A voluminous and ever-expanding research literature points to the general effectiveness of psychotherapy (Lambert & Ogles, 2004). Through the use of controlled clinical trials, psychotherapy researchers have identified many empirically-supported treatments for specific clinical phenomena (Roth & Fonagy, 2005). The extant research also suggests that, with just a few exceptions, different therapy modalities yield comparable clinical effects (Lambert & Ogles, 2004). From a glass-half-full perspective, the field has made impressive strides in legitimizing the power of psychosocial interventions. Furthermore, as reflected in the long-standing “Dodo bird” interpretation of the comparative outcome literature, it can be asserted that everybody has indeed won and all should have prizes.
From a glass-half-empty perspective, psychotherapy’s general effectiveness is tempered by its clear limitations. For example, effective treatments for some psychological conditions have yet to be established, and the generalizability to everyday practice of treatments tested in controlled efficacy contexts remains tenuous. Furthermore, even when provided the “gold standard” treatment for a particular condition, some patients fail to respond, only partially respond, or respond but relapse; others drop out of treatment or even deteriorate (Lambert & Ogles, 2004). Thus, it seems that the Dodo verdict can be reconsidered to suggest that all therapies can be improved (Castonguay, Reid, Halperin, & Goldfried, 2003).
Although efforts toward improvement can come in many shapes and sizes, some scholars have argued that improvement may perhaps best be achieved through (a) theoretical humility and openness to the contributions of other (and perhaps historically incompatible or rival) orientations (e.g., Castonguay et al., 2003), and (b) treatment modifications based on process research (e.g., Grawe, 1997). Such approaches preserve the field’s advances in empirically supporting certain treatment packages, while inherently recognizing the complexity of change and the need to move forward creatively in refining treatments to both embrace and address such complexities.
Both of the aforementioned pathways to improvement reflect a specific model of psychotherapy integration (see Norcross & Goldfried, 2005). The former captures assimilative integration, or the attempt to improve an established system of psychotherapy by carefully considering the potential contributions of other systems. The latter captures common factors integration, which focuses on the conceptual and empirical contributions of pantheoretical and pandiagnostic therapeutic ingredients, and the inevitable influence of the momentary and dynamic context on the treatment process.
Among others, process research has persuasively implicated two common factors in the change process – the therapeutic alliance and patient expectations (see Castonguay, Constantino, & Holtforth, 2006; Greenberg, Constantino, & Bruce, 2006). To date, a facet of my research program has been directed at developing, systematizing, and experimentally testing alliance and expectancy-based treatment modules as a means to improve the efficacy of a particular empirically-established treatment (i.e., cognitive therapy; CT) for a specific condition (i.e., adult major depressive disorder; MDD). Although CT is an efficacious treatment for depression, recent estimates from controlled trials suggest that over half of MDD patients do not remit at posttreatment (De Matt, Dekker, Schoevers, & De Jonghe, 2006), while an even higher percentage fail to maintain lasting improvements (Hollon et al., 2005). Thus, there is room for significant improvement in CT for depression, and there remains a pressing need to refine treatments to increase response and decrease relapse. My collaborators and I have embarked on two such efforts. The alliance-based effort focuses on incorporating humanistic and interpersonal alliance-rupture repair strategies into standard CT and, thus, fits the assimilative and common factors models. The expectancy-based effort focuses on incorporating into CT systematic and responsive efforts to foster, manage, and change patients’ treatment expectations, which follows the common factors pathway. Below I describe each research line, including (a) a brief review of process findings that led to the treatment development, (b) a brief overview of the treatment module, and (c) a summary of our preliminary research to date.
Therapeutic Alliance Process Research and Integrative Cognitive Therapy
Pantheoretically defined, the therapeutic alliance reflects the quality of the patient-therapist working collaboration and affective bond (Bordin, 1979). As reflected in our own reviews and process-outcome studies (e.g., Constantino, Arnow, Blasey, & Agras, 2005; Constantino, Castonguay, & Schut, 2002), the alliance is a well-established predictor of treatment success across a variety of psychotherapies and presenting problems. The alliance not only predicts outcome, but it also provides a dynamic context for the implementation and utility of other interventions. For example, in a study of CT for depression, Castonguay, Goldfried, Wiser, Raue, and Hayes (1996) found that strict therapist adherence to prescribed CT techniques in the context of an alliance rupture was negatively related to outcome. Inspired by these findings, and guided by an assimilative integration model, Castonguay (1996) developed Integrative Cognitive Therapy (ICT), which is an approach to depression that remains grounded in CT but systematically incorporates humanistic and interpersonal strategies for identifying, addressing, and repairing emergent alliance ruptures.
Based on the contributions of Burns (1989) and Safran and Segal (1990), ICT presupposes that CT therapists can be more effective in dealing with alliance strains by exploring the source of the difficulty (including their own contributions) rather than increasing their adherence to core CT interventions. In this vein, the ICT manual outlines a 3-step rupture-repair sequence in which the therapist: (1) Invites the patient to discuss his negative reaction to the therapy or therapist; (2) Empathizes with the patient’s feelings and invites additional emotional disclosure in the service of understanding, respecting, and validating the patient’s subjective experience; and (3) Disarms the patient’s antagonism, anger, and/or other negative feelings by acknowledging his or her own contribution to the rupture. Such action promotes a restoration or enhancement of the collaborative working relationship, at which time the therapist then resumes standard CT techniques.
