The aim of the study was to investigate if (a) substituting total sedentary time or long sedentary bouts with standing or various types of physical activity and (b) substituting long sedentary bouts with brief sedentary bouts; is associated with obesity indicators using a cross sectional isotemporal substitution approach among blue-collar workers.
A total of 692 workers from transportation, manufacturing and cleaning sectors wore an Actigraph GT3X+ accelerometer on the thigh for 1–4 working days. The sedentary (sit and lie), standing, walking, and moderate to vigorous physical activity (MVPA) time on working days was computed using validated Acti4 software. The total sedentary time and uninterrupted sedentary time spent in brief (≤5 mins), moderate (>5 and ≤30 mins), and long (>30mins) bouts, were determined for the whole day and during work and non-work time separately. The obesity indicators, BMI (kg/m2), waist circumference (cm) and fat percentage were objectively measured. Isotemporal substitution modelling was utilized to determine the linear association with obesity indicators of replacing 30 min of total sedentary time or long sedentary bouts with standing, walking or MVPA and separately replacing 30 min of long sedentary bouts with brief sedentary bouts.
Workers [mean (standard deviation, SD); age = 45.1 (9.9) years, BMI = 27.5 (4.9) kg/m2, %BF = 29.6 (9.5), waist circumference = 94.4 (13.0) cm] sat for 2.4 hours (~32% of the measured time, SD = 1.8 hours) across the day during work period and 5.5 hours (~62% of the measured time, SD = 1.5 hours) during non-work period. Most of the sedentary time was accrued in moderate bouts [work = 1.40 (SD = 1.09) hours] during work and in long bouts during non-work [2.7 (SD = 1.4) hours], while least in long sedentary bouts during work [work = 0.5 (SD = 0.9)] and in brief sedentary bouts [0.5 hours (SD = 0.3)] during non-work. Significant associations with all obesity indicators were found when 30 min of total sedentary time or long sedentary bouts were replaced with standing time (~1–2% lower) or MVPA (~4–9% lower) during whole day, work, and non-work periods. The exception was that a statistically significant association was not observed with any obesity indicator when replacing total sedentary time or long sedentary bouts with standing time during the work period. Significant beneficial associations were found when replacing the long sedentary bouts with brief sedentary bouts (~3–5% lower) during all domains.
Replacing total sedentary time and long sedentary bouts, respectively, not only with MVPA but also standing time appears to be beneficially associated with obesity indicators among blue-collar workers. Additionally, replacing long sedentary bouts with brief sedentary bouts was also beneficially associated with obesity indicators. Studies using prospective design are needed to confirm the findings.
Citation: Gupta N, Heiden M, Aadahl M, Korshøj M, Jørgensen MB, Holtermann A (2016) What Is the Effect on Obesity Indicators from Replacing Prolonged Sedentary Time with Brief Sedentary Bouts, Standing and Different Types of Physical Activity during Working Days? A Cross-Sectional Accelerometer-Based Study among Blue-Collar Workers. PLoS ONE 11(5): e0154935. https://doi.org/10.1371/journal.pone.0154935
Editor: Pedro Tauler, University of the Balearic Islands, SPAIN
Received: November 25, 2015; Accepted: April 21, 2016; Published: May 17, 2016
Copyright: © 2016 Gupta et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
The worldwide prevalence of obesity has almost doubled since 1980 . Based on recent estimates in the European Union, 30–70% of adults are obese  with particularly high obesity rates among workers in lower socioeconomic groups, such as blue-collar workers [3–6]. Obesity is associated with an increased risk for all-cause mortality , the metabolic syndrome, diabetes, cardiovascular diseases, cancer, arthritis and disability [8–10].
One of the potential risk factors for obesity is prolonged sedentary time which has increased in western countries with many adults spending as much as 70% of their waking hours sitting . Sedentary time refers to the time spent on activities (i.e., sitting and lying) that do not increase energy expenditure substantially above resting level (1.0–1.5 METs) [12, 13]. Among office workers, high rates of sitting have been observed (on average 75% of working hours using objective, postural-based measures) [14–16]. Among blue collar workers, the sitting time at work is typically lower, but high levels of sedentary time have been observed during non-work time [17, 18]. Also, due to the nature of work tasks (e.g. long transportation, assembly line work or surveillance) and organizational aspects (e.g., low decision latitude) among blue-collar workers, they may have limited autonomy to reduce or break up prolonged sedentary time during work.
Epidemiological studies have found a positive association between total sedentary time and obesity [19–21], even after statistical adjustment for the level of moderate to vigorous physical activity (MVPA) [22–26]. Recent similar studies have also addressed the importance of sedentary time patterns (how sedentary time is distributed across the day) with respect to obesity [27, 28]. Some experimental and epidemiological studies have found long sedentary bouts to be associated with, for example, biomarkers of cardio-metabolic health [29, 30]. In contrast, brief sedentary bouts are not considered a health hazard to the same extent [27–29, 31].
An emerging method to investigate the associations between sedentary time and health outcomes is isotemporal substitution analysis. This method calculates the estimates of the association when replacing time spent on one activity with another, while keeping the total time fixed (isotemporal) [32, 33]. Thus this method helps to answer a very important question, how to spend our discretionary time per day for minimizing health risk and achieving optimum health. It has currently been used to explore the association between various types of physical activity or sedentary behaviors and depression, cardiovascular disease, and weight gain [30, 32, 34, 35]. This question cannot be answered with statistical models which adjust for classical confounders and wear time (i.e., ‘single activity’ models) or models which also adjust for other sedentary behavior and physical activities (i.e., ‘partition’ models) as described in a study by Mekary et al . In the study by Healy et al , the sedentary time patterns (e.g., prolonged and non-prolonged sedentary bouts) were investigated among adult patients with type 2 diabetes, but not in specific working populations, such as blue-collar workers. The isotemporal substitution analysis is particularly interesting in this population as it provides theoretical estimates of the benefits of “rotating” job tasks and thereby shifting exposures among sedentary and various types of physical activity.
