A case-control study was conducted in Riyadh, Saudi Arabia form September through December of 2017. Cases were overweight or obese children and matched controls were normal weight children between the ages of 5-9 years attending school from kindergarten to fifth grade. Children and their parents were enrolled in the study to examine a broad range of factors related to being overweight or obese in childhood. The study was conducted according to the ethical standards of the institutional and national research committee. Informed consent was obtained from all participants. The study was approved by the Institutional Review Board (IRB) at the sponsoring institution.
A total of 492 children were enrolled in the study (246 cases and 246 controls) exceeding the required sample size of 288 participants (95% confidence, and 80% power, 1:1 ratio for cases and controls, and a least detected an odds ratio of 2.0). Participants were recruited from 22 public schools (10 elementary schools for girls, 7 elementary schools for boys, and 5 co-educational kindergartens). Coded questionnaires and consent forms were sent to the parents and all those who signed the consent forms and returned the filled questionnaire to the school administration within 2-3 days were considered for the study. Any children whose parents reported a major health condition (such as heart disease or diabetes) or any type of disability were excluded from the study sample. We attempted to recruit between 11 to15 cases from each school, matched with the same number of controls from the same school. For every case student, a control student was matched by class grade, gender, age, and socioeconomic status. The height and weight of all the children were measured, their Body Mass Index (BMI) was calculated, and they were classified as underweight, normal weight, overweight, or obese according to the Centres for Disease Control and Prevention (CDC) guidelines.  Underweight children were excluded from the sample for the purpose of this study.
A valid and reliable self-administered questionnaire was used in the study. The instrument was composed of material from previously conducted studies and published in the literature and from widely used valid and reliable scales such the Children’s Eating Behaviour Inventory (CEBI) [7-13]. The questionnaire collected information about parents’ demographic characteristics (level of education, employment status, monthly household income, and residence type), their self-reported weight and height, and whether they had any morbidities. In addition to questions regarding child feeding practices since infancy (breastfeeding history, consumption of fruits and vegetables, consumption of fast food, soda intake, other eating behaviours, and purchase of junk food), the instrument included questions regarding the child’s daily physical activity, frequency of daily use of screens (i.e., tablets, computers, smart phones, and television), and the type of behavioural reward system applied by the parents. Parental attitude regarding childhood obesity and how they perceived their child’s weight status (underweight, normal weight, overweight, or obese) were also included in the instrument. The developed questionnaire was translated into Arabic by a professional translator and reviewed by two bilingual research assistants for appropriateness of the terms used and inconsistencies. A group comprising of two nutritionists, two health educators, a paediatrician, and a public health practitioner reviewed the Arabic version of the instrument for content validity, relevance, and suitability of the different measures. Face validity was assessed by asking fifteen individuals to complete the survey and report any encountered difficulties. Questionnaire items were revised according to all recommendations and further tested among 25 parents of school-aged children for a final check of overall clarity. Test-retest reliability was established, and 90% agreement was reported among two administrations of the survey within a 10-day interval to the same group of participants (N=30). Internal consistency of the instrument was determined as Cronbach’s alpha coefficient of 0.79.
Questionnaire responses were coded, entered, and analyzed using the statistical software package STAT13 (College Station, Texas). Descriptive statistics were calculated for all study variables including demographic characteristics and socioeconomic status, parental calculated BMI, and children’s calculated BMI. Logistic regression analysis was conducted to examine the individual and combined effects of the study variables on the likelihood of being overweight/obese cases versus normal weight controls. Odds ratios and 95% Confidence Intervals were estimated for all probable risk factors for overweight/obesity in this population. The significance level was set at P <0.05 for all analyses.