There is a longstanding belief that personality represents a structure that is stable over time, and changes, if at all, very slowly. Nonetheless, clinical and empirical evidence suggests that in patients with some Axis I disorders, the rate of personality disorders using DSM criteria decreases after treatment, suggesting that personality as assessed by phenomenological systems is state-dependent.
An alternative to the DSM phenomenological system of conceptualizing personality is the dynamic concept of character, that is, a predictable pattern of both adaptive and pathological defense mechanisms, and personality organization comprised of object relations, ego strengths, and superego development. Data from this study address the hypothesis that defense mechanisms and personality organization remain relatively stable in patients treated for Axis I disorders, irrespective of clinical improvement. Patients meeting DSM-IV criteria for major depressive disorder (MDD) entered randomized, controlled medication trials. Defensive functioning was evaluated with the Defense Style Questionnaire (DSQ) [Bond et al., 1983: Arch Gen Psychiatry 40:333-338], and personality organization was assessed with the Inventory of Personality Organization (IPO; Clarkin et al., unpublished), both at baseline and at the completion of the clinical trial. Data were analyzed for whether an individual's pattern of defense mechanisms and personality organization were stable over time regardless of response to treatment of MDD. The question was also asked whether a predominant pattern of defense mechanisms or level of personality organization predicts response to treatment or dropout rate. Among treatment responders, nonresponders and drop-outs, baseline DSQ scores were similar except for "image-distorting" defenses, which were significantly more prevalent among drop-outs compared to responders (P = .016). Post-treatment DSQ values revealed a significant decrease in "maladaptive" defenses (P = .01) in the entire sample, while intermediate and "adaptive" defenses remained unchanged. This same pattern was found to hold true in treatment responders. When comparing treatment responders and nonresponders at the end of the trial, medication responders used significantly less "maladaptive" defenses than did nonresponders (P = .003), and had a significantly higher, or healthier level of "overall defensive functioning" (P = .04). Baseline and post-treatment IPO values did not show significant differences. Results of the study address the question of whether there are personality characteristics that are enduring and that can be appreciated irrespective of an Axis I disorder. Depression and Anxi-