Effect of Therapist Self Disclosure on Therapeutic Alliance and Drop Out

The following research was conducted and presented by Dr. Reeh
at the Canadian Psychological Association’s
annual conference in 2011, Toronto, Canada

Abstract
This study investigated whether or not therapist self-disclosure (TSD) moderated the relationship between perceived therapeutic alliance (TA) and dropout from residential substance abuse treatment. The research participants were 132 men attending residential substance abuse treatment. Four groups of men were randomly assigned to 1 of 4 experimental conditions. One of 2 statements was read to the participants that described the TA as weak or strong. Afterwards, participants watched 1 of 2 DVDs depicting 2 males role-playing an intake session. One DVD included TSD. The TSD was edited out of the other DVD. After watching the DVD, participants responded to the question of whether or not they would continue treatment with the depicted therapist. Results from hierarchical logistic regression indicated that weak TA was a good predictor of dropout. Therapist Self Disclosure did not significantly affect research participants’ opinions regarding whether or not they would continue treatment with the depicted therapist. Therefore, results from the present study do not support views that TSD should be used or avoided.

Does Therapist Self-Disclosure Moderate the Relationship
Between the Therapeutic Alliance and Dropout?

The average dropout rate from residential substance abuse treatment is approximately 50% (Sayre et al., 2002; Wierzbicki & Pekarik, 1993), which is a significant concern given that dropout is associated with poorer outcomes. For example, people who discontinued their attendance of 12-step meetings were three times more likely to abuse substances than those that continued (Kelly & Moos, 2003). People on probation who terminated their substance abuse treatment early were more likely to be rearrested within a four year period than people who completed treatment (Huebner & Cobbina, 2007). More specifically, Huebner and Cobbina (2007) found that 67% of people who dropped out and 37% of people who completed treatment were charged with new crimes. Therefore, finding factors that contribute to dropout so that dropout can be averted is of paramount importance given that people who complete treatment are more likely to become or to return to being healthy contributing members of society.
Researchers have generally found that building a strong therapeutic alliance is one factor that tends to assist in averting dropout (Horvath, 2001; Johansson & Eklund, 2006; Meier, Barrowclough, & Donmall, 2005, Saatsi, Hardy, & Cahill, 2007). For example, after reviewing 11 studies, Meier et al.’s (2005) key conclusion was that early therapeutic alliance consistently predicted engagement and retention in drug abuse treatment. They found moderate effect sizes that explained about 5% to 15% of the variance between therapeutic alliance and dropout.
Horvath (2001) found conflicting results in his meta-analysis of therapeutic alliance and dropout. Four of the studies he reviewed found that a weak therapeutic alliance at intake or after the first session was a good predictor of dropout. More recent studies, including a meta-analysis by Meier et al. (2005) have supported Horvath’s results that people who drop out of therapy tend to rate the therapeutic alliance as weaker than people who complete therapy (Johansson & Eklund, 2006; Saatsi, Hardy, & Cahill, 2007). Horvath also found two studies (Florsheim et al., 2000; Joyce & Piper, 1998) in which a high initial therapeutic alliance was related to dropout. Horvath speculated that research participants in these two studies may have had unrealistic and unfilled expectations. Other researchers found that therapeutic alliance did not predict retention or dropout (Barber et al., 2001, Brocato & Wagner, 2008; Sauer, Lopez, & Gormley, 2003).
One specific way that therapists have sought to develop the therapeutic alliance is through therapist self-disclosure (TSD). Most of the researchers that studied therapeutic alliance and TSD found that TSD enhanced the therapeutic alliance (e.g., Audet, 2004; Bedi, Davis, & Arvay, 2005; Burkard, Knox, Groen, & Perez, 2006; Hanson, 2005). One pair of researchers found that TSD and the therapeutic alliance were unrelated (Kelly & Rodriguez, 2007). Myers (2004) found an interaction between TSD and the therapeutic alliance. When Myers’s research participants rated a session in which a mock therapist disclosed personal information in the context of a weak alliance, they rated the therapist less favorably. The converse also occurred: When the therapist disclosed personal information with a client in which there was a strong alliance depicted, clients rated the therapist more favorably (Myers, 2004).
A specific type of TSD that is commonly used in residential substance abuse treatment centers is TSD that involves substance abuse counselors disclosing to their clients that they have struggled with addictions (Personal Communication, J. Lighterwood, January, 2010). Although this disclosure is common, there was no research to date that specifically addressed whether or not addictions therapists’ disclosures that they struggled with addiction affected their clients’ decisions to drop out of treatment. On the one hand, clients may believe that therapists who have experienced problems with addiction may be better able to understand them and thus better able to experience empathy. Therefore, clients may be more likely to complete treatment with such therapists. On the other hand, clients may believe that therapists who have experienced problems with addiction are less able to help them because clients may believe that these counselors have issues of their own to resolve. In this case, clients may be more likely to drop out of therapy with this type of therapist. Not having direction from empirical sources is therefore a problem because substance abuse counsellors are left to wonder what the best course of action is concerning whether disclosing addictions problems to clients can affect the relationship between the therapeutic alliance and dropout.
