This was a cross-sectional study conducted in Kuala Lumpur, Malaysia from March to July 2019. Kuala Lumpur’s preschool areas were selected as statistics from the Social and Welfare department showed it ranked second highest in child maltreatment cases, with 6.5% involving caregivers to the child [17].
Utilising Open Epi, the estimated sample size for this study was determined as n = 262, based on effect sizes from similar studies [38–40] (Supplementary Table 1). Thirty-six preschools were randomly selected from a sampling frame provided by the Kuala Lumpur Social and Welfare Department (138 preschools) for the study. All childcare providers aged 18 years and above working in preschools serving children under four were eligible to participate. Participants with acute psychiatric illness were excluded and referred for further assessment at the nearest clinic.
Written consent was obtained from each eligible respondent before enrolment. The data collection involved face-to-face interactions with the study subjects, utilising standardised questionnaires covering sociodemographic details, the 5-item Santa Clara Strength of Religious Faith Questionnaire (SCSRFQ-5), the Center for Epidemiological Studies-Depression Scale (CES-D), and the Job Content Questionnaire (JCQ) (Supplementary Questionnaires). The survey took approximately 20 minutes to complete.
The sociodemographic questionnaire includes age, gender, ethnicity, marital status, education level, income status, family history of mental illness, chronic illness, and stressful life events. Participants indicate “yes” or “no” to major events causing emotional distress, such as assault, loss of a loved one, childhood abuse, serious marital problems, family problems, and injury or accidents. Participants who answered “yes” to any life events were taken as positive for life events in the binary category [ 41]. For past medical illnesses, participants were asked about the presence of chronic disease for the past year or more. The positive answers were cross-checked with the participant’s home-based treatment card by the participant herself. They were guided by a list of chronic diseases adopted from a study by Kader Moideen et al. [42]. Participants assess the strength of religiosity using the 4-point Likert scale Santa Clara Strength of Religious Faith Questionnaire (SCSRFQ-5). Each of the five items is scored from 1 to 4 (1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly agree), with analysis based on the sample median cut-off [43]. The current study yielded a Cronbach’s alpha coefficient of 0.829.
The CES-D, comprising 20 items in English and Malay, assessed subjects’ current depressive symptoms based on Radloff’s 1977 publication [44]. General population-based validation studies revealed a strong Cronbach alpha coefficient ranging from 0.85 to 0.90 [44–45]. Participants reported depression-related symptoms from the past week, with each item scored from 0 to 3 (0 = Rarely or None of the Time, 1 = Some or Little of the Time, 2 = Moderately or Much of the Time, 3 = Most or Almost All the Time). Scores ranged from 0 to 60, and a score of 16 or higher indicated significant clinical depressive symptoms.
This study used English and Malay Job Content Questionnaires (JCQ) to measure job strain and psychological exposure. The Malay Version of the tool was validated with Cronbach’s alpha coefficients of 0.75 (decision latitude), 0.50 (job demand) and 0.84 (social support) [46]. The JCQ assesses psychosocial job domains, including skill discretion, demand, decision-making, supervisor support, co-worker support, and insecurity. Participants rated questions on a 4-point scale (1 = strongly disagree, 4 = strongly agree), and reversed coding was employed for questions on job insecurity. The summation of scoring for each participant was based on the formula (Supplementary Table 2). The level of job strain is interpreted based on the Karasek Job Strain Model (Supplementary Fig. 1). Job demand components measure the psychological stressors concerning fulfilling workload, unexpected tasks, and job-related personal conflicts. Meanwhile, job decision latitude refers to the worker’s potential to control tasks and conduct during the working day [47].
The scoring disclosure was done privately in a separate room towards the end of the sessions. All the participants with high scores were given a referral letter for further assessment at the nearest government clinics or hospitals. Participants received a confidential envelope with a referral letter, participation certificate, and contact details for additional counselling support through the local Befriender’s platform.
Statistical Analysis
Statistical analyses were conducted using IBM SPSS Statistics (version 24.0, IBM Corp LP, Armonk, NY, USA). The preliminary data exploration was conducted to identify the centrality, dispersion, normality, and missing data. Variables with < 5% missing data were included. Symptomatic depression was assessed using CES-D (scores ≥ 16 = depressed or < 16 = not depressed). The univariate analysis involved students’ independent sample T-test or Mann-Whitney U-test for continuous variables. Meanwhile, Pearson’s Chi-square or Fisher’s Exact test was used for categorical variables. Multivariable logistic regression was performed with the Hosmer-Lemeshow model development strategy to determine the final model. Effect modification first-order main effects of explanatory variables were checked using the likelihood-ratio test. The final model was assessed for assumption compliance, multicollinearity, and model fitness test. The rule of thumb recommended by Paul et al. (2013) [48]. was applied to ensure the effectiveness of the Hosmer-Lemeshow test. Model sensitivity was evaluated with the area under the receiving operating characteristics (AUROC) curve. A significance level of P-value < 0.05 was applied.