Familial and lifestyle factors related to physical activity in elementary school students: a cross-sectional study based on a nationally representative survey in Japan

This study was a secondary analysis of the national Sports-Life Survey conducted by the Sasagawa Sports Foundation in June-July 2019. Survey participants were selected by a stratified two-stage random sampling method with a set sample size of 2,400 participants. For geographical stratification, all prefectures in Japan were divided into 10 regions. The regions were then separated into 4 groups according to city size. Based on the population size of children aged 4 to 11 years in each region, 225 areas (205 city and 20 town or village) were established. Finally, 2,400 participants (8 to 16 participants/area) were allocated, representing the population composition by city size. The number of survey points and the number of samples by district obtained by the above sampling process are presented in the Supplemental Table. With reference to the national basic resident registry, participants were sampled from the specified survey areas by the equal interval sampling method. Participant extraction corresponded to the assigned age based on population composition in the respective areas. After mailing the survey request postcards, the questionnaires written in Japanese were hand-delivered and later collected by survey conductors using a drop-off and pick-up method, which has been validated to provide a higher response rate in population-based surveys [21]. In the survey of youths aged 4 to 11 years old, investigators also used the individual interview method in the presence of parents. Individuals who were capable of comprehending the Japanese language and completing the questionnaire in Japanese face-to-face with the investigator (including those with a family interpreter) were eligible. Of the selected 2,400 participants, 1,538 (64.1%) responded, of which 1,197 were elementary school students. Ethics approval for the use of analyzed data was obtained from the Sasagawa Sports Foundation. This study was approved by the institutional review board of Shinshu University School of Medicine (authorization number: 5416).

We evaluated the children’s preference for PA using an original questionnaire item simply asking, “Do you like physical activity?” The participants were asked to respond on a 4-point Likert scale of yes, somewhat yes, somewhat no, or no. In our analysis, we classified the answers as “like” (yes or somewhat yes) or “dislike” (somewhat no or no). Participation in school and club sports including lessons and league or team sports was identified using the following question for the parents of participants: “Is the child currently enrolled in school sports clubs, sports clubs as extracurricular activities, private sports clubs, or community sports clubs?” The parents answered “yes” or “no”. We assessed MVPA using the Japanese version of the WHO Health Behaviour in School-aged Children survey, one of the most comprehensive sources of data for school-aged student PA levels [22]. The validity of the Japanese version has already been verified in a study using accelerometers [23]. Before answering questions, a definition of PA was provided as “Physical activity is any activity that increases your heart rate and makes you out of breath some of the time. Physical activity can be done in sports, school activities, playing with friends, or walking or cycling to school. Some examples of physical activity are running, brisk walking, rollerblading, cycling, dancing, swimming, soccer, and basketball”. The participants were asked the following question: “Over the past 7 days, on how many days were you physically active for a total of at least 60 min per day?” The participants provided a number from 0 to 7 days (Supplemental figure).

The covariates in our models were selected a priori based on previously published literature [16‐18, 24]. The participants were asked questions on gender, age, and grade. Their parents provided additional information on such sociodemographic factors as annual household income and family members by choosing the most suitable response from a set of predetermined options. Parents also answered questions regarding the exercise habits of themselves and their spouse (never, rarely, sometimes, and frequently) and their exercise habits with their children (never, rarely, sometimes, and frequently), as well as the children’s frequency of eating breakfast per week (every day, 5–6 days, 3–4 days, 1–2 days, or never), defecation frequency (every day, once every 2 days, once every 3 days, less than once every 3 days, or irregular), bedtime, wake-up time, and weekday screen time on electronic devices, including computers, tablets, smartphones, handheld gaming consoles, and television, during discretionary time (i.e., non-school use).

We classified the items of school grade as grade 1–2, grade 3–4, or grade 5–6, residential area as metropolitan area in Tokyo, ordinance-designated city area including 20 cities, city with a population of 100,000 or more, city with a population less than 100,000, or town/village, and annual household income as < 4,000,000 JPY, 4,000,000–7,999,999 JPY, or ≥ 8,000,000 JPY) [25], and family structure as two-generation household or household of three or more generations. We defined “skipping” as non-daily breakfast consumption or defecation. Based on the recommendations of the American Academy of Pediatrics, screen time was classified as < 2 h or ≥ 2 h [26]. Weekday sleep time was calculated from the reported bedtime and wake-up time and categorized as < 8 h or ≥ 8 h [17].

We evaluated 2 main outcomes: participation in organized sports and MVPA level. MVPA level was categorized using the cut-off points of ≥ 5 days/week or ≤ 4 days/week based on MVPA reference values of at least 60 min/day in accordance with previous studies [1, 18, 23]. Considering the differences in gender for lifestyle and PA participation, Fisher’s exact test or the chi-square test were performed to compare variables between sexes. Multiple logistic regression analysis was performed on the factors related to the students who participated in organized sports. We also adopted multiple logistic regression analysis to identify factors related to adequate MVPA as dependent variables. Crude and adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated while controlling for covariates. Spearman’s rank correlation coefficient was used to check for multicollinearity of covariates. Hosmer–Lemeshow testing was employed to assess the goodness-of-fit of the models. All statistical analyses were performed using SPSS statistical software version 28 (SPSS Inc., Chicago, Illinois). A P value of < 0.05 was considered statistically significant.