There are two main reasons why conducting naturalistic studies of the therapeutic alliance and TSD (specifically, that the counsellor has struggled with addiction issues) in residential substance abuse facilities is difficult. Firstly, measuring the therapeutic alliance as it occurs in therapy is usually difficult because most clients who have not dropped out usually rate the therapeutic alliance with the top two most favourable ratings, thus creating lack of variance (Horvath, 2001). Secondly, clients of residential substance abuse tend to have contact with several treatment personnel (some of which may have disclosed addiction struggles and some of which may not have disclosed or have had addiction problems) and tend to have heterogeneous reasons for departing from such a facility. Hence, if a person drops out, it is difficult to determine the reasons for doing so. Therefore, the most practical way to study whether or not TSD moderates the relationship between the therapeutic alliance and dropout is to study it using an analogue design in which TSD, therapeutic alliance, and dropout can be isolated and measured.
In their review of the literature, Hill and Knox (2001) noted that the majority of the analogue studies reported that clients rated therapists positively when therapists disclosed moderately personal information. In general, research participants in these studies reported that TSD was helpful and that it facilitated greater involvement of their emotions (Hill & Knox, 2001).
There were only two studies (Barrett & Berman, 2001; Kelly & Rodriguez, 2007) available in which researchers either manipulated or observed the effect of different levels of TSD in actual therapy sessions. Barrett and Berman (2001) instructed one group of doctoral student therapists to provide TSD that was at an intimacy level similar to that of their clients. For instance, if a client disclosed relationship difficulties, the therapist was instructed to disclose similar experiences. The other group of doctoral students was instructed to not disclose any personal information. Instead, they were instructed to reflect requests for personal information back to the client. For example, they could explore the reasons why the client might want to know personal information about the therapist. Barrett and Berman found that clients whose therapists used TSD reported lower levels of symptom distress. They also found that clients liked the disclosing therapists more than those that did not disclose.
Kelly and Rodriguez (2007) surveyed psychiatric hospital patients. They found that TSD and symptom reduction were unrelated. They suggested that one particular intervention, such as TSD, was not likely to cause measurable change in therapy outcome.
Meier et al. (2006) studied therapist variables that affected therapeutic alliance in a large and unique sample of mostly male, unemployed, antisocial people who injected heroin daily. Meier et al. found that these clients rated the therapeutic alliance with counselors that had addiction issues as stronger than with counselors who did not have addiction issues. Unfortunately, it is not known whether or not the counselors that had addiction issues disclosed this information to their clients. Meier et al. also found that this group of clients rated the therapeutic alliance with more experienced substance abuse counselors and counselors with formal training less positively than with less experienced and less trained counselors.
The purpose of the present study was to provide clarity for substance abuse treatment providers by exploring whether or not the use of TSD (the therapist disclosing that he struggled with addiction and sought treatment) moderated the relationship between the perceived therapeutic alliance and dropout expectancy from substance abuse treatment.
Method
Setting and Sample
Research participants from three 90-day residential substance abuse treatment facilities near the West Coast of British Columbia, Canada participated in this study. These facilities provided individual and group therapy using cognitive behavioral and emotion-focused therapy for adult males.
Using a power analysis, it was determined that a total sample size of at least 128 people (32 people in each of four conditions) would be needed to reach a .80 level of power with an alpha level of .05, with three degrees of freedom in the regression equation (variables: TSD, therapeutic alliance, and cross product) and an expected medium-sized interaction effect (e.g., partial r-square value of .06, equivalent to f2 = .064) based on previous research (e.g., Horvath, 2001).
To be eligible for this study, participants needed to meet the treatment centers’ admission criteria, which were that they were males older than 18 years with a severe addiction who had detoxified for at least 72 hours, and who were not psychotic or severely mentally challenged. Also, participants needed to be willing to participate without any external motivation.
Materials
Introductory Statements
