• at-risk mental state;
  • attachment;
  • early psychosis;
  • premorbid social adjustment;
  • psychosocial treatment


Aim: Insecure attachment may influence vulnerability to and outcome of psychotic symptomatology. The present study examined whether attachment style predicted symptoms and functioning of at-risk mental state (ARMS) patients after 6 months of psychosocial intervention, over and above the effects of initial clinical severity and premorbid social adjustment (PSA).

Methods: Symptoms and functioning were assessed at baseline and 6 months later in 31 ARMS patients (mean age = 15.7). No patient received antipsychotic medication, but all engaged in intense psychosocial needs-adapted treatment. Clinicians (unaware of the aims of the study) rated attachment, PSA, symptoms, and functioning.

Results: Attachment was not related to baseline clinical severity. However, improvement in psychoticism was predicted by attachment (in particular by secure, preoccupied and dismissing) beyond the effects of baseline clinical severity and PSA. Secure attachment also predicted improvements in disorganization and functioning. Poor PSA predicted less improvement in disorganization and negative symptoms but did not impact psychoticism.

Conclusions: The three attachment prototypes that predicted improvement in psychoticism (secure, preoccupied and dismissing) share the existence of at least one positive psychological model (either about self or about others). It may be that the psychosocial intervention helped ARMS patients to disconfirm negative models and/or reinforce positive ones. Patients' attachment styles were not related to baseline clinical severity but impacted improvement of positive symptoms. These findings appear consistent with evidence that impaired self-esteem and dysfunctional self and others schemas constitute risk factors for reality distortion.