Clinical Psychology and Psychotherapy Clin. Psychol. Psychother. 14, 126134 (2007) Published online in Wiley InterScience (www.interscience.wiley.com).

DOI: 10.1002/cpp.526

Effect of Patient Dissatisfaction with the Therapist on Group Therapy Outcome

John S. Ogrodniczuk,1* Anthony S. Joyce2 and William E. Piper1
1 2

Department of Psychiatry, University of British Columbia Department of Psychiatry, University of Alberta

This study had three objectives concerning patients dissatisfaction with their psychotherapist. The rst was to compare patient dissatisfaction with the therapist in interpretive and supportive group therapies. The second was to investigate the relationship between dissatisfaction and treatment outcome. The third was to examine the quality of object relations (QOR) as a moderator of the effect of dissatisfaction on treatment outcome. The study utilized data from 107 patients with complicated grief who participated in a randomized trial of interpretive and supportive group therapies. Patients reported a greater dissatisfaction in interpretive therapy than in supportive therapy. Dissatisfaction was inversely related to improvement in supportive therapy, but not in interpretive therapy. Further, QOR moderated the effect of dissatisfaction on treatment outcome in supportive therapy. The ndings suggest that, for certain patients, dissatisfaction with ones therapist may have an adverse effect on ones ability to benet from group therapy. Copyright 2007 John Wiley & Sons, Ltd.

INTRODUCTION
There has been a proliferation of research on patient satisfaction with mental health services. Typically, satisfaction is considered as a measure of outcome and a quality of care (Shipley, Hilborn, Hansell, Tyrer, & Tyrer, 2000). This is due to a shift in the eld towards broadening the scope of evaluation of therapeutic services and recognition that satisfaction has the potential to inuence a patients decision about future treatment (Lee, 2005) and recommendation of services to others (Kersnik, 2003). In addition, patient satisfaction has been demonstrated to be an important factor in the care process, inuencing intervention efcacy (Hasler et al., 2004) and consumer behaviour such as compliance (Gellaitry et al., 2005) and service

* Correspondence to: Dr John S. Ogrodniczuk, Department of Psychiatry, University of British Columbia, Suite 420-5950 University Boulevard, Vancouver BC, V6T 1Z3, Canada. E-mail: [email protected]

utilization (Tehrani, Krussel, Borg, & MunkJorgensen, 1996). Research on patient satisfaction has usually focused on patients attitudes towards their treatment in general. Curiously, satisfaction with the direct service provider, i.e., the therapist, has been studied infrequently. The few studies that have examined patients satisfaction with their therapists have focused on how it may inuence patients decisions to remain in treatment. Findings from these studies have demonstrated that patient dissatisfaction with the therapist is associated with an increased likelihood of prematurely terminating treatment (Hunsley, Aubry, Verstervelt, & Vito, 1999; Pekarik, 1992). Unfortunately, therapists appear to signicantly underestimate the impact of patient dissatisfaction with them, often citing this as the least likely reason for patients to terminate treatment prematurely (Hunsley et al., 1999). Few studies have considered the effect of patient satisfaction (or dissatisfaction) with the therapist on treatment outcome. Those that have studied this issue have indicated that a greater satisfaction with

Copyright 2007 John Wiley & Sons, Ltd.

Patient Dissatisfaction the therapist was associated with better outcome (Conte, Buckley, Picard, & Karasu, 1994; Conte, Ratto, Clutz, & Karasu, 1995; Oei & Shuttlewood, 1999). A signicant limitation of these studies, however, was the concurrent assessment of satisfaction with therapist and treatment outcome, which typically occurred after treatment ended. This introduces the possibility that patients ratings of satisfaction with their therapists were confounded by their impressions of how well they did in treatment or vice versa. Assessing satisfaction prior to outcome, preferably early in therapy, would facilitate an understanding of the possible causal relationship between satisfaction and outcome. The current study assessed patient dissatisfaction with the therapist early in the process of therapy (after 4 weeks of group therapy) and assessed treatment outcome at the end of therapy (after 12 weeks of group therapy). The relationship between the two was examined. Dissatisfaction was assessed in order to be consistent with the literature on dropping out, which focuses on dissatisfaction rather than satisfaction. Conte and her colleagues (1994, 1995) examined a variety of therapist characteristics and behaviours that may contribute to a patients satisfaction with the therapist in individual psychodynamic psychotherapy. These studies indicated that satisfaction was greatest with therapists who were encouraging, likeable, respectful, and active and who provided advice. We are not aware of any study that has compared patients ratings of satisfaction with their therapists in different forms of therapy. It is possible that treatments (and therapists) that differ on features like therapist activity and provision of advice may differ in the extent to which they satisfy patients. Thus, the present study examined patients dissatisfaction with their therapists in two different forms (interpretive, supportive) of group psychotherapy. These two forms of therapy differed considerably on features such as therapist activity and provision of advice. Not all patients respond to dissatisfaction with their therapist in the same manner. Among the many factors that may shape a persons reactions to dissatisfaction or disappointment in a close relationship is the persons quality of object relations (QOR). QOR is dened as a persons lifelong tendency to establish certain kinds of relationships with others, ranging from primitive to mature (Azim, Piper, Segal, Nixon, & Duncan, 1991). Theoretically, external relations, through the processes of internalization, result in the formation of intrapsychic structures (object and self representaCopyright 2007 John Wiley & Sons, Ltd.

