Current Research

Our research focuses broadly on patient, therapist, relational, contextual, and measurement-based characteristics/processes that influence and/or result from psychosocial treatments, and on the development and refinement of therapeutic interventions that address pantheoretical principles of clinical change and augment (or replace) traditional treatment packages. We are keenly interested in developing empirically grounded skills on which psychotherapists can be trained to negotiate effective therapeutic relationships, influence patients' psychotherapy-related expectations, respond to patients' change ambivalence and self-strivings, foster corrective experiences, and generally facilitate adaptive therapy outcomes. Moreover, in growing acknowledgment of the important role of the provider in affecting therapy outcomes, we are also very interested in understanding differences between therapists in their general effectiveness, as well as the relative strengths and weaknesses that therapists possess when treating different types of mental health problems, treating patients with different salient identities, and/or using different types of therapy processes/interventions within their own caseloads. Understanding such between- and within-therapist effects, respectively, can help therapists capitalize on their strengths and redress their relative weaknesses. This knowledge can also enhance personalized, patient-centered care and decision making. 

Although various treatments for various psychological conditions have, on average, achieved some empirical support, response and relapse rates generally remain sobering. Furthermore, there remains a relatively limited understanding of how therapy works or why it fails do to so in some instances, for some people. The field also knows very little at this stage about what accounts for some therapists being generally more effective than others, some therapists treating certain mental health conditions effectively and other conditions ineffectively, some therapists treating patients with certain identities effectively and other identities ineffectively, and some therapists using certain treatment processes to better effect than other processes. All told, there is ample room for improvement in understanding, refining, and delivering all forms of therapy, as well as understanding, leveraging, and addressing therapist effects, and we adopt multiple research designs to contribute to the research base in these important areas. These designs include quantitative process research, randomized controlled treatment trials, quantitative meta-analyses, qualitative research, community-based participatory studies, and lab-based clinical analogue studies.

Psychotherapy process research attempts to understand factors that promote or hinder therapeutic change. This type of work is a central focus in our lab in the context of both controlled clinical trials and naturalistic, community-based care. For example, we study the prediction of posttreatment outcomes in various treatments (e.g., interpersonal psychotherapy [IPT], cognitive behavioral therapy [CBT], integrative therapy) for various conditions (e.g., depression, generalized anxiety disorder [GAD]) from patient characteristics/actions (e.g., baseline clinical or demographic characteristics, treatment beliefs, in-session interpersonal behaviors), therapist characteristics/actions (e.g., personal characteristics, in-session interpersonal behaviors), and dyadic processes (e.g., relational convergence, process attunement), measured both statically and dynamically over time. Importantly, we test these predictor effects at both the between-patient (within-therapist) and between-therapist levels. We also parse process variables into their coexisting parts of within-patient, “state"-like fluctuations over treatment (which may reflect improvement as a function of therapy) and between-patient, “trait"-like differences in patients’ average process levels (which may reflect the contributions of preexisting abilities to therapeutic gain).

We also study patient, therapist, and dyadic characteristics as correlates of key baseline or in-session variables that have been shown to relate to adaptive treatment outcomes (e.g., alliance quality, patient outcome expectation, therapist credibility). We are also interested in understanding therapist-level factors (e.g., humility, facilitative interpersonal skill) that differentiate clinicians whose patients reliably achieve better versus worse treatment outcomes, as well as those that may explain (e.g., theoretical orientation, personality characteristics) why some therapists use certain treatment processes (e.g., alliance quality, theory-specific interventions) to greater therapeutic effect than other processes. Also related to therapist performance, we examine therapists' self-perceived strengths and weaknesses, and how these perceptions relate to actual strengths and weaknesses (as determined by patient-reported, multidimensional routine outcome measures).

Further, we study potential mechanisms of the relations between in-session processes and treatment outcome; that is, the pathways through which process variables (e.g., alliance quality, patient outcome expectation) relate to posttreatment outcome. We also examine moderators of process-outcome correlations, again at multiple levels of analysis, as well as moderators (e.g., patient baseline aptitudes/traits) and mediators (e.g., in-session transactions) of comparative treatment effects on outcomes.

Our lab's interest in psychotherapy process and outcome has also led to more fine-grained, micro-analysis of patient-therapist interactions. One vehicle for such work is the application of the Structural Analysis of Social Behavior (SASB) to therapy transcripts and session audio/video. The SASB is a powerful lens for capturing specific helpful or hindering processes in the patient-therapist exchange. For example, we study dyadic processes on a moment-to-moment basis that differentiate known patient groups (e.g., good vs. poor outcome, high vs. low alliance convergence, more vs. less adaptive process following a key therapeutic event). 

