Abstract
This study found that despite differences in school instruction mode (in person, hybrid, or remote) and the influence of mental health early in the global COVID-19 pandemic, children’s participation in nonsedentary activity (outdoor play, leisure and extracurriculars) increased. Occupational therapy practitioners can play a role in reinforcing the importance of reengaging in activities and regular routines to promote health and well-being during challenging situations.
The global coronavirus disease 2019 (COVID-19) pandemic restricted children’s participation in play, leisure, and physical activity; in addition, it may have lasting impacts on health and development (Dunton et al., 2020; Loades et al., 2020; Moore et al., 2020; Viner et al., 2022). Limitations in activity participation are also linked to poor mood and mental health (Chien, 2022). Because most prior studies have used cross-sectional designs, it is unclear whether children were able to reengage in activity participation as schooling resumed and pandemic-related restrictions eased. Understanding the consequences of COVID-19 over time and how these influenced activity participation is critical for evaluating the overall impact on children’s health and well-being.
Variations in school instruction mode during the pandemic may have further complicated children’s return to everyday normal occupations. Nationally representative survey studies indicate that children who were in fully remote learning environments experienced greater levels of stress and other negative mental health outcomes compared with those receiving in-person instruction (Hertz et al., 2022; Verlenden et al., 2021). Combined with previous evidence before the pandemic (Sharfstein & Morphew, 2020), returning to in-person instruction modes would improve activity and play participation for most children. To our knowledge, the influence of school instruction mode on children’s activity participation has not been examined in the context of COVID-19.
Therefore, we conducted a longitudinal survey to evaluate the changes in children’s and adolescents’ activity participation during the initial phase of the COVID-19 pandemic. We examined both sedentary and nonsedentary activity; in addition, we hypothesized that children’s sedentary activity would be highest early in the pandemic and would decrease, whereas nonsedentary activity would increase over time (Nagata et al., 2020). Moreover, we examined the association between school instruction mode and changes in participation during the beginning of the 2020 to 2021 school year. We hypothesized that children returning to in-person instruction would show greater improvements in nonsedentary activity participation compared with those who remained partially or completely in remote instruction modes.
Method
Participants and Design
A convenience sample of parents was recruited as child proxies, representing children and families in 29 states. Parent participants were included if they were English-speaking adults ages 18 yr or older who (1) were the parent or legal guardian of and (2) had lived with at least one school-age child (ages 5–18 yr) for the past 6 mo. Responses were anonymous, the study presented minimal risk, and no protected health information was collected.
A longitudinal design was used, with two surveys administered approximately 3 mo apart. Survey links were distributed to personal and professional contacts and shared on community forums, university websites, and social media. The first survey (Wave 1) was released on June 24 and closed on August 24, 2020, and it was completed by 336 parents. Of this initial sample, 260 parents opted in to be contacted for a follow-up second survey (Wave 2), which was sent on October 1 and closed on October 31, 2020. The average time between completion of the two survey waves was 76 days (SD = 26). Survey questions were directed at parents as opposed to children themselves, allowing data collection from a larger sample in a shorter period and avoiding ethical considerations associated with consenting minors. The research protocol was reviewed and given exempt status from the Marquette University institutional review board, and a modified informed consent process was used.
Parent and Child Characteristics
A total of 143 parents participated and completed both surveys, resulting in responses for 208 children. Parents were mostly White (95.1%) and married (94.3%); in addition, they had household incomes >$50,000 (92%), had two or fewer children (92.4%), and resided in a midwestern state (58.4%; Table 1). About 80% of children of the parent respondents were age 13 yr or younger, and 71.6% attended a public school (Table 2). As of November 2020, of all school districts in the United States, 36% had reopened to fully in-person instruction modes, 45% were hybrid (45%), and 19% were fully remote (Return2Learn Tracker: https://www.returntolearntracker.net/). In the sample at the start of the 2020 to 2021 school year, 43.2% of children received remote instruction, 20.7% received in-person instruction, and 36.1% received hybrid instruction.
Family Characteristics of Parent Respondents
Note. Parent sample size is less than the child sample size because some parents reported for multiple children. Some parents did not report demographic data; therefore, not all frequencies sum to the total parent sample size.
Characteristics of Children Represented in Surveys
Note. Some parents did not report demographic data; therefore, not all frequencies will sum to the child sample size.