A limitation of previous studies using the isotemporal substitution approach is the use of count-based accelerometer thresholds to determine sedentary time . Both location of wear (for example hip) and usage of count-based accelerometer thresholds have been criticized for their inability to accurately differentiate sedentary postures from standing . Both of these limitations may lead to incorrect information about sedentary time patterns  and thus should not be used to investigate the effect of replacing the sedentary time with standing time on various health outcomes. Additionally, previous studies using the isotemporal substitution analysis to explore the relationship between various types of physical activity/sedentary behavior and health outcomes have almost exclusively focused on whole day measurements [30, 35]. It has been shown that work and non-work time differ in terms of sedentary patterns [38, 39] and also in terms of the barriers and facilitators for modifying these patterns. Thus, it is important to understand the contribution of each of these domains to the overall risk of obesity. Previous studies  have also not investigated the effect of replacing sedentary time with a detailed spectrum of physical activities (i.e., standing, walking, and MVPA instead of standing and stepping only) on obesity.
Thus, the aim of this study was to investigate the association of obesity indicators with total sedentary time and bouts of sedentary time (brief, moderate and long sedentary bouts) during whole day, work and non-work time among blue-collar workers. Furthermore, these associations were further explored when replacing total sedentary time and long sedentary bouts, respectively, with various types of physical activity, as well as replacing long sedentary bouts with brief sedentary bouts.
Material and Methods
Study design and population
Our study is a part of the Danish PHysical ACTivity cohort with Objective measurements (DPhacto). The main aim of DPhacto is to investigate the association between objectively measured physical activities at work and development of musculoskeletal pain among blue-collar workers . A detailed recruitment of the study population, inclusion and exclusion criteria and data collection is reported elsewhere .
This study includes the cross-sectional baseline data of DPhacto, collected from spring 2012 to spring 2014 at workplaces engaged in three different occupational sectors (cleaning, transport and manufacturing) in Denmark. In total, 2107 eligible workers from 15 companies were recruited in collaboration with a large labor union between December 2011 and March 2013. Workers were excluded if they were in a white-collar occupation (managers/administrative workers), pregnant, had a fever, or a band aid allergy.
All workers provided their written informed consent prior to participation. The present study was conducted according to the Helsinki declaration and approved by the Danish data protection agency and local Ethics Committee (The Capital Region of Denmark, H-2-2012-011).
All eligible blue-collar workers responded to a questionnaire about socio-demographic measures, lifestyle and health-related behaviours and participated in measurements of obesity indicators. Additionally, they were equipped with an Actigraph accelerometer (Actigraph GT3X+, Actigraph LLC, Florida, USA) placed on the right thigh  for four consecutive days (4 x 24 hours), including at least two working days to determine their sedentary and physical activities spectrum . They were also equipped with a short paper diary concerning their working hours, non-work time, time in bed (going to bed, to sleep, and getting out of bed), non-wear time, and time of reference measurement every day during the measurement period. Detailed pre-instructions given to the workers are mentioned elsewhere [17, 41, 43].
Objective measurements of sedentary time and types of physical activity.
The sedentary time and various types of physical activity were determined using a custom-made MATLAB program (Acti4), estimating time spent in various types of physical activity [moving (upright position that could not be classified as standing still or walking or running; for example, very small steps on the same spot), walking, running, stair climbing, and cycling] and body postures (sitting, lying and standing still) across the day(s) with a sensitivity and specificity of more than 94% and 99%, respectively, during semi-standardized conditions . During free living conditions (around 140 hours of measurements), the sensitivity and specificity of the Acti4 program for detecting the sitting posture was 98% and 93%, compared to pressure sensors placed in the back pockets .
Briefly, accelerometer data were low-pass filtered with a 5Hz 4th order Butterworth filter and split-up into 2 sec intervals with 50% overlap. Then, the occurrence of sedentary posture was defined as the posture in which the inclination of the thigh is above 45°. The daily reference measurements (i.e. standing in an upright position for 15sec on every measured day) were used to obtain the angle between the leg and accelerometer axis. The technical definitions of the identification of other activities using Acti4 program are reported in detail elsewhere .
Using the diary information about the start and end time of work and the time in and out of bed at night, the work and non-work periods were defined. All non-working days and non-wear periods were excluded according to previously reported criteria [17, 43]. As this study was intended to address working days only, only days with objective measurements during the working periods were included. Further, only workers with at least one valid day comprising at least 10 hours of wear time during waking hours were included in the analysis on whole day [17, 43, 44]. For the specific analyses of sedentary time during work and non-work domain, workers with at least one day of valid measurement of work and non-work time period were included; (a) a valid work time period: at least 4 hours of work time or 75% of the individual’s average reported working time, and (b) valid non-work time period: at least 4 hours of non-work time or 75% of the individual’s average reported non-work time [17, 43]. A day consisted of 24 hours starting from midnight to midnight, while whole day was defined as all waking hours during a day. Work was defined as the self-reported hours spent at the primary occupation, and non-work time was defined as the remaining time except time in bed during a day.
Subsequently, based on the measured periods of sedentary time on working days, total sedentary time (total time spent sitting or lying) in the following domains were retrieved; (a) whole day (total measured sedentary hours during the entire measurement period divided by the number of days), (b) work (total measured sedentary hours during working periods divided by the number of days) and (c) non-work time (total measured sedentary hours except time in bed during non-work time periods divided by the number of days). Total time periods spent ‘standing still’ and ‘moving’ were merged and divided by the number of days measured to calculate ‘standing time’. Total time periods spent ‘stair-climbing’, ‘running’ and ‘cycling’ were merged and divided by the number of measured days to calculate MVPA time. Total time spent walking was divided by the number of days to calculate ‘walk time’. ‘Time in bed’ was estimated based on the self-reported time in the diary for ‘going to bed to sleep and getting out of bed’ and was averaged across the days measured. Total wear/measured time during whole day and separately during work and non-work period were calculated by adding the registered measured time (excluding non-recorded periods which were less than ~1% of the total measured time) from accelerometers in each domain on all days divided by the number of measured days.
Exposure Variation Analysis of sedentary time. The sedentary time patterns were determined using Exposure Variation Analysis (EVA; ). Based on previous studies [16, 30, 39, 46], EVA was utilized to categorize uninterrupted sedentary bouts according to their duration, ‘long bouts’ (LB) (average time/day spent in uninterrupted sedentary bouts >30 min), ‘moderate bouts’ (MB) (average time/day spent in uninterrupted sedentary bouts >5 and ≤30min), and ‘brief bouts’ (BB) (average time/day spent in uninterrupted sedentary bouts ≤5mins) during whole day, work and non-work time .
Objective measurement of obesity indicators.