127 tions) that regulate and direct interpersonal behaviour (Blatt, Wiseman, Prince-Gibson, & Gatt, 1991). Object representations refer to the conscious and unconscious mental structures or schemata including cognitive and affective componentsof past interpersonal interactions. These schemata provide a template that affects the nature of subsequent interpersonal relationships. Derlega and colleagues (Derlega, McIntyre, Winstead, & Morrow, 2001; Derlega, Winstead, & Lewis, 1993) have suggested that the quality of a persons object relations plays a major role in responses to difculties in a relationship. Specically, they have suggested that patients with a positive internal working model of others and a history of rewarding interactions would be more likely to have constructive rather than destructive responses to perceived difculties in their relationship with the therapist. Azim and colleagues (1991) describe different levels of object relations that characterize responses to disappointment or difculties in a relationship. At the mature level, the person enjoys relatively trusting, give-and-take relationships characterized by mature behaviour and concern for others. There is a capacity to tolerate disappointment in others and to take risks to resolve relationship difculties. There is a paucity of emergency emotions (e.g., aggression or anxiety) in response to disappointment in a relationship or loss of a relationship. At the primitive level, the persons relationships are typically characterized by hurting and being hurt. The person perceives disappointment in a relationship as disapproval or rejection by a relationship partner and typically reacts with intense anxiety and affect. There is an inordinate dependence on the relationship partner, who provides a sense of identity for the person. Perceived threats to the viability of a relationship incite extreme, negative reactions by the person. Considering the role that a persons object relations play in shaping ones responses to relationship difculties, the present study investigated whether or not the QOR moderated the effect of dissatisfaction with the therapist. That is, we examined whether or not the effect of dissatisfaction on the outcome of therapy differed for patients with more mature and more primitive object relations.

SUMMARY OF OBJECTIVES
The present study had three objectives. The rst was to compare patient dissatisfaction with the therapist in interpretive and supportive group
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

128 therapies. Considering the therapist characteristics (e.g., active, forthcoming with advice) that have previously been found to be associated with satisfaction and their similarity with the features of supportive therapy, we expected to nd lower levels of dissatisfaction among patients in supportive therapy compared to that of patients in interpretive therapy. The second was to examine the association between dissatisfaction and treatment outcome. Considering previously found associations between dissatisfaction and certain negative events in therapy (i.e., premature termination, poor outcome), we expected to nd that dissatisfaction would be related to poorer treatment outcome in both forms of therapy. The third was to examine the role of QOR as a moderator of the effect of dissatisfaction on treatment outcome. We expected to nd that dissatisfaction with the therapist would have a more powerful impact on outcome for patients with more primitive object relations compared to patients with more mature object relations.

J. S. Ogrodniczuk et al. Interview for DSM-III-R (First, Gibbon, Williams, & Spitzer, 1990) and were validated by an independent clinical diagnosis assigned jointly by the intake assessor and a psychiatrist, both of whom saw the patient on the day of intake. A total of 73.8% of the patients received an Axis I diagnosis. The most frequent disorders were current major depression (54.2%) and dysthymia (8.4%). Axis II diagnoses were determined by the computer-administered Structured Clinical Interview for DSM-III-R Personality Questionnaire and by the Auto-Structured Clinical Interview for DSMIII-R (First, Gibbon, Williams, & Spitzer, 1991). A total of 55.1% of the patients received an Axis II diagnosis. The most frequent Axis II disorders were avoidant (26.2%), dependent (13.1%) and borderline (9.3%). A total of 38.3% of the patients received both Axis I and Axis II diagnoses. The average age of the patients was 43 years (Standard Deviation [SD] = 10.3, range = 1967). Seventy-seven percent were women. With regard to marital status, 41% were married or living with a partner, 26% were separated or divorced, 18% were widowed, and 15% were single and had never been married. Forty-seven percent were educated beyond high school and 52% were employed. Ninety percent of the patients were Caucasian. Many (72%) had received a previous psychiatric treatment, but few (15%) had a history of psychiatric hospitalization. The types of losses reported by the patients and their prevalence were parent (49%), partner (15%), child (9%), sibling (10%), friend (4%), grandparent (5%) and other (8%). The average time since the loss was 8.9 years (SD = 11.1, range = 0.2547.0).