Survey Instruments
Surveys were developed and distributed with QualtricsXM. Questions were designed and customized to collect comprehensive information about participation in sedentary and nonsedentary activities. Parents could enter data for more than one child, which allowed the acquisition of a larger sample of child-focused data. In the Wave 1 survey, parents were asked to consider their child after schools closed in the spring months of 2020 (typically the months of March and April). In the Wave 2 survey, parents were asked to consider their child at the beginning of the 2020 to 2021 school year. Ranges rather than specific dates were used to orient parents because the actual date and time of school closures may have varied across respondents.
Sedentary and Nonsedentary Activity Participation
Leisure and play participation are the primary occupations for school-age children and are important for physical and cognitive health as well as social development at any age (Ginsburg et al., 2007). Customized questions about the frequency (none, 0–1 hr, 2–3 hr, ≥4 hr) of daily activity spent in a typical week were used. Nonsedentary activities included participation in sports, leisure, extracurricular activities, and outdoor play. Sedentary activities included participation in indoor play and use of electronic devices (TV, computer or video game, cell phone) not for academic purposes. Broader questions for physical activities were selected rather than activity-specific questions (e.g., running, bicycling) to capture the entirety of activities in which the children participated.
Mood
Poor mental health is associated within reduced participation (Fossey & Scanlan, 2014). The construct and questions from the shortened Mood and Feelings Questionnaire (Thapar & McGuffin, 1998) were used to assess negative mood symptoms (e.g., unhappiness, loneliness) as a covariate in the analysis. This tool has strong validity for use among children and adolescents (Thapar & McGuffin, 1998) and high (0.84) test–retest reliability (Sund et al., 2001).
Instruction Mode
At Wave 1, all children were in a remote school setting. At Wave 2, parents reported which mode of instruction their children were receiving at the beginning of the 2020 to 2021 school year: remote only, in person only, or hybrid format (both in-person and remote components). Five parents responded “other” to the mode of instruction question and were excluded from the formal analysis.
Data and Statistical Analysis
Variable Coding
All activity questions were numerically coded to allow for quantitative statistical analysis (0 = none, 1 = 0–1 hr, 2 = 2–3 hr, and 3 = ≥4 hr). A total sedentary and nonsedentary activity sum score ranging from 0 to 12 was calculated. Mood questions were coded 0 (rarely), 1 (sometimes), 2 (often), and 3 (always). A total mood sum score ranging from 0 to 18 was calculated.
Statistical Analysis
Descriptive statistics and nonparametric signed rank tests were used to evaluate the changes in sedentary and nonsedentary participation from Wave 1 to Wave 2. Separate linear regression models were used to determine the factors associated with the total sum score for sedentary and nonsedentary participation. The dependent variable was the summed score at Wave 2. The independent variables were grouped in blocks: Block 1 included instruction mode (in person, hybrid, remote), and Block 2 included the summed activity participation score at Wave 1, socioeconomic status (household income level), negative mood, age, and COVID-19 severity (COVID-19 cases per 100,000 people on September 1, 2020, on the basis of zip code data). To consider the relationship between instruction mode with and without covariates, we entered variables beginning with Block 1 followed by Block 2 using a forward-entry method. We performed all statistical analyses using IBM SPSS Statistics (Version 28). The level of significance was set at α = .05 to determine statistically significant independent variables and covariates.
Results
From Wave 1 to Wave 2, nonsedentary activity participation significantly increased (Cohen’s d = 0.34, t[200] = 4.77, p < .001), and sedentary activity participation significantly decreased (Cohen’s d = 1.04, t[200] = −14.77, p < .001). Results of the linear regression for nonsedentary and sedentary activity at Wave 2 are shown in Tables 3 and 4, respectively. For nonsedentary participation, instruction mode was a significant factor along with the covariates of participation at Wave 1, negative mood, and COVID-19 severity. The directionality of the relationships indicated that higher nonsedentary participation at Wave 2 was associated with hybrid and remote instruction modes, higher participation at Wave 1, lower negative mood at Wave 2, and lower local COVID-19 severity. For sedentary participation, the covariates of participation at Wave 1, mood, and household income were significant factors; instruction mode was not a significant factor (p = .318). The directionality of these relationships indicated that higher sedentary activity participation at Wave 2 was associated with higher sedentary participation at Wave 1, higher negative mood, and lower household income. The maximum variance inflation factor was <1.55, indicating minimal collinearity among predictor variables.
