Identifying treatment moderators of a trauma-informed parenting intervention with children in foster care: Using model-based recursive partitioning

Abstract

Identifying treatment moderators of a trauma-informed parenting intervention with children in foster care: Using model-based recursive partitioning

Delivery and Impact of Foster Parent Training: Similarities and Differences for Relative and Non-Relative Foster Parents

“…The data for the study came from the Illinois Birth Through Three (IB3) Title IV-E waiver demonstration. The IB3 demonstration was a 5-year project whose purpose was to adapt and provide evidence-supported, trauma-informed parenting programs for both birth and foster parents of children under the state’s legal custody (Chung et al, 2021; Rolock et al, 2015). The demonstration focused on caregivers who had children aged three or younger, recognizing that young children with a history of traumatic experiences are likely to experience more severe emotional, behavioral and developmental problems (Cook et al, 2005).…”

“…The demonstration focused on caregivers who had children aged three or younger, recognizing that young children with a history of traumatic experiences are likely to experience more severe emotional, behavioral and developmental problems (Cook et al, 2005). Within the IB3 waiver demonstration, two adapted programs - Nurturing Parenting Program (NPP) and Child-Parent Psychotherapy (CPP) - were offered for birth and foster parents (Chung et al, 2021; Rolock et al, 2015). The current study focuses on only NPP for foster parents.…”

“…NPP was offered in a group format. Follow-up individualized coaching was also available, while its use was limited (Chung et al, 2021; Rolock et al, 2015).…”

Examining Financial Hardship and Caregiver Subgroups in Kinship Foster Placements: A Machine Learning Approach

“…To address Questions 2 and 3, this study utilized model-based recursive partitioning (MOB), a machine learning algorithm, and the "partykit" package for R version 4.3.1 to identify subgroups of caregivers based on the relationship between the number of DSS children that caregivers already had and their perceived ability to support them without additional DSS financial assistance. Other child welfare outcome studies have employed this tree-based machine learning method to identify smaller subgroups [65,66].…”

“…Our data-driven analytic approach was tested for parameter stability across nine partitioning variables (i.e., hypothesized moderators). If the algorithm found substantial between-group differences (i.e., parameter instability based on a 0.05 significance level), it repeatedly split the sample based on the variable that had the highest instability (i.e., lowest p-value), thus enabling subgroups to emerge, if any [65]. A minimum sample size of 10 per subgroup was specified to ensure a reasonable number of respondents per group and allow for meaningful between-group comparisons.…”