Study design and study participants
This cross-sectional study was carried out among 480 children/ adolescents from 2 to 18 years old and their parents who were covered by the health centers of Tabriz City Iran. The recruitment of the participants and the collection of data were carried out during Apr-Jul 2017. Samples were selected randomly using a multi-stage method, from the 88 rural health centers of the Tabriz area. The sample size was calculated based upon previously identified data regarding the relationship between sedentary behavior and demographic variables (OR = 2.36) . Considering 95% confidence level, 80% power, a two-tailed test and utilizing the PASS15 software, the sample size was estimated to be 240 cases. Taking into account the sampling design effect of two, the required sample size increased to 480 cases. At the time of study there were 88 urban active health centers in Tabriz city. Firstly, 30% of the urban health centers (26.4), in all 27 urban health centers were selected as study sample. Secondly, samples (parents, children) were randomly selected according to the sample size estimation (480) and the rate of population covered by each center. The samples were taken systematically based on the inclusion criteria (being residence of the Tabriz area, being interested in participating in the study, not having any physical and /or mental disorders), and the list of households.
All activities were harmonized and coordinated with the health center authorities. Invitations were sent to participants to take part in the study on a designated day. After stating the study goals and the process of participation, the participants were asked to fill in the questionnaires.
Data collection and measures
Data collection was performed by using several questionnaires; a demographic questionnaire consisted of the child sex and age, parent education and occupation, and family socio-economic status. The child age grouping was categorized as 2–4, 5–11, and 12 and above. The mother age grouping also was categorized as under 34, 35–44, 45–54, and 55 and above.
Socio-economic status of the families were determined by the valid family affluence scale (FAS) [33, 34]. It contained 14 common household properties: private bedroom, laundry, dishwasher, TV set, personal computer (PC), laptop, tablet, landline telephone, smartphone, and access to internet network. Answers were categorized into two levels: don’t have/not existing = 0 and have/existing = 1, and more. The scores ranged from 8 to 60, and were categorized into three levels of FAS: low = 8–25, medium = 26–42, and high = 43 and above.
The pattern of parent behavioral style in the present study was taken from Bjelland study , following two major variables: a. setting regulations, b. communicative style (b.1.threats and Punishment; b.2.logical communication and explaining reasons). Overall nine questions were used to assess parent style, three items for setting regulations, to control the time spent on watching TV, using computers and playing video games as well as smartphone. Six items were given for parent communicative style, to make children follow regulations through two ways of either punishment and threatening or logical style and explaining reasons. The answers were prepared based on likert scale (strongly agree = 5, agree = 4, no idea = 3, disagree = 2, strongly disagree = 1). The possible range for each scale was 3 to15. The higher score meant the higher degree of the parent behavioral style. That is, parents reported more regulations and more logical reasoning. A cut-off of 12 was used for determining ‘setting regulation, and applying a punishment style’s scores into two levels. Scores below 12 were considered to as not setting the regulation, or did not apply a punishment style, and 12 and above as setting the regulation or applied reasoning.
The guideline of recommended time for sedentary activities [12,35] was an indicator of assessing parent knowledge. Parent knowledge was measured using three questions: how much time experts recommend children spend on sedentary activities; under 2 years, 2–4 years, and 5 years and above. The answers were categorized based on six scales: not at all, 30 min/d, 60 min/d, 90 min/d, 120, and above. The parent answers for each age group recorded in two levels of ‘know’ and ‘don’t know’. For age groups: (under2), (2–4), and (5 and above), the correct answers were: (not at all), (less than 1 h/d), and (less than 2 h/d) respectively.
The sedentary behavior questionnaire was prepared based on existing studies [12, 29] to evaluate the frequency and time period of 10 sedentary behaviors during a week. Time period of each sedentary behavior listed according to min and hours (h) spent for each activity following the below pattern: 0 min, 30 min, 60 min (1 h), 90 min (1 h and 30 min), and so on. The sedentary behaviors of children/ adolescents were the mean period of sitting time (min/d) included watching TV and playing video games, working on the computer, playing games on computers, doing homework, studying, listening to music while lying down, day time sleep, and night time sleep on weekdays and on the weekend. The total time of the activities were estimated first included and then excluded the amount of night time sleep.
Activities such as watching TV, playing video and computer games, studying, and EMC were estimated individually. EMC included time spent on playing games on smartphone, talking on the phone, sending message, photo, file, and virtual communication. The time of activities estimated based on the guideline recommending < 120 min/d [10, 12, 35].
The self-administrated questionnaire was used for data collection from literate parents and children 10 years old and above, while for illiterate parents and children under 10 years old the questionnaire was filled in by the interviewer.
Statistical analysis methods
Descriptive statistic methods, frequency and percentages for qualitative variables and mean (M) and standard deviation (SD) for quantitative variables, were used. To determine statistical differences between groups, chi-square and independent t-tests were used.
Multivariate logistic regression analysis was used to determine the association between dependent variables (outcomes) and independent variables (predictors). Dependent variables were watching TV, playing computer games, communicating electronic media, and studying duration. Independent variables (predictors) were parent behavioral style, parent knowledge and FAS. The potential controlling variables were child sex, age, mother education, age, and job.
Analyzing data was performed individually for each dependent variable. Each dependent variable was divided based on time spent according to min/d: the recommended time under 120 min/d, recoded as 0; and 120 min/ and above considered as high risk time recoded as 1. Analyzing data was performed using SPSS 21, P-value lower than 0.05. The study was reported according to the ‘Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines’.