Introduction

The core of the relationship between the helper and the helped is a strong alliance, identified in many empirical studies as a condition of effective intervention and change (Wampold & Flückiger, 2023) and discussed on numerous occasions as a treatment-supporting mechanism (Zimmermann et al., 2021). Oriented towards a goal pursued through specific tasks in the presence of an adequately strong mutual bond, the alliance is not a homogeneous construct (Bordin, 1979; Prusiński, 2022; Paap et al., 2022) and is subject to complex development depending on many determinants (Folmo et al., 2021). This heterogeneity provokes the question of the weight of its components. Particularly interesting is the bond component, understood as the affectively experienced professional relationship and engagement resulting from the patient’s and the psychotherapist’s experience of one another, ensuring agreement and cooperation on goals and tasks (Prusiński, 2022). The contribution of bond to the total alliance, especially with relevant differentiating variables taken into account, has still been only moderately explored (Seewald & Rief, 2023). Bond is reported as an element of the therapeutic alliance, but the degree to which it saturates alliance structure is rarely specified. The present study addressed this research gap.

The Mutual Bond as a Key Element of the Alliance

With the gradual recognition of the relational theory of psychotherapy (Bordin, 1979; Nissen-Lie et al., 2021) and—in a broader context—with the building of increasingly effective institutional support solutions for recovering patients (Tyler, 2021), much attention is devoted to the value of bond in what essentially begins as an asymmetrical relationship between the patient seeking help and the psychotherapist offering it (in other types of relationships in the health service, respectively: between the patient and the functionally diverse medical staff; (Paap et al., 2022).

Compared to the psychotherapist, who is equipped with knowledge and skills and supported through supervision, the patient is in a difficult situation. When struggling with their health problems, patients are aware of their limitations and experience doubts about the causes and solutions of these problems. It is often difficult for them to decide on their own what goal to set, what direction the change should take, and what array of measures should be applied. This asymmetry leads to personal uncertainty, which—intensified by the ignorance of the consequences of various decisions—is the patient’s frequent, if not permanent, aversive experience (Pérez-Arechaederra, 2019). Additionally, what patients often bring into their relationship with the psychotherapist is their current maladaptive behavioral and cognitive profiles of functioning, pathological interpersonal patterns (Tanzilli et al., 2018), and dysfunctional attachment styles (Zack et al., 2015), whose clash with the psychotherapist’s reaction may cause the patient’s confusion, distance, and uncertainty.

The therapeutic alliance is important because it enables trust and openness to the therapeutic process despite the patient–psychotherapist asymmetry. It enables coping with the difficulties arising due to the patient’s pathological characteristics and the frequently dubious quality of institutional support (Howe et al., 2019).

A high-quality bond allows for more effective work, signaling to the patient that the situation is secure, remedial, predictable, and controllable and assuring them that the support they are receiving is adequate (Van den Bos, 2018). This eliminates personal uncertainty. A high-quality bond means that the patient feels the therapist respects, appreciates, and understand them rather than merely the problem they have brought, and that the patient perceives this strongly enough to let themselves work confidently and effectively during the therapy (Zimmermann et al., 2021).

A bond built with attention to the patient’s signals, based on respect and trust, is crucial because thanks to it the patient feels that they are a participant in the psychotherapeutic process— included in and collaborating on actions and decisions. They are heard out and their opinions are considered; they can share their thoughts and discuss their maladaptive behaviors (a psychologically significant effect of voice; Tyler, 2021). Importantly, just like the first normal relationship enables the child’s development, the therapeutic relationship, including the secure bond component, becomes a stable basis from which the patient can take the risk and rebuild their dysfunctional thought and affect regulation patterns (Wallin, 2014).

It is also worth stressing that in contemporary models of health behavior the optimization of patient’s self-interest in the form of recovery (the distributive effect) is a factor less central for the patient than it was expected to be (Carroll et al., 2008; Tyler, 2021). Health improvement is important, but not decisive for engagement in the process of change and treatment. The importance of relational factors is highlighted (Mentovich et al., 2014). The presence and feeling of bond make it possible to fulfill the need for self-determination and autonomy, because in this kind of relationship the patient receives the psychotherapist’s signals enabling them to experience what they perceive as significant: a sense of agency and choice (van Prooijen, 2009). Bond is also of autotelic value, being a manifestation of reverence and approval, which are essential components of human identity (Burdziej, 2018). In a healthy bond, moreover, the patient can experience the psychotherapist’s personal commitment. This increases their willingness to invest their effort and resources and reduces their personal helplessness, because it assures them of a high chance of achieving the goal.

