Therapeutic alliance mediates the association between personality and treatment outcome in patients with major depressive disorder

Journal of Affective Disorders 201 (2016) 137–144

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research paper

Therapeutic alliance mediates the association between personality and treatment outcome in patients with major depressive disorder Shauna C. Kushner a, Lena C. Quilty a,b, Amanda A. Uliaszek a,b, Carolina McBride a,b, R. Michael Bagby a,b,n a b

University of Toronto, Canada Centre for Addiction and Mental Health, Canada

art ic l e i nf o

a b s t r a c t

Article history: Received 9 September 2015 Received in revised form 30 April 2016 Accepted 13 May 2016 Available online 16 May 2016

Background: Patient personality traits have been shown to influence treatment outcome in those with major depressive disorder (MDD). The trait agreeableness, which reflects an interpersonal orientation, may affect treatment outcome via its role in the formation of therapeutic alliance. No published studies have tested this hypothesis in patients with MDD. Method: Participants were 209 outpatients with MDD who were treated in a randomized control trial. Mediation analyses were conducted to examine the role of therapeutic alliance in the association between pretreatment personality and the reduction of depression symptom severity during treatment. Separate models were estimated for patient- versus therapist-rated therapeutic alliance. Results: We found a significant indirect effect of agreeableness on the reduction of depression severity via patient-rated therapeutic alliance. Results were replicated across two well-validated measures of depression symptom severity. Results also partially supported indirect effects for extraversion and openness. Therapist ratings of alliance did not mediate the association between personality and treatment outcomes. Limitations: Patients were recruited as part of a randomized control trial, which may limit the generalizability of results to practice-based clinical settings. Due to constraints on statistical power, intervention-specific mediation results were not examined. Conclusions: These results highlight the importance of personality and the role it plays in treatment process as well as outcome. & 2016 Elsevier B.V. All rights reserved.

Keywords: Major depressive disorder Therapeutic alliance Five factor model Mediation Treatment response

1. Introduction Patient personality traits, as operationalized using the fivefactor model, are associated with treatment outcome in major depressive disorder (MDD; Bagby et al., 1995, 2008; Quilty et al., 2008). The five-factor model is a widely accepted hierarchical model of personality, and is composed of five higher-order dimensional trait domains–neuroticism, extraversion, openness-toexperience, agreeableness, and conscientiousness (Digman, 1997; McCrae and Costa, 2008). Investigators have documented positive associations between treatment responses for psychotherapy and agreeableness, a trait that encompasses interpersonal orientation; specifically, “agreeable” patient characteristics (e.g., trust, friendliness, warmth) have been associated with good treatment n Correspondence to: University of Toronto, Research Science, Bldg, SY-124, 1265 Military Trail, Toronto, ON, Canada M1C 1A4. E-mail address: [email protected] (R.M. Bagby).

http://dx.doi.org/10.1016/j.jad.2016.05.016 0165-0327/& 2016 Elsevier B.V. All rights reserved.

outcome, whereas disagreeable patient characteristics (e.g., distrust, manipulativeness, control-seeking) have been associated with poor treatment outcome (Gurtman, 1996; Ogrodniczuk et al., 2003). Results for pharmacotherapy have shown the opposite result, such that low variants of agreeableness (e.g., low trust and straightforwardness) were associated with lower posttreatment depression severity (Bagby et al., 2008). It has been theorized that agreeableness impacts treatment outcome through its influence on the therapeutic alliance (Miller, 1991; Zinbarg et al., 2008). For example, support for the influence of personality traits on clienttherapist alliance and treatment outcomes would underscore the value of conducting pre-treatment personality assessments, as this data would inform prognostic utility and treatment selection determinations. For example, personality assessments may help illuminate the optimal foci of therapeutic change efforts, supply patients with realistic expectations of therapeutic gains, match treatment selection to patient personality traits, and facilitate the development of self in therapy (Harkness and Lilienfeld, 1997).

138

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

Despite its potential implications for treatment, investigators have yet to test directly the influence of patient personality traits on treatment outcome via their impact on therapeutic alliance among patients with MDD. Therapeutic alliance refers to the collaborative bond between patients and therapists (Horvath and Luborsky, 1993; Krupnick et al., 1996; Martin et al., 2000). As with agreeableness, therapeutic alliance has been positively associated with treatment outcomes for both psychotherapy and pharmacotherapy (Castonguay et al., 1996; Falkenström et al., 2013; Klein et al., 2003; Krupnick et al., 1996; Orlinsky et al., 1994). Researchers have speculated that its effects on treatment outcome may be specious, in that therapeutic alliance represents a “proxy” for the influence of other variables, such as patient personality traits (Feeley et al., 1999; Klein et al., 2003; Orlinsky et al., 1994). Extensive research in personality psychology has substantiated agreeableness as a powerful predictor of positive interpersonal functioning (Roberts et al., 2007). Interpersonal processes within the therapeutic context have previously been operationalized according to the interpersonal circumplex, which characterizes interpersonal behavior across two bipolar dimensions: agency (dominance versus submission) and communion (warmth versus hostility; Kiesler, 1992; Wiggins, 1991). Agreeableness is positively associated with communion (DeYoung et al., 2013; McCrae and Costa, 1989; Wiggins and Pincus, 1994), and may likewise correspond to the complementarity influence of interpersonal behaviors in therapy (Kiesler, 1983, 1996). That is, highly agreeable patients exhibit greater warmth in therapy, which elicits greater warmth from clinicians, thereby facilitating the development of therapeutic alliance. Moreover, aspects of agreeableness (e.g., compassion, politeness, and cooperation) may be associated with a desire for affiliative bonding and social regulation (DeYoung, 2015). Therapeutic alliance may thus be a manifestation of a characteristic adaptation (i.e., a stable goal, interpretation, and/or strategy) of agreeableness; however, the personality antecedents of therapeutic alliance have yet to be demonstrated empirically. Ratings of alliance by highly agreeable patients may likewise be influenced by an underlying desire to provide kinder, more positive evaluations of their therapists. Yet, there has been little investigation examining the associations among personality traits, therapeutic alliance and treatment outcome. Indeed, most investigations of personality in the context of treatment for depression have focused upon treatment outcomes rather than process. Recently, Hirsh et al. (2012) found support for the mediation model linking agreeableness, therapeutic alliance and treatment outcome in a sample of patients with borderline personality disorder (BPD) treated with dialectical behavior therapy or treatment as usual. Results from this study revealed that associations between agreeableness and treatment outcome were mediated via therapeutic alliance, particularly in response to dialectical behavior therapy. Patients diagnosed with MDD and BPD notably exhibit discrepant five-factor model trait profiles, such that patients with BPD typically exhibit low agreeableness (Blais, 1997; Pukrop, 2002), whereas most studies have yielded no significant associations among agreeableness and MDD diagnoses or depression severity (Bienvenu et al., 2004; Rosellini and Brown, 2011). Nevertheless, preliminary evidence does support a positive association among patient agreeableness and therapeutic alliance in mixed outpatient samples (Bachelor et al., 2010; Coleman, 2006). It is currently unclear whether the associations among personality, therapeutic alliance, and treatment outcome are the same across diagnoses. The current investigation thus aimed to extend this research by examining whether agreeableness has a unique effect on therapeutic alliance in MDD, and whether this effect is specific to this personality trait.