Two introductory statements were used in this study. Both statements described a client’s first impressions of a therapist after an intake interview. One statement described the client’s first impressions of the therapist in a way that described a weak therapeutic alliance. More specifically, the statement explained that the client thought that the client’s and the therapist’s main goal of the interview was different. In addition, the statement depicted the client as not liking, trusting, or respecting the therapist. The other introductory statement described a strong therapeutic alliance. For example, the statement explained that the client thought that the client’s and the therapist’s main goal was the same and that the client liked, trusted, and respected the therapist.
Intake Interview DVDs
The stimulus materials for this research were two DVDs that depicted a section of an intake interview. The two DVDs differed in only one way. In one DVD, the therapist disclosed that he had struggled with alcohol addiction and that he sought treatment for it. The second DVD was the same as the first one with the disclosure edited out of it. The DVDs were approximately 5 minutes long. The DVDs involved two European Canadian males (ages 49 and 46) role-playing a scripted intake interview. The male actor, who role-played the therapist, had worked as a therapist for over twenty years.
Post-DVD Questionnaires
Participants were asked to complete a post-DVD questionnaire that contained a question that formed the dependent measure. The question was whether or not the participant would continue treatment with the therapist depicted in the DVD if he was the client in this session. The questionnaire also contained two items that served as manipulation checks. The questions inquired about whether or not the therapist and the client had a good bond and whether or not the therapist disclosed anything about himself. This study also included a demographic questionnaire that posed questions regarding age, income, living arrangements, and so on.
Procedure
The researcher sent flyers to invite men who were attending residential substance abuse to participate in a study about dropout from residential substance abuse facilities. The researcher also went to group rooms after a session to talk about the study and answer questions. Interested clients were asked to meet the researcher in another room (thus reducing a potential feeling of coercion). When the groups of men arrived in the alternate room, they were randomly assigned to one of four experimental conditions. The conditions were: (a) weak therapeutic alliance, no TSD, (b) weak therapeutic alliance and TSD, (c). strong therapeutic alliance, no TSD, and (d) strong therapeutic alliance and TSD.
After answering questions and going through informed consent procedures, the researcher introduced the DVDs by either reading the statement that depicted a strong or a weak alliance. The research participants watched one of two DVDs (with or without TSD). Afterwards, participants completed the post DVD questionnaires.
Results
Sample Description
A total of 171 men from three residential substance abuse treatment centers near the West Coast of Canada were invited to participate in this study. Of these 171 men, 132 men (77%) volunteered and consented to participate in this study. The men who participated in this study were between the ages of 19 and 64 years with an average age of approximately 35. Just under 13% of the research participants were of Aboriginal descent.
About 86% of respondents were not attending treatment due to legal sanctions whereas 4.5% of them were on parole, 3% were on probation, 2.3% were on Statutory Release, and 2.3% were on bail. About 75% of the research participants had never quit residential substance abuse treatment, whereas 13.6% had quit once before, 3.8% quit twice before, 2.3% quit three times before, 2.3% had quit four times before, and 3% quit five times or more.
Predictors of Dropout
The data for the criterion variable of the hierarchical logistic regression analysis consisted of yes/no answers to one question on the post-DVD questionnaire. The data were entered and analyzed using SPSS Version 18. In general, 45.5% of respondents indicated that they would want to continue therapy with the therapist depicted in the DVD, while 54.5% indicated they did not. About 75% of the research participants who were told that the therapist and the client’s bond was weak or strong indicated that they thought the client-therapist bond was as described, whereas 25% described it in the opposite direction than had been described. Also, 2.7% incorrectly indicated that the therapist disclosed something personal about himself when he did not.
Correlations between dropout, TSD, and therapeutic alliance were first conducted. A Spearman Rho correlation test revealed that therapeutic alliance and dropout were significantly and strongly related r(132) = .47, p < .001. Thus, 22% of the variance in dropout was explained by the therapeutic alliance. Correlations between dropout and TSD, and between therapeutic alliance and TSD were not significant.
Table 1 summarizes the results of a hierarchical logistic regression that predicted dropout using TSD, therapeutic alliance, and the product of TSD and therapeutic alliance as predictors: χ2(3, N = 132) = 33.902, p < .001, Nagelkerke R2 = .303, with classification results of 72.2% no, 75.0% yes, 73.5% overall. Therapeutic alliance was strongly related to dropout with a weak therapeutic alliance predicting dropout. TSD was not a significant predictor in this model nor was the product between TSD and the therapeutic alliance.