How Much Does Bond Contribute to the Structure of the Alliance?

The significance of the bond component in the therapeutic alliance has been theoretically substantiated (Bordin, 1979; Horvath, 2018), and its presence has been empirically supported many times (Paap et al., 2022). In empirical research, the bond factor is identified in the context of the alliance as a whole, as one of its dimensions associated with treatment outcomes (Moggia et al., 2022). Research has also revealed that especially the psychotherapist’s commitment to the therapeutic bond predicts the outcomes of psychotherapy (Del Re et al., 2021). The relationship becomes stronger especially when the psychotherapist is warm and friendly (Kornhaber et al., 2016). Different authors even point to warmth as an emotional component of the therapeutic alliance (Howe et al., 2019) and to the psychotherapist’s skills as an element of bond (Finsrud et al., 2022).

However, reports on what share bond actually has in the total structure of the therapeutic alliance as a non-specific curative factor are rare. Importantly, difficulties in indisputably estimating the contribution of bond to the total alliance sometimes stem from the lack of unambiguous empirical evidence of the contents of alliance structure. This is explained as a consequence of the fluid and unrestricted nature of the alliance (Horvath, 2018), the choice of different statistical methods of analysis only moderately allowing for comparisons of results, or the minimal replicability of studies on alliance structure caused by difficulties in replicating research samples homogeneous in terms of context variables (Paap et al., 2022).

Most studies confirm the presence of bond in alliance structure, consistent with Bordin’s model (Folmo et al., 2021; Killian et al., 2017; Prusiński, 2022). There is, however, also empirical evidence that supports one general dimension of alliance (Cirasola et al., 2020; Hsu et al., 2017—the study concerned early alliance between Sessions 1 and 3; Herrero et al., 2020) or a two-factor structure (Penedo et al., 2019) or does not provide clear information about the structure (Paap et al., 2022). Additionally, where bond is an extracted dimension of alliance, its share in the structure of the total alliance is often not reported directly (Killian et al., 2017).

Evidence of the magnitude of the contribution of bond to alliance structure can be based particularly on the results of exploratory factor analyses. Results from the analyses of simulated art therapy by Bat Or and Zilcha-Mano (2018) indicate that, of the three dimensions, bond was extracted with the lowest eigenvalue of 1.7 and explained the smallest percentage (12%) of the variance (compared, e.g., to 4.4 and 32% for tasks). In contrast, Soygüt and Işikli (2008), who examined psychotherapeutic patients reporting diverse life difficulties, found that bond explained a proportion of the variance (16%) similar tothat explained by goals (15%) and tasks (15%). In other analyses, concerning short-term cognitive behavioral psychotherapy of anxiety and mood disorder patients, Soygüt and Uluç (2009) extracted two factors of alliance, with bond contributing less to alliance structure (eigenvalue 4.79, explained variance 40%) than the main dimension of combined goals and tasks (eigenvalue 5.90, explained variance 49.19%). Therefore, although bond is an acknowledged factor in psychotherapy, it is difficult to clearly determine its weight. Moreover, empirical reports indicate that the quality of alliance—both global alliance and specific dimensions—varies depending on relevant context variables; for example, it changes as psychotherapy moves through successive stages: either increasing (Folmo et al., 2021) or having moments of rise and fall (Halfon et al., 2019). Explorations of the contribution of bond to alliance should therefore include relevant context variables.

The Structural Model of Alliance

The study was based on a widely known conception of the alliance, Bordin’s (1979) pantheoretical model. The term “pantheoretical” points to the universality of this model, whose elements are present in every therapeutic relationship. Therefore, according to its author and to researchers investigating alliance, this proposal allows for estimating the patient–therapist relationship in all psychotherapeutic modalities (Horvath, 2018).

In this conceptualization, alliance is called the working alliance, which highlights mutual agreement, communication, and cooperation in the patient–psychotherapist dyad. Alliance is understood as a triad: the goals agreed on, the tasks assigned, and the bond developed. With its emphasis on relationality, Bordin’s model is appropriate for assessing the contribution of bond to the therapeutic alliance.