1.1. The current investigation The present study tests directly the hypothesis that therapeutic alliance mediates the association between personality and treatment response in patients with MDD. In this research, we examined the associations among personality, therapeutic alliance, and treatment outcome in 168 patients with MDD who were treated as part of a randomized control trial. Although data from this sample has been included in earlier reports (Bagby et al., 2008; Bulmash et al., 2009; Kushner et al., 2009; McBride et al., 2006; Quilty et al., 2013), no previous work has examined the role of therapeutic alliance. Consistent with clinical research and theory (Hirsh et al., 2012; Miller, 1991; Zinbarg et al., 2008), we hypothesized that agreeableness would be positively associated with therapeutic alliance and negatively associated with symptom severity at the end of treatment. At the outset, there was no conceptual or theoretical basis to anticipate that the other five-factor model domains would influence treatment outcome via therapeutic alliance. Given mixed evidence for the convergence of patients’ and therapists’ ratings of therapeutic alliance (De Bolle, 2010; Krupnick et al., 1996; Tryon et al., 2007), we tested separate models using patient and therapists' ratings of alliance. In addition, we repeated analyses across two well-validated measures of depressive symptom severity.

2. Methods 2.1. Participants Participants were 146 outpatients (63.0% female) who participated in a large randomized control trial investigation conducted at a tertiary care psychiatric facility (see ClinicalTrials.gov, #NCT00744406—RMB Principal Investigator), which included three treatment conditions: antidepressant medication (ADM), interpersonal therapy (IPT), and cognitive behavioral therapy (CBT). Participant flow and demographic characteristics are displayed in Fig. 1 and Table 1, respectively. All participants were 18– 65 years of age, fluent in English, and provided informed consent. Patients were not obligated to disclose racial or ethnic information; however, the general impression was that the sample was predominantly Caucasian and Canadian-born, reflecting the broader patient base at the clinic.

Fig. 1. Participant flow.

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

Table 1. Participant demographic information.

Age Gender Male Female Marital status Married Single (previously married) Single (never married) Education Completed postsecondary Some postsecondary High-school or less Employment status Employed Unemployed Hollingshead occupation scale

ADM (n¼ 74)

IPT (n¼65)

CBT (n ¼70)

M/n

SD

M/n

SD

M/n

SD

40.94

12.69

39.69

13.01

40.77

12.24

27 47

24 41

24 46

27 16 27

27 11 24

26 18 26

43 12 15

36 19 7

52 12 6

45 17 5.87

48 7 5.73

50 12 5.90

2.19

2.01

2.43

Note: ADM ¼antidepressant medication; IPT ¼ interpersonal therapy; CBT ¼cognitive behavioral therapy. N ¼ 209.

2.2. Measures 2.2.1. Hamilton Depression Rating Scale (Ham-D) The Ham-D (Hamilton, 1960) is a 17-item semi-structured interview used to assess depressive symptom severity. It has long been considered the gold standard for assessing symptom severity (Williams, 2001; see also Bagby et al., 2004). All patients completed the Ham-D prior to commencing treatment and at posttreatment. 2.2.2. Beck Depression Inventory-II (BDI-II) The BDI-II (Beck et al., 1996) is a 21-item self-report instrument used to assess depressive symptom severity. It has shown excellent reliability and validity (Dozois et al., 1998; Quilty et al., 2010; Steer et al., 1997). All patients completed the BDI-II prior to commencing treatment. Subsequently, patients receiving psychotherapy completed the BDI-II at weekly therapy sessions and posttreatment; patients receiving ADM completed the BDI-II at weeks 1, 3, 5, 8, 12 and posttreatment. In the current investigation, we examined BDI-II ratings provided at pre- and posttreatment. 2.2.3. Revised NEO Personality Inventory (NEO PI-R) The NEO PI-R (Costa and McCrae, 1992) is a 240-item self-report questionnaire used to assess the traits of the five-factor model. The NEO PI-R has shown good reliability and validity in psychiatric patients (Bagby et al., 1999; Costa et al., 2005). Scoring for the NEO PI-R generates five trait domain scale scores (Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness). Patients completed the NEO PI-R immediately prior to treatment initiation (week 1) and following treatment completion. In the current investigation, we examined pre-treatment scores for the Neuroticism (T-score M ¼68.24, SD ¼ 8.63), Extraversion (T-score M¼39.86, SD ¼10.40), Openness (T-score M ¼54.00, SD ¼9.19), Agreeableness (T-score M ¼44.33, SD ¼10.91), and Conscientiousness (T-score M¼ 36.13, SD ¼ 10.26) scales. Relative to the normative standardization sample (Costa and McCrae, 1992), and consistent with previous empirical work (Tackett et al., 2008), depressed patients in the current sample reported higher Neuroticism (t(202) ¼30.10, p o.001, d ¼1.95) and Openness (t(202) ¼6.21, p o.001, d¼ .42), and lower Extraversion (t(202) ¼  13.89, p o.001, Cohen's d ¼ .99), Agreeableness (t (202) ¼  7.40, p o.001, d ¼.54), and Conscientiousness (t (202) ¼  19.27, p o.001, d ¼1.37).

