Parent-adolescent agreement in reported moderate-to-vigorous intensity physical activity during the COVID-19 pandemic

Among children and adolescents, moderate-to-vigorous intensity physical activity (MVPA; e.g., running, swimming, or bicycling) is associated with a myriad of health benefits, including lower risk of obesity, higher bone mass, and improved cardiometabolic health [1, 2]. Given the known benefits to child and adolescent health, the U.S. Department of Health and Human Services recommends that school-aged youth 6–17 years of age participate in at least 60 min of MVPA every day [3].

Population-based studies with youth often use self-report-based methods to assess MVPA and estimate the prevalence of those meeting physical activity guidelines. Report-based MVPA measures are generally simple to administer and inexpensive, and thus are commonly used [4, 5]. However, report-based MVPA measures are reliant on memories that are prone to fallicies in retrieval due to decay over time or interference of other memories. Memories may also be filtered by perceptions and biases. Specifically, participant-reported MVPA may be affected by recall [6] and social desirability bias: youth often overestimate their intensity and time spent on MVPA [7]. Use of a proxy respondent varies based on the age of the child with research suggesting that children aged 10 years and older are developmentally mature enough to provide reliable and valid estimates of MVPA [4, 8‐10]. To obtain physical activity estimates in younger children, reports from a parent or adult (e.g., teacher) have been used to estimate the child’s MVPA. Prior studies have reported on the agreement and correlation between parent and child accounts of children’s physical activity levels; however, a majority were in children under 10 years of age [4, 8, 10‐12]. Notably, levels of agreement between parent- versus child- reports of physical activity may vary between different populations dependent on the child’s age group. To obtain physical activity estimates in younger children, reports from a parent or adult (e.g., teacher) are relied upon to estimate the child’s MVPA. The literature has shown that parent report proxies are typically used for younger children’s activity levels [13‐16], while older children typically self-report their own physical activity levels [17‐25]. Findings regarding agreement between parent- and child- reports of physical activities have ranged from low to slight to fair in pediatric populations up until age 14 [12, 16, 26, 27]. Prior studies have suggested that using a self-report may be more reliable than parental proxy report in children over the age of 10 [27].

The sociodemographic correlates of discrepancies of parent- and youth- reported physical activity identified in the literature include parent gender [16], youth weight status [12, 16], youth age [28], youth gender [12], family cohesion [28], and family socioeconomic status [12]. These differences may also stem from the fact that parental monitoring of youth physical activity is limited during the time that youth spend at school and parents spend at work [29, 30].

With the onset of the COVID-19 pandemic, social distancing, school closures, and changes in access to community parks and recreational facilities have changed the way in which adolescents participate in MVPA and have led to more time spent at home [31, 32]. Although previous studies reported low agreement in MVPA between parents and adolescents [12, 16, 26, 27], the dyadic agreement between parents and adolescents may have improved during the pandemic due to increased time spent at home together as a result of stay-at-home orders, school and business closures, and travel restrictions [31, 32]. Conversely, with the cancellation of regular routines such as sports team practices and training, competitions, and in-person physical education classes [33], there may have been greater difficulty to recall time spent on MVPA given these activities are often more structured than recreational activity. The stay-at-home orders and other changes during the COVID-19 pandemic [32] allowed for a natural experiment to examine the parental proxy measures of adolescent MVPA in a unique time during which adolescents and their parents may have been in closer quarters with one another and there may have been limited sanctioned opportunities for physical activities.

The present study aims to describe the differences and agreement between parental reports of adolescent physical activity and adolescent self-reports of MVPA. We hypothesize that there will be some agreement between parent- and adolescent- reports of adolescent physical activity during the COVID-19 pandemic given the increased number of hours families spend together and due to fewer MVPA opportunities during the COVID-19 pandemic. Second, we aim to determine sociodemographic correlates of the parent-adolescent dyad physical activity reporting differences. We hypothesized that factors such as adolescent sex, race/ethnicity, education, and household income will be associated with differences in parent and adolescent physical activity reports.

In this sample, adolescents were aged 10–14 years old (mean 12.5 ± 0.9). Other sample sociodemographic characteristics were as follows: 50.0% were female; 38.9% were of racial/ethnic minority background; 10.0% had a parent with a high school education or less; 33.2% were in households earning less than $75,000 per year (Table 1).

Table 2 presents adolescent- reported MVPA, parent-reported MVPA, and parent-adolescent absolute difference scores by hours per day (h·d^− 1), days per week (d·wk.^− 1), and hours per week (h∙wk.^− 1). Adolescents reported engaging in a median of 2.17 [p25, p75 (0.00, 6.00)] hours of MVPA per week compared to 1.52 (0.29, 4.75) hours reported by their parents. Significant dyadic differences were found among each measure of MVPA, with absolute mean differences in parent- and adolescent- reports of 1.15 h·d^− 1 (95% CI 1.08, 1.21), 1.68 d·wk.^− 1 (95% CI 1.63, 1.72), and 4.89 h∙wk.^− 1 (95% CI 4.64, 5.15). Within-dyadic MVPA reports from parents and adolescents, correlations were poor to moderate among each duration assessment of MVPA (all p-values were < 0.001), with ICCs ranging from 0.19 to 0.48.

