With regard to improving the quality of care, a paradigm shift is currently advocated in medicine from disease-centred care towards patient-centred care. One of the fundamental requirements for practising patient-centred care is developing a therapeutic working relationship with the patient [1]. In the literature, this therapeutic working relationship is interchangeably referred to as therapeutic alliance, working alliance or helping alliance [1–4].

There is consensus about the three essential elements of the alliance based on Bordin’s concept of the Working Alliance [5]. The first two elements relate to the collaboration between the patient and clinician in reaching agreement on the treatment goals (goal dimension) and treatment strategies, or tasks, applied to achieve the goals (task dimension). The third element is the presence of an affective bond between the patient and clinician. An affective, emotional connection such as, for example, mutual trust and acceptance will favour collaboration and reaching consensus on treatment goals and treatment strategies [2–4, 6, 7].

The concept of the therapeutic alliance has become increasingly significant due to its positive and consistent association with treatment outcomes. This association was first identified in psychotherapy research [2–4] and more recently in research in the fields of general medicine [8–12] and physical therapy and rehabilitation [13–15]. To improve the quality of the therapeutic alliance, it is imperative for a clinician to understand which factors influence and enhance the quality of the therapeutic alliance.

A literature search (Additional file 1 and Fig. 1) was performed to identify the core components of the therapeutic alliance that determine its quality in relation to clinical outcomes in primary care settings. These were found to be: (1) empathy, (2) trust, (3) collaboration, (4) agreement on treatment goals and treatment strategies, and (5) patient-centred communication. Most research related to the association between empathy and trust with treatment outcomes, indicating the apparent relevance of the development of an affective bond as part of the alliance.

Empathy is a fundamental and crucial factor in the therapeutic alliance, and it is highly valued in relation to quality of care by both patients and clinicians [9, 16–19]. A recent systematic review found that an empathic behaviour on the clinician’s side diminishes anxiety and distress in patients and improves patient satisfaction, patient adherence and patient enablement in medical settings [9].

Trust also influences the affective bond and is defined as ‘the belief that the doctor is working in the patient’s best interest’. It is found to be associated with increased patient satisfaction, adherence to treatment and continuity of care [12, 20–22]. Surprisingly to some, clinicians’ behaviours that foster trust as perceived by patients are, above all, being comforting and caring, indicating that the affective characteristics of the clinician apparently provide the patient with more trust than do the technical competencies of the clinician [22].

Collaboration, or partnership, is considered another fundamental component of the therapeutic alliance [8, 23, 24] and is consistently associated with improved patient adherence and satisfaction with care [8, 25].

Although little research has been performed on the agreement on treatment goals and strategies component of the therapeutic alliance in primary care settings, it is also considered to be a fundamental part of the alliance [10, 11]. Positive outcomes are not very likely if there is no agreement on treatment goals and strategies between the patient and the clinician. Particularly in the case of chronic conditions, where self-management and self-efficacy are required from patients, consensus on treatment goals and strategies between the patient and the clinician have been found to improve treatment outcomes [26].

Finally, patient-centred communication is considered an essential and crucial component in the therapeutic alliance [11, 23, 27, 28], especially since it mediates the other core components of the alliance. Communication styles that help clinicians engage more with patients, and which facilitate patient participation, are associated with stronger therapeutic alliances [11].

Both understanding the concept of the therapeutic alliance and implementing the core components in clinical practice are essential in order to improve the quality of the working relationship. However, with respect to improving quality, it is important to know how patients and clinicians perceive the quality of the therapeutic alliance. Few studies have investigated this aspect, in particular in chiropractic care. The purpose of this study is, therefore, to describe the quality of the therapeutic alliance as perceived both by patients and by chiropractors in The Netherlands, and to determine whether patients and their chiropractors have similar perceptions of their encounter.

This study is the first to explore the quality of the therapeutic alliance between patients and chiropractors in The Netherlands as perceived by both patients and chiropractors. The measurement instrument used in this study assessed how both patients and chiropractors experienced collaboration in reaching agreement on treatment goals and treatment strategies to achieve their goals and how they perceived their affective bond. The results showed that, immediately after the third consultation, both patients and chiropractors perceived the quality of their working relationship overall to be very positive. Results on the quality of the three separate elements of the alliance showed that both patients and chiropractors slightly more often experienced the presence of the affective bond than collaboration in reaching consensus on treatment goals and treatment strategies. However, poor agreement was shown to exist between perceptions of patients and chiropractors of the quality of their alliance, both overall, and on the three elements of the alliance separately. Neither the chiropractors’ gender nor their years of experience, nor gender matching between chiropractor-patient pairs was shown to influence perceptions of patients on the quality of the alliance significantly.