139

2.2.4. California Psychotherapy/Pharmacotherapy Alliance Scales (CALPAS) The CALPAS is a 24-item questionnaire used to assess patient experiences in psychotherapy (Gaston and Marmar, 1991; Marmar et al., 1989) and pharmacotherapy (Weiss et al., 1997) sessions. The psychotherapy and pharmacotherapy versions of the CALPAS include the same item content, with minor wording differences (e.g., items on the psychotherapy version refer to the “therapist” and “therapy,” whereas items on the pharmacotherapy version refer to the “doctor” and “medication”). Both the psychotherapy and pharmacotherapy versions of the CALPAS can be used to calculate a total scale and four subscale scores: Patient Working Capacity; Patient Commitment; Therapist Understanding and Involvement; and Working Strategy Consensus. Patient- and therapist-report versions of the CALPAS are analogous; however, the therapist-report version of the CALPAS includes six additional items for measuring Hostile Resistance. Only those items that were common to both versions were included in the current study. Consistent with previous reports (Bachelor, 2013; Gaston, 1991; Hatcher et al., 1995), individual subscales on the patient-report version yielded poor internal consistency (coefficient α ¼.19–.63); the current analyses thus utilized the total CALPAS score (α ¼ .79 for patients; α ¼.89 for therapists). Patients and therapists completed the CALPAS at their third treatment session and at subsequent intervals thereafter (weeks 8, 12, and at posttreatment). In the current investigation, we used CALPAS scores obtained at their third and twelfth treatment session as indicators of therapeutic alliance in early and later stages of treatment, respectively. 2.3. Procedure The current investigation was approved by the institutional research ethics board. In total, 209 participants met DSM-IV criteria for a current major depressive episode (MDE) as assessed by the patient version of the Structured Clinical Interview for DSM-IV (SCID-I/P; First et al., 2002). Participants were excluded if they met criteria for bipolar disorder, schizoaffective disorder, or schizophrenia; current primary anxiety disorders, eating disorders, substance abuse or dependence; or antisocial or borderline personality disorder. Individuals who displayed active psychotic symptoms, suicidal or self-harm behaviors, organic brain syndromes and injuries, or concurrent active medical illnesses were also excluded. All eligible participants discontinued prior ADM for a minimum of two weeks before commencing the trial. Eligible participants were randomly assigned to ADM, IPT, or CBT. Stratified random sampling was used to ensure that each condition was matched according to sex and number of previous depressive episodes. Results from multivariate analyses of variance (MANOVA) revealed no significant mean-level differences across treatment conditions for baseline personality traits, depression severity, or early therapeutic alliance ratings (Pillai's trace¼1.05, F (18, 398)¼ 1.23, p ¼.236, η2 ¼.05). Participants randomized to pharmacotherapy were treated by six licensed psychiatrists and one clinical fellow. On average, each clinician for the ADM group treated 10.00 patients (SD ¼16.33). Participants randomized to psychotherapy received 16–20 sessions of manualized treatment. Participants randomized to IPT were treated by three licensed psychiatrists, two licensed social workers, and six clinical graduate-level trainees. Participants randomized to CBT were treated by two licensed clinical psychologists and six clinical graduate-level trainees. On average, each clinician for the IPT group treated 5.82 patients (SD ¼8.64) and each clinician for the CBT group treated 7.67 patients (SD ¼7.33). Treatment administration in each condition followed published guidelines (Greenberger and Padesky, 1995; Kennedy et al., 2001; Weissman et al., 2000). Treatment adherence in the psychotherapy conditions was assessed by a

140

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

licensed clinical psychologist using the Collaborative Study Psychotherapy Rating Scale (Hollon et al., 1988); following this procedure, all therapists in this study were deemed adherent to treatment protocols. 2.4. Statistical analysis Due to participant withdrawal prior to treatment initiation, NEO PI-R and early CALPAS data were unavailable for 35 participants (16.7%). An additional 34 participants (16.3%) terminated the treatment protocol prematurely. To preserve power, missing data were imputed using the expectation-maximization (EM) algorithm in SPSS. Little's Missing Completely at Random test for this imputation was not significant, χ2(987) ¼929.95, p¼ .902. Subsequent analyses were therefore based on the 209 patients who were randomized to treatment (ADM¼ 74, IPT¼70, CBT¼65).1 MANOVA results revealed significant mean-level differences for baseline personality traits, depression severity, and early therapeutic alliance ratings (Pillai's trace¼.96, F(1, 207)¼ 2.34, p ¼.016, η2 ¼ .10) across treatment completers versus non-completers, such that non-completers reported higher scores for NEO PI-R Neuroticism (F(1, 207)¼9.54, p ¼.002, η2 ¼.04), BDI-II depression severity (F(1, 207)¼ 7.20, p ¼.008, η2 ¼.03), and lower scores for NEO PI-R Agreeableness F(9, 199) ¼4.11, p ¼.044, η2 ¼.02). Independent samples t tests and χ2 tests were used to compare demographic and clinical variables for patients who withdrew versus patients who did not withdraw from the treatment protocol following randomization. There were no significant differences in demographic variables or baseline BDI-II ratings (all ps 4 .05); although patients who dropped out of treatment (M¼19.39, SD ¼3.49) had higher baseline Ham-D ratings than patients who completed at least three weeks of treatment (M¼ 17.90, SD ¼3.70; t(207)¼2.31, p ¼.02), the magnitude of this difference was small (d ¼.41; Cohen, 1992). Moreover, the observed difference in HamD ratings (Ham-Ddiff ¼1.49) do not meet thresholds for clinically meaningful differences in symptom severity (i.e., 43; Kirsch et al., 2008). Pearson correlation coefficients were used to evaluate the relative consistency for patient-therapist agreement on ratings of alliance. Results from these analyses showed that inter-rater agreement was r ¼ .28, p o.001 for early ratings of alliance, and r ¼.12, p ¼.088 for late ratings of alliance. Paired-samples t-tests were used to evaluate for potential mean differences in patients' and therapists' ratings of alliance and supplemented by calculating effect sizes (Cohen's d) to determine the magnitude of differences between patients’ and therapists’ ratings of alliance. Therapists provided significantly lower ratings of early alliance, t(202) ¼4.66, p o.001, d ¼.52, as well as late alliance, t(202) ¼ 4.29, p o.001, d ¼.57. A serial multiple mediation model was estimated to examine the influence of personality traits on patient-rated therapeutic alliance and treatment outcome. Separate models were estimated for each independent variable (five-factor model domains) and each dependent variable (posttreatment Ham-D and BDI-II scores). Separate models were also estimated for patient- and therapistreported alliance ratings. Ratings of early and late therapeutic alliance were conceptualized as serial mediators (e.g., pretreatment personality-early alliance-late alliance-treatment outcomes). Pre-treatment Ham-D and BDI-II scores were included as a covariate in all models to partial out the effects of baseline depression severity. Where applicable, models also included BDI-II scores 1 Analyses were also completed with the 142 participants for whom all data were available (ADM ¼41, IPT ¼52, CBT ¼ 49). These analyses yielded the same pattern of results and are available upon request.

