BACKGROUND
The concept of a “working alliance” (WA) originated in psychoanalytic theories that emphasized the significance of the relationship between therapist and client and its essential role in the therapeutic process. According to Bordin (Reference Bordin1979), this concept is a fundamental and universal variable across all types of therapies and in the treatment of various problems. It is comprised of three elements: (1) goals—a mutual understanding between the therapist and the client regarding the realistic goals and the target of the treatment; (2) tasks—an agreed-upon contract regarding the relevant and beneficial behaviors and cognitions required to achieve treatment goals; (3) bonds—the development of a personal and emotional relationship, based on mutual trust, acceptance, and confidence (Bordin, Reference Bordin1979; Horvath & Luborsky, Reference Horvath and Luborsky1993; Horvath & Greenberg, Reference Horvath and Greenberg1994). These elements determine the quality and strength of the working alliance and have a significant part to play in treatment outcomes, particularly during the initial stages of treatment, when the framework of therapy is established (Bordin, Reference Bordin1979; Horvath & Symonds, Reference Horvath and Symonds1991).
The WA concept emphasizes the interactive and mutual features of the relationship between patient and therapist. It assumes that the nature and quality of this relationship depend on the degree of concordance and agreement between both sides of the interaction and their willingness to collaborate toward a joint purpose, as well as on personal factors, such as their relationship with significant others, internal attributes, and models of attachment (Bowlby, Reference Bowlby, Greenspan and Pollock1989; Horvath, Reference Horvath and Horvath1994; Van Walsum et al., Reference Van Walsum, Lawson and Bramson2004).
Given its mutual aspects, studies have measured the effects of the WA on various treatment outcomes from the perspectives of both patient and therapist (Petry & Bickel, Reference Petry and Bickel1999; Busseri & Tyler, Reference Busseri and Tyler2004; Fitzpatrick et al., Reference Fitzpatrick, Iwakabe and Stalikas2005; Tryon et al., Reference Tryon, Blackwell and Hammel2007; Bethea et al., Reference Bethea, Acosta and Haller2008; de Bolle et al., Reference de Bolle, Johnson and De Fruyt2010).
In the medical setting, the effects of the WA on various physiological and psychological outcomes have been examined as well. However, to the best of our knowledge, they were solely investigated from the patient's perspective. For example, a study with cardiac patients during a rehabilitation program demonstrated that a strong WA with the staff, particularly agreement on tasks and goals, had a positive effect on cardiorespiratory fitness, weight loss, and self-reported mood (Burns & Evon, Reference Burns and Evon2007). Another study demonstrated the association between a WA with the physician and treatment attendance, as well as number of days of abstinence in medical-based interventions for alcohol dependence (Dundon et al., Reference Dundon, Pettinati and Lynch2008). Research in neurological patients demonstrated a significant association between a WA and patient satisfaction with their medical care and their adherence to the treatment regime (Fuertes et al., Reference Fuertes, Boylan and Fontanella2009). Among advanced cancer patients, a stronger WA was associated with better quality of life and greater acceptance of illness (Mack et al., Reference Mack and Block2009), as well as greater perceived social support, less grief due to cancer-related losses, and stronger adherence to medical regime (Trevino et al., Reference Trevino, Fasciano and Prigerson2013).
The aim of the current study was to emphasize the interactive nature of the WA by measuring the concept from both sides of the patient–medical staff interaction. This preliminary investigation may assist in estimating the different perspectives of the interaction and reveal the degree of agreement and understanding involved.
METHODS
Study Sample
The sample included physicians, nurses, and advanced cancer patients from the Division of Oncology at Rambam Health Care Campus in Haifa, Israel. Inclusion criteria for the patients were age over 18 years, willingness to sign an informed consent form, ability to complete the study's questionnaires, and Hebrew literacy. No exclusion criteria were utilized for medical staff.
Measurements
The Working Alliance Inventory–Short Revised (WAI–SR) is a self-report questionnaire that evaluates the three components of the working alliance (goals, tasks, and bonds), consistent with Bordin's model (Reference Bordin1979). It is a short version of the WAI (Horvath & Greenberg, Reference Horvath and Greenberg1989) and is composed of 12 items (Tracey & Kokotovic, Reference Tracey and Kokotovic1989; Hatcher & Gillaspy, Reference Hatcher and Gillaspy2006; Munder et al., Reference Munder, Wilmers and Leonhart2010). The Hebrew version of the questionnaire has been found to be valid and reliable (Rotman, Reference Rotman1999). In order to adjusting it to the medical context, the research team rephrased several items, similar to previous studies (Fuertes et al., Reference Fuertes, Mislowack and Bennett2007; Pegman et al., Reference Pegman, Beesley and Holcombe2011). The medical staff completed the “therapist” form of the questionnaire, regarding their various interactions with patients (rather than a specific patient). The patients completed the “client” form in two different versions, one regarding their physician and the other regarding their nurse, in order to estimate these different interactions.
Statistical Analysis
In order to reveal differences in the WA between the subgroups of physicians and nurses (senior and trainees), the Mann–Whitney nonparametric test was performed. The Wilcoxon paired nonparametric test was performed to reveal differences in the WA of patients regarding physicians and nurses. ANOVA with a Bonferroni post-hoc test was performed to reveal differences between the three study groups (physicians, nurses, and patients). A reliability test was carried out, and Cronbach's alpha was calculated. Data analysis was carried out using SPSS 21 statistical software. A p value less than 0.05 suggested statistical significance for all outcome measures.
RESULTS
Some 32 physicians completed the study, 15 of whom were interns who had been working for less than 10 years in oncology. Half were men (n = 17), with an average age of 43.67 years. A total of 39 nurses completed the study, 17 of whom had worked less than 10 years. Most were women (n = 28), and their average age was 44 years. In addition, 52 advanced cancer patients completed the study. All were on active chemotherapy treatment that had lasted 2–8 months. Most participants were male (65%), with a spouse (71%) and with at least one child (88%).