METHOD
Setting and Procedures
This study utilized data that were collected during a randomized controlled trial of two forms of group psychotherapy for patients with complicated grief (Piper, McCallum, Joyce, Rosie, & Ogrodniczuk, 2001). Patients were referred from a large psychiatric outpatient clinic of a university hospital if they had experienced a signicant death loss and if they had met the criteria for complicated grief. The criteria for complicated grief included (1) elevated scores on measures of grief symptoms; (2) signicant disturbance in social functioning; and (3) a duration of at least 3 months since the time of the loss. Written informed consent was obtained from all participants. Exclusion criteria included psychosis, substance abuse, active suicidal risk, organic mental disorder and antisocial personality disorder.

Therapists
The therapists were a 40-year-old male psychologist, a 41-year-old female social worker and a 40year-old female occupational therapist. They had substantial experience practicing group therapy (13, 14 and 10 years, respectively). A total of 16 groups were run in the clinical trial. The psychologist conducted four therapy groups and the other two therapists conducted six therapy groups each. Each therapist conducted interpretive and supportive groups concurrently.

Patients
The sample for the present study consisted of 107 patients who completed treatment and provided ratings of symptoms and functioning at pretherapy and post-therapy. All 107 treatment completers received diagnoses according to the DSM-III-R. Axis I diagnoses were identied by the computer-administered Mini-Structured Clinical
Copyright 2007 John Wiley & Sons, Ltd.

Therapies
Each patient received a form of group therapy that emphasized interpretive or supportive features.
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

Patient Dissatisfaction They were labelled interpretive therapy (N = 53) and supportive therapy (N = 54). The contractual and structural features were similar. The patient was scheduled for weekly 90-minute sessions for 12 weeks. Punctual attendance was emphasized. Each group was led by a single therapist. The therapist was paid by a third party (Canadian universal health care). The average number of sessions attended among treatment completers was 10.6 (SD = 1.3, range = 812). In interpretive therapy, the primary objective is to enhance the patients insight about repetitive conicts (intrapsychic and interpersonal) and trauma that are associated with the losses and that are assumed to serve as impediments to experiencing a normal mourning process. The patients are expected to begin each session and to assume responsibility for what follows. In the session, the therapist attempts to create a climate that is mildly anxiety arousing in order to allow defences, transference reactions and unconscious themes to emerge. In regard to technique, the therapist encourages the patients to explore conicts, which often involve uncomfortable emotions. Interpretations about sensitive topics, including transference, are often made. There is a focus on the here-andnow in the group, but patients contemporary relationships outside the group are also considered. The therapist attempts to link unconscious intrapsychic conicts with the more conscious interpersonal situation inside and outside the group. Unlike certain approaches to short-term, dynamic psychotherapy (e.g., Davanloo, 1978), our interpretive model did not encourage an aggressive confrontation of patients intrapsychic conicts and defences. However, a moderate interpretation of conict was emphasized. Similarly, our interpretive group therapy model did not follow extreme approaches to group therapy such as that advocated by Bion (1959), who recommended therapists to provide group interpretations exclusively, to avoid eye contact with patients and to refuse to repeat interpretations. While our model does advocate the use of group interpretation, it does not recommend its exclusive use. Also, our model encourages therapists to maintain eye contact with patients, to elaborate on interpretations when requested and to explore material that individual patients present, in addition to attending to group phenomena. In supportive therapy, the primary objective is to improve the patients immediate adaptation to their life situations. It is assumed that improvements in symptoms and social functioning can be achieved through the provision of support and
Copyright 2007 John Wiley & Sons, Ltd.