Fig. 2. Proposed serial multiple mediation model of association among personality traits, therapeutic alliance, and treatment outcome.

from week 3 as a covariate to partial out the effects of symptom improvement occurring prior to early ratings of alliance. Fig. 2 depicts the mediation model tested. The indirect effects thus provide estimates of the independent variable (patient personality) on the dependent variable (depressive severity) after accounting for the mediators (early and late therapeutic alliance). We used 5000 bootstrap resamples to provide stable estimates of the direct, indirect, and total effects. Confidence intervals were bound using the bias-corrected and accelerated method. Mediation was tested using Hayes' (2013) PROCESS modeling approach in SPSS 22 (model 6). This analytic approach permits the examination of both direct and indirect effects of personality and therapeutic alliance on treatment outcomes. Significant indirect effects were followed-up with multiple group mediation models in Mplus 6.1 to examine whether the mediated effects of personality on treatment outcomes via therapeutic alliance were further moderated by treatment type.

3. Results Descriptive statistics and correlation coefficients are listed in Table 2. Estimates of direct effects from mediation models using patient-rated therapeutic alliance are displayed in Table 3. Direct effects from serial mediation analyses showed that early CALPAS ratings were positively predicted by pretreatment Agreeableness, but not by Neuroticism, Extraversion, Openness, or Conscientiousness. In all models, early CALPAS ratings positively predicted late CALPAS ratings. Moreover, in all models, late CALPAS negatively predicted treatment outcome. Neuroticism was the only pretreatment personality trait that showed a direct effect on treatment outcome, such that higher Neuroticism predicted poorer treatment outcomes, as measured by posttreatment BDI-II scores (p ¼ .044); however, this effect was not replicated in models predicting posttreatment Ham-D scores. Estimates of indirect effects from all mediation models are displayed in Table 4. The indirect effect of Agreeableness on treatment outcome via early and late alliance was significant. No other significant serial mediation effects were observed for the remaining five-factor model domains. We did observe additional evidence for the indirect effects of Extraversion, Openness, and Agreeableness; however, these results varied depending on the time at which CALPAS ratings were obtained and our measure of depressive symptoms. For example, Openness showed significant indirect effects on posttreatment depression scores via late ratings of alliance. Extraversion similarly showed significant indirect effects on posttreatment depression scores via late ratings of alliance; however, this was only

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

141

Table 2. Correlations and descriptive statistics. 1

2

3

4

5

6

7

8

9

10

11

12

13

14

1. Ham-DT1 2. Ham-DT4 3. BDI-IIT1 4. BDI-IIT2 5. BDI-IIT4 6. CALPAS-PT2 7. CALPAS-PT3 8. CALPAS-TT2 9. CALPAS-TT3 10. NeuroticismT1 11. ExtraversionT1 12. OpennessT1 13. AgreeablenessT1 14. ConscientiousnessT1

– .21nn .45nnn .25nnn .18n .03 .27nnn  .09  .09 .05 .15n .02  .06 .12

– .25nnn .53nnn .88nnn –.21nn –.18n –.10  .06 .27nnn .02  .10 .15n  .02

– .45nnn .42nnn  .05 .16n  .02  .02 .29nnn  .02  .02  .09  .11

– .62nnn .00  .03 .01 .06 .26nnn .02 .01  .04  .17n

–  .20nn  .18nn  .11  .08 .34  .03  .08 .13  .08

– .45nnn .28nnn .15n  .09 .12 .08 .27nnn .02

– .02 .12 -.09 .22nn .17n .11 .13

– .72  .04 .07 .02 .11 .05

– .06 .02 .04 .05  .03

– .06 .02 .04  .03

– .46nnn .06 .21nnn

– .17n  .05

– .03

–

Mean SD α

18.11 3.47 .20

9.54 7.71 .83

30.95 7.73 .81

23.45 9.10 .90

14.97 11.49 .94

71.96 7.93 .77

75.51 6.87 .79

67.78 12.26 .89

71.51 11.97 .89

118.86 17.88 .89

90.99 18.97 .89

117.62 15.72 .86

117.19 16.36 .89

98.55 17.80 .89

Note: Subscripts denote time when measures were collected: T1 ¼ pretreatment; T2 ¼ therapy session 3; T3 ¼ therapy session 12; T4 ¼ posttreatment. Ham-D ¼ Hamilton Depression Rating Scale; BDI-II ¼ Beck Depression Inventory-II; CALPAS-P ¼ California Psychotherapeutic/Pharmacotherapy Alliance Scale–Patient Version; CALPAS-T ¼ California Psychotherapeutic/Pharmacotherapy Alliance Scale–Therapist Version. N ¼ 209. n

p o .05. p o .01. nnn p o .001. nn

the case for models predicting posttreatment BDI-II scores. In contrast, Agreeableness showed significant indirect effects on posttreatment depression scores via early ratings of alliance, but only for the model predicting posttreatment BDI-II scores. Models that showed significant indirect effects (i.e., Ham-D models for Openness and Agreeableness; BDI-II models for Extraversion, Openness, and Agreeableness) were followed up with multiple group mediation models. These analyses yielded no significant indirect effects for ADM, IPT, or CBT. Results were omitted for brevity, but are available upon request. Results for serial mediation models using therapist-rated therapeutic alliance yielded few significant direct effects. In all models, early CALPAS ratings positively predicted late CALPAS ratings (all ps o.001). Patient Neuroticism was the only pretreatment personality trait that showed a direct effect on treatment outcome, such that higher Neuroticism predicted poorer treatment outcomes, as measured by residualized BDI-II scores (p ¼.025); however, this effect was not replicated in models predicting residualized Ham-D scores. No direct effects were observed between patient personality and therapist-rated CALPAS scores (all ps 4.152). In addition, no significant indirect effects emerged. Full results for these analyses are available upon request.

