In this review, we set to explore the effect of working alliance on depression and anxiety in young people aged 14 to 24 years. Overall, a greater working alliance is associated with improvements in depression and anxiety, as previously shown in reviews focusing exclusively on adults [1, 2, 5, 34].
Although there is universal agreement on the importance of a functional WA, the exact mechanism by which a working alliance optimises reductions in anxiety and depression is currently unknown . It is theorised by both patients and therapists that improvements in the working alliance are associated with improvements in interpersonal relationships, self-esteem, positive coping strategies, optimism, adherence to treatment protocols, and emotional regulation [1, 2, 5, 15, 27, 37]. Improvements in negative psychosocial indices and treatment processes/factors attenuate anxiety and depression symptoms. Consequently, improved mental health can positively influence WA creating a positive feedback loop [1, 2, 5]. However, some studies did not find an association between working alliance and depression and anxiety . In recent research into psychotherapy with adults, there have been attempts to use advanced statistical modelling techniques to understand the WA clinical outcome relationship . However, heterogeneity in methodologies makes it difficult to conceptualise and model the association between WA and anxiety and depression [1, 2, 5]. For instance, in a systematic review by Baier et al. (2020), most studies retrieved (26/37) demonstrated that WA is a mediator to change. However, reverse causality and mediation by a third confounding variable could not be ruled out [1, 5].
Salient WA element
Improvements in bond and tasks were associated with improvements in interpersonal functioning [25, 35]. CMDs are sometimes associated with relationship problems; therefore, establishing a bond with the therapists can affect the client's social life . It is generally suggested that early WA is essential in the early stages of the therapeutic process; early bonding between the client-therapist is likely to improve the client's agreement on the task and subsequent adherence to treatment procedures by the client, which ultimately improves treatment outcomes . A functional WA improves the client's confidence and self-efficacy in building interpersonal relationships . Therapy sessions are the bedrock of social exposures and a source of social support that optimise clients' relational functioning . More critical, pre-treatment characteristics, including motivation, hope for change, and expectancy in therapy effectiveness, are precursors to forming a functional WA, which leads to improvements in clinical outcomes . Sentiments of our stakeholders' consultations support this proposition. One respondent said, "…I really had faith in my counsellor, they were very welcoming, and this increased my confidence in the treatment process….". In CBT, enhancing a client's ability to agree on tasks and assignments, including eliciting emotional engagement during therapy, is essential for forging a functional WA. However, despite forming a WA during sessions, if clients are not provided with coping skills, the mediation of WA in symptomatic relief is attenuated [2, 34].
In a naturalistic study, Webb et al. (2014) demonstrated that the task component was statistically associated with treatment outcomes after controlling for temporal confounders (patient expectations and prior symptom change) . That study demonstrates the importance of agreement on concrete tasks as being fundamental to changes in depression in CBT . The importance of tasks completion is further supported by Zelenchich et al. (2020) , who explored the effect of WA on anxiety and depression in youth with acquired brain injury. Facilitation by the therapists in completing tasks was linked to improved functioning and lower anxiety and depression . This is further supported by a systematic review of adult patients undergoing CBT for anxiety disorders; task agreement was more predictive of the WA therapeutic-outcome association, with bond/goals-outcome association equivocal .
Evidence is inconclusive regarding the requisite timing of working alliance on changes in mental health functioning . For example, some studies suggest that a more favourable early working alliance is associated with more significant symptom improvement [3, 5, 34, 14, 16, 21, 25–27, 30, 33], with others reporting a null association [2, 15]. The discrepancies have been attributed to differences in outcomes measures, the timing of WA assessment and differences in methodologies, i.e. sample sizes, heterogeneity in study participants, and study designs, amongst other methodological issues [2, 4, 5, 30, 33]. A meta-analysis exploring the WA therapeutic-outcome relationship in CBT for adults with depression revealed that early WA-outcome correlations are marginally lower than mid-and late assessments , thus the need for an early establishment of a WA to optimise treatment outcomes . Another meta-analysis also identified a reciprocal relationship between WA and symptom reduction early in therapy . Early WA was predictive of post-treatment outcomes optimised lower drop rates; this association was evident irrespective of baseline symptom severity and was optimised by greater levels of patients' engagement with treatment and treatment acceptance in the early stages of therapy . However, very few long-term studies have been conducted to assess the temporal WA therapeutic-outcome association. WA has been mainly studied using cross-sectional or clinical trials with short follow-ups [5, 15]. Hersoug et al. (2013)  explored the long-term effects of WA on 100 patients three years after receiving dynamic psychotherapy for anxiety, depression and personality disorders. This study showed that a functional WA was predictive of long-term changes in mental health outcomes. Furthermore, higher treatment expectations, less severe symptoms, and the ability to create mutually fulfilling relationships with others were predictive of a better functional WA . The temporal relationship between WA and treatment outcomes is not without controversy [4, 30, 33].
