Perinatal & Pediatric
Effectively targeting perinatal home visiting interventions: estimating heterogeneous treatment effects on infant hemoglobin, breastfeeding, and diet using sorted effects Siva Balakrishnan* Siva Balakrishnan Gary Darmstadt Yunwei Chen Sean Sylvia So Young Ryu Ann M Weber
Introduction
Heterogeneous treatment effects (HTE) are undertested in public health evaluations. We explored HTE of a stage-based home counseling intervention on exclusive breastfeeding (EBF), child hemoglobin and dietary diversity (DDS).
Methods
In a cluster randomized controlled trial, pregnant women and caregivers of infants <6 months of age were enrolled in intervention (n=486) or control (n=663) arms in rural Sichuan, China. Community health workers delivered educational modules tailored to participants’ stage of development for 12 months. A novel sorted effects method was used to estimate HTE. Individual conditional average treatment effects (CATE) were modelled by regressing baseline characteristics, treatment, and interactions thereof on outcomes. Characteristics of the 20% most and 20% least affected families were compared using multiple t-tests and adjusted p-values to identify those associated with HTE.
Results
Evidence of HTE on outcomes was strong with significant individual CATE on EBF and DDS among the 20% most affected families (Fig 1). Mothers with higher baseline caregiving knowledge (score difference between top and bottom quintiles: 0.83 SD) or who gave birth vaginally (prevalence difference: 0.17) saw greatest increases in EBF. Mothers pregnant at baseline (prevalence difference = 0.45) or with less social support (score difference: -0.77 SD) saw the most benefits on hemoglobin. Mothers with lower caregiving knowledge (score difference: -0.59 SD) had greatest increases in DDS.
Discussion
The evidence supports the use of stage-based curricula and targeting mothers from pregnancy to obtain the greatest increases in child hemoglobin and EBF. To improve overall effects on EBF, breastfeeding modules may need to adjust content for first-time mothers and those with low caregiving knowledge. Partnering with hospital staff may improve EBF, particularly after c-section. Engaging family members to strengthen maternal social support may improve infant DDS.