4. Discussion Our primary aim was to examine the associations between personality, therapeutic alliance, and treatment outcome in a sample of patients receiving one of three empirically supported treatments for MDD. Mediation analyses revealed significant indirect effects for agreeableness via therapeutic alliance. Results were replicated across two well-validated measures of depressive symptom severity. We also found support for indirect effects of extraversion and openness, but these were uniquely mediated through late ratings of alliance. This represents the first investigation to test empirically the hypothesis that patient personality influences treatment outcome via its influence on therapeutic alliance among patients with MDD (Bagby et al., 2008; Miller, 1991; Zinbarg et al., 2008). Our investigation supports the long-held assumption that

patient personality traits may enhance the benefit of therapeutic interventions via their impact on the patient-clinician bond (Miller, 1991; Zinbarg et al., 2008). These results have important implications for the utility of personality assessment in clinical practice, such as predicting prognosis and informing treatment planning efforts. Better understanding of factors that influence therapeutic alliance may be beneficial for targeting areas in need of improvement (DeRubeis et al., 2005). The current results also suggest that traits beyond Agreeableness may have relevance for the impact of alliance on treatment outcomes. Interestingly, we observed divergent results across traits, such that agreeableness showed indirect effects via early ratings of alliance whereas extraversion and openness showed indirect effects via late alliance. As such, agreeableness may have greater impact on the formation of alliance early on in treatment. In contrast, extraversion and openness may have greater bearing on the maintenance of alliance later in treatment. Future investigations should examine the utility of therapy techniques and modules geared toward improving therapeutic alliance among disagreeable patients. In particular, disagreeable patients may benefit from responsive and collaborative therapist techniques (e.g., exploration, reflection; Ackerman and Hilsenroth, 2003) and manualized treatments (e.g., motivational enhancement therapy, alliance-fostering psychotherapy) that may facilitate the development and maintenance of alliance (Crits-Christoph et al., 2006, 2009). The failure to replicate findings across patients’ and therapists’ ratings of alliance suggests that there may be qualitative differences in these informants’ reports. Stronger associations between patient-reported alliance, personality traits, and depressive symptoms may be expected due to shared method variance (i.e., measurement error in self-report data). Informant disagreements are common in the literature on therapeutic alliance, with clients reporting higher alliance than therapists (Tryon et al., 2007). Nevertheless, informant discrepancies have been shown to account for meaningful variance in important clinical outcomes. For example, higher patient-therapist agreement on early ratings of alliance predicts greater symptom reduction (Marmarosh and Kivlighan, 2012). Moreover, higher patient-therapist agreement also predicts lower depression severity at three months posttreatment (Laws et al., 2016). The current results are also

142

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

Table 3. Standardized direct effects of the influence of five factor model personality traits and therapeutic alliance on treatment outcomes. Early alliance

Ham-DT1 NeuroticismT1 CALPAST2 CALPAST3 Ham-DT1 ExtraversionT1 CALPAST2 CALPAST3 Ham-DT1 OpennessT1 CALPAST2 CALPAST3 Ham-DT1 AgreeablenessT1 CALPAST2 CALPAST3 Ham-DT1 ConscientiousnessT1 CALPAST2 CALPAST3

BDI-IIT1 BDI-IIT2 NeuroticismT1 CALPAST2 CALPAST3 BDI-IIT1 BDI-IIT2 ExtraversionT1 CALPAST2 CALPAST3 BDI-IIT1 BDI-IIT2 OpennessT1 CALPAST2 CALPAST3 BDI-IIT1 BDI-IIT2 AgreeablenessT1 CALPAST2 CALPAST3 BDI-IIT1 BDI-IIT2 ConscientiousnessT1 CALPAST2 CALPAST3

Late alliance

Treatment outcome

B (SE)

95% CI

B (SE)

95% CI

B (SE)

95% CI

Ham-DT4 Scores .07 (.16)  .04 (.04)

[  .24, .38] [  .10,.02]

.52nnn (.12)  .03 (.02) .38nnn (.05)

[.29, .76] [  .07, .02] [.27, .48]

.02 (.16) .05 (.03)

[  .30, .33] [  .01, .11]

.48nnn (.12) .05n (.02) .37nnn (.05)

[.29, .76] [.01, .09] [.27, .47]

.05 (.16) .04 (.04)

[  .26, .37] [  .03, .11]

.51nnn (.12) .06n (.03) .37nnn (.05)

[.28, .74] [.01, .11] [.27, .47]

.10 (.15) .13nnn (.03)

[  .21, .40] [.07, .19]

.52nnn (.12) .00 (.03) .38nnn (.05)

[.28, .75] [  .05, .05] [.27, .49]

.05 (.16) .01 (.03)

[  .26, .37] [  .05, .07]

.49nnn (.12) .03 (.02) .38nnn (.05)

[.26, .73] [  .01, .08] [.28, .48]

.54nnn (.15) .10nnn (.03)  .12 (.07)  .19n (.08) .57nnn (.15) .02 (.03)  .13 (.07)  .22n (.09) .58nnn (.15)  .03 (.03)  .13 (.07)  .20n (.09) .56nnn (.15)  .04 (.03)  .11 (.07)  .21n (.09) .58nnn (.15)  .01 (.03)  .13 (.07)  .21n (.09)

[.25, .84] [.04, .15] [  .25, .02] [  .36,  .02] [.27, .87] [  .04, .07] [  .27, .01] [  .39,  .05] [.27, .88] [  .09, .03] [  .27, .01] [  .37,  .03] [.26,.87] [  .10, .02] [  .25, .04] [  .38,  .04] [.28, .89] [  .07, .04] [  .27, .01] [  38,  .04]

BDI-IIT4 Scores  .05 (.08) .04 (.07)  .04 (.03)

[  21, .11] [  .09, .17] [  .10, .02]

.23nnn (.06)  .09 (.05) -.04 (.02) .39nnn (.05)

[.11, .35] [  .19, .01] [  .08, .05] [.29, .49]

 .06 (.08) .02 (.07) .05 (.03)

[  .22, .09] [  .11, .16] [  .01, .11]

.22nnn (.06)  .11n (.05) .06nn (.02) .38nnn (.05)

[.10, .34] [  .21,  .01] [.02, .11] [.28, .48]

 .06 (.08) .03 (.07) .04 (.04)

[  .22, .09] [  .11, .16] [  .03, .11]

.22nnn (.06)  .10n (.05) .06n (.03) .39nnn (.05)

[.10, .33] [  .20, .00] [.01, .11] [.29, .49]

 .04 (.08) .03 (.07) .13nnn (.03)

[  .19, .11] [  .10, .15] [.06, .19]

.21nnn (.06)  .10n (.05) .00 (.03) .40nnn (.05)

[.09, .33] [  .20, .00] [  .05, .05] [.29, .51]

 .07 (.08) .03 (.07) .01 (.03)

[.22, .09] [  .11, .16] [  .05, .07]

.22nnn (.06)  .09 (.05) .05n (.02) .40nnn (.05)