We also set out to explore if the therapist influenced WA. Greater professional experience is associated with better treatment outcomes and greater WA [17, 30, 31]. Our stakeholder consultations mirror outcomes from a study by Goldstein et al.  exploring comparability of anxiety/depression symptoms change, skills acquisition and WA between experienced and student therapists. In their study, TA was superior for clients treated by students than qualified therapists. Also, a study showed that videoconferencing is equally effective in treating anxiety and depression; a doctoral trainee psychologist was the therapist . Both alliance and skill acquisition were moderately correlated with therapeutic gains in changes in depression scores . However, our stakeholders' consultations were indeterminant; clients revealed that age was a potential determinant for establishing WA, with young people preferring to be treated by a similarly aged lay counsellor. A similar age counsellor was deemed likely to have the same experiences and relate more to a young person experiencing anxiety and depression. Other clients preferred to be seen by a more mature counsellor who could have more experience addressing the issues at hand. In problem-solving therapy and CBT, more experience appears helpful when the patient is still opening up; an experienced counsellor can use their clinical expertise to facilitate problem-solving in the client [17, 30, 31, 34].
The association between WA and anxiety/depression was the same across delivery modes, i.e., physical vs online therapy and individual vs group therapy [15, 40]. Evidence of the WA across physical formats is unequivocal in suggesting that a functional WA optimises in-person therapy outcomes, with evidence across digital platforms still evolving [1, 2, 4–6]. A pilot RCT (N=26) showed that videoconferencing clinically equalled in-person CBT, with client satisfaction and client- and therapist-rated WA comparable across the two groups . Andersson et al. (2012)  explored the association between WA and treatment outcomes in guided CBT in patients with anxiety, depression, generalised anxiety disorder and social anxiety disorder (N=174). The study showed high WA scores comparable to face-to-face therapies. They argue that a WA can still be formed in guided digital self-help despite a lack of physical contact over online interactions; the process of agreeing on goals and homework/tasks is essential for successful therapy outcomes . Unlike physical sessions, WA in guided digital therapy is a function of the clients' interaction with the therapist online and access to self-help materials/systems . However, there was no association between WA and clinical outcomes, despite the clients improving clinically; the null association require further research . Methodological limitations of this study, including a one-off measurement of WA and using an instrument developed for face-to-face therapies, could account for the null association, or maybe WA may not be an active ingredient for guided self-help modalities.
Most of the WA therapeutic-outcome association knowledge is derived from one-to-one therapy delivery. We also set to understand the effect of WA on group therapy in young people experiencing anxiety and depression. Evidence synthesized produced mixed findings. Group therapy was associated with increases in self-esteem, which had a moderating effect on both WA and depression . Group interactions, coupled with a secure working alliance, were associated with improved self-esteem and reduced depression . Furthermore, clients with more impaired relational experiences seem to benefit much more from group therapy, signifying a warm WA's potential impact on treatment outcomes . Group therapy among adults with mental health problems was also associated with an increase in WA in psychomotor therapies (body awareness and physical activity), with increases in collaboration the most salient predictor of changes in WA . However, a study by McEvoy et al. (2014)  exploring the relationship between interpersonal problems, WA, and outcomes following group (n=115) and individual (n=84), produced slightly different outcomes. In this study, individual therapy recipients reported greater WA pre-and post-treatment; the differences were statistically significant . Furthermore, in group therapy, severe pre-treatment anxiety/depression and interpersonal problems were associated with poorer WA and dropout than individual therapy . Some argue that when compared to individual therapy, the group therapy format may not be the most "conducive" platform for clients with severe pre-treatment interpersonal problems to form a functional WA, given their assumed difficulties to relate to other group members and the therapist(s) [2, 29].
A functional WA is considered active across psychotherapies [5, 33, 40]. Tschschke et al.  demonstrated that WA was essential in predicting clinical outcomes for behavioural, cognitive-behavioural, person-centred, and psychodynamic therapies. This proposition is further supported by systematic reviews and meta-analyses exploring the effects of WA on depression and anxiety for CBT [1–7].