129 problem solving. In the session, the therapist attempts to create a climate of gratication wherein patients can share common experiences and feelings and receive reinforcement for their efforts at coping. In regard to technique, the therapist uses guidance, problem solving and praise. The therapists were experienced in providing a variety of interpretive and supportive therapies. Nevertheless, each participated in a weekly group seminar and had conducted pilot groups before conducting groups in the study. The seminar continued throughout the project. The therapists followed technical manuals for interpretive group therapy and supportive group therapy for loss patients (McCallum, Piper, & Joyce, 1995; Piper, McCallum, & Joyce, 1995). The manuals described, illustrated and compared the technical emphases of the two forms of group therapy. Adherence to the technical manuals was monitored for all sessions by external observers using a version of the Interpretive and Supportive Technique Scale (Ogrodniczuk & Piper, 1999) that was adapted for group therapy. The evidence from the adherence ratings indicated that the two forms of therapy were well differentiated and conformed to the manuals.

Dissatisfaction with Therapist

Dissatisfaction with therapist was assessed using three items that were part of a larger 9-item scale that had been used in a study of group members reactions to their leader (Piper, Marrache, Lacroix, Richardsen, & Jones, 1983). The three items were: (1) I wish he (she) were more expressive of his feelings; (2) I wish he (she) paid more attention to me; and (3) I wish he (she) were more active. Items were rated on a scale that ranged from 1 (very little) to 6 (very much), with higher scores reecting greater dissatisfaction. Thus, patients reactions to the therapist were solicited, not merely their opinion of whether certain therapist behaviours were present or not. A dissatisfaction total score was derived using the mean of the three item ratings. The internal consistency of the three dissatisfaction items, represented by Cronbachs alpha, was 0.77. Patients rated their therapist after the rst third of therapy (i.e., after session 4).

QOR
QOR was assessed during a 1-hour, semistructured, pre-therapy interview and was scored
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

130 using a 9-point scale (Azim et al., 1991). The patients lifelong pattern of relationships was explored in reference to criteria that characterize ve levels of object relations: (1) primitive, (2) searching, (3) controlling, (4) triangular and (5) mature. The criteria refer to behavioural manifestations, regulation of affect, regulation of self-esteem and historical antecedents. Two psychologists and three psychiatrists served as raters. Rater reliability was assessed during three periods of the project. Each assessment involved the ve raters and 12 patients. The average intraclass correlation coefcient (2, 1) for the three assessments was 0.83.

J. S. Ogrodniczuk et al. and third objectives, we used three-step hierarchical regression analyses to test for the effect of dissatisfaction on treatment outcome, and also to test whether or not QOR moderated the effect of dissatisfaction. In the rst step of the regression analysis, dissatisfaction, QOR and form of therapy (interpretive or supportive) were included. The second step included the two-way interaction terms for dissatisfaction and QOR, dissatisfaction and form of therapy, and QOR and form of therapy. The interaction term for dissatisfaction and QOR tests whether or not QOR moderates the effect of dissatisfaction across both forms of therapy. The third step included a single variable representing the three-way interaction between dissatisfaction, QOR and form of therapy. This interaction term tests whether or not QOR differentially moderates the effect of dissatisfaction in the two forms of therapy. Each of the three outcome factors served as dependent variables.

Treatment Outcome
Treatment outcome was represented by three outcome factors (general symptoms, grief symptoms and target objectives/life satisfaction). These three factors were derived from a principal components analysis with orthogonal rotation of the outcome measures used in the clinical trial (Piper et al., 2001). The measures (questionnaire or interview) assessed grief symptoms, interpersonal distress, social role functioning, psychiatric symptoms, self-esteem, life satisfaction and physical functioning. The severity of disturbance for individual target objectives was also assessed. Residual change scores (pre-therapy to post-therapy) were calculated for each outcome variable. The scores represent the patients outcome at posttherapy after accounting for their status at pretherapy. For all three factors, higher scores represent a less favourable change in outcome. A detailed description of the individual outcome variables and of the principal components analysis can be found in Piper et al.

RESULTS
Possible Group or Therapist Differences
We did not nd any evidence of signicant differences in dissatisfaction ratings between the different groups or between the therapists who ran the groups in either the supportive therapy condition or the interpretive therapy condition.

Dissatisfaction Ratings in Interpretive Therapy and Supportive Therapy

We found that patients provided signicantly higher dissatisfaction ratings in interpretive therapy (M = 3.4, SD = 1.2) compared to supportive therapy (M = 2.5, SD = 1.2), t = 4.02, degrees of freedom (df) = 103, p < 0.001.