The weight, body fat percentage (Tanita, model BC418 MA) , and waist circumference of the workers were objectively measured. The waist circumference with millimeter precision was measured horizontally midway between the top edge of the hip and lower ribs using a measurement tape (Seca, model 201) . The average of the two measurements was recorded. Height was measured without shoes using a stadiometer (Seca, model 213) to the nearest 0.1 cm while weight and fat percentage (%BF) were measured without shoes and socks using the Tanita (model BC418 MA) bio-impedance segmental body composition analyzer , to the nearest 0.1 kg and 0.1%, respectively. Their BMI (kg/m2) was calculated as weight (kg) divided by height (m) squared.
The confounders were chosen a priori based on theoretical considerations and previous studies investigating the association between sedentary time and obesity [26, 27, 30, 49]. Sex and age of the workers were determined from each worker’s unique Danish civil registration number. Smoking behavior was determined from the question: “Do you smoke?” with 4 response categories summarized into three groups; smokers (yes daily; yes sometimes), ex-smokers (used to smoke, but not anymore) and non-smokers (have never smoked). Alcohol intake was determined by the item ‘How much alcohol did you drink during the last week’ with responses in number of units per week. Poor dietary habits were determined by averaging the responses of the following two items ‘How often do you usually eat/drink candy, ice cream, chocolate, soft drinks’ and ‘fastfood, pizza, burger, shawarma, etc.’ with four responses (daily, 3-4/week, 1-2/week, and rarely). Influence at work (decision authority) was determined by the 2-item scale from the Copenhagen Psychosocial Questionnaire . The responses were scored on a Likert scale with response categories ranging from 0 (never) to 5 (always). A composite scale measuring influence at work was constructed by calculating the mean rating of both items. For the analysis, this scale was recoded into a 0–100 scale, where a larger score represented a higher degree of influence at work . All confounders were treated as continuous variables except smoking behavior and sex which were treated as categorical variables.
All statistical analyses described below were performed for each of the three time domains, i.e., whole day, work and non-work time; and for each of the three indicators of obesity, i.e., BMI, %BF and waist circumference. The results of the associations between sedentary time variables and obesity indicators were presented as unstandardized regression coefficients per 30 minutes of modelled exposure. All models were adjusted for the potential confounders age, sex (reference: male), smoking (reference: non-smokers), alcohol intake, and dietary habits. The associations were tested using three linear regression models: ‘the single activity model’, ‘the partition model’ and ‘the isotemporal substitution model’.
First, four independent single activity models were fitted to assess the association between each independent sedentary time variable (i.e., total sedentary time and long, moderate and brief bouts) and each of the dependent obesity indicators using ordinary linear least-squares regression analysis, adjusted for total measured time in addition to the abovementioned confounders.
Second, the partition models were fitted by adjusting the single activity models for standing, walking, and MVPA time instead of total measured time. In the models containing long, moderate or brief sedentary bouts, these variables were entered simultaneously.
Third, the isotemporal substitution models were fitted to test the association with obesity indicators when replacing total sedentary time or long sedentary bouts (>30 minutes/day) with an equal amount of time spent on another activity (+30 min/day of standing, walking or MVPA time). Similarly, the effect on the indicators of obesity of replacing long sedentary bouts with short sedentary bouts was also estimated. In these models, all independent variables, except the variable of interest (total sedentary time or long sedentary bouts), plus the total measured time and confounders were entered into the model simultaneously. Thus, the regression coefficient of an included variable (brief sedentary bouts, standing, walking or MVPA) in the model represented the population-based average change in the indicators of obesity due to increasing 30 minutes/day of that variable and decreasing the 30 minutes/day of the variable of interest (total sedentary time or long sedentary bouts), keeping the time in all other modelled activities constant.
In all models, time spent on each activity in both domains was entered. The reason for this is that the time of activities in the domains are correlated  and therefore needs to be mutually adjusted for by being simultaneously entered in the model. To avoid overfitting, the combined time spent on activities in both domains except the variable of interest (i.e., total sedentary time or long sedentary bouts) was entered in the models.
The sensitivity of the results obtained from the isotemporal models was tested by; (1) additionally adjusting for influence at work, (2) additionally adjusting for the self-reported ‘time in bed’, and (3) additionally adjusting for shift work and job type of the workers in the primary analysis model.
In the analyses, no major multi-collinearity issues were detected (tolerance index >0.20, VIF values <5) . Visual inspection of P-P plots, histograms of standardized residuals, and scatterplots of standardized residuals against standardized predicted values indicated that assumptions of linearity were fulfilled, and that residuals were normally distributed and homoscedastic. Two outliers, indicated in boxplots, were removed from the analyses on waist circumference due their effect on the model fit.
Using the sample size calculation for linear regression, inclusion of at least 398 workers were sufficient to detect, with statistical power of 80% and alpha of 5%, a slope of 1 cm in waist circumference, 0.35 kg/m2 in BMI, and 0.40% in fat percentage attributed to 30minutes/day of total sedentary time and long sedentary bouts, assuming a standard deviation (SD) of 3 units per 30minutes/day. We considered these slopes to be meaningful based on previous studies [30, 53].
Fig 1 illustrates the flow diagram of the participants in the study. In total, 2107 workers were invited to participate of whom 901 blue-collar workers were eligible for participation. Of them, 692 were included in the analysis on whole day while 671 were included in the analysis of the work and non-work time domain.
Note: *Valid whole day, work and non-work time are defined in the ‘Material and Methods’ section.
For the analyses of whole day, work and non-work time, ~32,334 (2,023 valid days), ~13,521 (1,774 valid days) and ~15,550 (1,774 valid days) valid hours were measured, respectively. Participant characteristics and accelerometry-based measurements are described in Tables 1 and 2. Workers spent more time being sedentary during non-work time than during work. Overall during the whole day and non-work domain, they spent the majority of time being sedentary and the least time in MVPA. In the work domain, most time was spent in standing and least in MVPA. More specifically, during non-work time, the majority of the sedentary time was spent in long bouts and the least time in short bouts. During work time, the majority of the sedentary time was spent in moderate bouts and the least time in long bouts. Workers self-reported, on average, 5.9 (SD = 1.5) hours in bed per day (duration between going to and getting out of bed).
The results of the single activity models (shown in S1 Table) showed that total sedentary time was significantly positively associated with higher obesity indicators during whole day and non-work time, but not during work time. Similar results were observed for LB, while BB was significantly negatively associated with lower obesity indicators in all domains. No significant associations between MB and obesity indicators were found for any domain.
