Change and determinants of total and context specific sitting in adults: A 7-year longitudinal study

Kokoteksti

(1)UEF//eRepository DSpace Rinnakkaistallenteet. https://erepo.uef.fi Yhteiskuntatieteiden ja kauppatieteiden tiedekunta. 2020. Change and determinants of total and context specific sitting in adults: A 7-year longitudinal study Wennman, H Elsevier BV Tieteelliset aikakauslehtiartikkelit ©2019Sports Medicine Australia CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/ http://dx.doi.org/10.1016/j.jsams.2019.12.015 https://erepo.uef.fi/handle/123456789/8176 Downloaded from University of Eastern Finland's eRepository.

(2) Journal Pre-proof Change and determinants of total and context specific sitting in adults: A 7-year longitudinal study Heini Wennman, Tommi Härkänen, Maria Hagströmer, Pekka Jousilahti, Tiina Laatikainen, Tomi Mäki-Opas, Satu Männistö, Hanna Tolonen, Heli Valkeinen, Katja Borodulin. PII:. S1440-2440(19)30643-7. DOI:. https://doi.org/10.1016/j.jsams.2019.12.015. Reference:. JSAMS 2224. To appear in:. Journal of Science and Medicine in Sport. Received Date:. 16 August 2019. Revised Date:. 9 December 2019. Accepted Date:. 15 December 2019. Please cite this article as: Wennman H, Härkänen T, Hagströmer M, Jousilahti P, Laatikainen T, Mäki-Opas T, Männistö S, Tolonen H, Valkeinen H, Borodulin K, Change and determinants of total and context specific sitting in adults: A 7-year longitudinal study, Journal of Science and Medicine in Sport (2019), doi: https://doi.org/10.1016/j.jsams.2019.12.015. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier..

(3) Change and determinants of total and context specific sitting in adults: a 7-year longitudinal study. of. Heini Wennman a, Tommi Härkänen a, Maria Hagströmer b, c, Pekka Jousilahti d, Tiina Laatikainen d,e,f, Tomi Mäki-Opas g, Satu Männistö d, Hanna Tolonen a,. ro. Heli Valkeinen h, and Katja Borodulin a,i. a) Public Health Evaluation and Projection Unit, the Finnish Institute for Health and Welfare, Helsinki, Finland; PoB 30, 00271, Helsinki, Finland. -p. b) Department of Health Promoting Science, Sophiahemmet University, Stockholm, Sweden. c) Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden. re. d) Public Health Promotion Unit, National Institute for Health and Welfare, Helsinki, Finland. lP. e) Institute of Public Health, University of Eastern Finland, Kuopio, Finland. f) Joint municipal authority for North Karelia social and health services (Siun sote), Joensuu, Finland g) Department of Social Sciences, University of Eastern Finland (UEF), Kuopio, Finland. rn a. h) Ageing, Disability and Functioning Unit, the Finnish Institute for Health and Welfare, Helsinki, Finland i) Age Institute, Helsinki, Finland. Jo u. Corresponding author Heini Wennman. heini.wennman(a)thl.fi. Word count: 4630 (incl. in text references, excl. acknowledgements) Abstract word count: 248.

(4) of. Number of tables: 3 Number of figures: 0. ro. Number of supplementary tables: 3. -p. Abstract. re. Objectives: To assess the stability and determinants of total and context specific sitting in a follow-up of adults. Design: Longitudinal study. lP. Methods: Participants in the DILGOM cohort (n=3735, men 45%), reported daily sitting in five contexts (work-related, in vehicle, at home by the TV, at home at the computer, and elsewhere) in 2007 and 2014. Sociodemographic background, lifestyle and health were assessed in 2007. Total sitting comprised. rn a. the sum of context specific sitting. Changes in, and determinants of context specific sitting, stratified by baseline age into young middle-aged (<53 years); late middle-aged (53-68 years) and older-aged (>68 years) were estimated by generalized linear mixed models. Results: In 2007, total daily sitting was 7 hours 26 minutes, 6 hours 16 minutes, and 6 hours 3 minutes in young middle-aged, late middle-aged and older-aged. Jo u. groups, respectively. Over 7 years, total sitting decreased on average by 26 minutes. Sitting at the computer increased by 7 -17 minutes. The late middle-aged group also increased sitting by the TV, and decreased total, work-related, vehicle and elsewhere sitting. Occupational status determined context specific sitting, but somewhat differently in young and late middle-aged groups. Poor self-rated health determined less work-related and more sitting by the TV in the young, whereas good health determined less work-related sitting in the late middle-aged group..

(5) of. Conclusions: Self-reported sitting is a fairly stable behavior, with the exception for the late middle-aged group, where all context specific and total sitting changed significantly. Occupational status and health determined changes in sitting; however, somewhat differently by age group.. ro. Keywords: context; determinants; DILGOM 2007; Finland; longitudinal study; sedentary behavior; sitting Practical implications. a population approach to reduce sitting is recommended because sitting is a fairly stable behavior. . transfer to retirement imposes changes to sitting but further studies are needed to understand in more detail the compensation of work sitting by other. -p. . re. daily activities. it is important to intervene in excess sitting in those with poor health in the middle age. . intervention programs could launch less sedentary ways of screen-based activities, as sitting at the computer at home is increasing across the. lP. . Introduction. rn a. population. Jo u. Sitting is the most common form of sedentary behavior and it occurs in varying contexts (work, transportation, leisure time) throughout the day. A sedentary lifestyle threatens public health 1, but more data on the stability and determinants of daily sitting across different contexts are needed to inform and implement actions for preventing excessive sitting. Adults in different countries around Europe, in the United States and Australia report sitting around 6 to 10 hours a day 2-4 and based on cross-sectional data, the level of daily sitting time has been quite stable in many populations for the past 10 years 4, 5. Assessing only total daily sitting time limits the understanding of health risks associated with excessive sitting and masks changes that may occur in context specific sitting times. Repeated time-use data have shown changes in the distribution of leisure sedentary time, while the total leisure sedentary time remained.

