Over the past decade there has been plenty of good news for people with borderline personality disorder. First, the acute symptoms of the disorder, characterized by marked emotional dysregulation, suicidality, self-harm, and impulsivity—long believed to be chronic in nature—have been shown to remit naturally without treatment, with patients experiencing substantial reduction in symptoms far earlier than previously assumed (1). Second, carefully crafted psychotherapeutic treatments have been developed and have been shown in randomized controlled trials to be more effective than treatment as usual at relieving acute symptoms such as self-harm and suicide attempts. The new treatments, which are not widely available, are an unlikely cause for the apparent change in prognosis over the past two decades. The change may have been because general psychiatric treatment, or treatment as usual, delivered by well-meaning mental health professionals was actually harmful and has now improved or is no longer so easily available (2). No treatment trials have reported on negative outcomes, so the specialist treatments may primarily minimize harmful interventions, perhaps by carefully formulating the patient's problems according to a focused theory and offering crisis planning to prevent uncoordinated care.

In this issue, McMain et al. (3) add further important findings to this literature. First, their article adds persuasive evidence that well-structured general psychiatric treatment, which targets the acute symptoms of borderline personality disorder and bears little resemblance to the earlier unstructured treatment as usual, is as effective as branded specialized treatments (in this case, dialectical behavior therapy) against those symptoms when delivered in the context of an organized research study. McMain et al. previously reported (4) that general psychiatric management was as effective as dialectical behavior therapy on all outcomes at the end of 1 year of treatment. Now it appears that there are no differences in their findings at the 2-year follow-up, which is a more robust test of any treatment.

But this equivalence of outcomes between well-organized treatments may not be particular to dialectical behavior therapy. Every time a named specialized treatment has been compared with an alternative well-structured general psychiatric intervention that is organized around and specific to the supposed underlying pathology of borderline personality disorder, differences in outcomes have been either nonexistent or at best only moderate. In a randomized controlled trial, Clarkin et al. (5) compared two different specialist treatments, transference-focused psychotherapy and dialectical behavior therapy, and one generalist treatment, supportive psychotherapy, which was organized around clinical problems specific to borderline personality disorder. The study found that outcomes across the three treatments were “generally equivalent.” In another randomized controlled trial, Chanen et al. (6) compared cognitive analytic therapy with well-organized good clinical care for adolescents with borderline personality disorder or borderline traits. Good clinical care and cognitive analytic therapy were equally effective, with significant improvements across a range of clinical outcome measures. Bateman and Fonagy (7) compared mentalization-based treatment with structured clinical management and found that both were effective treatments and that structured clinical management was superior in the initial months at reducing self-harm.

The proponents of the named therapies may argue about the limitations and shortcomings of trials in which a well-organized comparator treatment is shown to be equally effective. But there are few limitations to the McMain et al. study. The general psychiatric management was provided by an expert group that was keen to work with borderline patients, and there may have been some overlap between therapist attitudes and the techniques used in the two treatment groups as a result of cross-contamination over time. An understanding of mechanisms of change and information about moderators of outcomes would be more informative to clinical care, but this study does not tell us much about either. Nor does it tell us which patients got better, whether they remained better, and which did not. We have very limited knowledge about moderators and mechanisms even for dialectical behavior therapy, the best-studied treatment. Neacsiu et al. (8), examining the role of dialectical behavior therapy skills in improving treatment clinical outcomes, found, unsurprisingly, that participants treated with dialectical behavior therapy reported using three times more behavioral skills by the end of treatment than those assigned to a comparison treatment. But the use of dialectical behavior therapy skills mediated the decrease in suicide attempts and depression and the increase in control of anger, and it partially mediated the decrease in nonsuicidal self-injury over time, which suggests that skills acquisition and practice may be an important mechanism of change in dialectical behavior therapy. The illusion of an explanation is created, but it falls when we generalize and ask how the patients treated with general psychiatric management in this study benefitted equally from treatment and maintained their improvement during follow-up.

What makes people with borderline personality disorder improve given so many disparate treatments using apparently contrasting interventions to achieve similar outcomes? Do all these treatments have something in common? What makes a treatment “specialized” for borderline personality disorder? All therapies for borderline personality disorder, including those in the McMain et al. study, share certain characteristics, and these elements rather than the specific techniques of treatment may be responsible for their effectiveness. These therapies 1) provide a structured manual that supports the therapist and provides recommendations for common clinical problems; 2) are structured so that they encourage increased activity, proactivity, and self-agency for the patients; 3) focus on emotion processing, particularly on creating robust connections between acts and feelings; 4) increase cognitive coherence in relation to subjective experience in the early phase of treatment by including a model of pathology that is carefully explained to the patient; and 5) encourage an active stance by the therapist, which invariably includes an explicit intent to validate and demonstrate empathy and generate a strong attachment relationship to create a foundation of alliance.

The message is becoming clear—most people with borderline personality disorder need specialized treatment that is primarily structured and organized around their core symptoms. Unfocused treatment is inadequate. However, the findings from this study that show continued functional impairment suggest that our current theoretical understanding of borderline personality disorder and the associated therapeutic packages is inadequate. We may have taken our eyes off the ball.

Zanarini et al. (9), in a 10-year prospective follow-up study, found that while substantial reduction in symptom severity is achievable, good social and vocational function are more difficult to attain with or without treatment. McMain et al. confirm this troubling finding by showing that patients with borderline personality disorder still showed marked functional impairment at 2-year follow-up despite well-organized treatment. This finding was heralded by earlier studies. In the longest follow-up to date of randomized controlled studies, Bateman and Fonagy (10) found that patients still had functional impairment 8 years after entry into the trial. This finding was further confirmed by Davidson et al. (11), who found at 6-year follow-up that only one-fifth of patients showed improvement in affective disturbance, and their quality of life remained poor. It is possible that clinicians have been so concerned about acute life-threatening symptoms of the disorder that they have neglected the interpersonal dysfunction that also lies at the core of the patient's problems. Patients function in a social and relational environment, and their sensitivity to attachment processes may make them especially vulnerable to long-term functional impairment. This would suggest that family and social interventions over the longer term might be more helpful, for example, by providing more intermittent treatment over longer periods.

The way forward is not a mix-and-match system, using techniques as and when the clinician thinks is appropriate. This would deliver unstructured treatments without coherence. We should generate an increasingly coherent theory of the disorder, underpinned by an understanding of mechanisms of change, and translate this into a carefully crafted therapeutic package. Only then is it likely that patients with borderline personality disorder will have a better chance of functional improvement. Part of the benefit of treatment for people with borderline personality disorder comes from the experience of being involved in a carefully considered, well-structured, and coherent interpersonal endeavor. But we need to do more.