When adjusted also for other sedentary behavior and physical activities in the model, beside classical confounders (see Methods section), the results of the partition models (shown in S2 Table) showed similar results to the single activity models for all analyses. However, the association between total sedentary time and %BF did not remain significant during whole day [B(95%CI) = 0.09(-0.11 to 0.29) %, P = 0.36] and non-work time [B(95%CI) = 0.19(-0.04 to 0.43) % P = 0.11]. Moreover, the association between brief sedentary bouts and obesity did not remain significant for BMI during work [B(95%CI) = -0.34(-0.77 to 0.08) kg/m2 P = 0.11].
Table 3 shows the results of the isotemporal substitution analyses indicating the ‘substitution association’ with obesity indicators when replacing 30 minutes/day of sedentary behaviors with 30 minutes/day of other activities during the whole day. Table 4 shows the corresponding results for work and non-work time domains separately. The estimates of the obesity indicators decreased significantly when total sedentary time or long sedentary bouts were replaced with standing time (~1–2% lower) or MVPA (~4–9% lower; equivalent to a meaningful amount of ~3.8–4.2 cm waist circumference, ~1.9–2.7% fat percentage and ~1.1–1.4 kg/m2 BMI per 30 minutes/day) during whole day and non-work time. Similar results were obtained when long sedentary bouts were replaced with brief sedentary bouts (~3–5% lower; equivalent to a meaningful amount of ~2.6–2.7 cm waist circumference, ~1.4–1.6% fat percentage, and ~0.8–0.9 kg/m2 BMI per 30 minutes/day) during whole day and non-work time. During work time, similar results were observed for brief sedentary bouts (~3–4% lower) and MVPA (~3–6% lower), but not for standing time. Interestingly, replacing total sedentary time or long sedentary bouts in any domain with increased walking time was not significantly associated with any of the indicators of obesity.
Results of the sensitivity analyses
Adjustment for influence at work did not change the estimates of the primary analyses of replacing 30 minutes of total sedentary time with standing, walking or MVPA. Similar results were observed for replacing long sedentary bouts with standing, walking, MVPA or brief sedentary bouts, except that the results of replacing 30 minutes of long sedentary bouts during work with MVPA became marginally non-significant for %BF [B(95%CI); -1.42 (-3.02 to 0.19)%, P = 0.08) and BMI (-1.03(-2.15 to 0.09) kg/m2, P = 0.07) (results not shown).
Similar estimates to the primary analysis using isotemporal substitution were also observed when adjusting for time in bed (results not shown), except that the results of replacing 30 min of long sedentary bouts during work with MVPA became marginally non-significant [-0.92(-1.88 to 0.05) kg/m2, P = 0.06] for BMI.
Adjustment for job type and shift work did not change the estimates of the primary analyses using isotemporal substitution, except that the results of replacing total sedentary time with standing became significant [-0.26(-0.46 to -0.06) %, P = 0.01] for fat percentage and replacing long sedentary bouts with standing became non-significant [-0.45 (-1.15 to 0.25) cm, P = 0.21] for waist circumference during work.
Discussion and Conclusion
The most important finding of the present cross-sectional study is that among blue-collar workers beneficial associations with indicators of obesity were observed by theoretically replacing time spent on sedentary behavior with standing or higher intensity physical activity. Beneficial findings were also observed when time in long sedentary bouts was reallocated to short sedentary bouts.
The amount of sedentary time in this study was similar to other studies among blue-collar workers [17, 18, 39, 43]. The workers sat for ~50% of their waking time with more sedentary time being spent during non-work (~62% of total non-work time) than work (~32% of the working time). This might reflect a compensation for not having sufficient opportunities to recover from strenuous work tasks. During whole day, workers spent the most time in long and moderate sedentary bouts and the least in brief sedentary bouts. They spent about 38% of the time standing (with ~25% standing still), ~12% walking and ~1% MVPA, which compares well with previous research on blue-collar workers  except for walking being slightly lower in this population (~26% in the previous study ).
Workers who not only spent more time in long, but also less time in brief sedentary bouts had higher indicators of obesity and vice versa, even after adjusting for potential confounders and wear time (the single activity model) and other activities performed (the partition model). For example, spending 30 minutes more in long sedentary bouts during non-work time was associated with 0.98 cm (~1%) increase in waist circumference, while spending 30 minutes more in brief sedentary bouts during non-work time was associated with a decreased waist circumference by 2.94 cm (~3%) (see S2 Table), keeping the other sedentary behavior and physical activities constant. The results of brief sedentary bouts correspond well with previous studies on the association between breaks in prolonged sitting (e.g., transitions from a sedentary to an active state lasting ≥1min) and obesity indicators [27, 28]. The brief sedentary bouts could be considered as a ‘proxy’ for ‘breaking up’ long sedentary bouts by various physical activities (i.e., intermittent physical activities). It is possible that frequent interruptions of prolonged sedentary bouts facilitate lipid metabolism and glucose removal from the blood due to intermittent muscle contractions [31, 54], which may, in the long term, decrease the probability of becoming obese. However, this hypothesis requires further investigation on metabolic variables in longitudinal studies. Nevertheless, these results suggest that the distribution of sedentary time patterns, in addition to total sedentary time, may be important for obesity.
During work, however, significant associations between total sedentary time or long sedentary bouts and obesity indicators were not observed, which is in contrast to some studies [55–58], but not all [59, 60]. One explanation for the conflicting results could be differences in the study population and types of measures used in the studies. In studies where a significant positive association between sedentary time at work and obesity have been observed, workers were very sedentary at work (sat for as much as 6 hours during work or 60% of the working time) compared to our study [55, 57]. Also most of these studies used self-reported measures (telephone interviews  and questionnaires [56–58]) to assess sedentary time, which could lead to misclassification of the workers’ occupational sedentary time  due to reporting bias.
Our findings suggest that replacing 30 minutes of total sedentary time or long sedentary bouts with 30 minutes of MVPA during a day has a considerable beneficial meaningful association with obesity during all domains, which confirms the previously established beneficial role of MVPA in prevention and management of obesity [62, 63]. However, beneficial associations were also found when total sedentary time and long sedentary bouts, respectively, were replaced with standing, and when long sedentary bouts were replaced with brief sedentary bouts. Although, these results on replacing sedentary time with standing time are not considered clinically meaningful in our study, they are consistent with previous studies , where replacing long sedentary bouts with light physical activity during whole day resulted in lower BMI  and waist circumference . During work, these results were similar except for replacing total sedentary time or long sedentary bouts with standing time during a day. Some previous studies have indicated a beneficial effect of standing at work [64–67]. However, most of these studies were either conducted in office workers only  or utilized self-reported standing measures . Therefore, we need future studies to verify the association between standing at work and obesity among blue-collar workers.
