Childhood Maltreatment and Adulthood Depression Severity: Using Structural Equation Modelling to Uncover Maltreatment-Symptom Specificity

Abstract

Major depressive disorder (MDD) is highly heterogeneous in its etiology, symptom presentation, course, and treatment response. Despite knowing this for decades, however, clinicians and researchers have generally treated MDD as homogenous by relying on total symptom severity scores and linking them to broadly defined etiological factors (e.g. childhood adversity). To truly understand the complex associations between MDD etiology and symptom presentation, general etiological factors need to be deconstructed into their distinct components and independently linked to individual symptom dimensions of MDD. This thesis project aimed to untangle heterogeneity in MDD etiology and symptom presentation by investigating symptom dimensions uniquely associated with childhood emotional, physical and sexual maltreatment. MDD symptom dimensions were identified with exploratory factor analysis (EFA) of Hamilton Rating Scale for Depression (HRSD) responses, using three archival datasets covering 243 community-sampled depressed participants. The data-driven measurement model of the HRSD was evaluated using confirmatory factor analysis (CFA) alongside a four-factor model taken from a previous meta-analysis of HRSD factor structure. Using structural equation modelling (SEM), associations were estimated between HRSD factors and childhood maltreatment (CM) variables from the Childhood Experience of Care and Abuse (CECA) retrospective clinical interview. A five-factor measurement model of the HRSD was identified, which statistically outperformed the four-factor theoretical model. No structural models were formally retained due to poor model fit with the sample data, meaning that the CM-MDD associations observed in this investigation were inconclusive. Although inconclusive, these results suggested that MDD symptom specificity does vary across the different CM types and that some MDD symptoms are associated more strongly with certain CM types than others. Initial evidence suggests that all symptoms captured by the HRSD are more severe in individuals with a history of sexual maltreatment, that anxiety and anhedonia-like symptoms are worse in individuals with a history of physical maltreatment, and that core depression symptoms are worse in individuals with a history of emotional maltreatment. To determine which of these specific associations generalize beyond the current sample, however, this SEM analysis must be cross-validated with an additional sample of community-residing depressed individuals.