Client and therapist characteristics
Patient optimism regarding the potential for improved symptoms is linked to better WA and more favourable treatment outcomes, underscoring therapists' need to build realistic expectations and optimism that therapy will be effective [2, 19, 30]. Also, patients with good attachment histories, adaptive attachment styles and developed social skills are more likely to forge good relationships with the therapists, thus improving WA . More critically, positive relations, characterised by an ability to develop a stronger bond between a patient and a therapist, are essential. It creates trust and safety, which spills over to the agreement of goals and subsequent completion of agreed tasks . Our stakeholders' consultations identified empathy from the therapist as an essential element across patients. However, evidence from a systematic review exploring the active ingredients of CBT in adult anxiety disorders produced mixed results . Furthermore, studies are needed to better understand critical windows for good WA to impact therapy outcome and any moderating effects on the empathy-outcome association . The exploration is essential given that a meta-analysis by De Re et al. (2012) demonstrated that therapists' characteristics hugely contribute to the WA formation regardless of patient diagnosis, research design, and WA measurement .
McLeod & Weisz (2005)  carried out a study to explore the relationship between WA and treatment outcomes in youth (mean age; 10.3 (SD 6.2 years) with anxiety and depression in an outpatient setting. Their study showed that better child–therapist alliance and parent–therapist alliance during treatment predicted greater reductions in internalising (anxiety and depression) symptoms at the end of treatment. Given that children rarely volunteer to engage in therapy, with parents usually deciding to get involved, a functional WA between therapist(s) and both parents and children is necessary for optimising treatment outcomes . Using a naturalistic study design, Webb et al. al. (2014) also explored the association between WA and changes in symptomatology in an inpatient setting. For patients with anxiety and depression (N=103) receiving combined CBT and antidepressants in a psychiatric?? facility . A functional WA was associated with a decrease in depression. Also, patients with optimism (greater treatment expectations) were likely to form greater WA, subsequently improving treatment outcomes . It seems reasonable to conclude that the WA-outcome association is independent of the setting in which treatment is provided.
Most studies have primarily focused on the relationship between WA and treatment outcomes for standalone psychotherapies. Stunk et al. (2012) explored the relationship between WA, adherence and symptom change in 176 randomised clients receiving combined cognitive therapy and antidepressants for depression in the US . A positive WA is associated with symptom change early in therapy; furthermore, only the task sub-scale was associated with symptom change. However, multivariate analyses showed that only the task subscale remained the statistically significant predictor after controlling for therapist skill and adherence to cognitive therapy. Taken together, the study showed the importance of agreeing on goals and provision of homework to influence both WA and subsequent symptomatic changes in combined therapy .
In relational dynamic psychotherapy and group CBT, the association between WA and treatment outcomes seems unclear. Some studies suggest that a functional WA does not seem to optimise treatment in patients with personality disorders and relationship problems [20, 21]. Patients with relationship problems may have challenges connecting with the therapist, which may cause a poor WA, and subsequent poor treatment outcomes . Also, clients with greater relationship difficulties are likely to be more dependent on their therapist; this may lead to challenges in developing their problem-solving capabilities, in turn influencing the ability to forge a functional WA and treatment outcomes [25, 29]. Conversely, the greater reliance on the therapist may cause an improved alliance, specifically the client and therapist bond .
Furthermore, patients with fewer relationship difficulties may be more realistic in treatment outcomes. They may have better appreciated how difficult it may be to attain any meaningful change . Among patients, a functional WA does not seem to play a huge role in treatment success; instead, other non-specific factors (e.g. adherence, symptoms severity) seem to influence treatment effectiveness . More research is needed to explore contexts where WA can be harmful or circumstances under which a functional WA can deter patients' functional recovery. Conversely, ruptures in WA can lead to decreased treatment expectancy, which may negatively affect adherence, thus ultimately reducing treatment effectiveness. However, the evidence concerning this is limited, and more research is needed for definitive conclusions .
Although our review suggests the positive impact of strong WA in treating depression and anxiety, the generalisation of our findings may be limited. First, we did not formally assess the risk of bias in each study. The scoping review aimed to summarise the relationship between WA and mental health outcomes in young people aged 14-24 years. Future systematic reviews and meta-analyses are warranted. Second, most of the studies were from high-income countries, and their applicability across different settings could be limited; we only retrieved a solitary study from Kenya . There is a need for context-specific studies to explore the effect of WA on anxiety and depression, given the potential influence of culture on WA, as evidenced by our stakeholder consultations .
Third, very few retrieved studies were exclusively done in young adults in the 14-24 age, which may potentially limit external validity. Fourthly, albeit the heterogeneous measurement in the WA , there is a need for psychometric evaluation studies to standardise WA measures from diverse perspectives, i.e., patient-, observer- and therapist perspectives .