Approach to Analysis
Prior to addressing the objectives of the study, we investigated the possibility that differences in dissatisfaction ratings could be attributed to the group to which the patient belonged or to the therapist who led the group. We used one-way analysis of variance to check for these possibilities. These analyses were conducted for the interpretive therapy and supportive therapy samples separately. With regard to the studys rst objective, we used a t-test to examine the difference in patient dissatisfaction ratings in interpretive group therapy and supportive group therapy. With regard to the second
Copyright 2007 John Wiley & Sons, Ltd.

Dissatisfaction, QOR and Treatment Outcome General Symptoms Outcome

We found evidence for the three-way interaction between dissatisfaction, QOR and form of therapy (F[1, 85] = 4.52, p < 0.05). This nding indicated that in supportive therapy, QOR signicantly moderated the effect of dissatisfaction on treatment outcome. In supportive therapy, dissatisfaction had an increasingly negative effect on treatment outcome as patients QOR scores decreased (i.e., became more primitive). In interpretive therapy, dissatisfaction had a minimal relationship with treatment outcome regardless of the patients level of QOR.
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

Patient Dissatisfaction

131 To further clarify the interaction ndings, we calculated a t-test to determine whether low-QOR patients provided higher dissatisfaction ratings than high-QOR patients in supportive therapy. We found that low-QOR patients (M = 2.2, SD = 1.3) and high-QOR patients (M = 2.7, SD = 1.0) did not differ signicantly in their dissatisfaction ratings (t[50] = 1.3, p > 0.20]. Thus, it is not the case that dissatisfaction had a particularly adverse effect on the outcomes of low-QOR patients because of higher reported levels of dissatisfaction among low-QOR patients compared to high-QOR patients.

Grief Symptoms Outcome

We did not nd evidence for the three-way interaction between dissatisfaction, QOR and form of therapy, although it did approach signicance (p = 0.08). However, we found a signicant two-way interaction effect for dissatisfaction and form of therapy, t(85) = 2.83, p < 0.01, indicating that dissatisfaction was inversely related to favourable outcome in supportive therapy but unrelated to outcome in interpretive therapy.

Target Objectives/Life Satisfaction Outcome

We found evidence for the three-way interaction between dissatisfaction, QOR and form of therapy (F[1, 85] = 8.35, p < 0.01). As with the general symptoms outcome factor, the nding indicated that QOR signicantly moderated the effect of dissatisfaction on treatment outcome in supportive therapy. In supportive therapy, the negative effect of dissatisfaction on outcome increased as patients QOR scores decreased. In interpretive therapy, dissatisfaction had a minimal association with outcome regardless of the patients level of QOR.

DISCUSSION
The present study examined patient dissatisfaction with the therapist in two different forms of group psychotherapy. Although patients reported greater levels of dissatisfaction with their therapist in interpretive group therapy compared to supportive group therapy, dissatisfaction appeared to have a minimal impact on outcome in interpretive therapy but a considerable impact on outcome in supportive therapy. The effect of dissatisfaction on outcome in supportive group therapy was moderated by the patients level of object relations. Dissatisfaction had a particularly powerful negative effect on outcome for patients with more primitive object relations. While dissatisfaction also had a negative inuence on outcome for patients with more mature object relations, the association was smaller and was not statistically signicant. We were not surprised to nd that patients reported greater levels of dissatisfaction with their therapist in interpretive group therapy compared to supportive group therapy. It is not unusual for patients to express discontent about the style of interpretive therapists. The interpretive therapist assumes a reective, non-gratifying style in an attempt to create a climate of tolerable tension and deprivation wherein conicts can be examined in the here-and-now experience. Patients sometimes react to this style by demanding that the therapist be more active, self-disclosing, and forthcoming with advice and guidance. These are aspects of the therapists behaviour that were assessed with our dissatisfaction measure. Conversely, the supportive therapist assumes an active, gratifying style in order to provide structure and encouragement to patients. The therapist uses self-disclosure (when appropriate) for the sake of modelling adaptive behaviours. Advice and guidance gure prominently as means of resolving immediate stressors
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

Clarifying the Interaction Findings

In order to clarify the nature of the three-way interaction ndings (dissatisfaction QOR form of therapy), we dichotomized the supportive therapy sample into low-QOR (QOR score 4.0) and high-QOR (QOR score > 4.0) groups and then examined the correlation between dissatisfaction and outcome for each group. As shown in Table 1, the ndings indicated that the negative association between dissatisfaction and improvement was substantially larger among low-QOR patients compared to high-QOR patients.
Table 1. Correlations between dissatisfaction and outcome among low-QOR and high-QOR patients in supportive group therapy Dissatisfaction r Low-QOR (N = 22) General symptoms Grief symptoms Target objective/life satisfaction High-QOR (N = 28) General symptoms Grief symptoms Target objectives/life satisfaction 0.45 0.72 0.46 0.30 0.36 0.10 p 0.045 0.001 0.032 0.135 0.075 0.597

Note: Higher scores on the outcome factors indicate worse outcome. QOR = quality of object relations.

