AbstractThe International Classification of Diseases 11th edition (ICD-11) has introduced a new trauma-based diagnosis of Complex PTSD (CPTSD). It has been debated within both research and clinical practice communities, whether CPTSD is distinct from other trauma-related disorders, particularly posttraumatic stress disorder (PTSD) and borderline personality disorder (BPD). Suggestions of symptom overlap between CPTSD and BPD, observations of relatively high rates of comorbidity between PTSD and BPD, and a recognition of shared risk, common across all three disorders, have led many to challenge both the validity and utility of this new trauma diagnosis. The present thesis aimed to assess the validity of CPTSD, specifically in relation to both PTSD and BPD. Part one of the thesis (chapter 3 and chapter 4) attempted to show that CPTSD can be reliably distinguished from PTSD while part two attempted to show that CPTSD can be distinguished from BPD (chapter 5 and chapter 6). A range of latent variable modelling techniques were employed to analyse data derived from a selection of international trauma-exposed population samples. The samples under investigation included: (i) an Irish treatment-seeking sexual trauma sample, (ii) a U.S. refugee sample, (iii) a U.S. sexual trauma sample, and (iv) a trauma-exposed Israeli sample.
First, confirmatory factor analysis was employed to determine the most accurate dimensional representation of PTSD and CPTSD symptoms using primary data gathered from an Irish, sexual-trauma treatment-seeking sample (chapter 3). The results revealed that the latent dimensional structure of both PTSD and CPTSD symptoms was consistent with the ICD-11 diagnostic proposals that have differentiated between each construct. Second, factor mixture modelling was employed to determine whether a distinct CPTSD symptom profile could be identified among a subgroup of U.S. refugees using data from a large-scale population based epidemiological survey (chapter 4). The findings revealed that it was possible to identify distinct subgroups of individuals that endorsed a distinct PTSD symptom profile and a distinct CPTSD symptom profile. Third, latent class analysis was employed to determine whether a distinct CPTSD symptom profile could be identified in the context of BPD symptoms using population-based data from a U.S. epidemiological survey subsample that had experienced sexual-trauma (chapter 5). Results revealed that a distinct sub-group of the sample was characterised by a CPTSD symptom profile. Moreover, BPD symptoms did not emerge independently of PTSD or CPTSD symptoms. Finally, using data from a trauma-exposed Israeli population sample, the latent dimensional structure of PTSD, CPTSD, and BPD symptoms was investigated using confirmatory bifactor modelling (chapter 6). The best fitting latent dimensional representation of PTSD, CPTSD, and BPD symptomology was consistent with a leading dimensional model of psychopathology (i.e. the HiTOP model). In this model, all symptoms shared a common latent dimension which was assumed to reflect a shared vulnerability to psychopathology. Importantly, with the inclusion of this common latent dimension, it was possible to show that CPTSD (and PTSD) symptoms were uniquely explained by a distinct CPTSD dimension further evidencing the validity of this new trauma-based construct.
The findings from this thesis provided support for the validity of CPTSD in the context of both PTSD and BPD measurement. CPTSD is a distinct trauma-based construct that is identifiable dimensionally and among distinct subgroups in the population. Moreover, CPTSD manifests independently from that which is common to other psychiatric constructs (i.e. general psychopathology vulnerability ‘P’). Identifying the most accurate conceptualisation and classification of trauma-based psychopathology (and psychopathology in general) will lead to significant advances in both clinical research and practice.
|Date of Award||May 2020|
|Supervisor||Jamie Murphy (Supervisor) & Mark Shevlin (Supervisor)|
- Complex PTSD
- Latent variable modelling