Following the seminal study by Linehan et al. (1) comparing dialectical behavior therapy with treatment as usual, there have been a number of randomized, controlled trials establishing the efficacy of various cognitive and psychodynamic treatments for borderline personality disorder (2—6). Most, but not all, of these studies have included some short-term follow-up assessment (2, 4, 7—10). Given the entrenched and chronic nature of borderline personality disorder, long-term follow-up is central for establishing the significance of these treatments. However, the time frames in these follow-ups have been relatively short, between 6 and 18 months, leaving the long-term efficacy of these treatments unclear.

Even more problematic is that the actual outcomes for these studies have generally been mixed. For example, whereas the overall results of an outpatient psychotherapy study by Linehan et al. are suggestive of the value of dialectical behavior therapy, results from their naturalistic follow-up of patients in dialectical behavior therapy were uneven (7). At the 6-month follow-up, there were no differences between dialectical behavior therapy and the treatment-as-usual groups in the number of days hospitalized; at the end of a 1-year follow-up, there were no differences between groups in number of days hospitalized or in self-destructive acts. Additionally, a 6-month follow-up from the study by Verheul et al. (11) found no differences between dialectical behavior therapy group and the treatment-as-usual control group on impulsive behavior, parasuicidality, or alcohol and both soft and hard drug use (12). Finally, in the recent randomized, controlled trial by Linehan et al. (9) comparing dialectical behavior therapy with community treatment by experts, the authors found that at the 1-year follow-up, there were no differences between the dialectical behavior therapy and community treatment by experts groups in terms of parasuicidality or crisis service use. Additionally, when results from the treatment year and follow-up period were combined, patients in dialectical behavior therapy were half as likely to make a suicide attempt as patients in the community treatment by experts group (9); however, this finding disappeared when the follow-up period was examined alone (Lynch TR, July 2004, conference presentation). Taken together, these findings suggest variable maintenance of treatment effects and ongoing impairment in functioning in patients who may have initially experienced symptom relief.

Accordingly, the article by Bateman and Fonagy in this issue (13) showing the long-term maintenance of treatment gains for patients with borderline personality disorder has particular significance for the field and is destined to become a seminal study in the annals of psychiatry. What makes their study unique and especially valuable is their long-term follow-up of patients with borderline personality disorder in a well-characterized treatment at a clinically meaningful interval—8 years after random assignment and 5 years after the end of treatment. As the authors note, their study has a number of advantages. The original study was an efficacy study of clinically referred patients; there was a well-characterized control group; there were few exclusion criteria; and they had a low dropout rate. Additionally, follow-up data were available from centrally organized medical records concerning self-harm, suicide attempts, and inpatient episodes on all patients. Thus, the loss of information due to incomplete self-report data was small. The findings are impressive: 8 years after initial random assignment and 5 years after the completion of any mentalization-based therapy, those treated with mentalization-based therapy not only showed statistical superiority in reduced suicidality, service use, and medication use and increases in global and vocational functioning, but they also exhibited a remarkable level of clinical change (only 13% met criteria for borderline personality disorder compared to 87% of those in the treatment-as-usual group).

The findings from the study by Bateman and Fonagy significantly increase our confidence in the value of their approach as a long-term efficacious treatment for borderline personality disorder and firmly cement mentalization-based therapy as a viable treatment in the existing armamentarium for borderline personality disorder. Having different treatment options is important because, given the heterogeneity of borderline personality disorder, it is unlikely that any one treatment will be useful for all patients (14). For example, although the randomized, controlled trial by Linehan et al. (1) of dialectical behavior therapy was a breakthrough for the research on the treatment of borderline personality disorder, the response rate or recovery rate in dialectical behavior therapy appears to be about 50%. Specifically, in their initial study, they reported that 45% of patients had one or more hospitalizations and/or one or more parasuicidal episodes during the last 4 months of the treatment year (15). Bohus et al. also found that only 50% of the patients in dialectical behavior therapy had good outcomes (14). Clearly, dialectical behavior therapy is an efficacious treatment when compared with treatment as usual, maybe even compared with community treatment by experts. Nonetheless, a significant portion of individuals receiving dialectical behavior therapy are not improving, and these individuals might be better served in different treatments, such as cognitive-behavioral therapy, mentalization-based therapy, transference-focused psychotherapy, schema-focused psychotherapy, or even community treatment by experts. Future research will need to examine more fully the interaction between treatment and patient characteristics in order to determine "what treatment, by whom, is most effective with this individual, with that specific problem, under which set of circumstances" (16).