A Bland-Altman plot (Fig. 1) further illustrated levels of within-dyad agreement between adolescent and parent-reported MVPA (h∙wk.^− 1), revealing that the level of agreement between individual parent-adolescent dyads varied with a range of difference scores between − 38.5 to 83.86 h per week. Difference scores were located closer to the line of perfect agreement when average reported MVPA within dyads was closer to 0 h per week. Based on the calculated lower and upper bounds of agreement, 95% percent of differences in MVPA reports were expected to lie between − 12.0 and 26.1 h per week. Therefore, overall agreement was found with only 241 of 4841 (4.8%) of points lying outside the 95% interval. Sensitivity analyses stratified by sex, race, and income are shown in the Supplemental Appendix.

Table 3 summarizes parent-adolescent MVPA difference scores by sociodemographic characteristics. Significant differences in parent-adolescent MVPA by sociodemographic characteristics were found for parent-adolescent dyads by household income, with high income households having higher differences in parent-adolescent reports (difference score = 0.36 [p25, p75 (− 1.19, 3.04 h∙wk.^− 1)]) compared to low-income households (difference score = 0.00 [p25, p75 (–1.15, 2.42 h∙wk.^− 1)]). No significant differences in median MVPA hours per week within sex, race/ethnicity, or parent education level were found in this sample.

To our knowledge, this is the first large, national study examining adolescent- and parent- reports of adolescent MVPA in the U.S. While parent and/or adult proxies are commonly used in younger children (aged < 10 years), the addition of the COVID RRR Survey to the ABCD study provided the novel opportunity to evaluate this methodological approach for use in adolescents during COVID-19 pandemic, when stay-at-home orders provided more opportunities for time with family members. We describe discrepancies between parent- and adolescent- reported physical activity, while also exploring reporting differences by several sociodemographic factors. We found that if an adolescent had no or little MVPA, the parent was generally accurate in reporting that their child was non-active. However, if the adolescent was active, particularly with very high levels of MVPA, there was less agreement between the parent's and youth's report. When examining sociodemographic factors associated with discrepancies in reported MVPA, there was a greater difference among MVPA parent and youth reporting among adolescents with higher MVPA, particularly in families with income > $75,000.

The finding that there was greater parent-adolescent agreement with lower levels of MVPA supports prior research results that indicate higher agreement on reporting of sports and outdoor activities among the lowest quartiles of reported activity [26]. In the context of the pandemic, non-active adolescents may be more likely to be home and engage in sedentary activities, such as recreational screen time, that can be better monitored by parents. It is also possible that there may be less activities to recall for both non-active adolescents and their parents, increasing the agreement of both reports (i.e., less measurement error). We initially hypothesized that there would be moderate to high parent-adolescent agreement in MVPA reporting during the pandemic for these reasons. Prior studies examining parent-adolescent agreement in physical activity among overlapping age groups reported Kappas of 0.11 (low) to 0.41 (fair). Our overall frequency agreement (ICC of 0.28) is within this range; however, the statistics were different (ICC vs Kappa) and the physical activity measure was slightly different (MVPA vs outside play, outdoor activities, leisure sports, organized sports, respectively) [12, 16, 26].

Conversely, if the adolescent is active, particularly with very high levels of MVPA, there is less agreement between the parent and the adolescent report, similar to previous findings [26]. On the adolescent side, this may be due to cognitive processes related to memory that make it harder to recall physical activity occurrence, frequency, and duration based on the structure of survey questions [44]. On the parent side, this may be due to the parent not knowing what their child is doing, particularly when the child is doing activities outdoors or overtraining. Active adolescents may be doing activities outside of the home with peers, teams, or individually of which parents may not be aware. For example, despite COVID-19 restrictions, some sports teams have maintained outdoor or distanced training routines. Moreover, at high levels of training for sports teams or competition, adolescents may overtrain [45] and may not want their parents to know how much they are training. Similarly, with eating disorders being exacerbated during the COVID-19 pandemic [46], more young people may be at engaging in excess covert exercise as a way to cope [47].

Overall, we found that adolescents reported more MVPA (h∙wk.^− 1) than parents, similar to previous findings [16]. Adolescents may overestimate their intensity and time spent on MVPA [7]. Our findings also demonstrated greater discrepancies in subgroups with higher MVPA levels, such as adolescents from households with higher income. During the pandemic, adolescents from low-income neighborhoods may have been more affected by restricted access to safe outdoors spaces for MVPA [48], leading to lower MVPA levels and less parent-adolescent reporting discrepancies. Conversely, adolescents attending private schools were more likely to have in-person schooling during the pandemic [49] which could have allowed for more MVPA options and a greater parent-adolescent reporting discrepancy due to time away. One prior study of 9–12-year-olds similarly found that children of high socioeconomic status were more likely to report a higher frequency of outside play than their parents compared to children of low socioeconomic status [12]. Although we initially hypothesized that parent monitoring could have been improved due to more time at home from stay-at-home orders [31, 32], parents who are working from home may not have the flexibility to monitor their children while performing work duties and parents who are essential workers continue to go to in person work.