This study examined relationships between psychosocial stress and distress with sedentary behaviors and physical activity in two different samples. We chose the Perceived Stress Scale and the Kessler-10 Distress Scale because these questionnaires are widely used among both the general population and among pregnant women, which allows for comparison with other studies. These measures encompass many similar constructs but cannot be directly compared among the two samples. Similarly, we cannot directly compare objectively measured sitting time and steps per day in the Montreal cohort to self-reported time spent in sedentary leisure activities and estimated energy expenditure in leisure and transport activities in the CCHS. However, comparison of the general patterns provides a broad overview of relationships between psychosocial stress with sedentarity and physical activity that helps to contextualize results and that points to areas for future research.
4.1. Sedentary Behavior
Sitting time in the Montreal cohort averaged more than 8 h per day. Similar accelerometer-based studies in the US showed estimates of 424 min (around 7 h) per day in sedentary behavior, with only minor variations by trimester [7
]. Others showed estimates of 12.4 h per day at gestation week 18 and 12.9 h per day at week 35 [8
]. Studies in Spain were similar, with women spending 10.0 h per day in sedentary behavior, and no notable differences between weekday (10.1 h/day) and weekend (9.9 h/day) estimates [24
]. Calculation of the percent of time spent in sedentary behavior depends on the bout length of interest, which complicates the direct comparison of accelerometer-measured sedentary behaviors across studies [25
]. Nevertheless, results of the studies cited above are consistent in showing that pregnant women spend a large percentage of their waking hours in sedentary behavior.
Self-reported participation in leisure-time sedentary behaviors in the CCHS was also consistent with other studies, averaging 3.3 h per day. This is similar to our estimates of screen sedentary behavior among men and women in the general population, which ranged from around 18 to 25 h per week (2.6–3.6 h/day) and varied by gender and sociodemographic characteristics [16
]. Among pregnant women, studies in the US that assessed responses to 2 questions on time spent watching TV and sitting quietly performing other activities showed an average of 2.6 h of sedentary behavior per day [26
]. Studies in Spain using the Sedentary Behavior Questionnaire (SBQ) showed that women self-reported around 8.7 h of sedentary behavior per day [24
]. The SBQ assesses 11 sedentary behaviors including leisure (e.g., television, playing a musical instrument), transport (e.g., driving or traveling in a motor vehicle), rest, and work activities, so estimates are higher than for leisure activities alone.
In both samples, our results show relationships between psychosocial stress and sedentary behavior, and these relationships persisted when controlling for important sociodemographic covariates. In the Montreal cohort, greater perceived stress was associated with greater sitting time at 16–18, 24–26, and 32–34 weeks pregnancy, with small to moderate effect sizes. Similarly, greater psychological distress predicted greater leisure-time sedentary behavior among pregnant women in the general Canadian population. Effect sizes were much smaller in the CCHS, highlighting the great deal of variability at the nationwide level. Furthermore, imprecisions in self-reported estimates of leisure-time physical activity might weaken the relationships observed. Nevertheless, results highlight overall that the level of sedentarity among pregnant women in Canada is high, and that stress and distress represent risk factors for these patterns.
These relationships are consistent with results among adults in other countries. For example, among Belgian adults, symptoms of psychological distress predicted greater sitting time [27
], and among US Latino adults, chronic and lifetime stressors predicted sedentary behavior [28
]. Similar results have been observed for screen sedentary behavior. Studies among Australian women in socioeconomically disadvantaged neighborhoods showed that psychological distress predicted increased television viewing [29
], and studies among Puerto Rican adults living in the US showed that those with the highest perceived stress had greater television viewing time [15
]. Our own studies among Canadian adults showed positive relationships between psychological distress and sedentary behavior among all adults, with variations by sociodemographic characteristics such as ethnicity, immigration status, and education [16
]. However, we identified very few similar studies among pregnant women. Researchers in Brazil observed no associations between perceived stress during pregnancy and low levels of physical activity derived from self-report responses on the International Physical Activity Questionnaire. In fact, lower levels of anxiety were associated with higher rates of physical inactivity. The authors note that women’s physical activity in the setting revolves largely around childcare and housework, which might be a source of anxiety [30
]. Results highlight the need for more studies of relationships between mental health and sedentary behavior among pregnant women in different settings.
4.2. Physical Activity
In the Montreal cohort, steps per day averaged 7878 in early pregnancy and 6273 in late pregnancy. These figures are consistent with accelerometer data among a sociodemographically diverse sample of women at less than 25 weeks of pregnancy in Norway, which showed step counts ranging from 7718 to 9603 depending on ethnicity [31
]. Similar studies among Caucasian women in Spain assessed at 16 weeks’ gestation showed an average of 7745 steps per day [32
]. Systematic reviews highlight that both the frequency and intensity of physical activities tend to decline over the course of pregnancy [33
], consistent with the trend toward decline from Evaluation 1 to Evaluation 3 observed in our study.
