Study design and setting:
This was a cross-sectional study conducted among the two selected universities students (i.e., University of Rajshahi and Jahangirnagar University) during March 2020 through an online survey. The University of Rajshahi and Jahangirnagar University is the second and fourth largest universities in Bangladesh respectively, with students coming from the different parts of the country. A total number of 55276 students are studying in these two universities (38495 in University of Rajshahi, and 16781 in Jahangirnagar University).
Sample size calculation and sampling method:
The sample size was calculated from the prevalence estimate using the formula:
![](data:image/png;base64,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)
where, n = number of the samples; z = 1.96 for 95% confidence interval (CI), p = “best guess” for prevalence and d = precision of the prevalence estimate. We did not find any existing data on the prevalence of migraine among the university students of Bangladesh. However, a previous study from the neighboring country India reported the prevalence as 14.12% (10), which was considered as p (best guess) value (p=0.1412) for calculating sample size for the present study, and the formula provided that 2337 sample would be the required size. Assuming a 10% non-response rate, a total of 2600 university students were approached. The convenient sampling method was used to include the participants in this study as those who had the social media id and personal relation with the recruited volunteers had the chances of enrollment in the study.
Data collection procedure: A self-administered online survey form created in Google forms were used to collect data from the participants. The survey link was posted in a regular interval of one week in the internal social media groups of the university students and an open request was placed by the team of investigators to fill-up the form. Also, 20 volunteers from different departments of these universities were employed to circulate the survey link among their student networks, in addition to regular posting in the above-mentioned social media groups. They were instructed to be inclusive, open, and circulate it periodically for maximum reach. Login with email and providing student ID number was mandatory for limiting single response. Email addresses of the participants were collected upon proper clarification and informed consent for the reliability of the data. The study was conducted following the Checklist for Reporting Results of Internet ESurveys (CHERRIES) guidelines (11).
Data collection instruments:
The survey questionnaire comprised of four parts: (i) socio-demographics, lifestyle and behavioral factors related data, (ii) headache-related data, (iii) impact of headache, if migraine was present using the Headache Impact Test (HIT-6), and (iv) presence of anxiety and depressive symptoms using two psychometric scales (the GAD-7 for assessing anxiety disorder and the PHQ-9 for assessing depressive symptoms). As the ID MigraineTM and HIT-6 scales were not used before among Bangladeshi population, these were not validated in Bangla. We have used back translation method for translating the tools after proper consent from the developers under supervision of a team of three consultant neurologists of the Department of Neuromedicine of RMCH (Rajshahi Medical College & Hospital). A pretest of the questionnaire was done in the Department of Neuromedicine od RMCH among 30 diagnosed migraine patients by a consultant neurologist.
Part 1: Socio-demographics, lifestyle and behavioral factors:
Socio-demographic information was collected during the survey by asking questions concerning age, gender, study year, monthly family income, marital status, height, and weight. Lifestyle and behavioral factors included fast food intake (frequency per week), amount of physical exercising (days per week for at least 30 minutes a day), smoking habits (yes/no), alcohol intake (yes/no), substance abuse (yes/no), and sleep quality measured by the Pittsburgh sleep quality index (PSQI) which is an appropriate screening tool for measuring sleep dysfunction in both clinical and non-clinical samples. The PSQI score was categorized as poor (PSQI score >5) and good (PSQI score £5) (12).
Part 2: Headache related data:
Participants were initially evaluated by the question “Did you have two or more headaches in the last 3 months?” Those who responded ‘yes” were considered as the subjects with potentially troublesome headaches and further screened using the ID MigraineTM test. The ID-MigraineTM test is a three-item self-administered screening tool, developed by Lipton et al. (2003) (13). It consists of three questions regarding problems related to migraines over the last three months: 1. Did you feel nauseated or sick in your stomach with your headaches? 2. Did light bother you when you had a headache (a lot more than when you do not have headaches)? and 3. Did your headache limit your ability to work, study or do what you needed to do for at least 1 day? A test-diagnosis of migraine headache is made by at least two positive responses. ID MigraineTM has been validated using the International Classification of Headache Disorders (ICHD) criteria in different studies with a pooled sensitivity of 0.84 and a specificity of 0.76 (14).
Headache related data were collected from the participants who were screened as positive for migraine. These included intensity of headache (measured on a four-point scale where 0 = no headache; 1 = mild headache; 2 = moderate headache; 3 = severe headache recommended for use in migraine research by the International Headache Society) (15), frequency of headache during the past month, associated symptoms of headache, characteristic of headache (unilateral, bilateral, pulsating, and throbbing), frequency of analgesic use during the past month, frequency of healthcare facility visit during past 12 months, migraine triggers, and family history of migraine.
Part 3: Impact of headache:
Headache Impact Test (HIT-6) was used to measure the impact of migraine headaches among the ID MigraineTM positive participants. The HIT-6 is a brief and easy to use instrument, developed by Kosinski et al. (2003) (16) to measure the adverse headache impact and to use in screening and monitoring patients with headaches in both clinical practice and clinical research. It consists of six items regarding problems related to headache (i.e., When you have headaches, how often is the pain severe?) with a five-point Likert scale ranging from 6 (Never) to 13 (Always). The total score was obtained by the summating raw score from each contract ranging from 36 to 78, with greater scores indicating a severe impact. In present study, the headache impact severity was categorized into four classes based on total sores of HIT-6: little or no disability (≤49), mild disability (50–55), moderate disability (56–59), and disability (≥60). This scale is suggested as a reliable and valid tool for measuring headache impact in migraine (17).
Part 4: Anxiety and depressive symptoms:
Patient Health Questionnaire (PHQ-9: The PHQ-9 is one of the most psychometrically sound and robust screening tools, developed by Spitzer et al. (1999) (18) which is one the most widely used instruments for assessing depressive disorder globally including Bangladesh (19,20). This scale consists of nine items regarding problems related to depression symptomatology over the past two weeks (e.g., “Thoughts that you would be better off dead, or of hurting yourself in some way?”) with a four-point Likert scale ranging from 0 (Not at all) to 3 (Nearly every day). The total score was obtained by the summating raw score from each contract ranging from 0 to 27. In the present study, those scoring moderate to severe (≥10) were classed as having depressive symptoms (21).
Generalized Anxiety Disorder (GAD-7): The GAD-7 is one of the most psychometrically sound and robust screening tools, developed by Spitzer et al. (2006) (22) and used in different countries including Bangladesh for assessing anxiety disorder (23,24). The scale comprises seven items regarding problems related to anxiety symptomatology over the past two weeks (e.g., “Feeling afraid as if something awful might happen?”) with a four-point Likert scale ranging from 0 (Not at all) to 3 (Nearly every day). The total score was obtained by the summating raw score from each contract ranging from 0 to 21. In the present study, those scoring moderate to extremely severe (≥10) were classified as having anxiety disorder positive (22).
Statistical analysis:
Statistical analysis was conducted using SPSS version 24.0. Descriptive statistics was performed for categorical variables (i.e., frequency and percentage), and for continuous variables (i.e., mean and standard deviation). Chi-square test and t-test were used in case of categorical variables and continuous variables, respectively to investigate the relationship between dependent and independent variables. Binary multiple logistic regressions were performed with a 95% confidence interval to determine the significant associations between categorical dependent and independent variables. Multicollinearity problems among independent variables in multiple logistic model was checked by standard error (SE) suggested by Chan (25). The association of variables was considered statistically significant if the two-sided p-value was less than 0.05.