The Present Study

The key problem addressed in this empirical study is the question of how significant bond is in the total therapeutic alliance. Although the status of bond as a dimension of alliance is seldom challenged in the literature, it is not clearly specified what share it has in the alliance structure. Despite the lack of empirical evidence clearly estimating the size of the contribution of bond to the therapeutic alliance as a whole (Paap et al., 2022), its presence is rarely questioned in the literature, and from the theoretical point of view bond occupies a unique place in the structure of the therapeutic alliance and is of special significance (Howe et al., 2019; Pérez-Arechaederra, 2019). In this study, the contribution of bond was analyzed in groups of psychotherapists and patients homogeneous in terms of relevant context variables. Most importantly, the early and advance phases of psychotherapy were distinguished.

Due to the significance of the alliance for treatment success, the following hypothesis was formulated: H1: The patient–therapist bond has a predominant contribution to the structure of both early and advanced professional therapeutic alliance or at least a high contribution to that structure, similar to the contributions of the other two components.

Method

Participants

The study included 415 psychotherapist–patient dyads. For the purpose of hypothesis testing, in accordance with the guidelines provided in the literature (Mellado et al., 2017) and with the methodology chosen (Prusiński, 2023), participants were divided into two groups according to the phase of psychotherapy.

The group in the early phase (Sessions 1–5; M = 3.23, SD = 1.08) included 147 psychotherapists: 86 women and 61 men, aged 27 to 71 years (M = 45.33, SD = 10.85). All psychotherapists had higher education, 89.8% of them were doing or had completed a postgraduate specialization course in psychotherapy. Additionally, 38.1% held a certificate from a psychotherapeutic society, and 76.2% had worked in their profession for more than 5 years. Psychotherapists represented the psychodynamic (25.1%), cognitive behavioral (27.2%), systemic (32.7%), and humanistic integrative (15%) approaches.

The individuals undergoing psychotherapy were 151 patients: 71 women and 80 men, aged 19 to 67 years (M = 35.77, SD = 10.39), most of them with higher education (79.4%) and living in towns with a population above 20,000 inhabitants (80.1%). Three-fourths of patients (73.5%) were in a romantic relationship.

The group in the advanced phase of psychotherapy (Sessions 6–72; M = 26.42, SD = 15.38; the values for this variable were the highest for the psychodynamic and systemic approaches, which accounted for a considerable proportion of the sample) included 208 psychotherapists: 139 women and 69 men, aged 26 to 63 years (M = 41.94, SD = 8.16). They had higher education, 95.2% of them were doing or had completed a postgraduate specialization course in psychotherapy. Moreover, 64.4% of the psychotherapists held a certificate from a psychotherapeutic society, and 53.8% had worked in their profession for more than 5 years. Psychotherapists worked in accordance with the psychodynamic (24%), cognitive behavioral (51.9%), systemic (13.5%), and humanistic integrative (10.5%) approaches. The subjects undergoing psychotherapy were 264 patients: 145 women and 119 men, aged 18 to 80 years (M = 33.46, SD = 10.99), most of them with higher education (42.4%), and lived in towns with a population above 20,000 inhabitants (83.7%). The majority of the patients (62.9%) were in a romantic relationship.

Treatment

The treatment administered was individual psychotherapy. Diagnosing the patient based on ICD-10 was the psychotherapist’s responsibility.

Patients in the early phase suffered from adjustment (25.2%), anxiety (14.6%), depressive, affective, and mood (20.5%), personality (33.8%), eating (2.6%), and addiction (3.3%) disorders. Sessions lasted 30 to 65 min (M = 51.59, SD = 6.32); in 76.8% of cases, they took place once a week.

Patients in the advanced phase suffered from adjustment (16.7%), anxiety (16.7%), depressive, affective, and mood (37.5%), personality (7.6%), eating (4.2%), addiction (16.2%), and posttraumatic (1.1%) disorders. Sessions lasted 40 to 90 min (M = 52.22, SD = 6.89); in 40.2% of cases, they took place once a week. The groups in which independent analyses of bond were performed were relatively similar in terms of many sociodemographic and treatment context variables.

Measurement of Variables

Therapeutic Alliance

Alliance quality was assessed using the full version of the WAI (Prusiński, 2022). The WAI has two versions: completed by the patient (WAI-PA) and by the psychotherapist (WAI-PT). Each version consists of 36 analogous items operationalizing Bordin’s construct of working alliance, rated on a 7-point Likert scale as true or untrue about cooperation in a given patient–psychotherapist dyad. The measure yields a separate score for each of the three subscales and the overall working alliance quality score. Each subscale consists of 12 items: 6 positive and 6 negative ones. The WAI also enables a reliable estimation of alliance through combined patient and therapist ratings (WAI-SUM), eliminating measurement distortions resulting from alliance being underrated or overrated by one of the evaluators (Prusiński, 2022). In this study, analyses were based on the WAI-SUM.