Copyright 2007 John Wiley & Sons, Ltd.

132 for patients. Patients typically nd the style of the supportive therapist appealing. One might expect that higher levels of dissatisfaction would have a more negative effect on treatment outcome. However, this was not generally true in the present study. Although patients reported greater levels of dissatisfaction in interpretive therapy than in supportive therapy, dissatisfaction was found to have a minimal impact on outcome in interpretive therapy but a considerable impact in supportive therapy. This suggests that dissatisfaction is not necessarily a universally negative phenomenon. The context (supportive or interpretive therapy) in which patients experience dissatisfaction is important to consider. Furthermore, we found that dissatisfaction had a particularly negative effect on outcome among low-QOR patients in supportive therapy. Considering that low-QOR patients did not report signicantly higher levels of dissatisfaction than high-QOR patients in supportive therapy, it appears that absolute levels of dissatisfaction are not commensurate with poor outcome. Rather, the patients characteristic reaction to a disappointment in close relationships seems to be a deciding factor that moderates the effect of dissatisfaction on the outcome of therapy. Low-QOR patients have a tendency to experience very strong negative emotions and ruminations about the destruction of a close relationship when disappointed with their relationship partner, e.g., a therapist (Azim et al., 1991). This may fuel intense and prolonged feelings of anger towards their relationship partner. However, they also have a deep fear of separation, a desperate desire to be accepted, and are overly dependent. This may hold in check the intense resentment and anger they have towards their relationship partner when disappointed and lead to a redirection of these feelings towards the self (Mikulincer & Shaver, 2005). This self-directed anger may be exacerbated by their poor self-esteem, which could result in them blaming themselves for their partners negative behaviour. As a result, low-QOR patients may react to dissatisfaction with their therapist with a complex mixture of resentment, hostility, anger, self-criticism, sadness and depression (Mikulincer & Shaver, 2005). Although such feelings towards the therapist are addressed in supportive therapy, the therapists attempt to neutralize them may not be successful. These feelings would likely intensify over the course of the therapy. The tendency of low-QOR patients to become preoccupied with or to ruminate about such feelings would interfere
Copyright 2007 John Wiley & Sons, Ltd.

J. S. Ogrodniczuk et al. with the work of therapy, resulting in a poor outcome for these patients. High-QOR patients, on the other hand, are typically able to tolerate some disappointments in relationships because they generally perceive that their relationship partners are well intentioned and that upsetting transactions are temporary. That is, the belief that their relationship partner is kind and well meaning prevails despite periodic disappointments. They also have a capacity to take risks to resolve difculties in relationships. Because of their tolerance of disappointment in relationships and of their capacity to address disappointment in a constructive manner, they are less affected by dissatisfaction with their therapist. These explanations are consistent with ndings from research on personal relationships, which have demonstrated that people with secure attachment patterns (similar to those with a high QOR) tend to respond to stress in a relationship in a constructive manner aimed at maintaining the relationship (Mikulincer & Shaver, 2003). Conversely, people with insecure attachment patterns (similar to those with a low QOR) use distancing strategies to cope with disappointment in a partner or display intense feelings of despair and depression in response to disappointment (Mikulincer, 1998). Interestingly, Rholes, Simpson, and Orina (1999) found that people with anxious (i.e., insecure) attachment patterns suppressed their angry feelings towards their relationship partner in supportseeking exercises (presumably because of their strong need of reassurance from their partner). However, after support was no longer needed or available, the angry feelings surfaced. This may help explain why low-QOR patients in our study remained in therapy despite their disappointment with their therapist and why, despite remaining in therapy, they experienced less favourable outcomes. That is, they suppressed their angry feelings in order to maintain a relationship with their therapist, but once the relationship ended, their anger surfaced and was reected in their outcome ratings. Dissatisfaction may have had a minimal impact in interpretive group therapy regardless of the patients level of QOR because the patients impressions of the therapist are directly addressed and worked on via an exploration of the transference and the use of the here-and-now relationship with the therapist to resolve chronic interpersonal difculties. Generally speaking, transference refers to the patients affective and behavioural reactions to the therapist that are, in part, determined by the
Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

