There has been considerable interest in the concept of the therapeutic relationship (also described as the therapeutic alliance or treatment alliance) since Freud's conceptualisation of transference and countertransference early in the 20th century. Part of the reason for this level of interest undoubtedly relates to the therapeutic relationship having been almost universally viewed as one of the most critical constituents of psychoanalytic, cognitive, narrative, solution-focused and schema therapy approaches (Reference Beck, Rush, Shaw and Emery2–Reference Young, Klosko and Weishaar6). However, the relevance of the therapeutic relationship goes beyond that of psychological interventions and is becoming more widely recognised as important in engagement and retention of people in biological treatments for disorders as diverse as oncology and diabetes Reference Pritchard-Jones, Dixon-Woods, Naafs-Wilstra and Valsecchi(7,Reference Anderson and Funnell8).

Compelling evidence is accumulating, which shows that the quality of the therapeutic relationship is a significant predictor of clinical outcome across a number of disorders (Reference Crits-Cristoph, Baranackie and Kurcias9–Reference Teyber, McClure, Snyder and Ingram11). For example, measures of therapeutic alliance, such as the Working Alliance Inventory Reference Horvath and Greenberg(12) and the California Psychotherapy Alliance Scale Reference Gaston(13), have shown the therapeutic relationship to be a reliable indicator of outcome in depression and affective disorders, with better relationships being correlated with better clinical outcome and reduced likelihood of drop out Reference Keijsers, Schaap and Hoogduin(10). Blatt et al. Reference Blatt, Sanislow, Zuroff and Pilkonis(14) analysed data from a large National Institute of Mental Health (NIMH) study on people with depression and concluded, ‘… therapeutic gain … is significantly influenced by interpersonal dimensions of the treatment process—by patient and therapist capacity to establish a therapeutic relationship’ (p. 1277).

Strauss and Johnson Reference Strauss and Johnson(15) found that a strong therapeutic relationship predicted fewer negative attitudes to medication, and led to people with bipolar disorder experiencing less manic symptoms over a 6-month follow-up. In another bipolar disorder sample, people who were more satisfied with their clinicians were found to adapt better to their diagnosis, coped better with symptoms and reported feeling less ashamed or angry than those who were less satisfied with their clinicians Reference Hirschfeld, Lewis and Vornik(16). Similarly, Frank and Gunderson Reference Frank and Gunderson(17) found that people diagnosed with schizophrenia, who were rated by their clinicians as having good therapeutic alliance, were less likely to drop out of treatment, had better medication adherence and obtained better functional outcomes. Among people attending treatment for substance abuse, the therapeutic alliance has been found to impact significantly on retention, completion of therapy and both clinical and functional outcomes Reference Bethea, Acosta and Haller(18).

Keeley et al. Reference Keeley, Storch, Merlo and Geffken(19) reported that the therapeutic relationship was also a significant predictor of outcome in obsessive-compulsive disorder, with ratings of the alliance by both patients and therapists being correlated with outcome in the expected direction. Keeley and colleagues suggested that the mechanism for this may be through ‘… persuasion and social influence’ (p.125), which impacted on the person's likelihood to complete tasks that in turn led to behavioural and cognitive changes. Fakhoury et al. Reference Fakhoury, White and Priebe(20) found that a strong therapeutic relationship predicted fewer rehospitalisations in people with severe mental health problems and who were new to a clinical service. The alliance therefore is a key clinical component that may modulate treatment outcomes in settings beyond that of formal psychotherapy.

In contrast to the studies reported earlier, a critical review by Meier et al. Reference Meier, Barrowclough and Donmall(21) found only a modest relationship between the therapeutic relationship and retention and outcome in people with substance use problems. This led to an editorial by Carroll Reference Carroll(22) hypothesising that the importance of the therapeutic relationship may vary across different disorders and possibly across different therapy models. Specifically, Carroll noted that the therapeutic alliance might be particularly challenging when working with people with substance abuse, as the clinician may also be the person who limits the patient's access to medication. Carroll suggested further that in cognitive-behavioural therapy (CBT), the therapeutic relationship might play less of a role in outcome than for some other models, as there are other significant aspects to CBT including skill acquisition and cognitive change. In contrast, for models that rely more heavily, if not solely, on the therapeutic relationship, if this is not strong, therapy is likely to be ineffective. Carroll concluded, ‘In effect, the presence of at least a minimally positive alliance may be a necessary, but by no means sufficient component of CBT and other effective therapies’ (p.267).