[.10, .34] [  .19, .01] [.00, .09] [.29, .50]

.24nn (.09) .65nnn (.07) .09nn (.03)  .19n (.08)  .21n (.10) .29nn (.08) .67nnn (.07) .00 (.03)  .19n (.08)  .24n (.10) .28nn (.09) .68nnn (.07)  .04 (.04)  .19n (.08)  .22n (.10) .28nn (.09) .67nnn (.07)  .03 (.04)  18n (.09)  .23n (.10) .29nn (.09) .68nnn (.07) .03 (.03)  .19n (.08)  .25n (.09)

[.06, .41] [.51, .79] [.02, .16] [  .35,  .02] [  .40,  .01] [.12,.46] [.53,.82] [  .06,.07] [  .36,  .03] [  .43,  .04] [.11, .46] [.53,.82] [  .12, .03] [  .36,  .03] [  .41,  .02] [.11, .46] [.53, .82] [  .10, .04] [  .35,  .01] [  .43,  .04] [.12, .47] [.54, .83] [  .03, .10] [  .35,  .02] [  .44,  .05]

Note: Subscripts denote time when measures were collected: T1 ¼pretreatment; T2 ¼therapy session 3; T3¼ therapy session 12; T4 ¼ posttreatment. Ham-D ¼Hamilton Depression Rating Scale; BDI-II ¼ Beck Depression Inventory-II; CALPAS ¼California Psychotherapeutic/Pharmacotherapy Alliance Scale. N¼ 209. n

p o .05. p o .01. nnn p o .001. nn

consistent with prior research indicating that patients' ratings of alliance offer superior predictive validity over therapists' ratings (Krupnick et al., 1996; although see De Bolle et al., 2010). Moreover, the results for patient-rated alliance were replicated across self-reported and clinician ratings of depression severity. As such, the current results may not be reducible to measurement error. Further research using objective ratings of patient personality and multi-informant ratings of alliance are needed to better disentangle the impact of patient personality on treatment processes and outcomes. The present study has some limitations. First, patients in the current sample were treated as part of a randomized control trial, which maximizes internal validity at the expense of external validity or generalizability. Random assignment and rigid exclusion

criteria may introduce differences in symptom severity, motivation, and baseline personality characteristics; research comparing depressed patients in a randomized control trial to those in a clinic-based study, however, observed few such differences (Kushner et al., 2009). Nevertheless, results from the current study suggest that patients who prematurely terminated treatment had higher depression severity and neuroticism scores, and lower agreeableness. Second, the lack of conditional indirect effects from posthoc multigroup mediation models may be due to insufficient statistical power to detect significant mediation effects within each treatment condition (Fritz and MacKinnon, 2007). Future studies that are sufficiently powered might further examine intervention-specific mediation results. Nevertheless, therapeutic alliance positively predicts outcome regardless of treatment

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

Table 4. Standardized indirect effects of the influence of five factor model personality traits and therapeutic alliance on treatment outcomes. B (SE)

143

variables for future research, as both may account for variance in the development of therapeutic alliance and its impact on treatment outcome (DeRubeis et al., 2005).

95% CI

References

Ham-DT4 Scores NeuroticismT1 -CALPAST2 -CALPAST3 -CALPAST2 - CALPAST3

.005 (.005) .005 (.005) .003 (.002)

[  .001, .019] [  .003, .021] [.000, .010]

ExtraversionT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .007 (.006)  .011 (.009)  .004 (.004)

[  .024, .001] [  .034, .000] [  .014, .001]

OpennessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .005 (.007)  .011 (.009)  .003 (.004)

[  .026, .004] [  .038, .000] [  .012, .002]

AgreeablenessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .014 (.010)  .001 (.007)  .011 (.005)

[  .038, .002] [  .016, .012] [  .023,  .003]

ConscientiousnessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .001 (.005)  .007 (.007)  .001 (.003)

[  .013, .007] [  .025, .003] [  .006, .005]

BDI-IIT4 Scores

NeuroticismT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

.008 (.008) .008 (.007) .003 (.003)

[  .003, .028] [  .001, .031] [  .001, .014]

ExtraversionT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .010 (.008)  .015 (.010)  .005 (.004)

[  .032, .001] [  .043,  .002] [  .016, .000]

OpennessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .007 (.010)  .013 (.010)  .003 (.004)

[  .033, .006] [  .041, .000] [  .015, .002]

AgreeablenessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .023 (.014) .000 (.008)  .012 (.006)

[  .057,  .003] [  .017, .014] [  .028,  .003]

ConscientiousnessT1 -CALPAST2 -CALPAST3 -CALPAST2-CALPAST3

 .002 (.007)  .012 (.010)  .001 (.003)

[  .017, .010] [  .039, .000] [  .008, .006]

Note: Specific indirect influences are displayed to three decimal places to enhance comparability. 95% confidence intervals that do not include “0” are significant and listed in bold. Subscripts denote time when measures were collected: T1 ¼pretreatment; T2 ¼ therapy session 3; T3 ¼ therapy session 12; T4 ¼posttreatment. Ham-D ¼ Hamilton Depression Rating Scale; BDI-II ¼Beck Depression Inventory-II; CALPAS ¼California Psychotherapeutic/Pharmacotherapy Alliance Scale. N ¼ 209.

modality (Krupnick et al., 1996). A final limitation is that we did not assess the individual characteristics of clinicians in this study, and thereby did not examine the impact of clinician characteristics and patient-clinician interactions. These represent important