(6) of. stable 6. In young Danish adults, both leisure time sitting and leisure time physical activity have increased 7 indicating that the younger generations may have adopted a physically active, yet highly sedentary lifestyle. Depending on the context of sitting, those with high sitting time are not the same people in regards. ro. to their sociodemographic, lifestyle and health characteristics. Full time employed adults often sit a lot at work and for transportation, the unemployed and retired spend more time sitting in front of the TV and highly educated adults sit a lot at the computer and at work 3, 8.. -p. Although cross-sectional data on sitting time and the correlates of sitting are extensively reported, population-based studies with a longitudinal design on. re. individual level change and determinants of daily sitting time, let alone context specific sitting time, are sparser. In two studies, total daily sitting decreased about 30 minutes over 2 9 and 9 years 10 respectively. Contrary, a small and non-significant increase in daily sitting time over 10 years was reported for. lP. Canadian adults, where maintaining sitting for more than 7 hours a day was more likely for men, the higher educated and employed 11. Furthermore, determinants of total sitting may vary according to age 10. In younger women, having a child, beginning work, and lowered income are related with decreased total sitting time, whereas returning to study or losing one’s job are related with increased time sitting. In middle-aged women retirement and lowered income. rn a. determined decreased daily sitting, while changes at work and the declined health of a nearby person were related with increased sitting 10. For context specific sitting, earlier studies report that self-reported sitting at work and sitting in vehicles decreases but leisure time sitting increases in relation. Jo u. to retirement 12, 13. However, not only stopping but also starting work may be related with changes in sitting for transportation 14. Those who have studied changes in TV time (although not specifically sitting), have observed less increase in TV time for the working and the higher educated over 4 years 15. Among men, higher age and becoming unemployed are related with increases in TV time whereas among women, higher age is related with increases, but becoming employed, a mother or normal weight are related with decreases in TV time over 10 years 16. It is important to identify groups who are vulnerable to increases in sitting time and to understand which groups may even decrease their sitting time. Our aim was to examine the change in self-reported total and context specific sitting time over a seven-year follow-up period from 2007 to 2014 among adults in.

(7) of. Finland and to contribute to the gap of knowledge regarding determinants of context specific sitting. We studied sociodemographic background, physical. -p. ro. activity, and self-rated health as potential determinants for total and context specific sitting in different age groups.. re. Methods. The data comprised the longitudinal cohort of Dietary, Lifestyle and Genetic Determinants of Obesity and Metabolic syndrome (DILGOM) Study 2007-2014.. lP. In 2007, 9957 randomly sampled adults aged 25 to 74 years in Finland were invited to the National FINRISK 2007 health examination survey. The participation rate was 63% (n= 6258) and all participants were invited to take part in the DILGOM Study conducted later the same year. In total, 5024 men. rn a. and women (84% of the eligible participants) participated in the DILGOM 2007 Study. In 2014, seven years later, all DILGOM 2007 participants were reinvited to a follow-up study and 3735 individuals (1687 men and 2048 women) participated. The complete study protocol for the FINRISK 2007 and the DILGOM 2007 and 2014 Study are explained elsewhere 17, 18. The DILGOM Study was approved by the Coordinating Ethics Committee of the Hospital. Jo u. District of Helsinki and Uusimaa. All participants gave a written informed consent. In both 2007 and 2014, participants completed extensive questionnaires concerning health, lifestyle and sociodemographic background. The current analyses included those DILGOM participants who took part in the baseline and the follow up and provided information on context specific sitting times in 2007, or in 2014, or both. Those participants with any context specific sitting time exceeding 720 minutes (12 hours) or total daily sitting exceeding 1440 minutes (24 hours) were excluded from analyses. The final data consisted of 3695 participants (99% of initial sample) with 3112 participants providing information on sitting in both 2007 and 2014..

(8) Sitting was reported in 2007 and in 2014 as hours and minutes by answering the following question: “How many hours on average do you sit on a weekday?. of. Mark 0 if not at all. a) During the workday in the office or equivalent” (hereafter SIT-work), “b) At home watching television or videos” (hereafter SIT-tv),. ro. “c) At home at a computer” (hereafter SIT-comp), “d) In a vehicle” (hereafter SIT-auto), “e) elsewhere” (hereafter SIT-else). This kind of question that assesses structured weekday sitting in different contexts has shown acceptable measurement properties 19. Missing information on context specific sitting. -p. times was replaced with 0 if the participant had reported sitting in at least one of the other contexts. The sum of the context specific sitting times represents the. re. total daily sitting time.. Marital status was assessed in 2007 as “What is your current marital status?” with response alternatives categorized into 1) married or cohabiting; 2). lP. unmarried; 3) separated or divorced; and 4) widowed.. Occupational status was assessed in 2007 by the following question: “What kind of work do you do most time of the year?” with answers grouped into 1). rn a. Factory, farmer, office worker or similar; 2) Unemployed, student or homemaker; 3) Retired. For sub-analyses, occupational status was dichotomized into working (factory, farmer, and office worker) and not working (including unemployed, student, homemaker and retired). A single question regarding participation in working life during the past 12 months in 2007 (answers yes or no) was used to describe the presence of recent work activity. In 2014 there. Jo u. was a question addressing the year of old-age retirement and disability retirement, respectively, by which those who retired after 2007 were identified. Education was assessed in 2007 as total years of schooling. The years of schooling were divided into thirds representing low, mid and high, in each birth cohort year of the participants 20.. Annual net household income (before tax deductions) was assessed in 2007 with nine predefined categories ranging from less than 10,000 euros to more than 80,000 euros. The responses were grouped into 1) less than 30,000 euros, 2) 30,000 to 60,000 euros, and 3) more than 60,000 euros..