Replacing prolonged sedentary time with brief sedentary bouts or standing may be more feasible than replacing it with MVPA among blue-collar workers who are already engaged in highly strenuous work tasks. We estimated that replacing ~45 minutes of long sedentary bouts with brief sedentary bouts or ~137 mins of long sedentary bouts with standing will have equal effect on obesity as replacing 30 min of long sedentary bouts with MVPA. In relative terms, ~45 min increase in brief sedentary bouts and ~137 min increase in standing corresponds to 69% or 37% increase in daily time spent in these activities, and may be more feasible among blue-collar workers than increasing 30 minutes in MVPA which corresponds to a 250% increase in their daily MVPA. Thus, the replacement of sedentary time with other activities such as standing and brief sedentary bouts may be relevant and practical in this cohort as long as the sufficient time is reallocated. Nevertheless, the observed results suggest that even a modest change such as replacing 30 minutes of long sedentary bouts with standing or brief sedentary bouts may reduce obesity among blue-collar workers. However, for these workers who may require sedentary time as ‘rest’ period between high strenuous working tasks, it might be appropriate to shift between job tasks to ensure reduction in prolonged, unbroken sedentary bouts instead of aiming to reduce total sedentary time. More research is needed to determine which activity-replacement, in practice, is most feasible for achieving sustainable favorable changes in obesity among blue-collar workers without affecting their work demands-recovery balance.
Interestingly, we did not find any beneficial association with obesity indicators of replacing total sedentary time or long sedentary bouts with walking time. These findings are in contrast to a growing literature on the positive effect of light physical activities on health outcomes including obesity [30, 35]. Indeed, previous studies using isotemporal substitution analysis have shown that walking or light physical activities are important for health outcomes such as weight, BMI , waist circumference  and cardiovascular diseases  among diabetic adults or the general population. The findings are intriguing, and should be investigated further among blue-collar workers specifically.
Strength and limitations and methodological considerations
A major strength of our study is the study population of blue-collar workers, being homogeneous with respect to being employed in blue collar work, but offering a great variation in sedentary time and various types of physical activity. We also used the validated software, Acti4 to assess the exposure. It is capable of discriminating between different types of physical activity and body postures with excellent sensitivity and specificity . Previous studies using isotemporal substitution analysis have recommended this method which is based on posture identification instead of count-based accelerometer thresholds for calculating objectively measured sedentary and physical activity time . Another strength of the study was the separation between prolonged and brief sedentary bouts using exposure variation analysis (EVA). EVA is a versatile generic approach for quantifying the level and frequency of activities, as demonstrated by the use of EVA for analyses of, e.g. muscle activity , working postures  and physical activity intensities . Although we did not have objective measurements of sleep duration, we adjusted for self-reported time in bed which did not substantially change the overall results of the primary analyses.
The main limitation of the study is the cross-sectional study design, which does not allow for inferences about causal associations with obesity of replacing prolonged sedentary activities with brief sedentary bouts and various types of physical activity. Thus, future prospective studies are needed to assess the strength and direction of this association in different domains. Since our study included a sample of Danish blue-collar workers only, our results may not be generalized to the general population of blue-collar workers in Denmark, let alone in industrialized countries in general. Furthermore, the results may be affected by residual confounding from medication usage, prevalence of chronic diseases, working status, family history, and activity patterns during commuting to work. The adjustment for influence at work was only performed in the sensitivity analyses due to a large amount of missing values in the variable (~30%), although adjusting for this variable did not change the results of the primary analysis. Another limitation is that the use of traditional statistical analyses may not be suitable due to the constrained nature of the data used in this study . This requires future investigation using compositional data analysis, specially designed for constrained data , to verify the results obtained in our study. Although adjustment for other health risk behaviors such as poor diet, smoking, and high alcohol intake did not change the results obtained in our study, these health risk behaviors were common in this cohort and thus future studies should investigate the association of these risk factors, beside activity levels, which may also be associated with obesity among these workers. Although, defining a day from midnight to midnight is a common practice in previous research , it may have some consequences on the overall distribution and amount of time spent sedentary and in physical activities. This is reflected in the short average ‘time in bed’ duration in this sample, where time in bed on the last measured day is stopped at midnight. Thus, future studies using some other definitions of a ‘day’ are needed to verify the results of our study. It is also possible that the observed strength of the associations between sedentary behaviors and indicators of obesity are different within different categories of obesity. However such analysis requires a large sample size and thus warrants a future investigation. The fluctuation or variation in work and non-working hours within a day could also contribute to the obesity among these workers which needs to be investigated upon in future studies.
In conclusion, total sedentary time and uninterrupted long sedentary bouts were positively associated with obesity indicators during whole day and non-work time while brief sedentary bouts were negatively associated with obesity indicators during all domains. Our findings indicate that beneficial health effects may be gained by replacing total sedentary time and uninterrupted long sedentary bouts not only with MVPA, but also with standing among blue-collar workers. Similar benefits may also be gained by replacing uninterrupted long sedentary bouts with brief sedentary bouts among these workers. However, these interpretations of the results need to be confirmed in future studies using a prospective study design.
S1 Dataset. This dataset contains two data files, ‘whole day’ and ‘work and non-work’.
The data file ‘whole day’ contains data for investigating the association between sedentary patterns during whole day and obesity indicators (N = 692) and the data file ‘work and non-work’ contains data for investigating the association between sedentary patterns during work and non-work time domain and obesity indicators (N = 671).
S1 Table. Results of single activity* models to estimate the association between sedentary time variables and obesity during whole day (n = 692), work (n = 671) and non-work (n = 671) time domain among blue-collar workers from Danish PHysical ACTivity cohort with Objective measurements (DPhacto).
S2 Table. Results of the partition* models to estimate the association between sedentary time variables and obesity during whole day (n = 692), work (n = 671) and non-work (n = 671) time domain among blue-collar workers from Danish PHysical ACTivity cohort with Objective measurements (DPhacto).