Whereas the results of this study are encouraging in that patients felt that their interaction with their chiropractor was very good, they do not permit chiropractors to sit back and relax. The therapeutic alliance is a dynamic and developing process of collaboration between the patient and clinician and its strength often fluctuates during the course of care [2, 29]. Understanding how clinician and patient factors, such as their present mood, preoccupation with personal issues and severity of symptoms, as well as situational factors, such as excessive workload, lack of time, waiting time, and delays in improvement, may cause fluctuations in the strength of the therapeutic alliance over time is imperative for chiropractors. To maintain a positive alliance between the patient and the chiropractor, both parties have to demonstrate a commitment to collaborate during the course of the treatment for however long that takes [2, 29]. Important to note in this respect is that although previous research has confirmed a positive and consistent association between therapeutic alliance and clinical outcomes, the term ‘association’ does not indicate that there is a causal relationship between the two [36]. However, a recent study in the field of physical therapy and rehabilitation found a clear dose–response effect between the therapeutic alliance and clinical outcomes. Outcomes were better when interventions were combined with enhanced therapeutic alliance applications, compared to a limited application of factors that have been shown to enhance the therapeutic alliance [14].

The small differences between patients’ views and chiropractors’ views might disguise the fact that there was poor agreement between these perceptions. The results showed that chiropractors and patients had very different perceptions of the same working relationship. A lack of agreement between patient views and therapist views is consistent with the literature in psychotherapy. It was suggested that patient perceptions are more predictive for clinical outcomes compared to therapist perceptions, on the basis that patient perceptions were shown to remain more stable over time compared to therapists’ perceptions [4, 6, 29]. However, no studies were conducted to test this hypothesis.

One could argue that the poor agreement between patients’ and chiropractors’ views is the most important finding of this study for clinical practice. Chiropractors should be aware of this possible dissonance and should verify their views with the patient, especially concerning views on setting treatment goals and treatment strategies. Dissonance in agreement on treatment goals and treatment strategies can jeopardise patient management, especially in the case of chronic disease, where self-management and self-efficacy from patients are essential [26].

Somewhat surprisingly, the results in the present study did not demonstrate any difference between the varying years of experience as a chiropractor on patients’ perceptions of the quality of any of the elements of the therapeutic alliance. In a study in psychotherapy, which used the same version of the WAV-12 as the present study [29], interactions with respect to agreement on treatment strategies with therapists with over 20 years of experience were perceived significantly less positive compared to interactions with less experienced therapists (10–19 years). Besides a possible difference in education, it was postulated by the authors that a more experienced therapist may more frequently work on ‘an automatic pilot’, and spent less time and energy on discussing treatment strategies as to how to achieve the treatment goals with their patients. Working on the ‘automatic pilot’ may also be true for more experienced chiropractors. At the same time, as a result of patients being informed better and having their own ideas about health, experienced chiropractors may have consciously or unconsciously adapted their patient management to accommodate for more collaboration, thereby improving the therapeutic alliance.

Studies on empathy and communication in relation to the therapeutic alliance [24, 37] found that female physicians displayed more empathy and active collaboration than male physicians, both of which behaviours are considered to improve the strength of the alliance. Although there might be differences in displaying empathy and active collaboration between male and female chiropractors, this study showed that patients’ perceptions of the quality of the alliance did not differ for male and female chiropractors. This lack of difference might partly be explained by the fact that chiropractic care is considered as alternative medicine in The Netherlands and is still fairly unknown to many patients. Chiropractors will have to make an effort to explain and promote their care and might, therefore, be more motivated to engage and collaborate with the patient, independent of their gender. Educational differences between medicine and chiropractic might play a role as well. Historically, chiropractic has always advocated a holistic approach, in which patients are attended to in their whole identity as opposed to the former disease-centred approach in medicine. Chiropractors might, therefore, be more experienced in practising patient-centred care, which involves empathic and collaborative behaviour. Their training may thus diminish any gender differences in displaying empathy that was seen amongst male and female physicians. Furthermore, many chiropractors in The Netherlands decided to become a chiropractor because of their personal experience with chiropractic as a patient. It has been suggested that to improve empathic behaviour a clinician should be a patient himself [18]. Having started as a patient may be an advantage for many chiropractors, both female and male, in displaying empathy.