Although considerable research has focused on the importance of the therapeutic relationship, debate continues as to how it should be defined, measured and its precise mechanism of action on the therapeutic process Reference Elvins and Green(23). Green Reference Green(24) noted, ‘It has been argued that the therapeutic alliance has really little face validity as an entity and is largely a confound of patient ratings reflecting levels of psychopathology and clinician ratings reflecting bias about the effectiveness of therapy’ (p.430). Furthermore, despite perhaps appearing self-evident and intuitive to clinicians, the therapeutic alliance is often referred to somewhat loosely and non-specifically.

Bordin Reference Bordin(25) proposed a definition of the therapeutic alliance that remains one of the most widely used, and focuses on three key interconnected concepts; that of the bond between patient and clinician, the patient's willingness to undertake tasks and agreement on shared goals. However, although such factors have been commonly described as important components of the therapeutic relationship, identifying the relative impact of each of these, as well as that of other potentially important aspects, such as being able to focus on the patient's strengths, the development of the relationship over time, and even the use of touch Reference Shattell, Starr and Thomas(26), remains elusive.

The development of a good therapeutic relationship has been described as being more difficult with some disorders than others. For example, establishing a good therapeutic alliance can be particularly challenging in disorders such as anorexia nervosa Reference Sloan(27), in which a strong drive for autonomy and denial may be commonplace Reference George(28). Similarly, gaining and maintaining a strong therapeutic relationship can be challenging with people diagnosed with personality disorders Reference Zinbarg, Uliaszek and Adler(29) given that factors such as a previous history of abuse can lead to increased mistrust and hopelessness on the part of the patient.

It can also be difficult to establish and maintain a positive therapeutic relationship when working with people with bipolar disorder, for whom control and independence may be particularly important. Kahn Reference Kahn(30) acknowledged this difficulty stating that ‘Mania turns the therapeutic relationship upside down’, adding that mania is a disorder ‘… that a patient finds so pleasurable and a psychiatrist so frustrating that neither feels any zeal for talking to the other’ (p.230). Goodwin and Jamison Reference Goodwin and Jamison(31) also acknowledged this potential challenge in a quote by a patient who stated, ‘The endless questioning finally ended. My psychiatrist looked at me, there was no uncertainty in his voice. “Manic depressive illness”. I admired his bluntness. I wished him locusts on his lands and a pox upon his house. Silent, unbelievable rage. I smiled pleasantly. He smiled back. The war had just begun’ (p.746).

There are factors in current practice that can also marginalise the therapeutic alliance. These include clinical and administrative loads, the impact of managed care in some systems and the emphasis on evidence-based treatments, which might favour pharmacological management alone or the use of brief, focused psychotherapy. In such treatment settings, sessions can be pressured, restricting the establishment of an optimal therapeutic relationship.

Despite some of these potential challenges, Ackerman and Hilsenroth Reference Ackerman and Hilsenroth(32) identified a number of key components associated with a positive therapeutic alliance, many being drawn from the client-centred psychotherapy literature Reference Rogers(33). These include the therapist's expression of accurate empathy, and his/her ability to express themselves clearly, to connect with the person, to be flexible, warm, genuine, respectful, friendly, trustworthy, interested, alert, competent and to work collaboratively. Techniques that were identified as impacting positively on outcome, regardless of the psychotherapeutic model, included exploration, depth, identifying past successes, accurate interpretation, being active in therapy and acknowledging the patient's experience.