(9) Occupational, commuting and leisure time physical activity were assessed in 2007 with three different questions, respectively, described in detail elsewhere 20.. of. Briefly, the questions were: ”How demanding is your work physically?”, “How many minutes do you walk, ride on a bicycle or otherwise exercise to get to. ro. work?” and “How much do you exercise and stress yourself physically in your leisure time?”. Responses regarding occupational physical activity were categorized as: “Sedentary”, “Light”, and “Moderate-to-high”, regarding commuting physical activity as: “None”, “Less than 30 minutes a day” and “30. -p. minutes or more a day”, and regarding leisure time physical activity as: “None”, “Low” and “Moderate-to-high”.. re. The participants rated their health status in 2007 as good, rather good, moderate, rather poor, or poor. Three groups (good, moderate, and poor) were formed by combining the first two alternatives and the two last alternatives.. lP. Age in 2007 served both as a grouping variable and as covariate in the analyses. In order to distinguish the effect of retirement on the results, the participants were stratified into three age groups with reference to the common retirement age-range of 63 to 68 years. The youngest age group would include participants. rn a. who, based on their age, would not retire before 2014. This group (hereafter “young middle-aged”) included participants younger than 53 years of age in 2007. The middle age group (hereafter “late middle-aged”) comprised participants aged between 53 and 68 years in 2007 and in whom age-related retirement was likely to take place during the follow-up period. The oldest age group (hereafter “older-aged”) included participants aged more than 68 years in 2007, of. Jo u. whom almost all were retired already in 2007.. Background characteristics were presented as frequencies by age group. Overall change in total daily sitting time was calculated as the difference between total daily sitting time in 2007 and 2014 for all participants with data from both time-points (n=3112). According to the distribution of the change in total sitting time, participants were grouped into those with more than 1 SD decrease in total sitting; those with more than 1 SD increase in total sitting; and those with a change within ±1 SD in total sitting time. The association of sociodemographic factors, physical activity and self-rated health with the change in total.

(10) of. sitting was studied using generalized logit models for nominal outcome, where the group within ±1SD change in total sitting served as the reference category. The models were adjusted for age.. ro. The average context specific sitting times by year and corresponding changes by age group were estimated using generalized linear mixed modeling with a t distribution with 3 degrees of freedom of the outcome as the distributions of the sitting variables are skewed. The models were adjusted for age. The. -p. interactions between sociodemographic factors, physical activity and self-rated health and study year were tested for every context and for total sitting,. re. adjusting the models for age. Two fully adjusted models including occupational status (dichotomized into working and not working) and self-rated health, were performed, adjusting for age, gender, education, household income, and self-rated health (when not the determinant), respectively. All analyses were. lP. performed using SAS statistical software (v.9.3.). Results. rn a. Descriptive characteristics of participants in 2007 (n=3695) by age group are presented in supplementary table 1. The young middle-aged were more often women, had high household income and rated their health good compared to the late middle-aged or the older-aged. Among the late middle-aged, 28% retired after year 2007, but in the young middle-aged only 2% and in the older-aged 0.7%, respectively. Compared to those who participated in both the 2007 and the. Jo u. 2014 assessments (n=3735), the non-participants (n=1289) were younger (mean age 51 years, SD 15 vs. 53 years, SD 13), more often men (50% vs. 45%), retired (35% vs. 32%), had low household income (47% vs. 39%) and poor self-rated health (10% vs. 6%).. Table 1 describes the average change in total and context specific daily sitting over the seven follow-up years by age group. The changes in daily context specific sitting times were mostly small and non-significant, around 15 minutes per day in most contexts. Larger changes (-62 minutes and -51 minutes) were observed in SIT-work and in total sitting among the late middle-age group. In relation to baseline level sitting, all statistically significant changes were greater.

(11) of. than 10%, except for SIT-work among the young middle-aged. SIT-comp increased significantly among all age groups, by 13, 17 and 7 minutes for the young. ro. middle-aged, the late middle-aged and the older-aged, respectively (table 1).. Supplementary table 2 presents the age adjusted p-values for all the determinant and study year interactions, as well as the direction of the observed change (+. -p. = increase, - = decrease or ns = not significant). The results are presented by context and for each level of the determinant. Supplementary table 3 shows the. re. adjusted results for changes in context specific and total sitting by dichotomized occupational status (working, not working) stratified by age groups.. lP. In the young middle-aged, occupational status determined the change in context specific sitting times (supplementary table 2). Adjusted models showed that participants with factory, farming or office occupation decreased SIT-work and increased SIT-comp, while the unemployed, student, homemaker or retired decreased SIT-tv and SIT-else and increased SIT-auto significantly over the follow-up (supplementary table 3). Occupational physical activity determined. rn a. change in SIT-work, SIT-else and total sitting, where moderate-to-high occupational physical activity was associated with increase in SIT-work and total sitting, but occupational sedentary time with decrease in SIT-work, SIT-else and total sitting (supplementary table 2). Commuting physical activity determined change in SIT-work, with decrease among those with active commuting. Having children younger than 7 years in 2007 determined change in SIT-work, SIT-. Jo u. auto and total sitting, with increase in total sitting time if having young children and decrease in SIT-work when not having young children. The gender and time interaction was significant for total sitting, but the change was nonsignificant in both men and women per se. Self-rated health determined change in SITwork, SIT-auto, SIT-tv and total sitting. Poor health was associated with increase in SIT-tv, but with decrease in SIT-work, SIT-auto and total sitting.. In the late middle-aged, occupational status significantly determined change in total sitting and in context specific sitting (supplementary table 2). Fully adjusted models showed that factory, farming or office occupation was related with significant decrease in SIT-work, SIT-auto and in total sitting, but increase.

