Study location and study population
The study was carried out among in-school adolescents (10 to 19 years old) in Ile-Ife, which is a semi-urban town in Southwestern part of Nigeria. Adolescents who were acutely ill, had chronic illnesses that could affect their weight (like sickle cell disease anaemia) and those with disabilities that made them unable to stand were excluded from the study.
Sample size and sampling technique
The sample size was calculated to get an absolute precision of ± 5% using STATCALC on the Epi-Info software. The proportion of expected outcome was taken as 37.2% which was the proportion of in-school adolescents with obesity from a previous study in Ile-Ife, with an acceptable margin of error of 5%. The calculated sample size was 359, but was adjusted for an anticipated non-response of 10%, giving a sample size of 400. Four hundred adolescents were therefore recruited from 6 secondary schools in Ile-Ife using multi-stage sampling technique. Two Local Government Areas (LGAs) were selected from the 5 in Ile-Ife using simple random sampling technique (balloting method) at the first stage. At the second stage, 3 schools each were selected from the list of schools in the selected LGAs. The number of respondents to be selected in each of the schools was determined using proportional allocation. At the third and final stage, the respondents were selected using stratified random sampling technique, with stratification along the line of the different classes.
Research instruments and data collection methods
A pre-tested structured questionnaire was used for data collection using the assisted self-administered method. The dietary patterns were assessed using 92-item quantitative food frequency questionnaire (QFFQ), while the activity patterns were assessed using the physical activity questionnaire for older children and adolescents by Kowalski et al[42], which has been validated and used among similar age group in Nigeria.[43] The instruments for anthropometric measurements were be the Seca® electronic bathroom weighing scale (SECA GmbH & Co, Germany) for measuring weight in 0.1 kilograms (kg). Height was measured to the nearest 0.1 meter using the stadiometer (Leceister® Height Measure, Seca, UK). The anthropometric measurements were done according to standard protocols recommended by the International Society for the Advancement of Kinanthropometry.[44]
Data Management
Data were analyzed using IBM SPSS version 23. Descriptive analysis of all the variables measured were first done, and the categorical variables were reported as frequencies and proportions/percentages, while the continuous variables were reported as means ± standard deviation. At bivariate level, cross-tabulations were done to test for associations between the different categorical variables (in line with the objective of the study) using the chi-square test. Fisher’s exact test was used when there was an expected value was less than 5. T-test for 2 independent samples was used to compare the means of the continuous variables between the 2 categories of the dependent variable (Obese/Not obese). Logistic regression was used to control for confounders and to identify the predictors of obesity out of the independent variables that were significantly associated with obesity at the bivariate level.
Overweight and Obesity were determined using BMI-for-age Z-scores from the WHO reference charts > + 1 to + 2 and obese > + 2 respectively, and the 2 groups represented obesity in this study.
The responses to the questions on activity patterns were scored, and each of the sections was scored over 5. Afterwards, all the scores from the different sections were scored over 5. The scores were then categorized into < 2, 2.00–3.99 and ≥ 4 for low, moderate and high physical activity patterns respectively.[42]
Principal Component Analysis (PCA) was done to reduce the dimension of dietary intake to a small number of dietary patterns. Factors were retained and interpreted for further analysis based on their natural interpretation, visual inflections of the scree-plot of eigen-values construction (Fig. 2) and the percentage of total variance explained. The reliability of the factor analysis was verified using the Kaiser–Meyer–Olkin (KMO) test with a sampling adequacy of 0.9 and Bartlett’s test of sphericity significant at p < 0.001.
Ethical considerations
Ethical clearance was obtained from the Ethical Review Committee, Institute of Public Health, Obafemi Awolowo University, Ile-Ife. Permission for the study was obtained from the Local Inspector of Education (L.I.E.) of the selected local government and the Management of the selected schools. Written informed consent was obtained from parents and the adolescents who were 18 years or above, while assent was obtained from respondents younger than 18 years. All information gathered was kept confidential and participants were identified using only serial numbers.