Ackerman, S.J., Hilsenroth, M.J., 2003. A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clin. Psychol. Rev. 23, 1–33. Bachelor, A., 2013. Clients' and therapists' views of the therapeutic alliance: similarities, differences and relationship to therapy outcome. Clin. Psychol. Psychother. 20, 118–135. Bachelor, A., Meunier, G., Laverdiére, O., Gamache, D., 2010. Client attachment to therapist: relation to client personality and symptomatology, and their contributions to the therapeutic alliance. Psychotherapy 47, 454–468. Bagby, R.M., Costa, P.T., McCrae, R.R., Livesley, W.J., Kennedy, S.H., Levitan, R.D., Levitt, A.J., Joffe, R.T., Young, L.T., 1999. Replicating the five factor model of personality in a psychiatric sample. Personal. Individ. Differ. 27, 1135–1139. Bagby, R.M., Joffe, R.T., Parker, J.D.A., Kalemba, V., 1995. Major depression and the five-factor model of personality. J. Personal. Disord. 9, 224–234. Bagby, R.M., Quilty, L.C., Segal, Z.V., McBride, C.C., Kennedy, S.H., Costa, P.T., 2008. Personality and differential treatment response in major depression: a randomized controlled trial comparing cognitive-behavioural therapy and pharmacotherapy. Can. J. Psychiatry 53, 361–370. Bagby, R.M., Ryder, A.G., Schuller, D.R., Marshall, M.B., 2004. The Hamilton depression rating scale: has the gold standard become a lead weight? Am. J. Psychiatry 161, 2163–2177. Beck, A.T., Steer, R.A., Brown, G.K., 1996. Beck Depression Inventory Manual, 2nd ed. Psychological Corporation, San Antonio, Texas. Bienvenu, O.J., Samuels, J.F., Costa, P.T., Reti, I.M., Eaton, W.W., Nestadt, G., 2004. Anxiety and depressive disorders and the five-factor model of personality: a higher- and lower-order personality trait investigation in a community sample. Depress Anxiety 20, 92–97. Blais, M.A., 1997. Clinician ratings of the five-factor model of personality and the DSM-IV personality disorders. J. Nerv. Ment. Dis. 185, 388–393. Bulmash, E., Harkness, K.L., Stewart, J.G., Bagby, R.M., 2009. Personality, stressful life events, and treatment response in major depression. J. Consult. Clin. Psych. 77, 1067–1077. Castonguay, L.G., Goldfried, M.R., Wiser, S., Raue, P.J., Hayes, A.M., 1996. Predicting the effect of cognitive therapy for depression: a study of unique and common factors. J. Consult. Clin. Psychol. 64, 497–504. Cohen, J., 1992. A power primer. Psychol. Bull. 112, 155–159. Coleman, D., 2006. Client personality, working alliance and outcome: a pilot study. Soc. Work Ment. Health 4, 83–98. Costa, P.T., Bagby, R.M., Herbst, J.H., McCrae, R.R., 2005. Personality self-reports are concurrently reliable and valid during acute depressive episodes. J. Affect. Disord. 89, 45–55. P.T., Costa, R.R., McCrae, 1992. Revised NEO Personality Inventory: Professional manual. Psychological Assessment Resources, Odessa, Florida. Crits-Christoph, P., Gibbons, M.B.C., Crits-Christoph, K., Narducci, J., Schamberger, M., Gallop, R., 2006. Can therapists be trained to improve their alliances? A preliminary study of alliance-fostering psychotherapy. Psychother. Res. 16, 268–281. Crits-Christoph, P., Gallop, R., Temes, C.M., Woody, G., Ball, S.A., Martino, S., Carroll, K.M., 2009. The alliance in motivational enhancement therapy and counseling as usual for substance use problems. J. Consult. Clin. Psych. 77, 1125–1135. De Bolle, M., Johnson, J.G., De Fruyt, F., 2010. Patient and clinicians perceptions of therapeutic alliance as predictors of improvement in depression. Psychother. Psychosom. 79, 378–385. DeRubeis, R.J., Brotman, M.A., Gibbons, C.J., 2005. A conceptual and methodological analysis of the nonspecifics argument. Clin. Psychol. Sci. Pract. 12, 174–183. DeYoung, C.G., 2015. Cybernetic big five theory. J. Res. Personal. 56, 33–58. DeYoung, C.G., Weisberg, Y.G., Quilty, L.C., Peterson, J.B., 2013. Unifying the aspects of the big five the interpersonal circumplex and trait affiliation. J. Personal. 81, 465–475. Digman, J.M., 1997. Higher-order factors of the big five. J. Personal. Soc. Psychol. 73, 1246–1256. Dozois, D.J.A., Dobson, K.S., Ahnberg, J.L., 1998. A psychometric evaluation of the Beck Depression Inventory–II. Psychol. Assess. 10, 83–89. Falkenström, F., Granström, F., Holmqvist, R., 2013. Therapeutic alliance predicts symptomatic improvement session by session. J. Couns. Psychol. 60, 317–328. Feeley, M., DeRubeis, R.J., Gelfand, L.A., 1999. The temporal relation of adherence and alliance to symptom change in cognitive therapy for depression. J. Consult. Clin. Psych. 67, 578–582. First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 2002. Structured Clinical Interview for DSM-IV-TR Axis I Disorders, research version, patient edition. (SCID-I/P). Biometrics Research, New York State Psychiatric Institute, New York. Fritz, M.S., MacKinnon, D.P., 2007. Required sample size to detect the mediated effect. Psychol. Sci. 18, 233–239. Gaston, L., 1991. Reliability and criterion-related validity of the California Psychotherapy Alliance Scales—patient version. Psychol. Assess. 3, 68–74. Greenberger, D., Padesky, C.A., 1995. Mind Over Mood: A Cognitive Therapy Treatment Manual for Clients. Guilford Press, New York.