(12) of. in SIT-tv and SIT-comp. The unemployed, student, homemaker or retired significantly decreased SIT-else, SIT-auto and total sitting, and increased SIT-comp over the follow-up (supplementary table 3). Supplementary table 2 further shows that occupational physical activity determined change in SIT-work, SIT-. ro. auto, SIT-else and total sitting, where all levels of occupational physical activity were related with decreased SIT-work and SIT-auto, and occupational sedentary time was associated with decrease in SIT-else and total sitting. Also commuting physical activity determined change in SIT-work and total sitting,. -p. with decrease in both contexts for all levels of commuting physical activity. Leisure time physical activity determined change in SIT-auto, SIT-tv and SITcomp, where all levels of leisure time physical activity were associated with decreased SIT-auto and increased SIT-comp and sedentary and light physical. re. activity with increased SIT-tv. Household net income and education determined change in SIT-work and total sitting and income also determined SIT-else. All. lP. levels of income and education were related with decrease in SIT-work and total sitting, and low income was associated with decrease in SIT-else. Marital status determined change in SIT-comp and SIT-else, where all marital status groups increased SIT-comp, but only the separated and widowed decreased SITelse. Gender determined change in SIT-tv with increase among women. Self-rated health determined change in SIT-work and SIT-else, and both good and. rn a. moderate health was related with decrease in SIT-work and poor health with decrease in SIT-else.. In the older-aged, gender determined change in SIT-tv and marital status determined change in total sitting. Men increased SIT-tv and the cohabited decreased. Jo u. total sitting whereas separated increased total sitting (supplementary table 2). Occupational status determined change in SIT-work, where those with factory, farming or office occupation significantly decreased SIT-work (supplementary table 3).. The fully adjusted associations between self-rated health and changes in context specific sitting time by age group are presented in Table 2. In the young middle-aged, only those with poor self-rated health significantly decreased SIT-work (mean -78 minutes, 95% CI for mean [-115, -40] minutes), and SIT-auto (-25 minutes, [-42, -8] minutes) and total sitting (-73 minutes, [-119, -26] minutes), and increased SIT-tv (+32 minutes, [15, 50] minutes). Among the late.

(13) of. middle-aged, decrease in SIT-work was significant in those with good or moderate self-rated health, and decrease in SIT-else was significant in those with. ro. good or poor self-rated health, respectively.. Table 3 presents the results from the logistic regression models including all participants with both baseline and follow-up total sitting time. The average. -p. change in total daily sitting was -26 minutes (SD 192 minutes). Women were more likely than men to have increased total sitting by more than 1 SD (i.e. an increase in total sitting larger than 166 minutes) over seven years. Compared to the young middle-aged, the late middle-aged were less likely to have increased. re. total sitting time but more likely to have decreased total sitting time by more than 1 SD (i.e. a decrease larger than 218 minutes) over seven years (Table 3).. lP. Baseline occupational status of being unemployed, student or homemaker, as compared to being in factory, farming or office work, was related with almost twofold odds for a large increase in total sitting time over seven years. The retired were less likely than the factory, farming or office work group to decrease total sitting time substantially over the follow-up. Also, mid compared to high education was associated with increased odds for more daily sitting, whereas. Discussion. rn a. low compared to high education associated with less daily sitting over seven years.. Jo u. The changes in total and context specific sitting times over the seven-year follow-up period were mostly small and non-significant. The average change in total daily sitting time was 26 minutes for the full sample. However, among the late middle-aged, SIT-work and total sitting time decreased by nearly an hour a day, and sitting in all other contexts also changed significantly over the follow-up. In all three age groups SIT-comp increased from 2007 to 2014. The impact of baseline occupational status was strongly reflected in the results, with most changes in context specific sitting determined by occupational status in both the young middle-aged and the late middle-aged. Also other determinants closely related to occupational status, such as occupational and commuting physical activity or education and household income, were mainly related with change in SIT-work and total sitting, and partly with SIT-auto. Interestingly,.

(14) of. poor self-rated health determined less SIT-work, SIT-auto and total sitting, but more SIT-tv among the young middle-aged, whereas good and moderate health. ro. determined decrease in SIT-work in the late middle-aged.. The average decrease in total daily sitting was modest and varied from 1 minute in the young middle-aged, to 51 minutes in the late middle-aged. In line with. -p. earlier findings, sitting seems to be a relatively stable behavior. One large, prospective population cohort with 37 000 adults aged 45-years and older similarly observed a 24 minute decrease in total daily sitting time, although in a shorter follow-up time (4 years)9. Also in smaller population-based samples and. re. constituting only women, decreases in total sitting time of the same magnitude have been reported 21, 22. Some have found marginal, albeit not statistically. lP. significant increases in total sitting time over a ten year follow-up when assessing sitting in predefined time categories 11.. The changes in context specific sitting times did not appear much larger than the changes in total sitting. However, when relating to the baseline level, the. rn a. changes were at least moderate (>10%) in many contexts, but mainly then among the late middle-aged and the older-aged. Comparison to previous studies is limited, as only few studies have assessed context specific sitting times as the outcome. Some have used time spent sedentary as the outcome, which does not explicitly mean sitting, and others ask only for time spent in a certain domain such as time spent in front of a screen. Among Chinese adults non-occupational. Jo u. sedentary time including activities such as reading, drawing, watching TV, DVDs, videos and surfing and chatting via the internet increased by 12 to 14% from 2004 to 2011 23. In Australian full time employed adults, aged 15-years or older, the average work-related sitting time remained between 227 and 233 minutes and leisure time sitting on a typical workday between 187 and 206 minutes between 2007 and 2014 4. Among Danish working-aged adults, an 18% increase in work-related sitting between 1990 and 2010 was evident only for people with high socioeconomic status 24. Smaller population-based studies (n < 1000) have reported increases in average daily TV time, by 4 minutes over four years 15 or increased screen time by two hours over ten years 25..