The reliability analysis is detailed in Table 1. Among physicians, the total reliability was average (0.653)—lowest in tasks (0.339), and highest in emotional bond (0.723). Among nurses, the total reliability was high (0.868), lowest in tasks (0.446), and highest in emotional bond (0.702) as well. Among patients, both versions, the physician and the nurse, were high (0.939 and 0.946, respectively).
Table 1. Reliability analysis
The data analysis (Table 2) revealed that physicians rated their overall WA with patients as relatively high (5.6), with senior physicians rating it higher than interns (5.8 vs. 5.5). All rated the emotional bond highest. No significant difference was found between seniors and interns in their perception of the working alliance.
Table 2. Median scores on the WA and its subscales
Nurses rated their overall WA with patients as relatively strong (5.3), with senior nurses rating it higher than trainees (5.8 vs. 5.3). All rated the emotional bond highest. A significant difference was found between seniors and trainees on all WA scales.
The comparison between the rating of the WA by physicians and nurses revealed no significant difference. Patients rated their WA with physicians and nurses as strong (6.56 and 6.67, respectively), as well as all its subscales. No significant differences were found between the WA and its subdomains of treating physicians and nurses (WA, p = 0.56; goals, p = 0.89; tasks, p = 0.34; emotional bond, p = 0.57).
The comparison between the rating of the WA by physicians and patients revealed a significant difference on all its subscales. The same results were demonstrated with nurses.
A comparison between the three study groups revealed a significant difference between physicians and patients on the WA (0.06), goals (0.03), and bond (0.02), which were rated higher by patients. Nurses and patients were significantly different on the goals scale. No significant difference was found on the tasks scale.
DISCUSSION
The aim of this preliminary study was to explore the concept of the WA from the different perspectives of the patient–medical staff interaction in order to emphasize its mutual aspects and to characterize these interactions in terms of mutual agreement and understanding. To the best of our knowledge, this is a distinctive investigation of the concept in the medical setting, since it was solely measured from the patient's point of view.
The findings demonstrated a difference in WA rating between senior staff members and trainees, both among physicians and nurses. This can be explained by the higher level of experience among senior members, which contributes to their personal and professional confidence. However, it can also indicate their difficulty in relating to personal doubts and difficulties, unlike trainees, who might feel more comfortable talking about these issues, given their level of training and experience (Moreau et al., Reference Moreau, Goldgran-Toledano and Alberti2004).
The medical staff rated the “bonds” subscale highest, while patients rated “goals” at the highest level. Previous studies have demonstrated that dealing with patients' emotional and personal issues is highly challenging for staff members (Maguire, Reference Maguire2002; Levin et al., Reference Levinson, Lesser and Epstein2010); hence, this finding can indicate their need to compensate for their difficulties by overestimating the emotional aspects of the interaction. On the other hand, it can also indicate the growing understanding regarding the importance of the relational factor in medical care and the trend toward patient-centered medicine (Mead & Bower, Reference Mead and Bower2000).
Comparing the three study groups, a significant difference was demonstrated between physicians and patients. Patients rated the WA, its goals, and bonds higher than did the physicians. This difference was observed in previous studies in the psychotherapy field, with clients generally giving higher ratings to the WA than therapists (Bachelor & Salame, Reference Bachelor and Salame2000; Ogrodniczuk et al., Reference Ogrodniczuk, Piper and Joyce2000; Cecero et al., Reference Cecero, Fenton and Nich2001; Hilsenroth et al., Reference Hilsenroth, Peters and Ackerman2004; Fitzpatrick et al., Reference Fitzpatrick, Iwakabe and Stalikas2005; Tryon et al., Reference Tryon, Blackwell and Hammel2007). The current finding can indicate patients' willingness to represent the interaction with their physicians from a positive and ideal perspective, given the importance they ascribed to it, especially for advanced cancer patients (Hillen et al., Reference Hillen, de Haes and Smets2011).
Studies in the field of psychotherapy have demonstrated that convergence between client and therapist perspectives on the WA were found to be associated with a higher-quality therapeutic process and better therapeutic outcomes (Rozmarin et al., Reference Rozmarin, Muran and Safran2008; Marmarosh & Kivlighan, Reference Marmarosh and Kivlighan2012). In addition, a positive rating of the WA by the therapist was related to better treatment outcomes (Petry & Bickel, Reference Petry and Bickel1999; de Bolle et al., Reference de Bolle, Johnson and De Fruyt2010). Future studies should investigate whether these correlations exist in the medical setting as well, and if they can predict better outcomes for the oncological treatment.
CONCLUSIONS
This preliminary study compared the concept of the WA between advanced cancer patients and medical staff. Its findings demonstrated different perspectives in these interactions. Given previous studies in the field of psychotherapy, the current findings may have significant implications on various treatment outcomes. Future studies should investigate the concept to better understand its interactive aspects within the medical setting.
LIMITATIONS
We are aware of several limitations of the current study. First, it included a small sample of patients and medical staff from a single hospital department. Second, the WAI was not completed by pairs (one patient–one nurse/physician), which limits the possibility of comparing the different perceptions of the same interaction. Finally, the questionnaire had low reliability among physicians. This can be attributed to their level of concentration while completing the questionnaire, but it can also raise a question regarding the validity of the concept of the WA, originated from the theoretical framework of psychotherapy, in the medical world.
SIGNIFICANCE OF THE STUDY
An awareness of the working alliance in patient–staff interaction may improve the quality of treatment given to patients who are confronting cancer.
CONFLICTS OF INTEREST
The authors state that they have no conflicts of interest to declare.
