Background and Aims
The crucial role of the therapeutic relationship in the change process and outcome, as well as the therapist’s significant contribution to that relationship (Del Re et al.
2012), is increasingly acknowledged in psychotherapy research (Norcross and Wampold
2011a). Different patients require not only different treatments but also different therapeutic relationships (Norcross and Wampold
2011b). The personality of the therapist does influence the psychotherapeutic process. A recent systematic literature review (Lingiardi et al.
2018) confirmed what every clinician knows: the therapist’s subjective characteristics influence results of psychodynamic psychotherapists. According to this review, therapists’ interpersonal functioning, reflective and introspective capacities, and specific personality characteristics showed the strongest evidence of a direct effect on treatment outcomes. It can be assumed that these therapists’ variables were complementary (opposite) to the patients’ inabilities.
Decades of psychotherapy research suggest that patient–therapist match accounts for outcome beyond single patient or therapist variables. This issue touches upon the challenging question of the nature of matching. Accordingly, matching what does matter? Matching more objective or more subjective characteristics? Convergent match based on similarities or complementary match based on opposites? The research findings are inconsistent.
Most studies of patient–therapist match try to demonstrate that similar sex, socioeconomic status, ethnicity, or more generally, a shared cultural background with similar attitudes, values, and believes influence the early therapeutic relationship, session content, continuation rates, treatment satisfaction, and outcomes (e.g., Cabral and Smith
2011; Ibaraki and Hall
2014; Reis and Brown
1999). However, contrasting convergent and complementary patient–therapist match might be too simplistic. As early concluded by Beutler et al. (
1991), optimal pairings comprise therapists who share similar humanitarian and intellectual values with their patients, but have discrepant views of personal safety and the value of interpersonal intimacy and attachment.
Attachment to the therapist has been conceptualized as an important ingredient in the therapeutic relationship (Mallinckrodt
2010). Furthermore, the therapist’s attachment style was demonstrated to affect the patient and the therapeutic relationship (Slade
2016). Following this line of research, several studies focused on matching the patient’s and the therapist’s attachment styles, most of them supporting the complementarity hypothesis (e.g., Bruck et al.
2006; Petrowski et al.
2011). Other studies concluded that convergent attachment patterns further good outcomes for severely disturbed or high-avoidant patients (Farber and Metzger
2009; Wiseman and Tishby
2014).
The complementarity hypothesis was also supported by early studies of the patient–therapist match on personality variables, assessed using the MMPI and the Omnibus Personality Inventory (Dougherty
1976), applying the type A and type B dichotomy (Anderson and Carter
1982), or the Check List of Interpersonal Transactions (Kiesler and Watkins
1989). Other studies supported the similarity hypothesis, such as matching cognitive structures (measured by the Interpersonal Discrimination Test; Hunt et al.
1985), similarity on the Five-Factor Personality Inventory (Coleman
2006), or fitting personality types (assessed with Self-Directed Search; Taber et al.
2011).
Despite different conceptualizations and operationalizations, the studies of personality match hint at the dimensions of mode of relating/affiliation and autonomy/self-boundaries. These are also the polarities of experience, central to Blatt’s (
2008) empirically anchored theory of personality development, psychopathology, and the therapeutic process. According to his double helix model, the main mechanism of change in psychoanalytic therapy is reactivation of normal developmental processes and new internalizations in the context of the therapeutic relationship (Blatt et al.
2008; Luyten and Blatt
2013). This involves alternating sequences of
gratifying involvement (attachment) and
experienced incompatibilities (separation, or some disruption to a gratifying relationship), leading to more mature forms of relatedness and a more integrated sense of self (Behrends and Blatt
1985; Blatt and Behrends
1987). Development of the sense of self (the introjective line) leads to increasingly mature levels of interpersonal relatedness (the anaclitic line) that, in turn, facilitates further differentiation and integration in the development of the self (Blatt and Luyten
2009; Luyten et al.
2013). Psychological health involves both a meaningful identity and meaningful attachments, i.e. a balance between differentiation and relatedness, autonomy and intimacy. Still, most individuals, also within the normal range of psychological development, have an inclination towards either the relatedness dimension or the self-definition dimension. In contrast, different forms of psychopathology reflect an exaggerated and distorted preoccupation with one or the other of these developmental dimensions (Luyten and Blatt
2013). The anaclitic configuration is connected with difficulties in close relationships and attachment anxiety, while the introjective configuration is connected with excessive demands for achievement and perfectionism, and with attachment avoidance (Luyten and Blatt
2013). Disturbances in each developmental line may lead to the same symptoms, but require different treatments. Introjective depression, based on the sense that “I am a failure,” responds to classical psychoanalysis, with the therapist as a listener, helping to elicit growth in an independent sense of self. Anaclitic depression, based on the feeling that “I am not worthy of love,” is effectively treated by a more assertive therapist, guiding the formation of relationships (Blatt et al.
