We conducted a cross-sectional study to measure the association between sociodemographic factors and mental well-being. Bangladeshi citizens were invited to participate in a study via social media during the onset of quarantine. Approximately 2500 people received the invitation from March 28 to April 7, 2020. Of potential online connections, 1523 participated, giving a 61% overall response rate. 1404 responses were considered for analysis after distinguishing them based on inclusion and exclusion criteria. Inclusion criteria include healthy adults who are between the ages of 16 to 60 years old. Pregnant women and participants having chronic illness were excluded.
The questionnaire included a variety of demographic variables, such as age, gender, marital status, current living location, occupation, current working condition, and education status. Current working conditions were classified by not working, working from home, both working from home and in-person to office. The occupational groups were split into business, government, healthcare, housewife/unemployed, non-government (i.e. private), and students.
Outcome measure: Warwick Edinburgh Mental Well-being Scale
The Warwick-Edinburgh Mental Well-being Scale (WEMWBS) was used throughout as it provided an appropriate measure of mental well-being. WEMWBS was found to be easy to complete, clear, and unambiguous; it is also popular with practitioners and policymakers [12,16]. The WEMWBS measures subjectively perceived wellness, and psychological function on a 5-point Likert scale over 2 weeks (14–70 points, a higher score means better mental health). As the scale was not invented through mental illness screening methods, there is no cut-off point . The questionnaire was used in various studies and validated in several languages and settings [11, 13-14]. Before beginning the survey a user license for using WEMWBS was obtained (Registration ID: 517150559).
We analyzed data using a software R. Where an item in the WEMWBS is absent, it substitutes for the mean of other items in the domain. Otherwise, we deleted the participant from the analysis with the missingness of two or more than two items in the WEMWBS items. Sociodemographic characteristics were described by frequency (percent), mean, and standard deviation. An independent t-test, and one-way ANOVA, were used to determine the relationship between socio-demographic factors and well-being scores. Afterward, a multivariable linear regression model was used to control confounding variables (Model 1). A separate multivariable linear regression model was generated after adjusting the confounding variables to stratify the analysis by male and female groups (Model 2 and Model 3). Factors at the 5% level were defined as statistically significant. We evaluated the internal consistency of WEMWBS using Cronbach's alpha .