Prenatal Sociodemographic Factors Predicting Maltreatment of Children up to 3 Years Old: A Prospective Cohort Study Using Administrative Data in Japan.
Aya IsumiKunihiko TakahashiTakeo FujiwaraPublished in: International journal of environmental research and public health (2021)
Identifying risk factors from pregnancy is essential for preventing child maltreatment. However, few studies have explored prenatal risk factors assessed at pregnancy registration. This study aimed to identify prenatal risk factors for child maltreatment during the first three years of life using population-level survey data from pregnancy notification forms. This prospective cohort study targeted all mothers and their infants enrolled for a 3- to 4-month-old health check between October 2013 and February 2014 in five municipalities in Aichi Prefecture, Japan, and followed them until the child turned 3 years old. Administrative records of registration with Regional Councils for Children Requiring Care (RCCRC), which is suggestive of child maltreatment cases, were linked with survey data from pregnancy notification forms registered at municipalities (n = 893). Exact logistic regression was used for analysis. A total of 11 children (1.2%) were registered with RCCRC by 3 years of age. Unmarried marital status, history of artificial abortion, and smoking during pregnancy were significantly associated with child maltreatment. Prenatal risk scores calculated as the sum of these prenatal risk factors, ranging from 0 to 7, showed high predictive power (area under receiver operating characteristic curve 0.805; 95% confidence interval (CI), 0.660-0.950) at a cut-off score of 2 (sensitivity = 72.7%, specificity = 83.2%). These findings suggest that variables from pregnancy notification forms may be predictors of the risk for child maltreatment by the age of three.
Keyphrases
- risk factors
- mental health
- pregnant women
- preterm birth
- pregnancy outcomes
- healthcare
- young adults
- electronic health record
- cross sectional
- public health
- big data
- machine learning
- risk assessment
- density functional theory
- deep learning
- cancer therapy
- drug delivery
- chronic pain
- human health
- health promotion
- affordable care act