2010). Nonetheless, following Blatt’s model, the goals of psychoanalytic psychotherapy include enhancing the patient’s capacities for both being together and autonomy, both relationships and self-other differentiation.
However, this may be not only a matter of therapeutic technique, but also of the therapist’s personality orientation. Heinonen and Orlinsky (
2013) studied the interplay between therapists, personal identities, theoretical orientations, and professional relationships. They concluded that the therapist’s professional self, underlying interactions with patients, is rooted in the general self-experience in close personal relationships. However, it is possible that different aspects of the therapist’s personality are actualized with different patients. Accordingly, we still need studies focusing on the therapists’ personality configuration as actualized in relationship to their specific patients.
Based on these issues, the objective of the present study is to examine how the therapist’s personality configuration (anaclitic orientation on relatedness/affiliation or introjective orientation on self-definition/autonomy) is actualized, manifests itself and interacts with the patient’s personality configuration early in the therapeutic relationship. Furthermore, we examine the associations between different patterns of patient–therapist matching (convergent or complementary personality configurations) and outcomes at termination of psychotherapy with young adults. How are the therapeutic dyads distributed at baseline between different matching patterns? Do the different groups of matching patterns show different outcomes at termination (in terms of symptom reduction and of changes in the inner representational world)? Following previous studies (Werbart et al.
2017), we expected more pronounced improvements in the complementary groups.
Discussion
The present study is, according to our knowledge, the first attempt to explore matching of patient and therapist personality configuration in terms of Blatt’s (
2008) two-polarities model in relation to outcome. Most therapists were assessed as predominantly anaclitic, whereas most patients were assessed as introjective. Half of the therapeutic dyads fitted in the convergent pattern of matching and the other half in the complementary pattern. Analysis of between-group differences showed large effect sizes, in terms of symptom reduction at termination, in the two convergent groups (both oriented on relatedness or both oriented on self-definition) and small effect sizes in the two complementary groups (dyads with opposite orientations). More than twice as many patients in the convergent dyads showed reliable change at termination, as compared to the complementary dyads. Changes in the developmental level of representations of self and others displayed another pattern, with large effect sizes for anaclitic patients and small to medium effect sizes for introjective patients. Thus, these results indicate that the patient–therapist personality match at pre-treatment does matter for symptom reduction at termination, whereas changes in the developmental levels of representations of self and others seem to be dependent on the patient’s predominant personality configuration pre-treatment. This is consistent with the double helix model, claiming direct link between the polarities of self-definition–relatedness and the dimensions of differentiation–relatedness in representations of self and others, both in personality development and in the therapeutic process (Blatt et al.
2008; Luyten and Blatt
2013).
Accordingly, a previous investigation of changes in the anaclitic–introjective personality configuration following psychotherapy in the same patient group (Werbart et al.
2017) showed better balance between relatedness and self-definition post-treatment in the initially anaclitic patients, whereas this improvement was not significant in the initially introjective patients. However, no significant between-group differences could be found on outcome measures. The present study demonstrated better outcomes in terms of GSI and D–R Mother in the merged convergent group than in the merged complementary group. A comparison of these two preliminary studies suggests that the lack of differences in outcome between the anaclitic and the introjective patients might conceal consequences of the patient–therapist personality match at the outset of psychotherapy. Consequently, our study confirms that regarding the patient–therapist dyad as a distinct variable, as suggested by Silberschatz (
2017), may result in new and clinically highly relevant findings, not attainable when regarding the patient and the therapist variables separately.
Still, our study suggests a potential within-therapist effect, besides the effects of patient–therapist match. Six of the 10 therapists with more than one patient were assessed as having different personality configurations with different patients. These therapists could be more relatedness-oriented with anaclitic patients and more self-definition oriented with introjective patients. Thus, different personality configurations could be actualized in the same therapist early in the relationship with different patients. This implies that the therapists’ personality configuration, as it manifests itself in relationship to their patients, might differ from their configuration in private relationships (cf. Heinonen and Orlinsky
2013). Zilcha-Mano (
2017) differentiated between more stable, “trait-like,” and more interaction-related, “state-like,” tendencies to form satisfying relationships with others, the latter making alliance therapeutic. Our study demonstrates, accordingly, that the therapist’s “state-like” interpersonal stance plays a role. The therapists’ capacity to adjust their balance between relatedness and self-definition early in the therapeutic process to the patients’ “trait-like” personality configurations was connected with better outcomes. Furthermore, it is possible that the therapists who did not alter their interpersonal stance despite their patients’ personality configurations are those for whom their “trait-like” personality configuration had pervasive influence on their ways of being with patients, potentially leading to less optimal outcomes.
Contrary to our expectations, the present study showed more pronounced improvements in the convergent group of patient–therapist matching personality configuration, thus supporting the non-complementarity or similarity hypothesis (Taber et al.
2011; Wiseman and Tishby
2014) and contradicting the complementarity hypothesis (Bruck et al.
