Full Program »
104. The structure of psychopathology and relationship with poor sleep, self-harm, suicidality, risky sexual behaviour, and low self-esteem in a population sample of young Australians
Introduction: Multiple mental and substance use disorders commonly co-occur across adolescence. Indeed, there is a growing body of evidence highlighting the presence of a single general dimension of psychopathology, labelled the p-factor, that can account for the relationships across multiple manifest indicators in a hierarchical or higher order structure. However, relatively little evidence has emerged regarding the validity and utility of this model when explaining relationships between indicators of psychopathology and a range of clinical and behavioural factors that have been previously implicated across multiple disorders. The current study utilised a large cross-sectional general population survey of Australian youth to model this latent structure of psychopathology and examine the extent to which the overall model can account for specific relationships between psychopathology and poor sleep, self-harm, suicidality, risky sexual behaviour, and low self-esteem.
Methods: Data were from the Young Minds Matter survey, a stratified multistage area probability sample of Australian households where there was at least once child aged 4-17 years old. The current study focused on the youth subsample representing 2,314 young persons aged 13 to 17. Indicators of psychopathology and clinical/behavioural covariates were assessed using parent-report and self-report diagnostic interviews/questionnaires. A range of confirmatory factor models, latent class models, and factor mixture models were estimated to identify the best structure of psychopathology based on several fit statistics. A series of structural equation models were then estimated to examine direct and indirect effects associated with each psychopathology indicator and the clinical/behavioural covariates, controlling for age and sex.
Results: Across all examined models, a confirmatory factor model with three lower order factors, representing internalising, externalising, and thought disorder, and a single higher order factor representing general psychopathology evidenced the best fit. Indirect effects models indicated that the relationship between manifest indicators of psychopathology and clinical/behavioural factors were largely non-specific with relatively small and non-significant direct effects. Although, significant direct effects were found between major depression and self-harm, suicidality, and risky sexual behaviour. At the latent level, the general p-factor accounted for much of the relationship across lower-order factors and the clinical/behavioural covariates, except for internalising and poor sleep, internalising and low self-esteem, and externalising and risky sexual behaviours.
Discussion: For the majority of pairwise relationships, the higher order model adequately captures the total amount of variance. However, there was some evidence of significant or relatively large direct effects, indicating potential measurement non-invariance associated with the higher order model. These findings have several implications for understanding the relationship between psychopathology at various latent levels with clinical/behavioural covariates, including identification of potential targets for intervention and/or tailoring of prevention programs.