Determining essential dimensions for the clinical approximation of personality disorder severity: multi-method study

Abstract

Background

Decades of research on the dimensional nature of personality disorder have led to the replacement of categorical personality disorder diagnoses by a dimensional assessment of personality disorder severity (PDS) in ICD-11, which essentially corresponds to personality functioning in the alternative DSM-5 model for personality disorders. Besides advancing the focus in the diagnosis of PD on impairments in self- and interpersonal functioning, this shift also urges clinicians and researchers worldwide to get familiar with new diagnostic approaches.

Aims

This study investigated which PDS dimensions among different assessment methods and conceptualisations have the most predictive value for overall PDS.

Method

Using semi-structured interviews and self-reports of personality functioning, personality organisation and personality structure in clinical samples of different settings in Switzerland and Germany (n = 534), we calculated a latent general factor for PDS (g-PDS) by applying a correlated trait correlated (method – 1) model (CTC(M–1)).

Results

Our results showed that four interview-assessed PDS dimensions: defence mechanisms, desire and capacity for closeness, sense of self, and comprehension and appreciation of others’ experiences and motivations account for 91.1% of variance of g-PDS, with a combination of either two of these four dimensions already explaining between 81.8 and 91.3%. Regarding self-reports, the dimensions depth and duration of connections, self-perception, object perception and attachment capacity to internal objects predicted 61.3% of the variance of a latent interview-based score, with all investigated self-reported dimensions together adding up to 65.2% variance explanation.

Conclusions

Taken together, our data suggest that focusing on specific dimensions, such as intimacy and identity, in time-limited settings might be viable in determining PDS efficiently.