Measurement reliability is αWAI−SUM = 0.97 for the total score, and its values for the subscales are: αWAI−SUM−GOALS = 0.93, αWAI−SUM−TASKS = 0.93, αWAI−SUM−BONDS = 0.92. Confirmatory factor analysis (CFA) and the analysis of change over occasions showed that measurement using the WAI was valid (Prusiński, 2022).

Demographic Data and the Formal and Contextual Aspects of Psychotherapy

An elaborate demographic survey (DS) was administered to control for demographic (sex, age, education) and other relevant variables (psychotherapist’s modality, psychotherapist’s work experience, number of sessions, frequency of meetings, and session length). Patient and psychotherapist versions of the survey questionnaire were prepared. The latter included questions about the formal and contextual aspects of psychotherapy.

Procedure

Participants were recruited from private offices and health care centers. Expert sampling was applied, with processes selected to ensure the appropriate representation of different values of the relevant main and context variables. First, the researcher approached the psychotherapist to present the purpose of the study and the rules of participation. Then, with the psychotherapist’s consent, the patient was recruited; the aim of the study and the circumstances of participation were presented to him of her too.

All participants gave their informed consent, and the study was approved by the Ethics Committee at the author’s university. Participation was voluntary. The one-time measurement was always performed after a completed session. It began with alliance assessment, followed by a survey on sociodemographic and treatment context variables. Respondents received no remuneration for taking part in the study.

Data Analysis

Statistical analyses were performed in IBM SPSS 29 and IBM SPSS AMOS 29. First, using the internal consistency method, measurement reliability was assessed for the measures administered. Next, by computing descriptive statistics for the variables, the researcher determined the characteristics of patients, psychotherapists, and the treatment process. The hypothesis was tested through structural equation modeling (SEM) analysis.

Results

The Contribution of Bond to the Alliance Structure: SEM Results

The hypothesis on the contribution of bond to the structure of both early and advanced professional therapeutic alliance was tested using SEM. The analyzed structural model was not optimal, but it was an acceptable fit to the data, χ2(1182) = 3034,66, p < .001; RMSEA = 0.062, CFI = 0.84, TLI = 0.83, SRMR = 0.07.

All parameters of the variables in the model are subject to interpretation because all of them are statistically significant. The results of the analysis (see Table 1) indicate that in the early phase of psychotherapy the alliance was constituted to statistically significant degrees by the goals agreed on (β = 0.99, p < .001), the tasks assigned (β = 0.99, p < .001), and the development of bond (β = 0.89, p < .001). The value of standardized path loading for bond was only slightly lower compared to the other two components of alliance. Similarly, in the advanced phase of psychotherapy, the alliance was significantly constituted by goals (β = 0.99, p < .001), tasks (β = 0.99, p < .001) and bond (β = 0.96, p < .001). In this case, the contribution of bond was greater, and the standardized path loading practically did not differ from the remaining components of alliance.

Table 1 Structural model analysis: estimators of the contribution of each component to the structure of the therapeutic alliance
Full size table

The values of multiple correlation coefficient (R2) in the early phase of psychotherapy were the highest for goals (R2 = 0.99) and tasks (R2 = 0.99) and somewhat lower for bond (R2 = 0.80). As much as 80% of the variance in early alliance was explained by the development of bond. This is a high value, exceeding the accepted standards (R2 > 0.5), but the lowest among those considered, though moderately outlying. In the advanced phase of psychotherapy, the values of multiple correlation coefficient were as follows: R2 = 0.98 for goals, R2 = 0.99 for tasks, and R2 = 0.92 for bond. Likewise, bond explained a high percentage of the variance in alliance (92%), but lower than the remaining components. Importantly, the contribution of bond to explaining the variance in alliance increased between the early phase and the advanced phase.

Although no support was found for the contribution of bond to the alliance structure being predominant over that of the remaining components, bond did have a high contribution to the structure of both early and advanced professional therapeutic alliance, similar to the contributions of the other two components. The hypothesis was supported and accepted.