Conceived and designed the experiments: NG AH MBJ. Performed the experiments: MK AH MBJ. Analyzed the data: NG MH AH. Contributed reagents/materials/analysis tools: NG MH AH. Wrote the paper: MA NG AH MBJ MK MH.
- 1. WHO Media centre. Obesity and overweight 2015 [cited 2015 4th May]. Available from: http://www.who.int/mediacentre/factsheets/fs311/en/.
- 2. World Health Organisation. Data and statistics: The challenge of obesity—quick statistics 2015 [cited 2015 4th May]. Available from: http://www.euro.who.int/en/health-topics/noncommunicable-diseases/obesity/data-and-statistics.
- 3. Neovius K, Johansson K, Kark M, Tynelius P, Rasmussen F. Trends in self-reported BMI and prevalence of obesity 2002–10 in Stockholm County, Sweden. Eur J Public Health. 2013;23(2):312–5. pmid:23002231.
- 4. Gu JK, Charles LE, Bang KM, Ma CC, Andrew ME, Violanti JM, et al. Prevalence of obesity by occupation among US workers: the National Health Interview Survey 2004–2011. J Occup Environ Med. 2014;56(5):516–28. Epub 2014/04/01. pmid:24682108.
- 5. Yu Z, Nissinen A, Vartiainen E, Song G, Guo Z, Zheng G, et al. Associations between socioeconomic status and cardiovascular risk factors in an urban population in China. Bull WHO. 2000;78(11):1296–305. Epub 2001/01/06. pmid:11143189; PubMed Central PMCID: PMCPMC2560638.
- 6. Alves L, Azevedo A, Silva S, Barros H. Socioeconomic inequalities in the prevalence of nine established cardiovascular risk factors in a southern European population. PLoS ONE. 2012;7(5):e37158. Epub 2012/06/06. pmid:22666343; PubMed Central PMCID: PMCPMC3362583.
- 7. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA. 2013;309(1):71–82. Epub 2013/01/03. pmid:23280227.
- 8. Must A, McKeown NM. The Disease Burden Associated with Overweight and Obesity. In: De Groot LJ, Beck-Peccoz P, Chrousos G, Dungan K, Grossman A, Hershman JM, et al., editors. Endotext. South Dartmouth (MA)2000.
- 9. Poulsen K, Cleal B, Clausen T, Andersen LL. Work, diabetes and obesity: a seven year follow-up study among Danish health care workers. PLoS ONE. 2014;9(7):e103425. pmid:25068830; PubMed Central PMCID: PMC4113351.
- 10. Pi-Sunyer FX. Medical hazards of obesity. Ann Intern Med. 1993;119(7 Pt 2):655–60. pmid:8363192.
- 11. Owen N, Sparling PB, Healy GN, Dunstan DW, Matthews CE. Sedentary behavior: emerging evidence for a new health risk. Mayo Clin Proc. 2010;85(12):1138–41. pmid:21123641; PubMed Central PMCID: PMC2996155.
- 12. Ainsworth BE, Haskell WL, Herrmann SD, Meckes N, Bassett DR, Tudor-Locke C, et al. 2011 compendium of physical activities: A second update of codes and met values. Med Sci Sports Exerc. 2011;43(8):1575–81. pmid:21681120
- 13. Pate RR, O'Neill JR, Lobelo F. The evolving definition of "sedentary". Exerc Sport Sci Rev. 2008;36(4):173–8. pmid:18815485.
- 14. Wilmot EG, Edwardson CL, Achana FA, Davies MJ, Gorely T, Gray LJ, et al. Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis. Diabetologia. 2012;55.
- 15. Dunstan DW, Wiesner G, Eakin EG, Neuhaus M, Owen N, Lamontagne AD, et al. Reducing office workers’ sitting time: rationale and study design for the Stand Up Victoria cluster randomized trial. BMC Public Health. 2013;13.
- 16. Parry S, Straker L. The contribution of office work to sedentary behaviour associated risk. BMC Public Health. 2013;13. pmid:WOS:000318796500001.
- 17. Gupta N, Stordal CC, Hallman D, Korshøj M, Gomes CI, Holtermann A. Is objectively measured sitting time associated with low back pain? A cross-sectional investigation in the NOMAD study. PLoS ONE. 2015;10(3):e0121159. pmid:25806808
- 18. Vandelanotte C, Duncan MJ, Short C, Rockloff M, Ronan K, Happell B, et al. Associations between occupational indicators and total, work-based and leisure-time sitting: a cross-sectional study. BMC Public Health. 2013;13:1110. pmid:24289321; PubMed Central PMCID: PMC3879072.
- 19. Jakes RW, Day NE, Khaw KT, Luben R, Oakes S, Welch A, et al. Television viewing and low participation in vigorous recreation are independently associated with obesity and markers of cardiovascular disease risk: EPIC-Norfolk population-based study. Eur J Clin Nutr. 2003;57(9):1089–96. pmid:12947427.
- 20. Stamatakis E, Hirani V, Rennie K. Moderate-to-vigorous physical activity and sedentary behaviours in relation to body mass index-defined and waist circumference-defined obesity. Br J Nutr. 2009;101(5):765–73. pmid:18680630.
- 21. Blanck HM, McCullough ML, Patel AV, Gillespie C, Calle EE, Cokkinides VE, et al. Sedentary behavior, recreational physical activity, and 7-year weight gain among postmenopausal U.S. women. Obesity (Silver Spring). 2007;15(6):1578–88. pmid:17557996.
- 22. Katzmarzyk PT, Church TS, Craig CL, Bouchard C. Sitting time and mortality from all causes, cardiovascular disease, and cancer. Med Sci Sports Exerc. 2009;41(5):998–1005. Epub 2009/04/07. pmid:19346988.
- 23. Thorp AA, Owen N, Neuhaus M, Dunstan DW. Sedentary behaviors and subsequent health outcomes in adults a systematic review of longitudinal studies, 1996–2011. Am J Prev Med. 2011;41(2):207–15. Epub 2011/07/20. pmid:21767729.
- 24. Hu FB, Li TY, Colditz GA, Willett WC, Manson JE. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA. 2003;289(14):1785–91. Epub 2003/04/10. pmid:12684356.
- 25. Raynor DA, Phelan S, Hill JO, Wing RR. Television viewing and long-term weight maintenance: results from the National Weight Control Registry. Obesity (Silver Spring). 2006;14(10):1816–24. Epub 2006/10/26. pmid:17062812.
