Change Mechanisms in Empirically-Supported Treatments
Examining Components of Therapies for Borderline Personality Disorder
The treatment models that we have reviewed (i.e., Dialectical Behaviour Therapy, Mentalization-Based Therapy, Transference-Focused Psychotherapy, and Schema-Focused Therapy, Dynamic Therapies, and Cognitive-Behavioral Therapies) represent manualized packages that are tested in their entirety in randomized controlled trial studies. Randomized Clinical Trials (RCTs) are currently regarded as the gold standard in determining whether a treatment is efficacious for a particular disorder (Chambless & Hollon, 1998; Chambless & Ollendick, 2001). Although RCTs afford a relatively high level of control over extraneous variables, there has been an ongoing debate within the psychotherapy research literature regarding their validity and transferability to real-world clinical practice (Wampold, 2001; Westen, Novotny, & Thompson-Brenner, 2004). Without exploring this debate any further, it is important to note that RCT studies cannot reveal which particular aspects of a treatment model are most helpful in producing change. Only research that systematically examines the components and change mechanisms within therapies can properly address this issue, and such work is at a very early stage with respect to treatments for BPD.
Researchers have begun to assess which components of Dialectical Behavioural Therapy (DBT) are necessary and sufficient for change. Skills training, a central component of DBT provided in group settings, has been found by itself to be more effective than unstructured group therapy in reducing emotional and symptomatic distress (Soler et al., 2009). However, adding a DBT skills training group to a non-DBT individual treatment did not result in greater symptom improvement (cited in Linehan et al., 1993). Skill use within DBT programs has been found to predict reductions in BPD features (Stepp, Epler, Jahng, & Trull, 2008), and to play a mediating role in improvements in depression, suicide attempts and anger control, and self-harm (Neacsiu, Rizvi, & Linehan, 2010). Patients report feeling that mindfulness and distress tolerance skills are most helpful (Miller, Wyman, Huppert, Glassman, & Rathus, 2000), and most frequently used (Lindenboim et al., 2007; Stepp et al., 2008). Validation also appears to be an important component of DBT. One study found that patients with BPD and heroin dependence who received a validation intervention, combined with Narcotics Anonymous meetings, had equivalent symptom improvements and no treatment dropouts over one year, compared with a 36% dropout rate in standard DBT (Linehan et al. 2002). Patients in DBT who perceive their therapist as affirming and protecting also report less frequent episodes of self-harm (Linehan et al., 2006).
The above-mentioned efforts represent beginning steps towards a better understanding of change processes in the treatment of BPD. This is preliminary work, and the possibility remains that other factors that have not yet been assessed contribute to both the process components being examined as well as treatment outcome. The issue is made more challenging by the fact that, at this point in time, no single treatment has clearly been found to be superior to any other bona fide psychotherapy for BPD. Do certain treatments and their unique components work better for certain patients and their problems, or are common elements driving therapeutic progress across all treatment models? Each therapy appears to have certain merits and certain limitations. For example, DBT appears to be particularly beneficial regarding self-injury, yet it is also one of the most time- and resource-intensive approaches; TFP has demonstrated success in altering patients’ internalized attachment patterns, yet it is perhaps one of the most conceptually difficult approaches to learn. How is the clinician to choose from among these various treatments, each with promising research support?
One way to address this issue in clinical practice is to consider integrating components from different empirically supported treatments. Doing so may enhance the clinician’s capacity to target particular symptoms or problems with particular interventions, according to the patient’s unique presentation and clinical needs. The combining of different approaches is not without concern, such as the potential threat to treatment coherence (Chapman, Turner, & Dixon-Gordon, 2011). A carefully considered integrated approach may nonetheless increase the likelihood of change occurring along multiple therapeutic pathways. Combining elements from different treatment models can also be implemented according to different phases of the patient’s disorder. Livesley (2005) has proposed a general framework for integrating therapeutic approaches and techniques in a sequenced manner that conceptualizes patients’ personality difficulties in stages, balancing treatment priorities (e.g. safety) with amenability to interventions (e.g. behavioural symptoms vs core traits). Such a perspective might, for example, prioritize the teaching of emotion regulation skills over the interpretation of transference patterns for a patient who frequently uses emergency services to manage her dysphoria. Livesley’s approach also emphasizes the nonspecific components of therapy, or features of treatment that are present across different theoretical models (Livesley, 2005). These are discussed in the common features section of this website.