Another important implication of the findings of Bateman and Fonagy is their suggestion that longer-term, phase-based treatment may be particularly useful for helping patients with borderline personality disorder establish more durable changes. Along these lines, Howard et al. (17) suggested a three-phase dose-response model of psychotherapy in which patients initially experience remoralization (the initial boost experienced from the feeling that help is there), followed by remediation (symptom reduction) and finally rehabilitation (establishing adaptive ways of living, also conceived of as personality change). Remoralization is usually accomplished quickly, whereas remediation is more gradual and typically occurs between 3 and 8 months. Rehabilitation is quite gradual and can take years. Each phase may have different treatment goals, measurable by different outcome variables, and require different interventions. Although there is no direct evidence that patients in mentalization-based therapy achieved rehabilitation, the continued symptomatic improvement seen in the patients after the conclusion of treatment is suggestive of such change. Finally, it is important to note that the findings clearly support the notion that developing behavioral control need not be skill based but can occur through the development of mental skills.

The impressive outcomes aside, the next steps in treatment research need to include expanding the range of outcomes targeted by treatment. Although symptom change is central to the treatment of borderline personality disorder, and symptom change can improve the quality of one’s life and result in personality change (18), thus far, the data suggest that we need to work harder to move beyond symptom change and help patients improve the quality of their day-to-day lives. As Bateman and Fonagy note, although patients showed symptom improvement, and most even lost their borderline personality disorder diagnosis; they still experienced significant social and functional impairment (54% had Global Assessment of Functioning Scale scores under 60). The findings from this study are consistent with those of the Collaborative Longitudinal Personality Disorders Study (19) and the McLean Study of Adult Development (20), indicating that symptomatic and diagnostic improvement does not necessarily result in social and functional improvement. This finding raises the question: What kind of outcome can we expect in the treatment of borderline personality disorder? Linehan et al. coined the phrase a "life worth living" to describe the laudable goals of treatment. However, Linehan et al. noted in their early naturalistic follow-up that although the "subjects in the dialectical behavior therapy group acted better… they were still miserable," experiencing "moderate symptoms" and/or "generally functioning with some difficulty" and living lives of quiet desperation (p. 1775) (8). Tolstoy, in a letter to Valerya Aresenjev in 1856, said, "One can live magnificently in this world if one knows how to work and how to love"; of course, Freud also believed that being able to love and work—(in German) lieben und arbeiten—were important indicators of well-being. The question is how much the field has moved or can move beyond symptom change to this broader goal of satisfaction with life. It is in this spirit that we need to strive in developing efficacious treatments for our patients with borderline personality disorder. The outcomes to date are encouraging, but they are also sobering and suggest that while treated patients with borderline personality disorder are better off, they have not achieved the overarching positive life changes that we would like to offer our patients. Clearly, mentalization-based therapy, dialectical behavior therapy, and other treatments are able to help patients with borderline personality disorder obtain remission from symptoms more quickly than the natural course of the disorder. However, we are far from helping our patients achieve fulfillment in love and work.

Given that even the most potent of the available treatments are limited in their effects, helping many but not all patients and resulting in only partial and/or temporary recovery, research needs to focus on understanding the therapeutic mechanisms that lead to change in these patients. The benefits of doing so would be considerable. Understanding the mechanisms through which a treatment operates is likely to facilitate the development of more advanced treatments that will yield larger effects as active components are identified, intensified, and refined, whereas inactive or redundant elements could be discarded or the focus reduced. The results are likely to yield more potent and more efficient therapies. In addition, understanding mechanisms has the potential for enabling prescriptive decisions regarding which patients will benefit from particular treatments and for whom a treatment will be ineffective. Finally, identifying mechanisms may not only enhance treatment development and delivery but also advances our understanding of the nature of clinical disorders.

Follow-up studies are also an important way of examining mechanisms of change in the treatment of borderline personality disorder. For example, patients in Kernberg’s transference-focused psychotherapy showed significant changes in reflective function, a hypothesized mechanism of change in transference-focused psychotherapy as well as in mentalization-based therapy but those in dialectical behavior therapy or a psychodynamic supportive psychotherapy did not show such change (6). It will be important to see if the patients who showed changes in reflective function, regardless of treatment condition, are more likely to maintain the gains they made in treatment. If so, reflective function can be considered a central mechanism to be targeted by treatment as it would lead to more sustainable change in adaptive functioning posttreatment.

In sum, a theory is corroborated to the extent that we have subjected it to risky tests; the more dangerous the tests it has survived, the better corroborated it is. Bateman and Fonagy are to be commended for embarking on such a task. In one important respect, Bateman and Fonagy’s mentalization-based therapy has survived a dangerous test—that of long-term follow-up—and the findings will push the field forward. Nevertheless, more tests—such as further study of mechanisms of change and the challenge of helping patients achieve satisfaction in life—await mentalization-based therapy and all of us involved in treatment research for borderline personality disorder.