- 26. Honda T, Chen S, Kishimoto H, Narazaki K, Kumagai S. Identifying associations between sedentary time and cardio-metabolic risk factors in working adults using objective and subjective measures: a cross-sectional analysis. BMC Public Health. 2014;14:1307. Epub 2014/12/21. pmid:25526746; PubMed Central PMCID: PMCPMC4302076.
- 27. Healy GN, Dunstan DW, Salmon J, Cerin E, Shaw JE, Zimmet PZ, et al. Breaks in Sedentary Time: Beneficial associations with metabolic risk. Diabetes Care. 2008;31(4):661–6. pmid:18252901
- 28. Healy GN, Matthews CE, Dunstan DW, Winkler EA, Owen N. Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003–06. Eur Heart J. 2011;32(5):590–7. Epub 2011/01/13. pmid:21224291; PubMed Central PMCID: PMCPmc3634159.
- 29. Dunstan DW, Kingwell BA, Larsen R, Healy GN, Cerin E, Hamilton MT, et al. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care. 2012;35(5):976–83. Epub 2012/03/01. pmid:22374636; PubMed Central PMCID: PMCPmc3329818.
- 30. Healy GN, Winkler EA, Brakenridge CL, Reeves MM, Eakin EG. Accelerometer-derived sedentary and physical activity time in overweight/obese adults with type 2 diabetes: cross-sectional associations with cardiometabolic biomarkers. PLoS ONE. 2015;10(3):e0119140. Epub 2015/03/17. pmid:25775249; PubMed Central PMCID: PMCPMC4361561.
- 31. Peddie MC, Bone JL, Rehrer NJ, Skeaff CM, Gray AR, Perry TL. Breaking prolonged sitting reduces postprandial glycemia in healthy, normal-weight adults: a randomized crossover trial. Am J Clin Nutr. 2013;98(2):358–66. Epub 2013/06/28. pmid:23803893.
- 32. Buman MP, Winkler EA, Kurka JM, Hekler EB, Baldwin CM, Owen N, et al. Reallocating time to sleep, sedentary behaviors, or active behaviors: associations with cardiovascular disease risk biomarkers, NHANES 2005–2006. Am J Epidemiol. 2014;179(3):323–34. pmid:24318278.
- 33. Mekary RA, Willett WC, Hu FB, Ding EL. Isotemporal substitution paradigm for physical activity epidemiology and weight change. Am J Epidemiol. 2009;170(4):519–27. pmid:19584129; PubMed Central PMCID: PMC2733862.
- 34. Balboa-Castillo T, Leon-Munoz LM, Graciani A, Rodriguez-Artalejo F, Guallar-Castillon P. Longitudinal association of physical activity and sedentary behavior during leisure time with health-related quality of life in community-dwelling older adults. Health Qual Life Outcomes. 2011;9:47. pmid:21708011; PubMed Central PMCID: PMC3142200.
- 35. Buman MP, Hekler EB, Haskell WL, Pruitt L, Conway TL, Cain KL, et al. Objective light-intensity physical activity associations with rated health in older adults. Am J Epidemiol. 2010;172(10):1155–65. pmid:20843864
- 36. Hart TL, Ainsworth BE, Tudor-Locke C. Objective and subjective measures of sedentary behavior and physical activity. Med Sci Sports Exerc. 2011;43(3):449–56. pmid:20631642.
- 37. Atkin AJ, Gorely T, Clemes SA, Yates T, Edwardson C, Brage S, et al. Methods of measurement in epidemiology: Sedentary behaviour. Int J Epidemiol. 2012;41(5):1460–71. Epub 2012/10/10. pmid:23045206; PubMed Central PMCID: PMCPMC3465769.
- 38. Saidj M, Jorgensen T, Jacobsen RK, Linneberg A, Aadahl M. Differential cross-sectional associations of work- and leisure-time sitting, with cardiorespiratory and muscular fitness among working adults. Scand J Work Environ Health. 2014;40(5):531–8. pmid:WOS:000341342300011.
- 39. Hallman DM, Mathiassen SE, Gupta N, Korshoj M, Holtermann A. Differences between work and leisure in temporal patterns of objectively measured physical activity among blue-collar workers. BMC Public Health. 2015;15(1):976. Epub 2015/09/30. pmid:26415931; PubMed Central PMCID: PMCPMC4587719.
- 40. Healy GN, Winkler EA, Owen N, Anuradha S, Dunstan DW. Replacing sitting time with standing or stepping: associations with cardio-metabolic risk biomarkers. Eur Heart J. 2015;36(39):2643–9. pmid:26228867.
- 41. Jorgensen MB, Nabe-Nielsen K, Clausen T, Holtermann A. Independent effect of physical workload and childhood socioeconomic status on low back pain among health care workers in Denmark. Spine. 2013;38(6):E359–66. Epub 2013/03/16. pmid:23492977.
- 42. Skotte J, Korshoj M, Kristiansen J, Hanisch C, Holtermann A. Detection of physical activity types using triaxial accelerometers. J Phys Act Health. 2014;11(1):76–84. Epub 2012/12/20. pmid:23249722.
- 43. Hallman DM, Gupta N, Mathiassen SE, Holtermann A. Association between objectively measured sitting time and neck–shoulder pain among blue-collar workers. Int Arch Occup Environ Health. 2015;88(8):1031–42. pmid:25677207
- 44. Villumsen M, Samani A, Jørgensen MB, Gupta N, Madeleine P, Holtermann A. Are forward bending of the trunk and low back pain associated among Danish blue-collar workers? A cross-sectional field study based on objective measures. Ergonomics. 2014;58(2):246–58. pmid:25374330
- 45. Mathiassen SE, Winkel J. Quantifying variation in physical load using exposure-vs-time data. Ergonomics. 1991;34(12):1455–68. Epub 1991/12/01. pmid:1800110.
- 46. Straker L, Campbell A, Mathiassen SE, Abbott RA, Parry S, Davey P. Capturing the Pattern of Physical Activity and Sedentary Behavior: Exposure Variation Analysis of Accelerometer Data. J Phys Act Health. 2014;11(3):614–25. pmid:WOS:000334983700021.
- 47. Body Composition Analyzer BC-418. Instruction Manual [cited 2015 21 June]. Available from: http://www.tanita.com/es/bc-418/.
