Although it is common for patients with dissociative disorders to report a history of suicide attempts, there is very little data systematically comparing suicidality in patients with dissociative disorders versus patients without these disorders. The subjects in our study were 231 patients consecutively admitted to an inner-city, hospital-based outpatient psychiatric clinic. Eighty-two of these patients completed structured interviews for dissociative disorders, borderline personality disorder, and trauma history (dissociative disorders interview schedule) and for posttraumatic stress disorder and substance abuse (Structured Clinical Interview for DSM-IV). Patients receiving a dissociative disorder diagnosis were compared with nondissociative patients on measures of self-harm and suicidality. Presence of a dissociative disorder was strongly associated with all measures of self-harm and suicidality. When we focused on patients with a history of multiple suicide attempts, significant associations were found between several diagnoses (dissociative disorder; borderline personality disorder; posttraumatic stress disorder; alcohol abuse/dependence) and multiple suicide attempter status. When these diagnoses were entered in a logistic regression, a highly significant association remained for dissociative diagnosis and multiple suicide attempter status (odds ratio, 15.09; 95% confidence interval, 2.67-85.32; p = 0.002). Dissociative disorders are commonly overlooked in studies of suicidality, but in this population they were the strongest predictor of multiple suicide attempter status.

Dissociative disorders were highly prevalent in this clinical population and typically had not been previously diagnosed clinically. The high prevalence of dissociative disorders found in this study may be related to methodological factors (all patients were offered an interview rather than only those who had scored high on a screening self-report measure) and epidemiological factors (extremely high prevalence rates for childhood physical and sexual abuse were present in the overall study population).

No abstract