study gives initial insights into the gender-specific socio-demographic, behavioural and cognitive correlates of work-related sitting in the working German population. The present findings showed that in particular, higher educated men and women as well as young women need special attention when developing interventions to reduce prolonged work-related sitting times. Furthermore, our findings suggest considering increases in transport-related PA, especially in women, as well as promoting leisure-related PA for populations with long work-related sitting durations. Only weak associations with the cognitive correlates were found. Here, future research needs to address the specificity of psychosocial as well as environmental correlates and possible associations with specific domains of sitting to obtain more fundamental insights into these associations.
Competing interests
The authors declare that they have no financial or non-financial competing interests.
Authors’ contributions
BWS participated in the conception and the design of the present study and performed statistical analyses interpreted the data and wrote and drafted the manuscript. JB and SvS contributed to the analyses and interpretation of data and provided critical revision of the manuscript. IF participated in the conception and design of the study. All authors read and approved the final manuscript.