- 48. Han TS, Sattar N, Lean M. Assessment of obesity and its clinical implications. Br Med J. 2006;333(7570):695–8.
- 49. Ryde GC, Brown HE, Peeters GMEE, Gilson ND, Brown WJ. Desk-Based Occupational Sitting Patterns: Weight-Related Health Outcomes. Am J Prev Med. 2013;45(4):448–52. http://dx.doi.org/10.1016/j.amepre.2013.05.018. pmid:24050421
- 50. Pejtersen JH, Kristensen TS, Borg V, Bjorner JB. The second version of the Copenhagen Psychosocial Questionaire. Scand J Public Health. 2010;38:8–24. pmid:21172767
- 51. Chau JY, van der Ploeg HP, Merom D, Chey T, Bauman AE. Cross-sectional associations between occupational and leisure-time sitting, physical activity and obesity in working adults. Prev Med. 2012;54.
- 52. O’brien R. A Caution Regarding Rules of Thumb for Variance Inflation Factors. Qual Quant. 2007;41(5):673–90.
- 53. Eakin EG, Reeves MM, Marshall AL, Dunstan DW, Graves N, Healy GN, et al. Living Well with Diabetes: a randomized controlled trial of a telephone-delivered intervention for maintenance of weight loss, physical activity and glycaemic control in adults with type 2 diabetes. BMC Public Health. 2010;10(1):1–15.
- 54. Latouche C, Jowett JBM, Carey AL, Bertovic DA, Owen N, Dunstan DW, et al. Effects of breaking up prolonged sitting on skeletal muscle gene expression2013 2013-02-15 00:00:00. 453–60 p.
- 55. Yang L, Hipp JA, Marx CM, Brownson RC. Occupational sitting and weight status in a diverse sample of employees in Midwest metropolitan cities, 2012–2013. Prev Chronic Dis. 2014;11:E203. pmid:25412026; PubMed Central PMCID: PMC4241372.
- 56. Mummery WK, Schofield GM, Steele R, Eakin EG, Brown WJ. Occupational sitting time and overweight and obesity in Australian workers. Am J Prev Med. 2005;29(2):91–7. pmid:16005804.
- 57. Bennie JA, Pedisic Z, Timperio A, Crawford D, Dunstan D, Bauman A, et al. Total and domain-specific sitting time among employees in desk-based work settings in Australia. Aust N Z J Public Health. 2015;39(3):237–42. pmid:25545803
- 58. Chau JY, van der Ploeg HP, Merom D, Chey T, Bauman AE. Cross-sectional associations between occupational and leisure-time sitting, physical activity and obesity in working adults. Prev Med. 2012;54(3–4):195–200. Epub 2012/01/10. pmid:22227284.
- 59. Saidj M, Jorgensen T, Jacobsen RK, Linneberg A, Aadahl M. Separate and joint associations of occupational and leisure-time sitting with cardio-metabolic risk factors in working adults: a cross-sectional study. PLoS ONE. 2013;8(8):e70213. Epub 2013/08/13. pmid:23936391; PubMed Central PMCID: PMCPmc3735561.
- 60. Pinto Pereira SM, Ki M, Power C. Sedentary behaviour and biomarkers for cardiovascular disease and diabetes in mid-life: the role of television-viewing and sitting at work. PLoS ONE. 2012;7(2):e31132. Epub 2012/02/22. pmid:22347441; PubMed Central PMCID: PMCPmc3276501.
- 61. Kwak L, Proper KI, Hagströmer M, Sjöström M. The repeatability and validity of questionnaires assessing occupational physical activity—a systematic review. Scand J Work Environ Health. 2011;37(1):6–29. pmid:20802979
- 62. Donnelly JE, Blair SN, Jakicic JM, Manore MM, Rankin JW, Smith BK. American College of Sports Medicine Position Stand. Appropriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc. 2009;41(2):459–71. Epub 2009/01/08. pmid:19127177.
- 63. Saris WH, Blair SN, van Baak MA, Eaton SB, Davies PS, Di Pietro L, et al. How much physical activity is enough to prevent unhealthy weight gain? Outcome of the IASO 1st Stock Conference and consensus statement. Obes Rev. 2003;4(2):101–14. Epub 2003/05/23. pmid:12760445.
- 64. Stephens BR, Granados K, Zderic TW, Hamilton MT, Braun B. Effects of 1 day of inactivity on insulin action in healthy men and women: interaction with energy intake. Metabolism. 2011;60(7):941–9. Epub 2010/11/12. pmid:21067784.
- 65. Buckley JP, Mellor DD, Morris M, Joseph F. Standing-based office work shows encouraging signs of attenuating post-prandial glycaemic excursion. Occup Environ Med. 2014;71(2):109–11. Epub 2013/12/04. pmid:24297826.
- 66. Katzmarzyk PT. Standing and mortality in a prospective cohort of Canadian adults. Med Sci Sports Exerc. 2014;46(5):940–6. Epub 2013/10/25. pmid:24152707.
- 67. Thorp AA, Kingwell BA, Sethi P, Hammond L, Owen N, Dunstan DW. Alternating bouts of sitting and standing attenuate postprandial glucose responses. Med Sci Sports Exerc. 2014;46(11):2053–61. Epub 2014/03/19. pmid:24637345.
- 68. Ciccarelli M, Straker L, Mathiassen SE, Pollock C. Variation in muscle activity among office workers when using different information technologies at work and away from work. Hum Factors. 2013;55(5):911–23. Epub 2013/11/14. pmid:24218901.
- 69. Hooftman WE, van der Beek AJ, van de Wal BG, Knol DL, Bongers PM, Burdof A, et al. Equal task, equal exposure? Are men and women with the same tasks equally exposed to awkward working postures? Ergonomics. 2009;52(9):1079–86. Epub 2009/07/17. pmid:19606366.
- 70. Chastin SF, Palarea-Albaladejo J, Dontje ML, Skelton DA. Combined Effects of Time Spent in Physical Activity, Sedentary Behaviors and Sleep on Obesity and Cardio-Metabolic Health Markers: A Novel Compositional Data Analysis Approach. PLoS ONE. 2015;10(10):e0139984. pmid:26461112; PubMed Central PMCID: PMC4604082.
- 71. Choi L, Liu Z, Matthews CE, Buchowski MS. Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc. 2011;43(2):357–64. pmid:20581716; PubMed Central PMCID: PMC3184184.