Energy expenditure in the CCHS sample averaged 1.6 kcal/kg/day. This is consistent with observations among pregnant women in England, which showed mean leisure energy expenditure, based on estimations from self-report questionnaires, of 1.9 kcal/kg/day in the first trimester, 1.3 in the second trimester, and 1.2 in the third trimester [34
]. These values are consistent with moderate leisure activity levels (defined as daily energy expenditure of 1.5 to <3.0 metabolic equivalents) [35
Relationships between stress or distress and the measures of physical activity analyzed were less consistent than relationships with sedentary behavior. In the Montreal cohort, relationships between perceived stress and steps per day were significant in early and mid but not in late pregnancy. The lack of a significant relationship in late pregnancy would be expected to weaken any associations in analyses including women in all stages of pregnancy, and indeed, there were no relationships between distress and energy expenditure in the CCHS sample. Other population-wide studies of psychosocial health and physical activity have shown mixed relationships. For example, studies from the US showed that perceived stress was associated with less leisure time physical activity [15
]. Similarly, psychological distress predicted decreased likelihood of participating in medium or high levels of leisure physical activity among Australian women in socially disadvantaged neighborhoods [29
]. Our studies among Canadian adults showed very weak negative relationships between psychological distress and physical activity, with more marked negative relationships among adults with greater education and income [16
]. In general, results of the current study suggest that stress and distress might represent risk factors for less physical activity in early or mid-pregnancy, but not in late pregnancy. Other risk factors likely play a more important role in the decline in physical activity over the course of pregnancy and the low levels of physical activity observed in late pregnancy in many studies.
4.3. Strengths, Limitations, and Future Directions
Study strengths in the case of the Montreal cohort include the use of a validated questionnaire for assessing perceived stress, and objectively-measured sedentary behavior and physical activity. Furthermore, the assessment of perceived stress three times during pregnancy allowed us to highlight relationships over the course of pregnancy. Finally, the study is strengthened by the diverse sample, including immigrant women and women with low income who often face barriers in participating in research studies [37
]. Major limitations of the cohort study include the sample size, which is too small to test advanced statistical models. The current analyses were not corrected for multiple testing, and studies in larger cohorts are needed to replicate the results. Furthermore, the indicators of physical activity and sedentary behavior, chosen to allow comparison to other studies, do not take into account many characteristics of interest. For example, our measure of physical activity does not reflect the intensity of activity. Similarly, our measure of sedentary behavior includes total sitting time throughout the day, but we cannot account for the type of activity. Important differences have been noted in relationships between mental health and sedentary activities that are mentally active (such as office work) versus passive (such as watching television) [38
]. Finally, sedentary behaviors and physical activity might vary on weekend versus weekdays, but measures here represent primarily weekdays only.
Strengths of the CCHS analyses include the use of a validated questionnaire to assess psychological distress, and the large representative sample that permits generalization to the population level. The study is limited by the self-reported data collection, which likely underestimates sedentary behavior and overestimates physical activity compared to objective measures. Furthermore, our indicator of physical activity is based on estimated values of energy expenditure for common activities that, while widely used, provide only a basic estimate of actual leisure and transportation energy expenditure. For both sedentary behavior and physical activity, the lack of data on activities outside of leisure and transportation is a major limitation.
In both studies, our cross-sectional analyses do not allow us to identify causal relationships. Furthermore, we cannot control for a number of confounding variables associated with sedentary behavior and physical activity. Data on season of data collection, physical characteristics such as body weight and weight gain during pregnancy, medical conditions that might be associated with restricted activity levels, health behaviors such as dietary patterns and smoking, and pre-pregnancy characteristics are not available or comparable between datasets. We thus focused on sociodemographic covariates that could be compared between datasets and that have been associated with stress, sedentary behaviors, or physical activity in past studies. Physical, medical, and health behavior characteristics might mediate or moderate relationships between stress and sedentary behavior or physical activity and should be integrated into future studies. Such analyses might point to key characteristics that could be targeted to reduce the relationship between distress and sedentary behavior.
In addition to assessment of confounding variables, future studies would also benefit from testing relationships between stress and sedentary behaviors or physical activity in larger and more diverse samples. Patterns might be more pronounced, for example, among women with very high stress levels, among socially disadvantaged women, or in communities where opportunities to participate in leisure and transportation physical activity are limited. Such analyses might point to groups among whom relationships are particularly pronounced and thereby guide future interventions.