This study used the Korea Youth Risk Behavior Web-Based Survey (KYRBWS) from the Korea Center for Disease Control and Prevention to estimate national representative values and extrapolate the findings to the entire Korean population.
This study compared the prevalence of obesity among the 2016–2019 participants (before COVID-19 pandemic group) with the 2020–2021 participants (during COVID-19 pandemic group).
The primary outcome was the difference in obesity prevalence due to increased sitting hours per week for purposes other than study before and during COVID-19. We will define the prevalence of obesity as the proportion of a sample of obese adolescents in that year. Increased sitting time means an increase in annual sitting time for purposes other than study of adolescents. Subgroup analyses on obesity status were conducted according to biological sex, school grade, average sleep hours per week, whether to skip breakfast more than five days a week, smoking status, household income, academic level, region, and sitting hours per week for purposes other than study.
Data and study population
This study analyzed serial cross-sectional data from the KYRBWS, a survey of middle- and high-school students, to understand the current status and trends of health behaviors, such as smoking, drinking, physical activity, diet, mental health, awareness of damage and safety, and oral health of Korean adolescents. This survey is a government-approved statistical survey (approval number: 11758) and has been conducted annually since 2005. All participants provided informed consent to participate in the KYRBWS and were guaranteed anonymity and all methods were carried out in accordance with relevant guidelines and regulations.
A sample of middle- and high-school students representing the whole country was obtained using stratified multi-stage sampling, and students were surveyed anonymously during regular class time based on a self-filling web . Average sleep hour per week, skipping breakfast more than 5 days a week and Sitting hour per week for purposes other than study were answered based on the adolescents’ memory, and the household income, and GPA were answered by the subjective criteria of the adolescents. And we confirmed that sleep hour, skipping breakfast, smoking status, sitting hour, household income, GPA and region are factors associated with adolescent obesity. The risk of obesity decreases as sleep time increases and breakfast is not skipped [15, 16]. And individuals who smoke more have lower BMI compared to infrequent or non smokers . In addition, the shorter the sitting time, the higher the household income, the lower the risk of obesity, and the lower the GPA, the higher the BMI [13, 18,19,20].
Our study population consisted of adolescents aged 12–18 years from the KYRBWS 2016 to 2021. We excluded adolescents with missing monthly age, height, and weight information. Sitting time was added to the following questions: sitting time per week for purposes other than study. This includes watching TV, playing games, using the Internet, chatting, and sitting on a move. Obesity was assessed by measuring the BMI. It is one of the most widely used and recommended methods for determining the obesity status of children and adolescents . BMI was calculated by dividing the body weight in kilograms by the body height in meter square. Age- and sex-adjusted BMI Z scores were obtained using a Korean National Growth Chart . Adolescents were considered obese if their BMI was in the 95th percentile, and overweight was defined as follows: 85th percentile ≤ BMI < 95th percentile for age and sex.
The outcome variable of this study was the effect of changes in sitting time before and after the pandemic on adolescents with obesity behavior. Categorical and continuous variables were compared between the groups using the chi-squared test and t-test. Multiple logistic regression analysis was performed to examine which of the demographic and lifestyle factors including sitting hours per week for purposes other than study had the greatest influence on obesity prevalence. We also estimated the odds ratio (OR) and 95% confidence interval (CI) of the OR. Responses that had logical errors and those that were outliers were processed as missing values, and observations with missing data were excluded from the analysis. Statistical significance was set at P < 0.05. Because the KYRBWS data included multi-level sampling, layering, and clustering, we analyzed it by applying weights. Responses with logical errors or outliers were processed as missing values.