Data for Postpartum Depression Prediction in Bangladesh
Description
This dataset (n=800) explores risk factors for postpartum depression (PPD) in Bangladesh. Data was collected from postpartum women (birth within 24 months) across various locations (health complexes, hospitals, clinics, and rural areas and residential areas in major cities). Participation was voluntary and required informed consent. The dataset includes sociodemographic (age, residence, education, marital status, occupation, income), familial (husband's education/income, family type, household members, relationship with family), personal health (addiction, children, pregnancy history, abuse, depression history, chronic diseases), and neonatal health-related variables (pregnancy details, birth details, newborn health, postpartum feelings) variables. For depression screening, PHQ-2, EPDS, and PHQ-9 scores are included. The data was digitized after collection. This dataset can be used to investigate the prevalence of PPD, identify risk factors, and develop predictive models. Limitations include self-reported data, screening tools vs. diagnosis, and potential sampling bias. The dataset has shown, about 44% of the participants are in the risk of high level of postpartum depression, based on the EPDS scoring and 28% of the participants are in the risk of sever and moderately severe postpartum depression, based on the PHQ-9 scoring.
Files
Steps to reproduce
This dataset (n=800) for postpartum depression (PPD) prediction in Bangladesh was created using the following steps: 1. Questionnaire: Based on literature, the questionnaire, validated by experts (gynecologists, psychologists, psychiatrists), included sociodemographic, familial, personal/neonatal health variables. 2. Sample Size & Eligibility: Target sample size was calculated. Postpartum women (up to 24 months postpartum, live birth) were eligible. Women unable to respond due to serious illness were excluded. 3. Ethical Approval: Ethical approval was obtained. 4. Recruitment: Participants were recruited from health facilities and communities in major Bangladeshi cities and villages. 5. Consent: Informed consent was obtained before interviews. 6. Data Collection: Data were collected using a structured questionnaire via two methods: face-to-face interviews and an online form. Both methods ensured consistency in question format and response structure. 7. Depression Screening: PHQ-2, EPDS, and PHQ-9 were used. 8. Data Digitization and Translation: Data was digitized and translated from Bangla to English. 9. Missing Data: Data points with missing values were removed. 10. Data Cleaning: Data was cleaned and encoded using Python. 11. Dataset Creation: Final dataset compiled in CSV format.