Background
Sedentary behaviour, defined as any waking activity characterized by an energy expenditure of ≤ 1.5 metabolic equivalents and a sitting or reclining posture [
1], has increased considerably in countries at all levels of development [
2‐
4]. This increase is mainly attributed to changes in transport, entertainment and workplace environments, and has been linked to increases in obesity and related chronic diseases [
5,
6]. Disentangling the nature of this association is complicated by interactions with other lifestyle factors, such as dietary habits [
7,
8]. In particular, it is hypothesized that the association between television viewing and obesity may be due to the increased snacking behaviour associated with television viewing, rather than the action of sitting in front of the television [
7].
This hypothesis is derived from evidence for an association between television viewing and unhealthy dietary habits [
7,
9,
10]. However, the available evidence is strong for children and adolescents, but limited for adults [
7,
9]. Moreover, previous studies focused solely on television viewing, only one domain of sedentary behaviour. Other domains, including transport-related, work-related, and other leisure-time sedentary behaviour [
11], have not to date been associated with dietary habits. Nevertheless, there is a need to gain insight into any potential associations between these domain-specific sedentary behaviours and dietary habits, as many adults spend a substantial amount of time sitting in contexts other than domestic television viewing [
12‐
14].
Additionally, it may be that the association between television time and dietary habits does not apply to everyone. Research by Pearson et al. [
7] suggests that two demographic characteristics may moderate the association, namely age and gender. However, here too, evidence in adults is limited [
7].
To address these gaps in the evidence, the first aim of this study was to explore associations between domain-specific sedentary behaviours and dietary habits in adults. Second, it assessed whether any association differed by age and gender. A better understanding of these associations could help guide future health promotion interventions.
Discussion
Our first aim was to determine whether there are associations between domain-specific sedentary behaviours and dietary habits in adults. After controlling for socio-demographic factors, BMI and physical activity, eleven significant associations were found, of which nine showed that higher levels of work or leisure time (ie television time, computer time and other leisure sitting time) sedentary behaviour were associated with less healthy dietary habits. This finding may be important given that clustering of unhealthy behaviours has been found to have synergistic effects on health, which implies that the combined effects are more harmful than those from the sum of the individual unhealthy behaviours [
32]. However, although several statistically significant associations were found, they do not necessarily reflect meaningful associations. Both the large sample size of this study, which may have resulted in an over rejection of the null hypothesis, and the small effect sizes, call into question the clinical relevance of the significant sedentary behaviours and dietary habits associations. In view of this, the main focus of this discussion will be on the four associations that were significant at the .001 level.
These four associations all showed that television viewing was unfavourably related to dietary habits. This is noteworthy, as television time is only one of the five domain-specific behaviours that were included in this study. More concretely, spending more time watching television was related to less frequent consumption of fruit and vegetables, and more frequent consumption of sugar-sweetened beverages and fast food. These findings confirm previously reported results [
33‐
36], and could be explained by disrupted habituation to food cues [
37,
38] or by increased exposure to unhealthy food advertisements [
39,
40]. For example, Scully et al. reported that respondents were significantly more likely to eat fast foods for snacks at least once weekly if they usually watched commercial television for two or three hours/day compared to those who watched commercial television for less than two hours/day [
40]. Food advertisements may not only contribute to increases in unhealthy dietary habits by promoting unhealthy food options, they may also contribute to a reduction in consumption of healthy foods, such as fruit and vegetables, through potentially misleading messages about the nutritional value of food items [
41]. Furthermore, individual-level factors, such as attitudes and norms about healthy lifestyles, or a number of other aspects affecting motivation, may also partially explain these results. For example, results from Mata et al. [
42] suggest the existence of specific motivational “spill-over” effects across health behaviours during lifestyle change. The many associations between television time and unhealthy dietary habits seem likely to explain, at least in part, the adverse impact of television time on obesity and related chronic diseases [
43‐
45], and thus clarify the stronger negative influence of television time on health compared to total sedentary behaviour. In this way, results of this study also contribute to resolve the controversy surrounding the direct (ie the action of sitting in front of the television leads to obesity) or indirect (ie sitting in front of the television leads to obesity via dietary habits) influence of television time on obesity and related chronic diseases, by suggesting that the effects are rather indirect.