In an initial pilot investigation of ICT, Castonguay et al. (2004) found that ICT produced significantly superior outcomes than a wait-list control condition. As a follow-up, my research team conducted a pilot study (Constantino et al., 2008) to test further ICT’s efficacy and specificity by directly comparing it to standard CT. In this sense, the study employed an additive design, the strength of which resides in its high level of control across the independent treatment variable (e.g., Behar & Borkovec, 2003). With the exception of ICT’s additional rupture-repair strategies, the treatments were delivered according to the same CT protocol, thereby reducing the likelihood that outcome differences are a function of “nonspecific” factors and strengthening the causal inferences that can be made about the alliance manipulations. Furthermore, by using CT-as-usual as a control group, the additive design (a) transcends the specific versus common factors debate by assessing whether rupture-repair interventions work additively or synergistically to improve an already established treatment package, and (b) adds a cause-and-effect dimension to the alliance-outcome link. In our study, using clinicians-in-training, we found preliminary evidence that ICT could be distinguished from CT, and that ICT outperformed CT (small to medium effects) in terms of reducing depressive and global symptomatology. Furthermore, relative to CT, there were fewer dropouts, higher quality alliances, and higher perceived therapist empathy in ICT (medium to large effects). Thus, ICT may be considered a promising limited support treatment (Roth & Fonagy, 1996) worthy of future rigorous study.
Although preliminary, our emerging ICT findings suggest that psychotherapists should not only strive to foster good initial alliances with their patients, but also constantly assess for any deviations in the relationship climate. In the face of potential or actual alliance ruptures, clinicians should resist rigid adherence to the techniques they have been employing (e.g., standard CT interventions) and work through such relationship issues directly, openly, and nondefensively. The use of gentle probing, active listening, empathizing, and disarming may not only help to get the relationship back on track, but such metacommunication strategies may also promote a corrective relational experience (see also Safran & Muran, 2000).
Expectancy Process Research and the Expectancy Enhancement Treatment Module
Patients’ expectations have long been considered a common treatment factor (e.g., Frank, 1961). Outcome expectations refer to a prognostic belief that therapy will help, while process expectations reflect beliefs about what will transpire during therapy. As reflected in our own review and process-outcome studies (e.g., Constantino et al., 2007; Greenberg et al., 2006), expectations have been shown to be important contributors to adaptive during- and post-treatment outcomes. However, the expectation construct has been remarkably undervalued (Weinberger & Eig, 1999). Although many therapies include elements that address patient expectations in some manner, expectancy strategies are often neither explicit nor systematic. Moreover, in experimental treatment studies, expectations have been traditionally viewed as artifacts to be controlled – a perspective that now seems outdated. Thus, we have developed a treatment manual that outlines an explicit and systematic approach to enhancing patient expectations about therapeutic change and the treatment course.
The expectancy enhancement (EE) manual (Constantino, Klein, & Greenberg, 2006) addresses pre- and during-treatment expectations. Specifically, it comprises (a) an initial session EE interview to enhance patients’ outcome expectations and their expectations about the length and nature of treatment, (b) ongoing standard and reactive EE strategies, (c) general relationship strategies to be considered in light of patients’ expectations, and (d) a termination component that aims to enhance patients’ posttreatment expectations for maintaining their treatment gains. The present version of EE was designed as a companion manual to CT for depression. However, we suspect that such strategies can eventually be adapted for a wide range of clinical conditions and for other treatment modalities. We are currently conducting another pilot study, again utilizing an additive design, to foster the development of the EE manual and to test preliminarily its causative enhancement of standard CT. Although the outcome data are still forthcoming, we have been able to successfully train graduate trainees on the protocol, and they have been able to implement the treatment with good fidelity and competence.
Despite not yet having data on the specific efficacy of our EE manual, psychotherapy process research points consistently to the clinical importance of patients’ process and outcome expectations. Moreover, most psychotherapies involve some level of manipulation, exploration, challenge, and/or revision of patient expectations (Greenberg et al., 2006). Thus, clinicians should carefully assess patients’ expectations at the beginning of treatment in order to inform prognosis, case formulation, and treatment-planning. Regarding process expectations, clinicians may need to spend time socializing patients to the treatment process (e.g., typical role behaviors; duration), as well as checking in on patients’ met and unmet expectations as the therapy work unfolds. Regarding outcome expectations, clinicians should make a concerted effort to offer personalized hope-inspiring statements (e.g., “Your problems are exactly the type for which this therapy can be of assistance”) at the treatment’s outset, and to respond appropriately to hope-diminishing moments with both alliance-based sensitivity and expectation-enhancing strategies (e.g., reminding patients of depression’s recurrent nature; drawing on past successes) (Constantino et al., 2006).
The lines of research discussed above are representative of my overarching research program that focuses on understanding patient, therapist, and relational processes that influence the course and outcome of psychosocial treatments, and on the development and systematization of therapeutic interventions that address pantheoretical principles of clinical change. The overarching aim of the program is the development of empirically-grounded skills on which therapists can be trained to negotiate effective therapeutic relationships and to enhance patients’ treatment expectations. This focus on two key common factors adds a much-needed complement to the testing and training of theory-specific treatment techniques that have, to date, received much more empirical attention. Of course, the efficacy findings discussed above should be interpreted within their preliminary spirit. However, our hope is that when the jury returns, we will have uncovered two promising common factor treatment modalities that can be assimilated into CT to augment its effectiveness. If so, our work will have helped substantiate a glass-not-yet-full-but-still-full-of-promise perspective on psychotherapy outcome research.