Much of the research on the therapeutic relationship has focused on therapist factors that influence the alliance. However, it is important to recognise that the therapeutic relationship is bi-directional, and that the patient has a considerable role in the development of warmth, trust, respect and openness. Specifically, the patients’ openness to describing their difficulties and disclosure of personal information has been identified as being correlated with treatment outcome Reference Keijsers, Schaap and Hoogduin(10).

Whereas a more detailed account is provided in Macneil et al. Reference Macneil, Hasty, Conus, Berk and Scott(34), there are a number of pragmatic tips that can increase the treatment alliance. These include:

• • Tailoring the intervention to the person's stage of recovery and not assuming ‘motivation’ to change, that the person will fully comply with medication, or necessarily even wish to attend regularly. For many disorders, the clinician may need to ‘earn’ these.

• • Enquiring as to the person's previous experiences of treatment, both positive and negative, as this can be extremely valuable in informing potentially successful interventions.

• • Taking time to understand the ‘whole person’ rather than focus solely on pathology. This should include learning about the patient's strengths, with Schwartz and Flowers Reference Schwartz and Flowers(35) suggesting, ‘… it is essential that we are as methodical in our search for our clients’ strengths as we are in searching for the correct clinical diagnosis’ (p.30).

• • Clearly understanding the person's own explanatory model of their experience, as this often forms the basis for establishing shared goals, which should be expressed in a manner concordant with the person's understanding and beliefs.

• • The motivational interviewing literature Reference Miller and Rollnick(36) emphasises the importance of language, and advises that labels, which have the potential to be stigmatising such as ‘schizophrenia’, ‘anorexia’ or ‘alcoholic’, should be avoided in favour of shared descriptions of the person's situation.

• • Encouraging realistic hope and optimism can be essential in creating a shared treatment agenda. As Meyer et al. Reference Meyer, Pilkonis, Krupnick, Egan, Simmens and Sotsky(37) noted, ‘… patients’ engagement in therapy depends on their expectations of treatment effectiveness’ (p.1051). However, a balance should be reached as expressing false hope can also be damaging to the alliance, and is likely to disrupt the clinician's credibility.

• • Clinicians actually caring about their patients may appear obvious, but can be overlooked. Yalom Reference Yalom(38) advised therapists, ‘Let the patient matter to you’ (p.26).

• • Within realistic and pragmatic boundaries, which need to be made clear, availability of the clinician in times of crisis, and continuity of care are keys in building the alliance.

• • Ackerman and Hilsenroth Reference Ackerman and Hilsenroth(32) noted, ‘… ruptures are an expected part of the treatment process …’ (p. 29), but that these can provide ‘… fertile ground for patient change and an opportunity for deepening the alliance’.

• • Finally, the clinician should be aware that engagement is an ongoing process and although appearing particularly significant in the initial phase of contact, its importance does not diminish after the initial sessions. Therefore, ongoing vigilance and effort is required to maintain a good therapeutic relationship.

In conclusion, despite some debate around the concept and its definition, numerous studies have suggested that the therapeutic relationship has an important impact on treatment for a number of disorders. Green Reference Green(24) stated further that the therapeutic relationship ‘… can both be measured reliably and shown to be equally if not more powerful than treatment type in predicting treatment change’ (p.425). While we live in an era that somewhat justifiably emphasises the importance of evidence-based models, it appears that the specifics of one of the most important ‘active ingredients’ of treatment, the therapeutic relationship, remains somewhat elusive. Green Reference Green(24) offered a cautionary note regarding this, stating ‘… explicit attention to the process and value of the therapeutic relationship will wither unless it too receives research attention’. However, this ongoing search for clarification of its components and mode of efficacy should not stop clinicians from making considerable effort to establish and maintain a positive therapeutic relationship with their patients. Despite pressures to provide generic, standardised and internet-based interventions, we should perhaps be mindful of Erickson and Rossi's Reference Erickson and Rossi(39) words when they noted, ‘Each psychotherapeutic encounter is unique and requires fresh creative effort on the part of both the therapist and patient to discover the principles and means of achieving a therapeutic outcome’ (p.234). Although Erickson and Rossi were clearly referring to psychological therapy, we would suggest that this principle can be equally applied to any therapeutic encounter.