In contrast, associations with other sedentary behaviours are weaker, less clinically relevant, and less consistent. For example, computer time was related to both unhealthy and healthy dietary habits: spending more computer time was related to both eating fruit and drinking alcohol less frequently. Given the considerable difference between television and computer time on dietary habits, our findings support the importance of analysing these two behaviours separately, rather than using a summary construct such as ‘screen time’, used in some previous papers (eg [
46‐
48]). Other leisure sitting time was associated with two unhealthy dietary habits: more frequent consumption of both alcohol and fast food. However, the association with alcohol consumption was only significant in adults aged below 65 years, which is consistent with the fact that alcohol consumption decreases with increasing age [
49]. As far as we know, there are no studies on the association of other types of leisure sitting time on dietary habits. However, given that other leisure time activities form a heterogeneous grouping, including visiting restaurants, socializing with friends, or going to a pub, the positive association with alcohol consumption and fast food intake could be expected. Nevertheless, more research is needed to confirm these results. Finally, spending more time sitting for work was positively related to alcohol consumption and negatively related to fruit intake. A cross-sectional Australian study found that those who sat longer at work had greater psychological distress [
50], itself plausibly associated with higher alcohol consumption [
51]. However, evidence is lacking on both the direction of causality and generalizability. On the other hand, evidence of an association between sedentary behaviour and fruit intake has been conflicting. Contrary to our findings, Pereira et al. showed that higher sitting time at work was associated with higher fruit intake [
52]. The reason for these inconsistent results is unclear; however, it might be that other characteristics, such as type of work, may have influenced the relationship.
The second aim of this study was to assess whether any association differed by age and gender, as previous studies showed mixed results concerning the moderating role of age and gender [
7]. Our results did not provide evidence to support a potential moderating role of age or gender, given that only one significant moderating effect was found, ie age moderated the association between leisure sitting time and alcohol consumption. Consequently, most associations between domain-specific sedentary behaviour and dietary habits seem to be consistent across men and women, and across adults and older adults.
The main strength of this study was its innovative nature, as this study was the first to link different domain-specific sedentary behaviours to dietary habits. Examining domain-specific sedentary behaviour is important, as specific associations may be masked when analysing total sedentary behaviour. A second strength was the large sample size, which ensures adequate power to identify associations. A final strength was that the study sample was recruited from five urban regions in different European countries, which increases the external validity of our findings to Europe.
Despite these strengths, some limitations affect the validity and generalizability of the results. Firstly, information on both domain-specific sedentary behaviours and dietary habits was self-reported and therefore prone to social desirability and recall biases. Moreover, dietary habits were measured using single items, which may have reduced the accuracy. As both the predictor variables, the outcomes and the confounders were self-reported, observed associations may be the result of correlated error. Future studies should use more precise dietary assessment with quantitative assessment of dietary intake, preferably in combination with objective measurement instruments for sedentary behaviour. These objective measurement instruments (eg accelerometers, or inclinometers [
53]) should be combined with Global Positioning Systems and/or diaries to gain insight into domain-specific sedentary behaviours. Secondly, the cross-sectional study design does not allow for causal inferences between domain-specific sedentary behaviours and dietary habits. Finally, despite sending reminders, the response rate was relatively low, which may have resulted in a selection bias. Although there is a good representation of men (44 %) and women (56 %), lower (46.4 %) and higher (53.6 %) educated individuals as well as younger (from age 18 years) and older (up to age 109 years) adults [
31], it remains likely that generally more healthy people participated, suggesting that we may have underestimated domain-specific sedentary behavior and unhealthy dietary habits. Possible reasons for the low response rate include first the oversampling of low SES residents. Low SES residents have been shown to be less likely to participate in a health survey [
17]. However, as we aimed to have a heterogeneous sample with as many low SES residents as high SES residents, we decided to oversample the former, which is likely to have led to a lower overall response rate. Secondly, with regard to the absence of an upper age limit, we know that there may be attrition in surveys where older people are less likely to be able to complete a survey due to, for example, limited cognitive function, or vision impairment [
54]. In addition, the questionnaire was mainly administered online. Previous studies have indicated that Internet use drops off significantly after the age of 75 [
55], also potentially contributing to a lower response rate. Thirdly, the survey was relatively long. Participants spend on average 25.1 ± 12.4 min to complete the questionnaire, which contained 50 key questions on 30 pages. Finally, we recognize that, in an era of frequent opinion polls and market research, people may react to what they perceive as over-surveying (i.e. become fed up with surveys). Although each of these factors, on their own, may not have had a large impact, they all act to reduce the response rate so, in combination, the effect may be appreciable.
Acknowledgements
We would like to thank Hélène Charreire, Thierry Feuillet, and Maher Ben-Rebah (Equipe de Recherche en Epidémiologie Nutritionnelle, Université Paris 13) for their contribution to WP3 of the SPOTLIGHT project.