Introduction
Drop-out from psychotherapy is a significant concern for therapists and researchers alike. By dropping out of treatment early, clients may not have time to learn and practise the therapeutic strategies posited to promote successful outcomes. To better understand and ultimately reduce premature drop-out, it is important to identify factors that can alert therapists to those at greater risk early in treatment. One such factor, validation, communicates to clients the ways in which their behaviours, thoughts, and/or feelings make sense and are understandable given the current situation or their personal history (Linehan, Reference Linehan1993). A therapeutic environment that communicates acceptance and understanding of the client’s experiences is likely to facilitate client retention (Linehan, Reference Linehan, Bohart and Greenberg1997). Clients who believe they are understood and validated by their therapists may be more inclined to continue with treatment, even when sessions are difficult.
With this study, we examined the role of therapist validation early in treatment as a predictor of drop-out in three cognitive behavioural graduate training clinics. Using a longitudinal design, we assessed the predictive value of client- and therapist-rated validation in treatment drop-out. We hypothesized that more positive ratings of therapist validation would reduce the risk of drop-out above and beyond client baseline symptom severity. Additionally, we hypothesized that client ratings would be better predictors than therapist ratings.
Method
Participants
Clients
The sample was composed of 55 client–therapist dyads from three graduate training clinics at a large Midwestern university: a general cognitive behavioural clinic (gCBT), an anxiety and stress disorders clinic (ASDC), and a dialectical behaviour therapy clinic (DBT). From this sample, 40% (n = 22) were identified as drop-outs. However, due to missing data, the final sample for all analyses was limited to 50 clients (gCBT: n =14; ASDC: n = 20; DBT: n = 16), with a drop-out rate of 36%. Clients were primarily female (80.0%) and Caucasian (82.0%) with a mean age of 27.7 years (SD = 10.6).
Therapists
Therapists (n = 22) were doctoral level trainees with a range of experience from 0 to 986 clinical hours (mean = 291.5, SD = 317.7).Footnote 1 First year trainees were all supervised in the gCBT clinic (n = 9), and trainees with one or more years of experience were supervised in the ASDC (n = 9) or DBT (n = 4) clinics. Therapists were also primarily female (68.2%) and Caucasian (86.4%) with a mean age of 27.0 years (SD = 2.5). On average, each therapist treated 2.27 clients, with a mode of one client per therapist.
Treatment clinics
All three clinics are located within the in-house psychological services centre at a large Midwestern university and are intended to facilitate training in cognitive behavioural treatment strategies. All therapists received supervision in both individual and group formats on a weekly basis by a licensed clinical psychologist, each with extensive experience supervising doctoral students. Clients in all three clinics were university students or members of the community.
Measures
Symptom measures
Depressive symptoms were measured with the Beck Depression Inventory-2nd edition (BDI-II; Beck et al., Reference Beck, Steer and Brown1996). Anxiety symptoms were measured with the Beck Anxiety Inventory (BAI; Beck et al., Reference Beck, Epstein, Brown and Steer1988). Borderline personality disorder features were measured with the Personality Assessment Inventory-Borderline Features Scale (PAI-BOR; Morey, Reference Morey1991).
Therapist validation and invalidation
The Self-Reported Validation and Invalidation Scale (SRVIS) is a 9-item scale designed to assess perceived levels of validation and invalidation. It was created by this research team for a series of studies on the role of validation in both experimental and clinical trial contexts. Subscales of the SRVIS (i.e. validation and invalidation) have demonstrated high internal consistency in studies in which validation and invalidation were experimentally induced (α = .89 and .86, respectively). In the clinical trial version, each item assesses different validating or invalidating behaviours by the therapist during a given session and is rated from 0 (never) to 4 (almost always/always). The SRVIS includes items such as the therapist’s level of attention towards the client, responsiveness to the client’s emotions, pathologizing the client’s responses, and fragilizing the client. High scores on the SRVIS represent greater levels of validation, while low scores represent less validation and greater invalidation. In this study, both clients and therapists reported on their perceptions of in-session validation by the therapist. We calculated between-person composite reliability estimates using multi-level confirmatory factor analysis (Geldhof et al., Reference Geldhof, Preacher and Zyphur2014) to account for the nesting of sessions within clients. Reliability was good for both client-rated (ω = .88) and therapist-rated validation (ω = .87).
Procedure
Clients completed baseline measures prior to starting treatment, including a demographics questionnaire and PAI-BOR. BAI and BDI-II session scores were taken from the first available time point for each client as a baseline measure of symptom severity. Therapists also provided information relevant to demographics and prior clinical experience. Clients and therapists completed the SRVIS after four early treatment sessions (ranging from 3 to 7).
Therapy completion was defined as clients who reached a mutually agreed upon termination in collaboration with their therapist or attended a full 6-month treatment course without any significant gaps (defined as 4 weeks or more). Therapy drop-out was defined as clients who unilaterally discontinued treatment before the treatment protocol was completed/without therapist agreement or missed at least four consecutive weeks of treatment within the first 6 months.