Table 1

Demographic Information: Full Sample (N = 132)

Predictor
B
Wald2
P
OR
95% CI
Constant
-1.232
8.228
.004
.292

Therapeutic Alliance
1.558
7.653
.006
4.747
[0.444, 2.671]
TSD
-.021
.001
.972
.980
[-1.184, 1.142]
TSD x TA
1.045
1.661
.197
2.844
[-0.560, 2.650]
Note. OR = odds ratio; CI = confidence interval; Omnibus χ2(3, N = 132) = 33.902, p < .001; Hosmer and Lemeshow Goodness of Fit χ2(2, N = 132) = 000, p = 1.000; Cox and Snell R2 = .227; Nagelkerke R2 = .303; Classification results: 72.2% no, 75.0% yes, 73.5% overall.

Logistic Regression Analysis of TA and TSD on Dropout Expectancy (N = 132)

Predictor
B
Wald2
P
OR
95% CI
Constant
-1.232
8.228
.004
.292

Therapeutic Alliance
1.558
7.653
.006
4.747
[0.444, 2.671]
TSD
-.021
.001
.972
.980
[-1.184, 1.142]
TSD x TA
1.045
1.661
.197
2.844
[-0.560, 2.650]
Note. OR = odds ratio; CI = confidence interval; Omnibus χ2(3, N = 132) = 33.902, p < .001; Hosmer and Lemeshow Goodness of Fit χ2(2, N = 132) = 000, p = 1.000; Cox and Snell R2 = .227; Nagelkerke R2 = .303; Classification results: 72.2% no, 75.0% yes, 73.5% overall.