2006; Kiesler and Watkins
1989; Petrowski et al.
2011). However, our study does not answer the question why convergent personality configuration in therapeutic dyads is connected with more pronounced improvements. We can only speculate that matching interpersonal stance facilitates early “moments of meeting” (Stern et al.
1998), thus promoting working alliance and fruitful therapeutic work. Furthermore, our study is limited to the initial match. Hypothetically, the patient–therapist matching is not a static variable but an ongoing process. Our finding that some therapists can better than others adjust their interpersonal stance to the patient’s personality configuration, is consistent with Mallinckrodt’s (
2010) suggestion that effective therapists can offer their patients progressively changing series of relationships, creating a corrective emotional experience that promotes more adaptive functioning. Accordingly, a study of anaclitic and introjective patients in psychoanalysis suggested that the psychoanalytic technique has to be adjusted to the anaclitic and introjective patients’ different needs and defenses, in order to reactivate developmental processes (Werbart and Levander
2016). Thus, the complementarity hypothesis might be still valid later on in the therapeutic process, in the middle phase of working through, and has to be tested in further studies.
Strengths, Limitations and Future Directions
The present study contributes to our knowledge of the role of the patient’s and the therapist’s personality configuration in creating productive therapeutic dyads. Rather than studying match on directly observable, explicit characteristics (such as gender, socio-economic status, ethnicity, etc.) or the participants’ own subjective experiences of match, we focus on implicit, “deep” personality features, highly relevant for psychotherapy process, and in consequence, for treatment outcomes. Furthermore, no previous study investigated the therapists’ personality configuration (orientation on relatedness or self-definition) as actualized early in the therapeutic process. The need of studies focusing on effects of the patient–therapist match as a single variable has been repetitively stressed in psychotherapy research (cf. Taber et al.
2011; Zilcha-Mano
2017). Our approach, regarding the patient–therapist dyad as a distinct variable (Silberschatz
2017), made it possible to discover differences in outcomes between the convergent and the complementary therapeutic dyads, not visible in a previous comparison of outcomes for anaclitic and introjective patients (Werbart et al.
2017).
The main limitation of this study is its small sample size. The low statistical power limits the possibility of finding genuine between-group differences, at the same time as the positive findings are less likely to be true positive. Thus, this research should be conducted at larger scale. As a consequence of the naturalistic design, we cannot claim that the observed between-group differences were real effects of patient–therapist matching. Furthermore, the sample is not representative of young adult outpatients in other forms of long-term treatment or of more homogenous diagnostic groups. The treatment duration varied greatly and could potentially affect the outcome. In our statistical analysis, we entered duration as a covariate, thus controlling for its effects. Studying young adults in psychotherapy raises the question of maturational processes in this dynamic period of life and of spontaneous improvement. However, we may assume that these effects should be equal in the two groups.
Future studies, with larger number of therapists treating several patients, can address the between-therapists differences in adjusting their relational stance to their patients’ personality configurations. Lacking recordings of therapy sessions we could not study potential changes in the therapists’ adjustment (or maladjustment) to their patients’ changing dynamics between relatedness and self-definition during the course of treatment. The interviews with patients and therapists varied in quality and were not especially designed for PMAI ratings. Session recordings could be used for PMAI expert ratings, together with self-rating measures of the two dimensions of personality configurations, for example applying the Dysfunctional Attitude Scale or the Big Five Personality Test.
Clinical Implications
One implication of our study is the importance of early understanding of the patient’s predominant interpersonal stance in terms of focus on relatedness/emotional bond or self-definition/autonomy. Furthermore, it is the therapist’s task to observe which aspects of the therapist’s “trait-like” interpersonal stance are actualized and staged early in the therapeutic relationship. Therapists might expect to feel and react differently depending on whether their characteristic focus on relatedness or self-definition is convergent or complementary to the patient’s personality configuration. Blatts’ two-polarities model enables therapists to reflect on the dynamics in the therapeutic relationship, starting from the interplay between the patient’s and the therapist’s personality configuration. Being aware of both participants’ relational stance, the therapist might form an idea of the patient’s capability to establish an attachment to the therapist, the patient’s potential reactions to therapeutic boundaries and separations, and to different kinds of interventions. Based on this knowledge, it can be possible for the therapist to adjust to the patient’s way of relating and safeguarding own self-boundaries, thus potentially facilitating the change process and enhancing the outcome. The therapists might be more able to work through ruptures in the working alliance and potentially prevent negative outcomes by monitoring, from the beginning, the patient’s ways of being with the therapist, as well as their own ways of being with the patient. Measures of the patient’s and the therapist’s personality configurations can be included in systems for monitoring patient progress and for feedback-informed treatments (Lambert et al.
2005; Miller et al.
2015). The interplay between the patient’s and the therapist’s personality configurations together co-creates the unique situation within which change is possible. Our study suggests that this should be considered already in the beginning of therapy.