METHODS: We conducted a 9-month, parallel, multiarm, cluster-randomised controlled trial in 31 rural villages in Kishoreganj District, Bangladesh. Villages were randomly allocated to: group sessions ('group'); alternating groups and home visits ('combined'); or a passive control arm. Sessions were delivered fortnightly by trained community members. The primary outcome was child stimulation (Family Care Indicators); the secondary outcome was child development (Ages and Stages Questionnaire Inventory, ASQi). Other outcomes included dietary diversity, latrine status, use of a child potty, handwashing infrastructure, caregiver mental health and knowledge of lead. Analyses were intention to treat. Data collectors were independent from implementers.
RESULTS: In July-August 2017, 621 pregnant women and primary caregivers of children<15 months were enrolled (group n=160, combined n=160, control n=301). At endline, immediately following intervention completion (July-August 2018), 574 participants were assessed (group n=144, combined n=149, control n=281). Primary caregivers in both intervention arms participated in more play activities than control caregivers (age-adjusted means: group 4.22, 95% CI 3.97 to 4.47; combined 4.77, 4.60 to 4.96; control 3.24, 3.05 to 3.39), and provided a larger variety of play materials (age-adjusted means: group 3.63, 3.31 to 3.96; combined 3.81, 3.62 to 3.99; control 2.48, 2.34 to 2.59). Compared with the control arm, children in the group arm had higher total ASQi scores (adjusted mean difference in standardised scores: 0.39, 0.15 to 0.64), while in the combined arm scores were not significantly different from the control (0.25, -0.07 to 0.54).
CONCLUSION: Our findings suggest that group-based, multicomponent interventions can be effective at improving child development outcomes in rural Bangladesh, and that they have the potential to be delivered at scale.
TRIAL REGISTRATION NUMBER: The trial is registered in ISRCTN (ISRCTN16001234).
METHODS: This study used available under-five nutritional secondary data from the Demographic and Health Surveys performed in sub-Saharan African countries. The research used bagging, boosting, and voting algorithms, such as random forest, decision tree, eXtreme Gradient Boosting, and k-nearest neighbors machine learning methods, to generate the MVBHE model.
RESULTS: We evaluated the model performances in contrast to each other using different measures, including accuracy, precision, recall, and the F1 score. The results of the experiment showed that the MVBHE model (96%) was better at predicting malnutrition than the random forest (81%), decision tree (60%), eXtreme Gradient Boosting (79%), and k-nearest neighbors (74%).
CONCLUSIONS: The random forest algorithm demonstrated the highest prediction accuracy (81%) compared with the decision tree, eXtreme Gradient Boosting, and k-nearest neighbors algorithms. The accuracy was then enhanced to 96% using the MVBHE model. The MVBHE model is recommended by the present study as the best way to predict malnutrition in under-five children.