Study design
A cross-sectional study design was used to collect data on exposure and outcome variables from a group of madrasa students in the Mirpur area of Dhaka, Bangladesh between April and May 2022. In Dhaka, there are about a thousand madrasas18. We chose the Mirpur area purposefully because of the number of madrasas in the area, including Maktab, Hafizia, Qawmi and Alia. However, one-third of them, particularly Maktab, Hafizia and Qawmi madrasas, are not fully functional18. In terms of educational programs, environment, and facilities, the madrasas in this area are similar to those in Dhaka city. For this study, six madrasas were randomly selected from a total of 12 madrasas in the area. The data collection process involved the utilization of a structured questionnaire, which was designed using insights from prior research9.
Data collection
One teacher at each madrasa was appointed as a coordinator and given information about the research project and procedures for collecting the data. The coordinator informed the students and their parents about the study before data collection. Following, we used a simple random sampling technique to select participant from each selected madrasa using year of the course. From the register books of the madrasa, we were able to determine the participants’ year of the course. The principal investigator (the first author) and data collectors were present at each marasa when the students filled out the questionnaires. Data collectors stressed informant confidentiality, responded to questions, and read questions aloud for students with reading problems. Written informed consent was obtained from both the madrasa authorities and parents/guardians prior to the survey. In case of boarding and orphan students, madrasa authorities gave the consent on behalf of their guardians. Questionnaires were distributed and returned on the spot. Data collectors checked whether there were any missing items or unclear identification; if there were, the participants were asked to complete the information. A pilot survey was conducted among 20 students to investigate the capacity to comprehend the relevant techniques and potential trouble-some situations during interviews. The eligible study participants included those aged 8 to 18 years old who had been studying at madrasa for at least 6 months. Informed consent was obtained from all the study participants and was informed about the right to free to withdraw or opt out at any point without any penalty. If the students did not consent to data collection, they were not included in the study. The required sample size was 384 at 90% power, 95% confidence level, with anticipated 50% of students had poor HRQoL and margin of error was 5%. We then randomly selected at least 70 students from each of 6 madrasas. A total of 450 students were randomly invited to participate, but 31 students were deemed ineligible, resulting in 419 students receiving the questionnaire, yielding a response rate of 93%. Of them, 24 participants did not continue, leaving 395 participants who filled out the questionnaire. Finally, a total of 373 data were selected for analysis after eliminating incomplete and insufficient quality information of 22 data. The participation in the study was voluntary without any financial compensation. Data selection, identification and inclusion flow diagram process are presented in Supplementary Appendix 1.
Measures
Outcome variable
The quality of life was considered as the primary outcome measure in this study and the quantification of HRQoL was using the KIDSCREEN-10 index which contained 10 items. It is a concise, internationally validated screening tool that can be used to assess vulnerabilities in mental health and subjective well-being among 8- to 18-year-olds. The KIDSCREEN-10 is based on more KIDSCREEN-52 quality of life assessments, but maintains strong psychometric properties while offering an efficient alternative with just 10 items7. Details regarding the index, including its validity can be found elsewhere15. Answer to the first and ninth question had 4 categories: not at all, slightly, moderately, very-extremely. The rest of the questions had 5 categories: never, seldom, quite often, very often, always. Responses were coded so that a higher score indicates better HRQoL. The median score for the participants was 36 (33.0–38.0). Median split was used to categorize the finally summed score as: good (participants at or above the median score of 36) and poor (participants below the median score of 36). The index showed good internal consistency with Cronbach’s alpha of 0.778.
Independent variables
The study examined several independent variables, including demographic features such as age (in years), gender, orphan status, financial support, fathers’ education, fathers’ occupation, mothers’ education, mothers’ occupation, accommodation, academic year, body mass index [BMI] (BMI > 1SD as ‘overweight/obese’, BMI < − SD as ‘thinness’ and others are coded as ‘normal’ as per WHO)16, number of family members, and number of siblings. In addition, student’s orphan status, financial guardian, preference for madrasa study, library facility, and opportunity for mobile and computer use was considered independent variables. Furthermore, present illness, visiting doctor and reason for not visiting doctor, feeling safe at madrasa, bullying and place of bullying were independent variables.
Statistical analysis
The study used STATA 17 and jamovi (v 2.4) for data analysis. The normality of continuous variables was tested using the Shapiro–Wilk test. For descriptive statistics, continuous data were presented as mean (Standard Deviation; [SD]) or median (IQR) where applicable. Categorical variables were represented as counts and percentages. For inferential statistics, the authors used the Chi-square test, Fisher’s exact test for testing association between two categorical variables and Mann–Whitney U test for association between categorical and non-normal continuous variable. A logistic regression model with all significant variables identified by bivariate screening was developed to find the strength of association. A p-value < 0.05 was considered statistically significant.
Ethical approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki Declaration. The ethics committee of North South University, Bangladesh approved the study protocol (Ref-NSU/2021/102). For data collection in the madrasas, official permission was obtained from the respective authorities prior to the survey. Participants voluntarily participated in the study and had the right to withdraw from the study without a reason. All methods were carried out in accordance with relevant guidelines and regulations.
Declaration of AI usage
The authors used ChatGPT (version 3.5) during the preparation of this work to enhance the language and readability of the manuscript, as the authors are not native English speakers. After utilizing the tool, the authors thoroughly reviewed and edited the content as necessary and assume full responsibility for the manuscript’s content.