(15) Occupational status was strongly related with the changes in context specific and total sitting. Also, the determinants for SIT-work were often also identified. of. as determinants for total sitting. For working-aged people, work-related sitting takes up a major part of daily total sitting time 24, 26 and it is natural that. ro. changes in work-related sitting time contribute largely to the changes in total sitting time. In the current study, most of the changes in sitting times between 2007 and 2014 were observed in the late middle-aged who consisted of participants vulnerable for retirement with reference to their age in 2007. In 2014,. -p. almost one third among the late middle-aged reported that they had retired after 2007. For mid-aged women, retiring was the strongest predictor of decrease in total sitting time 10. Considering only non-occupational sitting, retirement increased total sitting time (self-reported), with significant increases in sitting in. re. front of the TV, at the computer and other sitting, but decrease in sitting in vehicle 12. The increase in total non-occupational sitting over retirement is found to. lP. be larger in women than men and in those with initially higher occupational sitting time 12. We also observed, that among all participants, students, homemakers and the unemployed at baseline were likely to increase total daily sitting over the follow-up. Contrary, in another study, young women who began to work or had an increase in income were found to decrease total sitting time 10. Although retirement and starting to work during the follow up cannot. rn a. completely be determined from the current questionnaire data, the results support the conclusion that those who leave working life reduce their work-related sitting, and therefore total sitting, the most. In addition, changes in SIT-auto may also display the transition out of or into working life, through effects in. Jo u. commuting prevalence.. SIT-comp was the only context besides SIT-work, where all age groups changed i.e. increased sitting time over the follow up. However, the reasons for increased SIT-comp may be different for the different age groups. Particularly in the young middle-aged, the observed decrease in SIT-work and increase in SIT-comp among the factory, farming or office work group can reflect increase in remote work at home over the years. Findings from other cross-sectional and prospective as well as from time-use studies, have likewise suggested that time spent at computer has increased over the past years, also among older people 6, 13. For retiring persons, a stronger increase in sitting at the computer has been observed for such mid-aged adults who have been engaged in a.

(16) of. sedentary rather than more physically active occupation 12. Thus, in mid-aged and older adults increased sitting at computer may be related with previous work-related computer use. In the future, people may further replace traditional sitting by the TV with sitting at the computer. Having portable devices like. ro. smart phones and tablets in peoples’ everyday lives can in that case be an opportunity also to reduce daily sitting time as such devices can be operated without. -p. being in sitting position.. Self-rated poor health determined less SIT-work and more SIT-tv among the young middle-aged, whereas better self-rated health determined a decrease in. re. SIT-work among the late middle-aged. In all age-groups, the result showed initially lower SIT-work and higher SIT-tv (cross-sectional result) in those with. lP. poor health compared to the ones in better health. Even if causality cannot be confirmed, the finding may reflect the important influence of health status on the ability to work and participate in working life and thereby also on sitting. If less SIT-work and more SIT-tv in young adults with poor self-rated health means leaving working life due to poor health, this could to a larger part be prevented unlike age-related retirement that also increases non-occupational sitting 12. In. rn a. French adults aged 35 to 65 years, better self-rated general health predicted lower screen-viewing time and total sitting time over 10-years 27. In Australian mid-aged adults (45-64 years), obesity (body mass index >30kg/m2) predicted an increase in total sitting time whereas sitting did not predict obesity, suggesting important reverse causality between health and sitting time 28. Obesity was related with increased sitting only among those who initially sat less. Jo u. than 8 hours and those not working or in part time employment at baseline 28.. Sitting has been announced as a risk-factor for subsequent poor health and even mortality 1, but the reverse causality and the moderating role of sociodemographic factors clearly needs more attention in future work. The scientific evidence is still too sparse for more specific public health guidelines regarding sitting and advising people to increase physical activity is still the best way when encouraging people to adopt a healthy lifestyle 29. Even if device-.

(17) of. based sitting data may be required for the development of quantitative guidelines regarding sitting, the variation in how different population groups allocate. ro. and change their daily sitting time in different contexts is important knowledge gained from epidemiological self-report data.. The current study has its limitations. Sitting was assessed by questionnaire and may be distorted by reporting bias. The participants were asked to report 0 if. -p. they did not sit in some of the predefined contexts. Nevertheless, there were missing data in the sitting questionnaire, and these were imputed with zeros in the analyses. The total sitting time computed based on sitting time in five contexts will inevitably leave out some time not fully covered by the predefined. re. contexts. Device-based assessment enables capturing the total sedentary or sitting time in more detail and detects smaller changes in daily sitting time. But,. lP. device-based assessment does not allow recognizing the context of sitting time. All determinants were assessed at baseline and they may also have changed during the follow up period which was not accounted for. Also, the relationships of smoking, alcohol, or diet with changes in sitting were not studied and this warrants further examination. Different occupational types could not be evaluated separately as they were all combined in one category. The current study. rn a. suffered from drop-out in three phases. First, of the 9957 initially invited adults, 63% participated in the FINRISK 2007 health examination survey. This participation rate is quite good and in line with such large epidemiological studies in Europe 30. Second, a small proportion of participants (n=1234, 20%) did not complete the DILGOM sub-study in 2007 and third, not all participated in the 2014 follow-up study. Our results are not generalizable to the population. Jo u. below 25 years or above 74 years, but do apply to a large proportion of adults in Finland.. The strengths of the study include a large population-based sample with a long follow-up. Sitting in different contexts was assessed in the same structured way at both time points, allowing for context specific comparison between time points. Furthermore, the questions specifically addressed sitting and not only context specific time or sedentary time in general. Sitting time was also assessed as a true continuous variable, allowing studying minute-level changes..