Analytic strategy
First, drop-outs and completers of therapy were compared on client and therapist demographic characteristics and client symptoms. Differences in drop-out rates were also examined between clinics. We planned to enter symptom measures (i.e. BAI, BDI-II and PAI-BOR) as covariates due to their likely influence on drop-out and any demographic variables that distinguished drop-outs and completers at the p < .10 level. To represent general patterns of validation, scores for both client- and therapist-rated validation across the four therapy sessions were averaged. All covariates and predictor variables were mean centred to ease interpretability of the intercepts. Using binary logistic regression, individual models were run for both client- and therapist-rated validation. We attempted to account for within-therapist correlation in drop-out using multi-level modelling. However, most therapists only treated one client and, as such, using these methods produced an uninterpretable model and the within-therapist effects were estimated as 0. Therefore, all analyses were run without accounting for nesting within therapists. To help account for the variation in therapists’ hours of experience, we ran individual models examining the interaction between therapist experience and both client- and therapist-rated validation.
Results
The overall rate of drop-out across all clinics was 36% (n = 18), with no differences in drop-out rates between clinics (χ2 (2) = .03, p = .99). There were no significant differences between drop-outs and completers in any client or therapist demographic variable or in measures of symptom severity; however, symptoms were entered as covariates into all models as planned.Footnote 2
Client-rated validation significantly predicted drop-out. With each unit increase in therapist validation, as reported by clients on average across these four sessions, the likelihood of dropping out of treatment decreased by 22% (B = –.25, SE = .12, OR = .78, p = .03). Therapists’ perceptions of their own use of validation did not significantly predict treatment dropout, p = .76. Baseline client symptom measures were not significant predictors in either model, p > .38.
There was a significant interaction between client-rated validation and therapist experience (i.e. number of clinical hours) (B = .001, SE = .0004, p = .04). We probed the interaction at the mean and one standard deviation above and below the mean on therapist experience (Fig. 1). Greater client-rated validation significantly reduced the risk for drop-out only for clients whose therapist had average (B = –.37, SE = .16, p = .02) or below average (B = –.63, SE = .23, p = .006) experience for this sample. However, client-rated validation was unrelated to drop-out for clients who had a therapist with above average experience (B = –.08, SE = .18, p = .66). The interaction between therapist-rated validation and therapist experience was not significant (p = .13). Baseline client symptoms remained non-significant in both models (p > .40). We also examined these models controlling for clinic because therapists in the gCBT clinic had less experience. All patterns of results remained the same. Clinic was not a significant predictor in either model (p > .48), nor did clinic moderate the relationship between client- or therapist-rated validation and drop-out (p > .22), suggesting that the interaction between client-rated validation and therapist experience in predicting drop-out is unrelated to differences between clinics. For further details, please see the extended report available in the Supplementary Material.
Figure 1. Moderation analysis between client-rated validation and therapist experience in hours predicting risk of drop-out.
Discussion
With this study, we aimed to investigate the relationship between validation from clients’ and therapists’ perspectives and drop-out in graduate training clinics. These results suggest that therapist validation, when rated from the perspective of the client in early sessions, predicted treatment drop-out above and beyond symptom severity. Finding ways to make clients feel more validated in therapy may help keep clients in treatment for the recommended length of time, which may have a positive impact on client outcomes. Therapists attending only to their own perceptions, without assessment of client views, may be unaware of clients at greater risk of drop-out. Furthermore, therapist experience moderated the relation between client-rated validation and drop-out such that validation no longer predicted drop-out for therapists with more experience. It is possible that validation is less important to drop-out for more experienced therapists because they are providing higher quality therapy (e.g. providing more specific strategies to help clients ameliorate their problems or symptoms). This would be a question for future research.
In this study, clients with more severe symptom presentations were no more likely to drop out than those with less severe symptom presentations; rather, it was validation, as assessed by the client, that predicted drop-out, which may have important implications for other clinical outcomes. Client ratings of therapist validation can be a simple method for alerting therapists early in treatment to risk of premature or unilateral drop-out. Replicating these findings in larger samples and with more experienced therapists and understanding ways to improve clients’ perceptions of validation would be important areas of future study.
Author ORCID
Kristen P. Howard 0000-0001-6711-633X
Acknowledgements
We would like to thank Ciara Stigen for her efforts in the development and data collection of this study. We would also like to thank all of the graduate student therapists and patients involved in the completion of this research.
Ethical statement
The authors of this study have abided by the Ethical Principles of Psychologists and Code of Conduct as set out by the APA. This study was approved by the Institutional Review Board of the Ohio State University (2011B0254).
Conflicts of interest
Kristen Howard, Erin Altenburger and Jennifer Cheavens have no conflicts of interest with respect to this publication.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1352465819000420
Comments
No Comments have been published for this article.