The general factor of psychopathology : precursors and consequences

Abstract

Comorbidity among different types of psychopathology is very common whether they are treated as discrete diagnostic categories or continuous dimensions. Empirical methods to describing these dimensions of dysfunction and their high co-occurrence have advanced our understanding and definition of mental disorders. These methods consistently show that different forms of psychopathology are correlated, resulting in the extraction of a general factor of psychopathology known as the p factor.

In this thesis, we performed multivariate analyses on the nationwide Swedish registers and the Swedish twin registers to examine how the general psychopathology relates to genetics, cardiometabolic complications and pain and suicidal behavior.

In study I, we explored whether psychiatric polygenic risk scores (PRS) could directly predict general and specific psychopathology. We modeled one general and seven specific factors based on childhood psychiatric symptoms, and one general and three specific factors based on adolescent psychiatric symptoms. We then regressed each general and specific factor onto ten psychiatric PRS simultaneously between and within twin pairs, the latter controls for indirect pathways (population stratification, assortative mating and dynastic effects). We found that PRS-general psychopathology associations did not appear attributable to indirect pathways, suggesting that genetics appeared to directly influence symptomatology.

In study II, we examined whether the increased risk of cardiometabolic complications for mental health conditions might be attributed to a general liability toward psychopathology or confounded by unmeasured familial factors. We identified general, internalizing, externalizing, and psychotic factors based on the comorbidity among psychiatric diagnoses and criminal convictions in young adulthood. We then regressed the cardiometabolic complications in middle adulthood on the latent general factor and three uncorrelated specific factors within a structural equation modeling framework in individuals and across sibling pairs. BMI and smoking were used as mediators among child-bearing females. We found that associations between individuals with mental disorders in early life and later long-term risk of cardiometabolic complications appeared attributable to a general liability toward psychopathology. Sibling analyses suggested that the elevated risk could not be attributed to confounds shared within families, and the associations could be partly mediated via lifestyle factors. Clinicians may consider lifestyle-based interventions to reduce the risk of cardiometabolic complications for patients with several mental disorders.

In study III, we investigated the link between chronic pain comorbidity and later suicidal behaviors. Based on nine self-reported chronic pain conditions, we identified three factors related to pain: one general pain factor, and two specific factors, which measure neck-shoulder pain and pain-related somatic symptoms respectively. And we applied a co-twin control model to control for familial confounding when regressing general and specific pain on suicidal behaviors. We found that general pain factor and somatic pain factor are associated with increased risk for suicidal behavior; but these associations appear to be mainly attributable to familial confounding. Clinicians may find it advantageous to assess pain comorbidity in addition to specific pain types. Nonetheless, addressing pain may not necessarily lead to a reduction in future suicidal tendencies, as the associations may be influenced by familial factors.