Patient Dissatisfaction patients previous relationships (i.e., feelings and behaviours associated with early relationship gures are displaced onto current relationship gures). The therapist addresses the transference to help the patient clarify which aspects of his or her feelings and reactions represent a recapitulation of unresolved conicts and which aspects belong to the real relationship with the therapist. The therapist helps the patient develop an awareness that transference reactions are inappropriate and do not belong to the present. Negative reactions to the therapist that do belong to the present are openly explored and worked through. Thus, although patients in interpretive group therapy may perceive their therapist in a somewhat negative light, the therapist purposefully addresses such impressions, which may attenuate the potentially negative inuence that dissatisfaction can have on the patients progress in therapy. The ndings of this study raise the clinical implication that therapists in supportive group therapy should attend to the patients perceptions of the therapist, particularly for those patients who have a history of poor relationships. While an examination of patients impressions of the therapist is typically not emphasized in supportive therapy, our results suggest that ignoring such impressions may lead to a poor outcome for some patients. Patients with more primitive object relations may be particularly likely to perceive the therapists behaviour as demanding, non-gratifying, or critical. These behaviours could be perceived as threatening to the relationship. The therapist needs to attend to such perceptions in order to diffuse the potentially negative responses that these patients typically have to dissatisfaction and disappointment in relationships. Regarding the limitations of the present study, our sample consisted of psychiatric outpatients who had lost a signicant person through death, met the criteria for complicated grief and received treatment that was intended to help them resolve their loss issues. It is unclear whether or not our ndings would generalize to other patients without complicated grief who receive a treatment that is focused on issues unrelated to loss. Another limitation concerns the extent to which one can generalize the ndings to other formats for psychotherapy. The ndings of the present study were based on data collected from patients who received time-limited group psychotherapy. Whether or not the ndings can be applied to individual psychotherapy remains to be determined. A third limitation was our use of a relatively simple scale
Copyright 2007 John Wiley & Sons, Ltd.

133 that assessed only three aspects of a patients dissatisfaction with his or her therapist. A more comprehensive scale that assessed a greater range of positive and negative therapist behaviours could have provided a greater insight into which aspects of the therapists behaviour gure most prominently in a patients perception of (dis)satisfaction with his or her therapist. Finally, although we attempted to maximize the timing between ratings of dissatisfaction and treatment outcome, we cannot make conclusions about causality. This is an important issue for future research to address. Patients should be instructed not to consider the improvement (or lack thereof) that they have made when rating the therapist. Further, symptom change at the time of the patients rating of dissatisfaction should be assessed and included in the statistical analyses (i.e., examine the association between dissatisfaction and outcome after taking into account the patients degree of improvement or worsening at the time of the dissatisfaction rating). In conclusion, the present study has provided evidence that suggests that a patients dissatisfaction with his or her therapist may have an important negative impact on the patients ability to benet from group therapy. The potentially negative effect of dissatisfaction depends on the treatment context (i.e., interpretive or supportive therapy) and the patients level of object relations (i.e., more primitive vs. more mature). The ndings are important in terms of increasing our understanding of how aspects of the therapy process may inuence the outcome of treatment.

ACKNOWLEDGEMENT
This research project was supported by Grant MT-13481 from the Medical Research Council of Canada.

REFERENCES
Azim, H.F.A., Piper, W.E., Segal, P.M., Nixon, G.W.H., & Duncan, S. (1991). The quality of object relations scale. Bulletin of the Menninger Clinic, 55, 323343. Bion, W.R. (1959). Experiences in group. New York: Basic Books. Blatt, S.J., Wiseman, H., Prince-Gibson, E., & Gatt, C. (1991). Object representations and change in clinical functioning. Psychotherapy, 28, 273283. Conte, H.R., Buckley, P., Picard, S., & Karasu, T.B. (1994). Relations between satisfaction with therapists and psychotherapy outcome. Journal of Psychotherapy Practice and Research, 3, 215221.

Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp

134
Conte, H.R., Ratto, R., Clutz, K., & Karasu, T.B. (1995). Determinants of outpatients satisfaction with therapists. Journal of Psychotherapy Practice and Research, 4, 4351. Davanloo, H. (1978). Basic principles and techniques in short-term dynamic psychotherapy. New York: Spectrum. Derlega, V.J., McIntyre, R., Winstead, B.A., & Morrow, G. (2001). A preliminary study of an attractionbarrier model of patients commitment and responses to dissatisfaction in psychotherapy. Psychotherapy, 38, 283296. Derlega, V.J., Winstead, B.A., & Lewis, R.J. (1993). Clients responses to dissatisfaction in psychotherapy: A test of Rusbults exit-voice-loyalty-neglect model. Journal of Social and Clinical Psychology, 12, 307318. First, M.B., Gibbon, M., Williams, J.B.W., & Spitzer, R.L. (1990). Mini-SCID. Toronto: Multi-Health Systems Inc. First, M.B., Gibbon, M., Williams, J.B.W., & Spitzer, R.L. (1991). SCID-II PQ and AutoSCID II. Toronto: MultiHealth Systems Inc. Gellaitry, G., Cooper, V., Davis, C., Fisher, M., Leake, D.H., & Horne, R. (2005). Patients perception of information about HAART: Impact on treatment decisions. AIDS Care, 17, 367376. Hasler, G., Moergeli, H., Bachman, R., Lambreva, E., Buddeberg, C., & Schnyder, U. (2004). Patient satisfaction with outpatient psychiatric treatment: The role of diagnosis, pharmacotherapy, and perceived therapeutic change. Canadian Journal of Psychiatry, 49, 315321. Hunsley, J., Aubry, T.D., Verstervelt, C.M., & Vito, D. (1999). Comparing therapist and client perspectives on reasons for psychotherapy termination. Psychotherapy, 36, 380388. Kersnik, J. (2003). Patients recommendation of doctor as an indicator of patient satisfaction. Hong Kong Medical Journal, 9, 247250. Lee, K.J. (2005). A practical method of predicting client revisit intention in a hospital setting. Health Care Management Review, 30, 157167. McCallum, M., Piper, W.E., & Joyce, A.S. (1995). Manual for time-limited, short-term, supportive group therapy for patients experiencing pathological bereavement. Unpublished manuscript. Mikulincer, M. (1998). Adult attachment style and individual differences in functional versus dysfunctional

J. S. Ogrodniczuk et al.
experiences of anger. Journal of Personality and Social Psychology, 74, 513524. Mikulincer, M., & Shaver, P.R. (2003). The attachment behavioural system in adulthood: Activation, psychodynamics, and interpersonal processes. In M.P. Zanna (Ed.), Advances in experimental social psychology (Vol. 35, pp. 53152). San Diego, CA: Academic Press. Mikulincer, M., & Shaver, P.R. (2005). Attachment theory and emotions in close relationships: Exploring the attachment-related dynamics of emotional reactions to relational events. Personal Relationships, 12, 149168. Oei, T.P.S., & Shuttlewood, G.J. (1999). Development of a satisfaction with therapy and therapist scale. Australian and New Zealand Journal of Psychiatry, 33, 748753. Ogrodniczuk, J.S., & Piper, W.E. (1999). Measuring therapist technique in psychodynamic psychotherapies: Development and use of a new scale. Journal of Psychotherapy Practice and Research, 8, 142154. Pekarik, G. (1992). Relationship of clients reasons for dropping out of treatment to outcome and satisfaction. Journal of Clinical Psychology, 48, 9198. Piper, W.E., Marrache, M., Lacroix, R., Richardsen, A.M., & Jones, B.D. (1983). Cohesion as a basic bond in groups. Human Relations, 36, 93108. Piper, W.E., McCallum, M., & Joyce, A.S. (1995). Manual for time-limited, short-term, interpretive group therapy for patients experiencing pathological bereavement. Unpublished manuscript. Piper, W.E., McCallum, M., Joyce, A.S., Rosie, J.S., & Ogrodniczuk, J.S. (2001). Patient personality and timelimited group psychotherapy for complicated grief. International Journal of Group Psychotherapy, 51, 525552. Rholes, W.S., Simpson, J.A., & Orina, M.M. (1999). Attachment and anger in an anxiety-provoking situation. Journal of Personality and Social Psychology, 76, 940957. Shipley, K., Hilborn, B., Hansell, A., Tyrer, J., & Tyrer, P. (2000). Patient satisfaction: A valid index of quality of care in a psychiatric service. Acta Psychiatrica Scandinavica, 101, 330333. Tehrani, E., Krussel, J., Borg, L., & Munk-Jorgensen, P. (1996). Dropping out of psychiatric treatment: A prospective study of a rst-admission cohort. Acta Psychiatrica Scandinavica, 94, 266271.

Copyright 2007 John Wiley & Sons, Ltd.

Clin. Psychol. Psychother. 14, 126134 (2007) DOI: 10.1002/cpp