The influence of personality on the treatment outcome of psychopathology

T. Widiger’s paper summarizes how personality and psychopathology relate to each other and the challenge of disentangling the various forms of inter-relationship. This commentary focuses on how personality may influence treatment outcome and choice of treatment in patients with psychopathology.

The research on the relationship between personality and treatment outcome is largely confined to one psychopathological condition, which is depression. Widiger notes that some authors claim that a change in self-description secondary to a mood disorder can represent an actual change in personality. He argues that it is then difficult to conduct research since depression and personality are no longer distinguished constructs. This commentary attempts to consider the relationship between depression and personality as distinct constructs in relation to treatment outcome.

It can be stated with some confidence that high scores on neuroticism or similar measures such as harm avoidance are associated with poorer outcomes in the treatment of depression 1. This result has been consistently reported in pharmacotherapy 2 as well as psychotherapy 3,4 trials.

The effects of personality disorders on the outcome of depression have been less robust. Although it is generally maintained that personality disorders have a negative effect on outcome, a review by Mulder 1 found that the best designed studies show little or no difference in outcome between depressed patients with and without personality disorders. A meta-analysis of pharmacotherapy studies also reported no significant difference in outcome 5. However, a larger meta-analysis using less stringent study criteria reported a modest but significant negative effect on treatment outcome in depressed individuals with a comorbid personality disorder 6.

Recent treatment trials, using more sophisticated designs to specifically study the relationship between personality and treatment outcome, have not reported a worse outcome. Kelly et al 7 measured the severity of personality disorder as well as depression and found that, while the severity of depressive symptoms at baseline predicted poorer outcome, the severity of personality disorder did not. Craigie et al 8 reported that, while increasing personality disorder complexity was related to slightly poorer end state functioning, there was no association between personality disorder complexity and poorer treatment outcome.

One problem is that the severity of personality disorder and the severity of depressive symptoms were strongly correlated in both studies. Unless the analyses correct for this, the effect of personality disorder on outcome will be over-estimated. De Bolle et al 9 went further contending that the statistical approach used to analyse data may contribute to the inconsistent findings. They showed in their study that, while single level regression analysis showed significantly worse treatment outcome among depressed patients with personality disorder, this worse outcome was no longer present after controlling for statistical heterogeneity and interdependency. We are left with the uncomfortable conclusion that better designed studies show the least effect of personality on depression outcome.

Clinicians are most interested in whether a depressed patient with a comorbid personality disorder would respond better to some treatments than others. In particular whether psychotherapy, either alone or added to antidepressant treatment, would produce a better outcome. The available evidence in this respect is a little counter-intuitive. In Newton-Howes et al’s 6 meta-analysis, there was a non-significant trend for pharmacotherapy to be superior to psychotherapies. Fournier et al 10 directly compared paroxetine and cognitive behaviour therapy in depressed patients with and without personality disorder. They reported a non-significant trend favouring paroxetine over the psychotherapy (66% vs. 44% met response criteria at 16 weeks). Kelly et al 7 reported no interaction between personality disorder states, problem solving treatment, group seminars or treatment as usual. Joyce et al 11 reported that comorbid personality disorders negatively affected treatment response to interpersonal psychotherapy but not cognitive behaviour therapy.

With regard to pharmacotherapy, there is also little specific guidance. There is some evidence that tricyclic antidepressants (TCAs) are less useful in individuals with comorbid Cluster B personality disorders 12. Perhaps of greater interest is that antidepressant treatment may produce improvements in neuroticism 13 and personality disorder pathology 14 in patients whose depression has shown only minimal improvement with treatment. In other words, the improvement in personality functioning was not dependent upon depression improvement.

In conclusion, the relationships among personality, psychopathology and treatment outcome are complex. In depression, higher neuroticism scores are con- sistently associated with poorer outcome. Comorbid personality disorders are generally but less consistently associated with worse treatment response. However, these findings may partially reflect correlations between depression severity and neuroticism and personality pathology. Depression could be understood as direct expression of neuroticism, as Widiger notes. Depression may also have links with specific personality disorders (e.g., Cluster C with low self-esteem and worry).

The limited literature appears to slightly favour antidepressants (with the possible exception of TCAs in Cluster B personality disorders) over psychothera- pies, and cognitive behavioural type ther- apies over less structured interpersonal psychotherapies. It is possible that anti- depressants produce improvements in personality pathology without much effect on an individual’s depressive symptoms. All this evidence encourages clinicians to actively treat depressed patients with personality pathology.