Linear Regression Summary for Nonsedentary Activities at Survey Wave 2
Note. Block 1: F = 23.82, R 2 = .12, p < .001; Block 2: F change = 10.03, R 2 change = .20, p < .001. β = standardized coefficient; CI = confidence interval; COVID-19 = coronavirus disease 2019; SE = standard error.
Reference category: in-person school.
Linear Regression Summary for Sedentary Activities at Survey Wave 2
Note. Block 1: F = 0.56, R 2 = .003, p = .455; Block 2: F change = 7.78, R 2 change = .19, p < .001. β = standardized coefficient; CI = confidence interval; COVID-19 = coronavirus disease 2019; SE = standard error.
Reference category: in-person school.
Discussion
Early in the pandemic, school closures were predicted to affect physical activity because of fewer opportunities for play and extracurricular activities (Guan et al., 2020; Margaritis et al., 2020; Shahidi et al., 2020). These forecasted problems were supported by the present survey and prior studies published since the onset of COVID-19 (Dunton et al., 2020; Jiao et al., 2020; Moore et al., 2020). The return to school in fall 2020 offered potential benefits for students to improve activity participation. Many variables, including school instruction mode, could affect how children returned to activity participation. To our knowledge, our prospective longitudinal survey is one of the few assessments to investigate changes in participation during the early pandemic and to examine key factors affecting participation.
The changes in participation from Wave 1 to Wave 2 demonstrated an overall increase in nonsedentary activity and a concomitant decline in sedentary activity. This finding suggests that activity participation may have normalized with the onset of the 2020 to 2021 school year and as COVID-19 restrictions eased. The relatively high levels sedentary activity (i.e., electronic device use) during Wave 1 is supported by other survey research and is likely a product of the restrictions on social interaction and mobility during the early months of the pandemic (Xiang et al., 2020). Although sedentary behavior is still a major concern for children and adolescents, our data suggest that these levels decreased as the pandemic eased and schools reopened. Moreover, reopening of schools and other organizations allowed for nonsedentary activity to increase through participation in clubs, organized sports, and general physical activities.
The primary hypothesis was that instruction mode when returning to school would be associated with return to nonsedentary participation during Wave 2. Instruction mode was a significant factor in the linear regression; however, the direction of the relationship was that children in remote and hybrid learning environments had greater nonsedentary participation than children who went to school in person. Thus, our data did not support the hypothesis that in-person school instruction was related to resumption of nonsedentary activity. One explanation for this unexpected result is that the return of regular school routines, rather than the specific school instruction mode, is a driving factor for why children showed increases in nonsedentary participation (Rosen et al., 2021). Another explanation is that easing of initial quarantine restrictions may have had a general effect on increasing opportunities for children to engage in nonsedentary activities. The COVID-19 severity level may be a proxy measure for the general impacts of the pandemic on the participants; however, there may still be regional differences in availability and format (in person vs. virtual) of extracurricular activities available to children that were not captured in the survey.
The regression analyses also showed that other variables—including level of participation at Wave 1, negative mood at Wave 2, COVID-19 severity, and household income—explained additional variance in sedentary and nonsedentary activity participation at Wave 2. For instance, children who maintained some level participation at Wave 1, had greater negative mood symptoms, and were part of lower income families had higher sedentary activity participation levels at Wave 2. The role of negative mood in modifying activity and engagement is important, indicated by work showing that mood stability during the initial phase of the pandemic was more affected in young adults and adolescents who experienced more stressors (Green et al., 2021; Rosen et al., 2021). Such stressors (e.g., difficulties with school and caring for family members) were not measured directly in our survey but are likely mediating factors related to how children and adolescents managed their physical and emotional states. Altogether, the combination of multiple significant factors in the regression emphasizes the complex multifactorial nature of activity participation.
