%0 Journal Article
%T Understanding forms of childhood adversities and associations with adult health outcomes: A regression tree analysis.
%A Perrins SP
%A Vermes E
%A Cincotta K
%A Xu Y
%A Godoy-Garraza L
%A Chen MS
%A Addison R
%A Douglas B
%A Yatco A
%A Idaikkadar N
%A Willis LA
%J Child Abuse Negl
%V 153
%N 0
%D 2024 Jul 17
%M 38761717
%F 4.863
%R 10.1016/j.chiabu.2024.106844
%X <B>BACKGROUND: </B>Empirical studies have demonstrated associations between ten original adverse childhood experiences (ACEs) and multiple health outcomes. Identifying expanded ACEs can capture the burden of other childhood adversities that may have important health implications.<BR><B>OBJECTIVE: </B>We sought to identify childhood adversities that warrant consideration as expanded ACEs. We hypothesized that experiencing expanded and original ACEs would be associated with poorer adult health outcomes compared to experiencing original ACEs alone.<BR><B>METHODS: </B>The 11,545 respondents of the National Longitudinal Surveys (NLS) and Child and Young Adult Survey were 48.9 % female, 22.7 % Black, 15.8 % Hispanic, 36.1 % White, 1.7 % Asian/Native Hawaiian/Pacific Islander/Native American/Native Alaskan, and 7.5 % Other.<BR><B>METHODS: </B>This study used regression trees and generalized linear models to identify if/which expanded ACEs interacted with original ACEs in association with six health outcomes.<BR><B>RESULTS: </B>Four expanded ACEs-basic needs instability, lack of parental love and affection, community stressors, and mother's experience with physical abuse during childhood -significantly interacted with general health, depressive symptom severity, anxiety symptom severity, and violent crime victimization in adulthood (all p-values <0.005). Basic needs instability and/or lack of parental love and affection emerged as correlates across multiple outcomes. Experiencing lack of parental love and affection and original ACEs was associated with greater anxiety symptoms (p = 0.022).<BR><B>CONCLUSIONS: </B>This is the first study to use supervised machine learning to investigate interaction effects among original ACEs and expanded ACEs. Two expanded ACEs emerged as predictors for three adult health outcomes and warrant further consideration in ACEs assessments.