Using latent class analysis to empirically classify maltreatment according to the developmental timing, duration, and co-occurrence of abuse types

Abstract

Background: Individuals can have vastly different maltreatment experiences depending on the types, developmental timing, and duration of abuse. Women and men may be differentially affected by distinct abuse patterns.

Objective: To examine whether maltreatment subgroups could be identified based on the types, developmental timing, and duration of abuse, and to determine their prevalence among a large, community-based sample. We also examined sex differences in associations of maltreatment subgroups with adverse health outcomes.

Participants and setting: Data came from 9310 women and men (95 % White) in the United States who responded to the Growing Up Today Study questionnaire in 2007 (aged 19-27 years).

Methods: Participants reported on physical, sexual, and emotional abuse occurring in childhood (before age 11 years) and adolescence (ages 11-17 years). We conducted latent class (LC) analyses using indicators for child and adolescent abuse. We examined associations of LCs with health outcomes using sex-stratified log-binomial models with generalized estimated equations.

Results: We identified five LCs characterized by: 1) no/low abuse (59 %), 2) child physical abuse (16 %), 3) adolescent emotional abuse (9%), 4) child and adolescent physical and emotional abuse (16 %), and 5) child and adolescent sexual abuse (1%). LCs were uniquely associated adult health outcomes among both women and men. Associations of LCs with eating disorder behaviors appeared stronger for men than women.

Conclusions: Individuals experience distinct patterns of maltreatment based on the types, developmental timing, and duration of abuse. These patterns are uniquely associated with adverse health outcomes in adulthood, and can be identified using LCA.

Keywords: Child abuse; Epidemiology; Latent class analysis; Population-based.