Discussion
The main finding of this study was that TSD did not significantly affect research participants’ opinions regarding whether or not they would continue treatment with the depicted therapist regardless of whether the therapeutic alliance was described as weak or strong. The other finding was that a perceived weak therapeutic alliance was found to be a good predictor of dropout from residential substance abuse treatment.
Results from the present study were not congruent with Myers (2004) who found that when the therapeutic alliance was depicted as weak and when the therapist disclosed very personal information, the participants rated the therapist less favorably than when the therapeutic alliance was described as strong and the therapist disclosed the same very personal information. However, Myers’s very personal information included disclosure that the therapist was abused and that the therapist still struggled with sequelae of his abuse. In the present study, the therapist disclosed that he struggled with addiction issues and that he attended treatment and resolved personal issues. The disparate results between these two studies may have occurred due to the dissimilarity of these disclosures as well as that both studies had dissimilar dependent measures. Myers’s research participants in the weak therapeutic alliance condition may have been negatively impacted by the information that the therapist had current struggles.
It is also interesting to note that Myers (2004) excluded data from research participants who incorrectly answered the manipulation check questions. For instance, if participants misperceived a weak alliance as a strong alliance or a strong alliance as a weak one, Myers excluded their data from the analyses. In the present study, three research participants answered the manipulation question regarding TSD incorrectly. They indicated that the therapist disclosed something personal about himself when they were in the condition that did not include TSD. In addition, 25% of the research participants incorrectly answered the manipulation check question regarding the therapeutic alliance. About 7% stated that the alliance was weak after they participated in the condition in which the alliance was described as strong. The other 18% indicated that the alliance was strong despite being told that the alliance was weak. Data were analyzed with and without the 27.3% of participants who incorrectly answered the manipulation check questions and the results did not differ between the groups. Therefore, the results described in the present study were from the whole sample.
Hanson (2005) found in her qualitative study that TSD fostered the alliance however in the present study, TSD had no significant impact. These differences may have occurred due to the different methods and populations used in these studies. Hanson conducted her study by asking open-ended questions to mostly female European American therapy clients. The present study involved an analogue design with mostly European Canadian men who were struggling with severe addictions. Although the men who participated in the present study may have also felt that TSD enhanced the therapeutic alliance, the present study did not capture this possible effect. Alternatively, given the severe problems and lifestyles of these men, TSD may not have affected the present study’s research participants as greatly as did the research participants in Hanson’s study.
Results from the present study, that therapeutic alliance predicts dropout, are congruent with Meier et al.’s (2005) literature review. Meier et al. also found that the early therapeutic alliance consistently predicted retention in drug abuse treatment. In addition, consistent with the present study, Horvath (2001) found that a weak therapeutic alliance at intake was a good predictor of dropout.
Strengths and Limitations
The main strength of this study, as a virtue of its design, was also its main weakness. The analogue design provided the ideal method and conducting this research in actual residential substance abuse treatment facilities provided ideal environments in which both therapeutic alliance and disclosure could be manipulated and measured.
Using an analogue design also had inherent weaknesses. The artificiality of the research and that research participants responded regarding what they would expect or intend to do might or might not generalize to actual situations. For example, what people expect or intend to do is often different from what they actually do.
