Psychodynamic Psychotherapy for Personality Disorders

In 1990, Gerald Klerman argued persuasively, and surely noncontroversially, that psychiatric patients have a right to effective treatment (1) . Klerman referred to litigation in which a patient hospitalized for 7 months at Chestnut Lodge held the hospital negligent for providing only intensive psychotherapy without adding antidepressant medication in spite of a marked worsening of his depressive condition. Klerman concluded that “the issue is not psychotherapy versus biological therapy but, rather, opinion versus evidence,” and he added that it is “regrettable that psychoanalysts and psychodynamic psychotherapists have not developed evidence in support of their claims for therapeutic efficacy” (p. 415).

In 2007, Klerman’s plea for the right to effective treatment is no less cogent, but his lament that “the efficacy of psychotherapies has not been addressed by NIH” (p. 414) no longer applies, at least to the treatment of depression. Since 1990, many clinical trials funded by the National Institutes of Health (NIH) have focused on psychotherapy as a treatment for depression, adding substantially to our menu of effective treatments for this disorder.

But how are we doing developing an evidence base concerning the efficacy of psychotherapy, often longer term, for severe and disabling personality disorders that either stand alone or co-occur with depression or other axis I conditions? Here, Klerman’s lament remains relevant, although we’re making progress. Interest (long overdue) in rigorous randomized, controlled trials of psychotherapy with this population has grown in the research community, and NIH funding for this work has increased. Leichsenring and Leibing (2) reported a meta-analysis of psychodynamic treatment and cognitive behavior therapy, concluding that both are effective treatments for personality disorders. However, only 14 psychodynamic studies and 11 cognitive behavior therapy studies (including dialectical behavior therapy) were found that used adequate methodology for inclusion; of these, three psychodynamic studies and five cognitive behavior therapy studies were randomized, controlled trials.

The refrain that absence of proof does not equal proof of absence (of efficacy) is most often heard regarding psychodynamic psychotherapy. Fortunately, labor-intensive randomized, controlled trials of long-term psychodynamic treatment are being carried out, such as the work of Bateman and Fonagy (3) and the work of Clarkin and colleagues (4) . However, numerous challenges remain. The number of subjects in each study is often small, intent-to-treat analyses are seldom reported, and head-to-head studies are rare. There is still much to learn about which treatment is best for which patient, what length of treatment is best, what level of care is best, what outcomes should be measured, and how durable the treatment gains are. And answers to these questions cannot rely solely on randomized, controlled trials but must include “real world” collaborative effectiveness studies, illustrated by the CATIE trials for schizophrenia and the STAR*D trials for depression.

Blatt and Auerbach (5) suggested, regarding psychodynamic measures of therapeutic change, that “sustained symptom remission, while essential to any successful treatment outcome, is secondary to and dependent on more basic changes in the personality structure” (p. 269). Of interest, this argument dovetails with recent findings from naturalistic longitudinal studies of personality disorders. The Collaborative Longitudinal Personality Disorders Study, funded by the National Institute of Mental Health (NIMH), now in its 12th year, demonstrated that, unlike the DSM-IV-TR generic definition of personality disorders as enduring and stable over time, more than half of the four personality disorders studied (schizotypal, borderline, avoidant, and obsessive-compulsive) showed a “remission” within the first 2 years of follow-up, defined as at least 12 consecutive months meeting no more than two diagnostic criteria (6) . These researchers suggested that personality disorders be reconceptualized as hybrids of stable personality traits and dysfunctional behaviors (symptoms) that fluctuate over time, perhaps correlated with environmental circumstances. For example, the most stable criteria for borderline personality disorder were affective instability and inappropriate intense anger; the least stable were frantic efforts to avoid abandonment and self-injury. A Borderline Personality Disorder Phenotypes Conference was held in October 2006 in New York, sponsored by the Borderline Personality Disorder Research Foundation, in which borderline personality disorder was considered from the point of view of stable core traits, heritable endophenotypes (e.g., affective instability and impulsive aggression) to clarify the stable trait structure of the disorder and to differentiate core traits from symptomatic behavior. Such efforts are already contributing to the early phase of rethinking the defining criteria for, and the classification system of, the personality disorders for DSM-V.

In 2004, Blatt and Shahar (7) stated that if “psychoanalytic treatment is to survive in the era of evidence-based medicine and managed care systems, empirical evidence is needed to demonstrate its unique nature and effectiveness” (p. 393), a sentiment reminiscent of Klerman’s 1990 advice and one that reassures us that this need continues to be recognized even if the wheel is turning very slowly. It is welcome, therefore, that in this issue of the Journal , Blatt and colleagues present a careful study of seriously disturbed young adults, most of whom had personality disorders, who were receiving intensive psychodynamically oriented hospital-based treatment. Building on a large body of earlier work, the authors utilized two trained judges to review extensive intake clinical records to indentify two patterns of psychopathology: an anaclitic configuration reflecting preoccupation with establishing and maintaining interpersonal relatedness and an introjective configuration reflecting preoccupation with establishing and maintaining a meaningful sense of self. The authors used projective psychological testing to evaluate the two groups of patients at two time points to assess change in three different types of thought disorder. The authors suggest that psychodynamic criteria such as these could identify subgroups of patients who respond differently to different therapeutic interventions, much as neurobiologic criteria can predict therapeutic response.

Intensive treatment of the sort described here is available today only to a small group of patients; however, these and other strategies may allow us to identify prototypic, core pathological traits that can guide treatment selection and serve as meaningful measures of change. The importance of unpacking the current defining criteria for the personality disorders, particularly to differentiate traits from symptoms, is now widely recognized, and careful process research such as the work of Blatt and colleagues will surely inform our progress.

Meanwhile, as clinical researchers study psychotherapy process and outcome, neuroscientists continue to teach us about the biological nature of psychotherapy viewed as a particular form of learning and memory. (A term now being heard in the halls of NIMH is “neuropsychotherapy.”) Kandel and coworkers (8) recently stated that “it is now clear that psychotherapy can induce robust changes in brain function that are detectable with neuroimaging” and that “several lines of evidence point to an important future role for neuroimaging in evaluating the mechanisms and outcome of psychotherapy” (p. 680). Because it is also clear that psychotherapy is a core evidence-based treatment for at least some of the personality disorders (e.g., borderline personality disorder) our hope for the future lies with the partnership of the psychotherapy researcher and the neuroscientist to study which treatment works for which patient—not only for retrospective understanding but, more important, for prospective selection of the right treatment for each individual patient based on brain imaging, molecular neurobiology, genomics, core psychological traits, and other critical factors transforming treatment planning for patients with severe personality disorders into a more exact science.