Extracting Vulnerable Populations by Classifying Resilience and Perceived Support in Parenting Mothers: An Explainable Machine Learning Analysis with a Web-Based Cross-Sectional Study (Preprint)

BACKGROUND One life event requiring the most extensive resilience and adaptation is parenting. However, resilience and perceived support in child rearing are diverse, and the actual situation is unclear, especially after public postpartum checkups. OBJECTIVE To explore the psychosocial status of mothers during the child-rearing period from newborn to toddler, with the classifier based on the data on resilience and adaptation characteristics of mothers with newborns. METHODS A web-based cross-sectional survey was conducted. The mothers with newborns aged to about one month (newborn cohort) were analyzed to construct an explainable machine-learning classifier to stratify the resilience and adaptation characteristics for parenting and detect vulnerable populations. Explainable k-means clustering was used because of its high explanation and applicability. The classifier was applied to mothers with infants aged two months to 1 year (infant cohort) and mothers with toddlers aged over one year to 2 years (toddler cohort). The psychosocial status, including depressed mood assessed by the Edinburgh postnatal depression scale, bonding assessed by the postpartum bonding questionnaire, and sleep quality assessed by the Pittsburgh sleep quality index between the classified groups, was compared. RESULTS The total of 1559 participants were included in three cohorts were populations with different characteristics, including parenting difficulties and psychosocial measures. The classifier, which stratified participants into five groups, was generated depending on the self-reported scores of resilience and adaptation in the newborn cohort data (N=310). The classifier determined the group with more incredible difficulty in resilience and adaptation to a child’s temperament and perceived support was the highest difficulty group with high incidences of problems with depressed mood (Relative Prevalence [RP] 5.87, 95% CI 2.77-12.45), bonding (RP 5.38, 95% CI 2.53-11.45), and sleep quality (RP 1.70, 95% CI 1.20-2.40), compared with the group with the lowest difficulty of perceived support. In the infant cohort (N=619) and the toddler cohort (N=461), the stratified group with the most difficulty had high incidences of problems compared with the lowest difficulty group with depressed mood (RP 9.05, 95% CI 4.36-18.80; RP 4.63, 95% CI 2.38-9.02), bonding (RP 1.63, 95% CI 1.29-2.06; RP 3.19, 95% CI 2.03-5.01), and sleep quality (RP 8.09, 95% CI 4.62-16.37; RP 1.72, 95% CI 1.23-2.42), respectively. CONCLUSIONS The classifier based on a combination of resilience and adaptation to the child's temperament and perceived support identified psychosocial vulnerable groups in the newborn cohort, the start-up stage of childcare. Psychosocially vulnerable groups were also detected in qualitatively different infant and toddler cohorts, depending on their classifier. The detected group in the infant cohort showed particularly high RP for depressed mood and poor sleep quality.