Integrating trauma, self-disturbances, cognitive biases, and personality into a model for the risk of psychosis: a longitudinal study in a non-clinical sample

Abstract

The hypothesis of the psychosis continuum enables to study the mechanisms of psychosis risk not only in clinical samples but in non-clinical as well. The aim of this longitudinal study was to investigate self-disturbances (SD), a risk factor that has attracted substantial interest over the last two decades, in combination with trauma, cognitive biases and personality, and to test whether SD are associated with subclinical positive symptoms (PS) over a 12-month follow-up period. Our study was conducted in a non-clinical sample of 139 Polish young adults (81 females, age M = 25.32, SD = 4.51) who were selected for frequent experience of subclinical PS. Participants completed self-report questionnaires for the evaluation of SD (IPASE), trauma (CECA.Q), cognitive biases (DACOBS) and personality (TCI), and were interviewed for subclinical PS (CAARMS). SD and subclinical PS were re-assessed 12 months after baseline measurement. The hypothesized model for psychosis risk was tested using path analysis. The change in SD and subclinical PS over the 12-month period was investigated with non-parametric equivalent of dependent sample t-tests. The models with self-transcendence (ST) and harm avoidance (HA) as personality variables were found to be well-fitted and explained 34% of the variance in subclinical PS at follow-up. Moreover, we found a significant reduction of SD and subclinical PS after 12 months. Our study suggests that combining trauma, cognitive biases, SD and personality traits such as ST and HA into one model can enhance our understanding of appearance as well as maintenance of subclinical PS.

Keywords: Cognitive biases; Personality; Psychosis risk; Self-disturbances; Trauma.

Fig. 1

Path analysis with self-transcendence. The model has satisfactory fit indices: (χ² (11) = 5.117, p = 0.925; RMSEA = 0.000 [90% CI = 0.000–0.029] p = 0.978, CFI = 1.00, TLI = 1.083, SRMR = 0.033). The bootstrapping estimate revealed a significant standardized indirect effect of traumatic life events through all other variables to subclinical positive symptoms II (β = 0.088, 95% CI = 0.043—0.147, p = 0.002). This model explained 33.9% of the variance in subclinical positive symptoms II and 33.2% in self-disturbances II. Different colours in the figure mark two parts of the model—one that refers to the mechanisms of self-disturbances and the other that indicates the association of subclinical positive symptoms and self-disturbances in 12-month follow-up based on their baseline measurement. *p < 0.05, **p < 0.01, ***p < 0.001, n.s. non-significant

Fig. 2

Path analysis with harm-avoidance. Results of path analysis suggested an acceptable model fit: (χ² (11) = 17.701, p = 0.089; RMSEA = 0.066 [90% CI = 0.000–0.121] p = 0.281, CFI = 0.971, TLI = 0.917, SRMR = 0.062). The bootstrapping estimate revealed a significant standardized indirect effect of traumatic life events through all other variables to subclinical positive symptoms II (β = 0.080, 95% CI = 0.038–0.137, p = 0.003). This model explained 33.7% of the variance in subclinical positive symptoms II and 32.9% in self-disturbances II. Different colours in the figure mark two parts of the model—one that refers to the mechanisms of self-disturbances and the other that indicates the association of subclinical positive symptoms and self-disturbances in 12-month follow-up based on their baseline measurement. *p < 0.05, **p < 0.01, ***p < 0.001, n.s. non-significant