(18) of. Conclusion Between 2007 and 2014, total daily sitting time decreased on average by 26 minutes. Sitting at home at the computer increased for all age groups between 7 to. ro. 17 minutes. The late middle-aged changed all context specific sitting, notably there was a decrease in work-related and total sitting and an increase in sitting at home at the computer and by the TV. The findings of our study, even though modest in magnitude, are important for tackling a sedentary lifestyle as the. -p. natural changes in context specific sitting times over time are small and determined somewhat differently in different age groups. Conflict of interest. re. Authors declare no conflict of interest.. lP. Acknowledgements. This study was funded by the Ministry of Culture and Education. References. rn a. (OKM/67/626/2018). Mäki-Opas, T was supported by the Academy of Finland (#303615).. Jo u. 1. Dunstan, D. W., Howard, B., Healy, G. N. & Owen, N. Too much sitting - A health hazard. Diabetes Res. Clin. Pract. 97, 368-376 (2012). 2. Harrington, D. M., Barreira, T. V., Staiano, A. E. & Katzmarzyk, P. T. The descriptive epidemiology of sitting among US adults, NHANES 2009/2010. J. Sci. Med. Sport 17, 371-375 (2014).. 3. Loyen, A., van der Ploeg, H P, Bauman, A., Brug, J. & Lakerveld, J. European Sitting Championship: Prevalence and Correlates of Self-Reported Sitting Time in the 28 European Union Member States. PLoS One 11, e0149320 (2016)..

(19) of. 4. Loyen, A. et al. Recent trends in population levels and correlates of occupational and leisure sitting time in full-time employed Australian adults. PLoS One 13, e0195177 (2018).. ro. 5. Milton, K., Gale, J., Stamatakis, E. & Bauman, A. Trends in prolonged sitting time among European adults: 27 country analysis. Prev. Med. 77, 11-16 (2015).. -p. 6. Chau, J. Y. et al. Temporal trends in non-occupational sedentary behaviours from Australian Time Use Surveys 1992, 1997 and 2006. Int. J. Behav. Nutr. Phys. Act. 9, 7-76 (2012).. lP. Danish adults. Scand. J. Public Health 41, 706-711 (2013).. re. 7. Aadahl, M. et al. Recent temporal trends in sleep duration, domain-specific sedentary behaviour and physical activity. A survey among 25-79-year-old. 8. Prince, S. A., Reed, J. L., McFetridge, C., Tremblay, M. S. & Reid, R. D. Correlates of sedentary behaviour in adults: a systematic review. Obes. Rev. 18, 915-935 (2017).. rn a. 9. Astell-Burt, T., Feng, X., Kolt, G. S. & Jalaludin, B. Is More Area-Level Crime Associated With More Sitting and Less Physical Activity? Longitudinal Evidence From 37,162 Australians. Am. J. Epidemiol. 184, 913-921 (2016). 10. Clark, B. K., Peeters, G. M., Gomersall, S. R., Pavey, T. G. & Brown, W. J. Nine year changes in sitting time in young and mid-aged Australian women:. Jo u. findings from the Australian Longitudinal Study for Women's Health. Prev. Med. 64, 1-7 (2014). 11. Gebel, K. et al. Patterns and predictors of sitting time over ten years in a large population-based Canadian sample: Findings from the Canadian Multicentre Osteoporosis Study (CaMos). Prev. Med. Rep. 5, 289-294 (2017). 12. Leskinen, T. et al. Changes in non-occupational sedentary behaviours across the retirement transition: the Finnish Retirement and Aging (FIREA) study. J. Epidemiol. Community Health 72, 695-701 (2018)..

(20) of. 13. Van Dyck, D., Cardon, G. & De Bourdeaudhuij, I. Longitudinal changes in physical activity and sedentary time in adults around retirement age: what is the moderating role of retirement status, gender and educational level? BMC Public Health 16, 112-4 (2016).. ro. 14. Busschaert, C., De Bourdeaudhuij, I., Van Cauwenberg, J., Cardon, G. & De Cocker, K. Intrapersonal, social-cognitive and physical environmental variables related to context-specific sitting time in adults: a one-year follow-up study. Int. J. Behav. Nutr. Phys. Act. 13, 2-1 (2016).. -p. 15. Ding, D. et al. Correlates of change in adults' television viewing time: a four-year follow-up study. Med. Sci. Sports Exerc. 44, 1287-1292 (2012). 16. Yang, X. et al. Long-term determinants of changes in television viewing time in adults: Prospective analyses from the Young Finns Study. Scand. J. Med.. re. Sci. Sports 28, 2723-2733 (2018).. lP. 17. Borodulin, K. et al. Cohort Profile: The National FINRISK Study. Int. J. Epidemiol. (2017). 18. Kanerva, N. et al. Adherence to the healthy Nordic diet is associated with weight change during 7 years of follow-up. Br. J. Nutr. 120, 101-110 (2018). 19. Marshall, A. L., Miller, Y. D., Burton, N. W. & Brown, W. J. Measuring total and domain-specific sitting: a study of reliability and validity. Med. Sci.. rn a. Sports Exerc. 42, 1094-1102 (2010).. 20. Borodulin, K. et al. Time trends in physical activity from 1982 to 2012 in Finland. Scand. J. Med. Sci. Sports 26, 93-100 (2016). 21. Uijtdewilligen, L. et al. Biological, socio-demographic, work and lifestyle determinants of sitting in young adult women: a prospective cohort study. Int. J.. Jo u. Behav. Nutr. Phys. Act. 11, 7 (2014).. 22. Gradidge, P. J., Norris, S. A., Munthali, R. & Crowther, N. J. Influence of socioeconomic status on changes in body size and physical activity in ageing black South African women. Eur. Rev. Aging Phys. Act. 15, 8. eCollection 2018 (2018). 23. Su, C. et al. Longitudinal association of leisure time physical activity and sedentary behaviors with body weight among Chinese adults from China Health and Nutrition Survey 2004-2011. Eur. J. Clin. Nutr. 71, 383-388 (2017)..