Discussion

This empirical study analyzed the contribution of an important factor regulating patients’ attitudes and behaviors to ensuring treatment effectiveness (Zimmermann et al., 2021)—namely, the contribution of bond to the total psychotherapeutic alliance. The source of variance was sought and the results were reported for two different phases of the psychotherapeutic process: early and advanced (Mellado et al., 2017). The study relied on two ratings of alliance (patient’s and psychotherapist’s), yielding a balanced alliance estimation (Prusiński, 2023).

Although it is pointed out in the literature that a positive and valuable bond is indispensable for a successful treatment process (Moggia et al., 2022), there is a lack of unambiguous empirical evidence of the magnitude of its contribution to the structure of the alliance between the helper and the person seeking psychotherapeutic help. Previous results vary, and the very presence of bond in alliance structure was not always supported (Cirasola et al., 2020). The results of this study provide clear and strong empirical evidence that bond is a crucial and permanent element of the relationship in the patient–psychotherapist dyad. Structural analyses revealed its presence both in the phase initiating psychotherapy and in the phase when the psychotherapeutic dyad engaged in increasingly intensive work. This may support the assumption that bond is a significant factor enabling the promotion of specific goals and engagement in tasks that make it possible to achieve these goals (Ekeblad et al., 2022).

Though slightly lighter in weight compared to the identical and very high weights of goals and task, the patient–psychotherapist affective bond in the early phase of psychotherapy, occupies an important place in alliance structure. The literature indicates that sometimes it is difficult to establish a strong bond from the very beginning of the psychotherapeutic process and that bond takes time to develop (Bordin, 1979; Ekeblad et al., 2022; Folmo et al., 2021); it turns out, however, that a good bond is of significance from the start.

Although the literature indicates that agreement on goals and the assignment of tasks are a process that takes time, and that goals themselves and the tasks that accompany them are often modified (Jennissen et al., 2020; Klajs, 2017), the question arises of for what purpose the value of bond is so prominent in the early phase of psychotherapy. Perhaps only the high value of bond makes it possible to effectively achieve the right goals and define the tasks. Or perhaps, apart from increasing the patient’s commitment to the newly discovered goals of psychotherapy, it also reduces the aversive experience of uncertainty (Pérez-Arechaederra, 2019). The size of the contribution of bond inspires the search for explanations of its value being so high, but this can currently be done only through speculations.

Importantly, when examining a different group of patients and psychotherapists, relatively similar in terms of controlled context variables but already in the advanced stage of psychotherapy, one will note not only that bond continues to make its high contribution to the alliance they build, but also that the weight distance between bond and the remaining two dimensions of alliance is smaller. This may indicate greater commitment to work on the understanding and clarification of goals and tasks in this phase of psychotherapy (Ekeblad et al., 2022).

The convergence of the weights of components supports the assumption of Bordin’s model (Bordin, 1979) that bond is important to the extent to which it supports therapeutic work (Zimmermann et al., 2021). Thus, in the advanced stage, when goals and tasks may have already been defined more precisely, there appears an appropriate value of bond supporting their achievement and performance. It should be noted that the convergence of weights was replicated and is similar to the distance between the components of alliance established in earlier empirical studies (Soygüt & Işikli, 2008; Soygüt & Uluç, 2009).

Finally, the estimation of high results for the bond component in the therapeutic alliance indicates one more important issue. Namely, in the patient–psychotherapist relationship, the burning need to understand the patient’s problem is always accompanied by a strong bond, which allows the patient to feel that the psychotherapist is not only close to the problem but also close to the patient. A high bond in the dyad is an important experience and a signal of trust in and respect for the specific person one has come to know (Del Re et al., 2021). The results of this study are evidence of the stability rather than fluidity of the contribution of bond to the structure of alliance regardless of the stage of psychotherapy (Horvath, 2018). They also provide clear guidelines on how to protect and improve the alliance: if bond is so important, then striving for its quality by building an empathic understanding of the patient with precisely communicated boundaries (Carroll et al., 2008) and by providing unconditional acceptance (Howe et al., 2019) will strengthen the alliance itself.

Limitations

Various limitations of this study should be noted. In future studies, samples should be larger to yield stronger empirical support for SEM structural models. It would also be important to track changes in the significance of alliance components depending on the stage of psychotherapy in the same sample. Even samples highly homogeneous in terms of context variables but not composed of the same subjects do not enable direct comparisons of results. It becomes necessary when identifying and analyzing the contribution of bond to take account of other potential moderators, such as psychotherapist’s work experience, patient’s pharmacotherapy, or patient’s disorder.