144

S.C. Kushner et al. / Journal of Affective Disorders 201 (2016) 137–144

Gurtman, M.B., 1996. Interpersonal problems and the psychotherapy context: the construct validity of the inventory of interpersonal problems. Psychol. Assess. 8, 241–255. Hamilton, M., 1960. A rating scale for depression. J. Neurol. Nerosurg. Psychiatry 23, 351–379. Harkness, A.R., Lilienfeld, S.O., 1997. Individual differences science for treatment planning: personality traits. Psychol. Assess. 9, 349–360. Hatcher, R.L., Barends, A., Hansell, J., Gutfreund, M.J., 1995. Patients' and therapists' shared and unique views of the therapeutic alliance: an investigation using confirmatory factor analysis in a nested design. J. Consult. Clin. Psych. 63, 636–643. Hayes, A.F., 2013. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach. Guilford Press, New York. Hirsh, J.B., Quilty, L.C., Bagby, R.M., McMain, S.F., 2012. The relationship between agreeableness and the development of the working alliance in patients with borderline personality disorder. J. Personal. Disord. 26, 616–627. Hollon, S.D., Evans, M.D., Auerbach, A., DeRubeis, R.J., Elkin, I., Lowery, A., Kriss, M., Grove, W., Tuason, V.B., Piasecki J., 1988. Development of a system for rating therapies for depression: Differentiating cognitive therapy, interpersonal psychotherapy, and clinical management pharmacotherapy, Unpublished manuscript. Horvath, A.O., Luborsky, L., 1993. The role of the therapeutic alliance in psychotherapy. J. Consult. Clin. Psych. 61, 561–573. Kennedy, S.H., Lam, R.W., Cohen, N.L., Ravindran, A.V., 2001. Clinical guidelines for the treatment of depressive disorders. IV. Medications and other biological treatments. Can. J. Psychiatry 46, 38S–58S. Kiesler, D.J., 1983. The 1982 Interpersonal Circle: a taxonomy for complementarity in human transactions. Psychol. Rev. 90, 185–214. Kiesler, D.J., 1992. Interpersonal circle inventories: pantheoretical applications to psychotherapy research and practice. J. Psychother. Integr. 2, 77–99. Kiesler, D.J., 1996. Contemporary Interpersonal Theory and Research: Personality, Psychopathology, and Psychotherapy. John Wiley And Sons, New York. Kirsch, I., Deacon, B.J., Huedo-Medina, T.B., Scoboria, A., Moore, T.J., Johnson, B.T., 2008. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the food and drug administration. PLoS Med. 5, 260–268. Klein, D.N., Schwartz, J.E., Santiago, N.J., Vivian, D., Vocisano, C., Castonguay, L.G., Arnow, B., Blalock, J.A., Manber, R., Markowitz, J.C., Riso, L.P., Rothbaum, B., McCullough, J.P., Thase, M.R., Borian, F.E., Keller, M.B., 2003. Therapeutic alliance in depression treatment: controlling for prior change and patient characteristics. J. Consult. Clin. Psychol. 71, 997–1006. Krupnick, J.L., Sotsky, S.M., Simmens, S., Moyer, J., Elkin, I., Watkins, J., Pilkonis, P.A., 1996. The role of the therapeutic alliance in psychotherapy and pharmacotherapy outcome: findings in the national institute of mental health treatment of depression collaborative research program. J. Consult. Clin. Psychol. 64, 532–539. Kushner, S.C., Quilty, L.C., McBride, C., Bagby, R.M., 2009. A comparison of depressed patients in randomized versus nonrandomized trials of antidepressant medication and psychotherapy. Depress Anxiety 26, 666–673. Marmar, C.R., Weiss, D.S., Gaston, L., 1989. Toward the validation of the California Therapeutic Alliance Rating System. Psychol. Assess. 1, 46–52. Marmarosh, C.L., Kivlighan Jr, D.M., 2012. Relationships among client and counselor agreement about the working alliance, session evaluations, and change in client symptoms using response surface analysis. J. Couns. Psychol. 59, 352–367. Martin, D.J., Garske, J.P., Davis, M.K., 2000. Relation of the therapeutic alliance with outcome and other variables: a meta-analytic review. J. Consult. Clin. Psychol.

68, 438–450. McBride, C., Atkinson, L., Quilty, L.C., Bagby, R.M., 2006. Attachment as moderator of treatment outcome in major depression: a randomized control trial of interpersonal psychotherapy versus cognitive behavior therapy. J. Consult. Clin. Psychol. 74, 1041–1054. McCrae, R.R., Costa, P.T., 1989. The structure of interpersonal traits: Wiggins's circumplex and the five-factor model. J. Personal. Soc. Psychol. 56, 586–595. McCrae, R.R., Costa, P.T., 2008. The five-factor theory of personality. In: John, O.P., Robins, R.W., Pervin, L.A. (Eds.), Handbook of Personality: Theory and Research, 3rd ed. Guilford Press, New York, pp. 159–181. Miller, T.R., 1991. The psychotherapeutic utility of the five-factor model of personality: a clinician's experience. J. Personal. Assess. 57, 415–433. Ogrodniczuk, J.S., Piper, W.E., Joyce, A.S., McCallum, M., Rosie, J.S., 2003. NEO-five factor personality traits as predictors of response to two forms of group psychotherapy. Int. J. Group Psychother. 53, 417–442. Orlinsky, D.E., Grawe, K., Parks, B.K., 1994. Process and outcome in therapy. In: Bergin, E., Garfield, S.L. (Eds.), Handbook of Psychotherapy and Behavior Change, 4th ed. Wiley, New York, pp. 270–376. Pukrop, R., 2002. Dimensional personality profiles of borderline personality disorder in comparison with other personality disorders and healthy controls. J. Personal. Disord. 16, 135–147. Quilty, L.C., De Fruyt, F., Rolland, J.P., Kennedy, S.H., Rouillon, F., Bagby, R.M., 2008. Dimensional personality traits and treatment outcome in patients with major depressive disorder. J. Affect. Disord. 108, 241–250. Quilty, L.C., Mainland, B.J., McBride, C., Bagby, R.M., 2013. Interpersonal problems and impacts: further evidence for the role of interpersonal functioning in treatment outcome in major depressive disorder. J. Affect. Disord. 150, 393–400. Quilty, L.C., Zhang, K.A., Bagby, R.M., 2010. The latent symptom structure of the Beck Depression Inventory–II in outpatients with major depression. Psychol. Assess. 22, 603–608. Rosellini, A.J., Brown, T.A., 2011. The NEO Five-Factor Inventory: latent structure and relationships with dimensions of anxiety and depressive disorders in a large clinical sample. Assessment 18, 27–38. Steer, R.A., Ball, R., Ranieri, W.F., Beck, A.T., 1997. Further evidence for the construct validity of the Beck Depression Inventory-II with psychiatric outpatients. Psychol. Rep. 80, 443–446. Tryon, G.S., Blackwell, S.C., Hammel, E.F., 2007. A meta-analytic examination of client-therapist perspectives of the working alliance. Psychother. Res. 17, 629–642. Weissman, M.M., Markowitz, J.C., Klerman, G.L., 2000. Comprehensive Guide to Interpersonal Psychotherapy. Basic Books, New York. Wiggins, J.S., 1991. Agency and communion as conceptual coordinates for the understanding and measurement of interpersonal behavior. In: Cicchetti, D., Grove, W.M. (Eds.), Personality and Psychopathology 2. University of Minnesota Press, Minneapolis, MN, pp. 89–113. Wiggins, J.S., Pincus, A.L., 1994. Personality structure and the structure of personality disorders. In: Widiger, T.A., Costa, P.T. (Eds.), Personality Disorders and the Five-Factor Model of Personality. American Psychological Association, Washington, DC, pp. 73–93. Williams, J.B.W., 2001. Standardizing the Hamilton depression rating scale: past, present, and future. Eur. Arch. Psychiatry Clin. Neurosci. 251, 6–12. Zinbarg, R.E., Uliaszek, A.A., Adler, J.M., 2008. The role of personality in psychotherapy for anxiety and depression. J. Personal. 76, 1649–1688.