(21) of. 24. van der Ploeg, H P, Moller, S. V., Hannerz, H., van der Beek, A J & Holtermann, A. Temporal changes in occupational sitting time in the Danish workforce and associations with all-cause mortality: results from the Danish work environment cohort study. Int. J. Behav. Nutr. Phys. Act. 12, 7-1 (2015).. ro. 25. Knaeps, S. et al. Ten-year change in sedentary behaviour, moderate-to-vigorous physical activity, cardiorespiratory fitness and cardiometabolic risk: independent associations and mediation analysis. Br. J. Sports Med. 52, 1063-1068 (2018).. -p. 26. Saidj, M. et al. Descriptive study of sedentary behaviours in 35,444 French working adults: cross-sectional findings from the ACTI-Cites study. BMC Public Health 15, 37-8 (2015).. re. 27. Omorou, A. Y. et al. 10-year cumulative and bidirectional associations of domain-specific physical activity and sedentary behaviour with health-related. lP. quality of life in French adults: Results from the SU.VI.MAX studies. Prev. Med. 88, 66-72 (2016). 28. Pedisic, Z. et al. High sitting time or obesity: Which came first? Bidirectional association in a longitudinal study of 31,787 Australian adults. Obesity (Silver Spring) 22, 2126-2130 (2014).. rn a. 29. Stamatakis, E. et al. Is the time right for quantitative public health guidelines on sitting? A narrative review of sedentary behaviour research paradigms and findings. Br. J. Sports Med. 53, 377-382 (2019).. 30. Mindell, J. S. et al. Sample selection, recruitment and participation rates in health examination surveys in Europe--experience from seven national surveys.. Jo u. BMC Med. Res. Methodol. 15, 7-4 (2015).. Table 1. Baseline 2007 and follow up 2014 context specific sitting times presented as mean minutes per day (95% CI), by age groups. The change between the years is shown as mean minutes with 95% confidence intervals and as a relative change (%). Context SIT-work. Young middle-aged (<53 y) 2007 2014 Change 214 204 -10 / -5%. Late middle-aged (53-68y) 2007 2014 Change 111 49 -62 / -56%. 2007 19. Older-aged (>68y) 2014 10. Change -9 / -47%.

(22) (-18, -2). (104, 118). (42, 56). 49 (46, 53). -1 / -2% (-5, 3). 38 (35, 41). 26 (24, 29). 113 (109, 117). -0.3 /-0.3% (-4, 3). 148 (144, 153). 54 (51, 57). +13 / +32% (9, 16). 33 (29, 36). 24 (21, 27). -3 / -11% (-6, 1). 445 (435, 456). -1 / -0.2% (-10, 9). (13, 26). (4, 17). (-17, -0.5). -12 / -32% (-15, -9). 22 (19, 25). 16 (12, 19). -7 / -32% (-10, -3). 164 (159, 168). +15 / +10% (11, 20). 190 (180, 200). 195 (185, 206). +5 / +3% (-7, 17). 49 (46, 53). +17 / +52% (13, 20). 27 (22, 33). 34 (29, 40). +7 / +26% (2, 12). 43 (39, 47). 34 (30, 38). -9 / -21% (-14, -3). 44 (36, 51). 31 (23, 40). -12 / -27% (-23, -2). 373 (363, 382). 322 (312, 331). -51 / -14% (-61, -41). 303 (289, 317). 287 (272, 302). -17 / -6% (-34, 1). ro. -p. re. (-70, -54). of. (195, 213). lP. (206, 223) age group*time p<0.0001 SIT-auto 50 (47, 54) age group*time p<0.0001 SIT-tv 113 (109, 117) age group*time p<0.0001 SIT-comp 41 (38, 44) age group*time p=0.023 SIT-else 27 (24, 29) age group*time p=0.054 Total sitting 446 (436, 456) age group*time p<0.0001. Jo u. rn a. Note 1: Models adjusted for age. The interaction between age group and time is estimated from separate models including the total sample while average change is estimated in models stratified by age group. Note 2: Bolded numbers indicates significant (p<0.05) change in sitting time. SIT-work: work-related sitting, SIT-auto: sitting in vehicles, SIT-tv: sitting at home in front of TV, SIT-comp: sitting at home at computer, SIT-else: sitting elsewhere, Total sitting: sum of context specific sitting times.