A limitation of this study was that only West Coast Canadian men participated in this study. Results from this study therefore may not generalize to men from other regions and cultures. It could be that men from other cultures may react to TSD differently than did the men in the present study.
Recommendations for Future Research
Although intuitively it seems that TSD would affect the relationship between therapeutic alliance and dropout, this study did not provide support for this belief. However, both these constructs (therapeutic alliance and dropout) are complex areas of research with possibly many influencing factors. Therefore, future researchers may want to explore the role of other therapeutic techniques (such as empathy, confrontation, making decisional balance sheets), in concert with TSD, so that the therapeutic alliance-dropout relationship can be better understood.
References
Audet, C.T. (2004). Client experiences of therapist self-disclosure. (Doctoral
dissertation). Available from ProQuest Dissertations and Theses database. (ISBN
0-612-96235-0)
Barber, J.P., Luborsky, L., Gallop, R., Frank, A., Crits-Cristoph, P. Weiss, R.D., …
Siqueland, L. (2001). Therapeutic alliance as a predictor of outcome and retention
in the national institute on drug abuse collaborative cocaine treatment study.
Journal of Consulting and Clinical Psychology, 69, 119-124. doi: 10.1037//0022-
006X.69.U19
Barrett, M. S., & Berman, J. S. (2001). Is psychotherapy more effective when therapists
disclose information about themselves? Journal of Consulting and Clinical
Psychology, 69, 597-603. doi: 10.1037//0022-006X.69.4.597
Bedi, R.P., Davis, M.D., & Arvay, M.J. (2005). The client’s perspective on forming a counseling alliance and implications for research on counselor training. Canadian Journal of Counseling, 39, 71-85.
Brocato, J., & Wagner, E.F. (2008). Predictors of retention in an alternative-to-prison
substance abuse treatment program. Criminal Justice and Behavior, 35, 99-119.
doi: 10.1177/0093854807309429
Hanson, J. A. (2005). Should your lips be zipped? How therapist self-disclosure and
non-disclosure affects clients. Counseling and Psychotherapy Research, 5, 96 –
104. doi: 10.1080/17441690500226658
Horvath, A.O. (2001). The alliance. Psychotherapy, 38, 365-372.
Huebner, B.M., & Cobbina, J. (2007). The effect of drug use, drug treatment
participation, and treatment completion on probationer recidivism. Journal of
Drug Issues, Summer, 619-641. doi: 0022-0426/07/03 619-642
Johansson, H., & Eklund, M. (2006), Helping alliance and early dropout from
psychiatric out-patient care. Social Psychiatry and Psychiatric Epidemiology, 41,
140-147. doi: 10.1007/s00127-005-0009-z
Kelly, J.F., & Moos, R. (2003). Dropout from 12-step self-help groups: Prevalence,
predictors, and counteracting treatment influences. Journal of Substance Abuse
Treatment, 24, 241-250. doi: 10.1016/S0740-5472(03)00021-7
Kelly, A. E. & Rodriguez, R. R. (2007). Do therapists self-disclose more to clients with
greater symptomatology? Psychotherapy: Theory, Research, Practice, Training,
44, 470 – 475. doi: 10.1037/0033-3204.44.4.470
Meier, P.S., Barrowclough, C., & Donmall, M.C. (2005). The role of the therapeutic
alliance in the treatment of substance misuse: A critical review of the literature.
Addiction, 100, 304-316. doi: 10.1111/j.1360-0443.2004.00935.x
Meier, P.S., Donmall, M.C., McElduff, P., Barrowclough, C., & Heller, R.F. (2006).
The role of the early therapeutic alliance in predicting drug treatment dropout.
Drug and Alcohol Dependence, 83, 57-64. doi: 0.1016/j.drugalcdep.2005.10.010
Myers, D.M. (2004). Therapist self-disclosure of countertransference: Necessity or
indulgence? (Doctoral dissertation). Available from ProQuest Dissertations and
Theses database. (UMI No. 3157562)
Saatsi, S., Hardy, G.E., & Cahill, J. (2007). Predictors of outcome and completion status in cognitive therapy for depression. Psychotherapy Research, 17, 185-195. doi:10.1080/10503300600779420
Sauer, E.M., Lopez, F.G., & Gormley, B. (2003). Respective contributions of therapist and client adult attachment orientations to the development of the early working alliance: A preliminary growth modeling study. Psychotherapy Research, Fall, 371-382. doi:10.1093/ptr/kpg027
Sayre, S.L., Schmitz, J.M., Stotts, A.L, Averill, P.M., Rhoades, H.M., & Grabowski, J.J. (2002). Determining predictors of attrition in an outpatient substance abuse program. American Journal of Drug and Alcohol Abuse, 28, 55-72.
Wierzbicki, M. & Pekarik, G. (1993). A meta-analysis of psychotherapy dropout. Professional Psychology: Research and Practice, 24, 190-195.

Leave a Reply

Your email address will not be published. Required fields are marked *

BizStudio-lite Theme by SketchThemes