One effective way to improve patient outcomes is to modify established psychotherapies based on process research findings. Consistent with this approach, our lab is involved in several lines of outcome research that are heavily informed by our own process findings that implicate the importance of responsively addressing (with pointed and timely interventions) clinical markers of patient resistance/change ambivalence, patient treatment expectations, patient-therapist alliance ruptures, and feedback from routine outcomes monitoring—a type of doing the evidence-based "right thing at the right time." Given that many studies related to these constructs are correlational, there is a pressing need for experimental designs that can better determine causal connections. Thus, our research uses randomized clinical trials to test explicit procedures for addressing patient ambivalence/resistance, cultivating patients' expectations about the process and outcome of treatment, repairing alliance ruptures, and using patient outcomes data (analyzed at the therapist level) to inform responsive treatment assignment—all as augmentations to treatment/assignment as usual.

When a research base matures, meta-analysis can be a powerful analytic tool for determining aggregate effects, heterogeneity of such effects, and moderators that might explain such heterogeneity. In this vein, we use meta-analyses to determine the aggregated effect of patient (e.g., outcome expectation, perceptions of treatment credibility) and process (e.g., alliance quality) variables on posttreatment outcomes. We also conduct meta-mediation analyses to help determine mechanisms of such associations.

Our lab also uses qualitative methods (e.g., consensual qualitative research, grounded theory, interpersonal process recall) to better understand how patients experience and understand the therapeutic change process, and what they value in their mental health care. We also qualitatively assess therapists' perspectives on effective (and ineffective) therapy practices, and how therapists make use (or not) of the extant psychotherapy research base.

Our work also engages various mental health care stakeholders to examine their experiences, needs, and values with respect to treatment seeking and decision making. For example, we examine people of color's experiences of mental health care (from treatment seeking to engagement), as well as their relative valuing of using different aspects of provider information (e.g., racial/ethnic identity, experience, historical effectiveness track records) to improve their treatment access and quality (and thereby reduce longstanding disparities in these domains). We also examine practicing clinicians' attitudes toward receiving performance feedback and using such personal effectiveness data to inform their future practice and training directions.

We use lab-based analogue designs to study correlates of empirically supported common treatment factors (e.g., patient preferences, patient expectations, patient perceptions of therapist empathy) and develop and experimentally test strategies for cultivating these factors (e.g., expectancy persuasion strategies).

In addition to our general interests in studying psychotherapy change processes and outcomes, we are developing and testing components of two conceptual psychotherapy models. The first is an expectancy-based approach to facilitating corrective experiences. This integrative approach privileges various patient expectancies as pantheoretical mechanisms of corrective experiences. The approach draws on the social psychological and interpersonal principles of positivity, verification, social influence, complementarity, and metacommunication. As we have noted, the model proposes that "through an evolving sequence of three primary therapist skills that are corrective (i.e., alliance development and negotiation, disconfirmation of the patient's restricted self, treading lightly in the course of change), patients will have one or more of three novel and unexpected experiences (i.e., experiential relearning of interpersonal exchange, incorporation of novel and previously unexpected information about self, and revision of self- and other-related emotion schemes). Through the interaction of the therapist's corrective skills and the patient's new experiences, the patient will develop new and more adaptive recapitulations and introjections (i.e., this is what gets corrected)" (Constantino & Westra, 2012). We have produced data to support various aspects of this model, and we are committed to further testing the model's components and its entirety.

The second model represents a form of psychotherapy integration, which we have termed context-responsive psychotherapy integration (CRPI). This paradigm proposes an if-then structure for therapists to respond to cues, or markers, in the therapy process or in the patient's psychological or biological makeup with context-relevant, empirically supported therapeutic strategies. The framework draws on both theory-specific and theory-common treatment factors. As we have noted, "This approach would move beyond viewing psychotherapy integration as some variation on blending treatment models and their respective techniques, to a fully context-responsive treatment approach. This contextualized model should center on systematized, flexibly manualized, and empirically tested modules for addressing specific psychotherapy process themes or scenarios ... This nomothetic approach can then be complemented by more idiographic principles derived from an understanding of basic research in personality and psychopathology (and other related areas) and how it applies to one's current patient and his or her traits, motivations, self-concept, affects, and so forth" (Constantino et al., 2013). To us, treatment could begin from any distinct approach/orientation and then shift into, and back out of, specific modular strategies based on readily available, empirically derived shift (or "departure") markers. To date, we have proposed multiple candidate markers of treatment processes or patient characteristics toward which our research has been directed. These include: patient change ambivalence/resistance, patient waning outcome expectation, patient measurement-based risk for deterioration, patient need for self-verification, patient-therapist alliance rupture, and therapist cultural misattunement. We have also expanded the model to contextually responding to patients prior to therapy, at the time of intake. Namely, patients can be assigned to therapists who have a track record of treating patients with similar identities and/or who present with similar mental health concerns. We have accumulated ample data to support various aspects of this CRPI framework, and we are committed to further testing the model's components and its entirety. Such work will inform its appropriate evolution over time.