(23) 21. Table 2. Sitting time as mean minutes / day (95% CI) in 2007 and 2014 by self-rated health in three age groups.. 2007. Moderate. 2014 2007. Poor. 2014 p for self-rated health p for self-rated health*time Self-rated health. Year 2007. Good. 2014 2007. Moderate. 2014 2007. Poor. 2014. SIT-work 136 (127, 146) 59 (50, 69) 102 (91, 114) 50 (37, 62) 76 (51, 101) 61 (35, 88). SIT-tv 142 (136, 148) 157 (151, 163) 152 (144, 159) 169 (161, 177) 186 (170, 203) 185 (167, 202). 0.001 <0.0001. <0.0001 0.141. SIT-work 32 (21, 43) 16 (4, 28) 26 (15, 37) 17 (5, 29) 21 (-4, 47) 29 (1, 56). SIT-tv 166 (149, 184) 180 (161, 198) 189 (172, 207) 196 (177, 214) 226 (185, 266) 180 (136, 223). 0.855 0.367. 0.045 0.080. Self-rated health Good. Year 2007 2014 2007. ur. Moderate. na. p for self-rated health p for self-rated health*time. 2014 2007. Jo. Poor. 2014. p for self-rated health p for self-rated health*time. SIT-else Total sitting 25 443 (22, 28) (431, 455) 23 445 (20, 26) (433, 457) 24 459 (18, 29) (439, 480) 25 463 (19, 31) (441, 485) 51 517 (38, 63) (471, 563) 38 444 (24, 51) (396, 492) 0.001 0.096 0.314 0.009. ro of. 2014. SIT-tv 112 (108, 117) 108 (103, 113) 116 (108, 123) 119 (111, 127) 140 (123, 158) 173 (154, 191) <0.0001 0.0002. -p. 2007. Good. SIT-work 216 (206, 226) 212 (202, 223) 217 (199, 235) 202 (183, 220) 199 (159, 238) 121 (80, 162) 0.016 0.001. re. Year. lP. Self-rated health. Young middle-aged (<53y) SIT-comp SIT-auto 39 51 (36, 43) (47, 55) 49 52 (45, 53) (48, 57) 39 63 (33, 45) (56, 70) 58 59 (52, 65) (51, 66) 64 63 (50, 77) (48, 79) 75 39 (61, 90) (22, 55) <0.0001 0.027 0.123 0.008 Late middle-aged (53-68y) SIT-comp SIT-auto 33 40 (29, 37) (36, 44) 51 28 (46, 55) (25, 32) 32 42 (27, 38) (37, 46) 47 27 (41, 53) (22, 32) 39 27 (27, 51) (17, 37) 60 24 (48, 72) (13, 34) 0.205 0.113 0.546 0.181 Older-aged (>68y) SIT-comp SIT-auto 38 23 (28, 48) (18, 28) 42 17 (32, 52) (11, 22) 28 23 (18, 38) (18, 28) 32 15 (22, 43) (10, 21) 16 30 (-6, 39) (18, 42) 42 23 (19, 66) (11, 36) 0.152 0.130. 0.337 0.896. SIT-else Total sitting 39 391 (34, 45) (378, 404) 32 327 (26, 38) (314, 340) 43 371 (36, 50) (355, 387) 37 329 (29, 44) (312, 346) 64 392 (48, 80) (358, 427) 23 351 (6, 40) (315, 387) 0.356 0.008. 0.330 0.123. SIT-else Total sitting 43 302 (29, 57) (278, 327) 34 288 (19, 48) (262, 314) 45 311 (31, 59) (286, 336) 31 291 (17, 46) (265, 317) 84 378 (52, 115) (321, 435) 36 309 (2, 70) (248, 370) 0.244 0.244. 0.148 0.365. Note 1: Models are adjusted for age, gender, education, and income. Bolded numbers indicate statistical significance (p<0.05)..

(24) 22. Table 3. Odds ratios for decrease in total sitting by more than 218 minutes (-1 SD) or increase in total sitting by more than 166 minutes (+1 SD) between 2007 and 2014 as compared with an average (between -218 to 166 minutes) change. Decrease more than 218 minutes in total sitting time (n=364). Between -218 and +166 minutes change in total sitting time (n=2334). Increase more than 166 minutes in total sitting time (n=414). Gender Men Women 0.98 (0.79-1.20). Ref. 1.29 (1.03 – 1.62). Age group Young middle-aged (<53y) Late middle-aged (53-68y) 1.59 (1.27 – 1.98) Older-aged (>68y) 0.83 (0.56 – 1.24) Occupational status. Ref.. Factory, farming, office work Student, homemaker, unemployed 0.80 (0.54 – 1.19) Retired 0.32 (0.23 – 0.45). Ref.. Net household income. Civil status. Self-rated health. na. Married or cohabiting Unmarried 1.24 (0.91 – 1.70) Separated 0.88 (0.59 – 1.29) Widowed 1.23 (0.73 – 2.06). ro of. 1.86 (1.34 – 2.59) 1.39 (0.99 – 1.96). 0.99 (0.74 – 1.32) 1.31 (1.01 – 1.69). Ref.. 0.77 (0.58 – 1.04) 0.95 (0.72 – 1.25). lP. ≤30 000 euros 30-60 000 euros >60 000 euros. re. Low 0.67 (0.51 – 0.89) Mid 1.04 (0.82 – 1.32) High. -p. Education. 0.58 (0.46 – 0.75) 0.74 (0.52 – 1.07). 1.35 (0.96 – 1.90) 1.33 (0.95 – 1.84). Ref. Ref. 1.21 (0.88 – 1.66) 1.09 (0.74 – 1.61) 1.08 (0.57 – 2.08). Jo. ur. Good Ref. Moderate 0.99 (0.78 – 1.26) 1.24 (0.97 – 1.59) Poor 1.47 (0.99 – 2.20) 1.01 (0.60 – 1.69) Note 1: Based on generalized logit models adjusted for age. Bolded numbers indicate p< 0.05..

(25)