Overall, sedentary and nonsedentary activities normalized during the early months of the pandemic. Although the benefits of these changes for children’s health and well-being were not directly measured, these results highlight the importance of regular play and leisure participation for children. Occupational therapy practitioners can amplify this message by helping parents and children in any schooling environment encourage nonsedentary participation, even during times of prolonged isolation or stress. Virtual and remote schooling may work better for a child’s formal education; however, these modes will require different strategies for ensuring that children are engaging in play and other activities compared with children returning to school in person. Offering education on developing healthy routines and promoting the use of age- and context-appropriate practical strategies and resources for teachers, parents, and other caregivers can limit the impact of the pandemic and other future instances of isolation.
Limitations
Several limitations affect the generalizability of the results. First, the retrospective and nonrandomized survey design precludes determination of causal relations about the effects of school instruction mode on children’s participation. Moreover, imbalances in demographic variables among subgroupings based on instruction mode may confound any observed relationships. Because of the anonymity promised in the survey, demographic factors could not be balanced prospectively. Other unmeasured factors, such as health of other family members or changes in income, could also affect children’s participation.
Second, most children came from affluent households with White parents. Although racial and socioeconomic diversity varies on the basis of geography, our study sample does not reflect the diversity of the entire U.S. population. Research suggests that both racial minority and low-income groups, particularly Black households, experienced a disproportionate burden of the pandemic (Fairlie et al., 2020; Tai et al., 2021); in addition, low-income and racial minority students were more likely to receive remote instruction modes and therefore experience greater mental health difficulties compared with White and higher income peers (Hawrilenko et al., 2021). In addition, although about 20% of parents reported that their child had a physical, intellectual, or developmental disability, a subanalysis based on these factors was not performed given the relatively small samples from the survey. Children with disabilities also may have had unique experiences because of a lack of therapy or educational services received in the school setting (Summers et al., 2021).
Third, some measurement bias was present in our survey instrument. We used a short list of directed questions based on single-survey instruments with limited depth and specificity for any one domain. Differences between response categories that use Likert scales may not be equivalent and could have varied interpretations across respondents. Moreover, parent assessment of children who attended school in person was limited to the time they were not in school; parents of children who remained at home may have had more information to report regarding their physical and emotional status.
Fourth, differences in geographic location of parent respondents could have had an effect but were not accounted for directly. Weather can vary during the spring months of March and April; therefore, outdoor activity levels could be influenced by lower temperatures and inclement weather in midwestern and eastern states. Instruction modes and general experiences of the pandemic could also vary geographically; however, general COVID-19 effects on outcomes were accounted for by including COVID-19 severity level as a covariate in the analysis.
Finally, the evolving nature of the COVID-19 pandemic presents a general challenge for studying its impact on children and adolescents. For instance, because of local changes in COVID-19 severity level, some children’s school instruction mode may have changed at times throughout the school year. Such rapid and unpredictable changes make the relationship between our study variables difficult to capture, even with longitudinal surveys. The aggregation of research conducted at different time frames of the pandemic may help to elucidate these dynamic relationships.
Implications for Occupational Therapy Practice
This study has the following implications for occupational therapy practice: ▪ The results support the body of evidence that children’s physical and mental well-being were affected during the COVID-19 pandemic and indicate the importance of promoting activity participation regardless of school modality (in person or virtual). ▪ Emphasizing nonsedentary activities for children is critical for long-term health, especially during difficult or stressful periods. ▪ The interplay between activity limitation and schooling should continue to be studied as schools reopen after the pandemic, especially among children at risk for activity limitations.
Conclusion
Participation in play, sports, and extracurricular activities contribute to a child’s physical and emotional development. These benefits likely outweigh the minimal risk in COVID-19 exposure and disease transmission because safety data suggest that reopening schools did not significantly increase COVID-19 community case rates (Ertem et al., 2021). Although the long-term effects of COVID-19 on physical and mental health continue to be elucidated, our findings demonstrate that, on average, children’s participation in nonsedentary activities increased as schools reopened, confirming and extending prior research on activity levels among children during the pandemic. Occupational therapy practitioners should advocate for the importance of regular play and leisure activity participation, facilitated through regular school routines, for children to enhance function and healthy development. Future research and resources should be targeted to identify the subset of children who have on-going participation limitations and to develop strategies to mitigate the long-term impact of the pandemic on physical and mental well-being.
Footnotes
Acknowledgments
We thank the Marquette University COVID-19 Research Initiative and the Office of Institutional Research and Analysis, as well as all the parent participants for completing this survey. Funding for this research was provided by Marquette University.