The target population for the present study was junior high school students from three selected cities in the eastern Mazandaran province, Iran, who were selected by multi-stage random sampling. All public schools in the three selected cities and their subordinate villages were listed separately for each city. Then, three girls' schools and three boys' schools were randomly selected for each city. Considering that the ratio of city to village population was 2 to 1, for each city, two schools from the city and one school from the dependent village were selected. Since there were at least two classes in each school for each grade, one class has been selected randomly and all students were given a questionnaire.
With respect to effect size = 0.04, power = 0.8 and α = 0.05, the sample size was calculated as 382 individuals. After considering a design effect = 1.5 for cluster sampling method, the number of needed subjects increased to 573. This figure was divided between 6 groups of girls and boys and across three educational grades in each city. Samples were randomly selected from these three educational grades in each of three selected cities and among public schools. In cases where the selected samples did not cover all students in one class, the questionnaire was provided to all students in that class to avoid a feeling of discrimination among the students. Finally, 923 students completed and returned questionnaires to be used for the further analysis.
Prior to the implementation, consent was obtained from the students' parents. In each class, the purpose of the study was explained to the students by one the authors (RK) and then the questionnaires were distributed among the students. They were asked to answer the questions carefully and return it. Further explanation has been provided by RK in case of any question regarding the content of the questionnaire.
The questionnaire included background characteristics of the respondents (age, gender, living in urban or rural areas), questions related to oral health-related behaviors (frequency of brushing, flossing, eating sugary snacks, smoking, and visiting a dentist), and spiritual health (level of agreement with twenty phrases based on the six point Likert scale). The 20-item questionnaire(Paloutzian & Ellison, 1982) was used to assess spiritual health. The reliability and validity of this questionnaire has been confirmed in a previous study(Abhari et al., 2018) on the Iranian adolescents. The answers to the questions were in the form of a 6-point Likert scale from 1 (completely disagree) to 6 (completely agree). The sum of the scores indicated each student’s level of spiritual health with a theoretical range of 20–120. For further analysis, quartiles of the possible scores (20–120) were calculated (Q1 = 20–45, Q2 = 46–70, Q3 = 71 = 95, Q4 = 96–120) and students were subdivided into four groups based on their spiritual health scores. Oral health-related behaviors were compared between two groups of students whose spiritual health scores were in the upper (Q4) and lower (Q2 & Q3) quartiles.
Descriptive statistics included frequency and percent of the adolescents belonging to each category. The Chi-square test served as statistical evaluation to assess differences in frequencies. To evaluate the factors related to the adolescents’ reported oral health behaviors, five similar multivariate logistic regression models were fitted to the data while the adolescents’ gender, age, living in urban or rural areas, and spiritual health score served as covariates. The corresponding odds ratios (OR) and their 95% confidence intervals (95% CI) were calculated. The goodness of fit of the models was assessed by the Hosmer-Lemeshow test.