Some of these reasons could also explain the unexpected lack of difference in patients’ perceptions of the therapeutic alliance between gender-matched and unmatched chiropractor-patient pairs. Other explanations for this lack of difference might be that patients who have a clear preference for a female or male chiropractor can often be booked with the chiropractor of their choice since many chiropractic clinics have both female and male chiropractors working in the clinic. Consulting a chiropractor of the gender of their preference may give patients a more positive stance in the working relationship. Recommendation by an acquaintance or family member might also positively influence the patient’s stance in the working relationship at the start of the treatment, irrespective of factors such as gender or age of the chiropractor. Although this could be pure coincidence, the freedom of the patient to choose a chiropractor of the gender of preference may explain why almost two-thirds of the analysed working relationships in this study involved gender-matched chiropractor-patient pairs.

There are several limitations to this study, which potentially compromise the external validity of this study. Firstly, participation in this study was on a voluntary basis. Chiropractors who advocate patient-centred care and have good communication skills could be overrepresented in this study. Secondly, questionnaires were sent to the participating chiropractors in advance, which allowed them to read the questions in advance, and as a result, allowed them to give their best performance to the patient with respect to the alliance. Thirdly, perceptions were observed at one moment in time and were, therefore, influenced by the mood of the participant at that particular moment. In addition, patients who did not trust their chiropractor not to open the envelope and read their perceptions may have reported more positive perceptions. Furthermore, since all questions were scored in the same direction, participants may have failed to pay close attention to the questions and may have answered all items in the same way. The use of a questionnaire, which was developed in the field of psychotherapy, might be a further limitation. In the absence of a better alternative, the Working Alliance Inventory (WAI) has been proven valid and usable in physiotherapy settings [32]. Nevertheless, participants in the present study found some questions confusing or ambiguous. Noteworthy in this respect is that more than 10 % of patients did not complete all questions as requested. Particularly issues related to the quality of the affective bond between the patient and chiropractor were left unanswered. Although no feedback was requested from the participants, several patients reported questions exploring the quality of the affective bond to be irrelevant or inappropriate. And even though it could be argued that these questions were not relevant, some questions are considered less applicable in a chiropractic setting. Similar responses were found in a physiotherapy setting, and the authors suggested to develop a more conceptually sound measure of the therapeutic alliance for use in physiotherapy [32]. Furthermore, several chiropractors commented on the choice of response scale. The current options (seldom to always) were found inappropriate and not applicable to rating perceptions, and suggestions were made using options as to how much the participant agreed with each statement (strongly disagree to strongly agree). This modification was applied in a study on perceptions of the quality of the therapeutic alliance in a primary care setting, along with some modifications in the phrasing of the statements, and proved useful [10]. As in any observational study, confounders could not be fully controlled for in this study. Specific patient or chiropractor characteristics could have influenced the results. Some patients might have had severe symptoms, unintentionally influencing their perceptions more negatively. Female and male chiropractors might have treated patients with different levels of symptoms, different levels of cognitive ability, and of different ages, which could all have influenced perceptions on gender and experience. These differences, however, reflect daily practice and were, therefore, considered acceptable.

One important recommendation for further research is to develop a uniform instrument to assess the quality of the therapeutic alliance, with an appropriate response scale that could be used in primary care settings. This uniform instrument should be translated and validated for use in different countries. The second recommendation for research is to further investigate the level of agreement on treatment goals and treatment strategies to achieve these goals between patients and chiropractors. In the present study, only perceptions of collaboration in reaching agreement were studied. It would be useful for clinical practice to conduct a study examining to what extent patients and chiropractors agree on treatment goals and treatment strategies, as was conducted for patients with diabetes and their physicians and showed poor levels of agreement [26].

The results of this study showed that both patients and chiropractors perceived their working alliances very positive. Contrary to what was expected, no significant differences were shown to exist in patients’ perceptions in relation to the chiropractors’ gender or years of working experience.

The most important finding with respect to clinical practice was that poor agreement was found between the perceptions of patients and chiropractors on the same working relationship. This dissonance in perceptions must be given serious consideration by chiropractors. In order to develop, improve and maintain a positive therapeutic alliance, chiropractors should, during the course of treatments, continue to collaborate with the patient and frequently verify their perception with the views of the patient, especially with respect to determining treatment goals and treatment strategies to achieve these goals.