Narrative Therapy: A Systematic Review of Efficacy and Effectiveness for Improving Biopsychosocial Outcomes in People With Cancer.

1
Background
According to the World Health Organisation, one in five people develop cancer in their lifetime [1]. Cancer and its treatment can severely impair biopsychosocial functioning, which incorporates biological, psychological and social health impacts [2]. Physical (e.g., pain, infertility, neurological dysfunction, fatigue), psychological (e.g., anxiety, fear of recurrence, and depression) and social challenges (e.g., change in social roles, stigma, and dependence on the health system and others) contribute to a cancer experience that is unique for each patient [3, 4, 5]. Due to the biopsychosocial burden associated with cancer, it is imperative to establish effective interventions to improve quality of life [6, 7].
Psychological intervention has been broadly demonstrated to reduce stress, depression, and anxiety, improve quality of life, and promote emotional well‐being in people affected by cancer [5, 8]. Narrative therapy (NT), developed by Michael White and David Epston, is a postmodern constructivist approach built on the understanding that individuals construct meaning via interpretive stories of their experiences and environment, which they hold as truth [9, 10]. The power of dominant societal discourses is pervasive to the forming of these stories and can result in disempowered, hopeless interpretations by those who are stigmatised by society [10]. NT utilises a collaborative approach where the therapist openly, non‐judgementally listens to a client's stories and uses questions to facilitate exploration. Narrative conversations do not typically follow a linear progression; rather a cyclical progression ensues which contains the following elements: (1) the therapist assists the client with mapping how the problem has been disrupting, dominating, or discouraging them, encouraging externalised descriptions of the problem; (2) during the sharing of problem‐saturated stories the therapist identifies moments of exceptional strength and competence and utilises these to inquire into alternative meanings for events; (3) the client is encouraged to construct a new, rich counter‐story associated with positive identity conclusions and alternative meanings discovered in therapy [9, 11]. It is imperative that a social environment which supports and affirms the emerging “new story” is facilitated to counteract negative discourses outside of therapy and utilise positive discourses to promote long‐term change [11].
A systematic review of psychosocial interventions for people with cancer found that multiple different interventions improved health and wellbeing [8]. Cognitive behavior therapy (CBT) based interventions were found to manage disease symptoms (pain, fatigue) and increase self‐efficacy, while meaning‐enhancing interventions and narrative‐based approaches resulted in a greater sense of meaning and quality of life (QOL). Therefore, the authors recommended that interventions be selected based on the specific needs of the patient. CBT is clinicians' predominant treatment of choice due to its strong empirical basis and proven effectiveness; however, CBT has been criticised for being less suited to diverse populations, for the perceived power differential between therapist and client, underpinning of European‐American values, insufficient consideration of social, cultural or environmental influences on beliefs, and minimal guidance regarding culturally competent delivery [12, 13]. Alternatives include supportive‐expressive group therapy, meaning centered psychotherapy, mindfulness‐based interventions and narrative therapy [8, 14]. While research into the effectiveness and optimisation of each of these approaches is important, NT is noteworthy in its cultural responsiveness [15], and connection with the oral traditions of many minority groups [16, 17, 18]. For example, a strong connection can be found between NT's story‐based practice and Aboriginal Australians' yarning, and shared meaning making [19].
Narrative approaches and perspectives have been utilised to some extent in narrative medicine and narrative nursing to support people with acute and chronic illness. In an exploration of narrative medicine, Charon [20] writes “without these narrative acts, the patient cannot himself or herself grasp what the events of illness even mean”. Chronic illness has been associated with dominant stories about pain, suffering, or being a burden, that detract from subordinate narratives that portray interwoven stories and multiple meanings [21]. Naming the problem (e.g., the illness, diagnosis, or symptoms) externalises the problem, removes blame or individual responsibility for the problem, and can lead to expansion of counter narratives beyond those of cancer or illness to incorporate agency, strengths, hope, dignity, and a richer sense of self [21, 22].
NT may also be well suited to individuals with cancer. Cancer is often associated with negative, anxiety inducing discourses surrounding the severity of the illness and challenges of recovery. The therapeutic model of NT endeavors to separate the patient from the problem and give them control of their story [23]. NT treats the diagnosis, and its assumptions, as another story over which the person has control, rather than as a part of who they are, providing an opportunity to differentiate themselves from the potential stigma associated with a cancer diagnosis [18, 24]. Peterson and colleagues suggested that it is imperative to revise stigmatised cancer narratives to promote coping and adaptation, and to avoid hopelessness [17]. Sharing personal narratives may help people with cancer reduce emotional distress through catharsis, finding meaning, and gaining understanding [25, 26]. Thus, the potential utility of NT in supporting people with cancer has yielded a growing literature [27].
Studies have reported positive therapeutic outcomes of NT for cancer affected populations [6, 28, 29, 30]. However, a published review incorporating all of these studies has not been conducted. Recent reviews have focused on patients with advanced cancer, narrative nursing, or qualitative studies reporting experiences of NT. Teo et al. [8] conducted a review of psychosocial interventions for advanced cancer patients that included two studies of narrative interventions, however these yielded non‐significant findings. In contrast, a meta‐analysis of eight studies examining narrative nursing interventions found significantly reduced self‐reported anxiety and depression in cancer patients [31]. Narrative nursing, similar to NT, involves registered nurses listening to and understanding patients' treatment stories and helping patients to find meaning during their regular nursing practice.
A systematic review of qualitative studies exploring the experience of cancer patients engaged in narrative practice yielded additional support for narrative approaches [27]. The results of seven studies were synthesized into the following themes: increasing trust in clinicians and their decision‐making, gaining hope, confidence, and power, establishing a shared social world when completed in a group setting, and facilitating reflection, self‐recognition and self‐realisation. Narrative practice was viewed favourably, however, the researchers acknowledged that patients had different preferences toward group/individual settings and that storytelling and reliving negative experiences can be a painful process.
Finally, a recent systematic review and meta‐analysis [32] assessed the effects of different psychological interventions for fear of cancer recurrence, a common and distressing concern for people affected by cancer. Through synthesising the results of 41 articles the authors found that NT was most likely to alleviate fear of cancer recurrence in the short term (SMD: −2.89, p < 0.05).
In summary, NT appears well‐suited for people affected by cancer because it offers a holistic approach which aims to promote hope and resilience by giving the individual control over their story and can be sensitively implemented with diverse populations [17, 23, 33]. However, systematic reviews and meta‐analyses have been unable to determine NT's efficacy and/or effectiveness for people with cancer due to a lack of high‐quality studies, a focus on qualitative outcomes, or being limited to a specific biopsychosocial outcome (e.g., fear of cancer recurrence).
This study aims to answer the following research questions.Is narrative therapy efficacious and/or effective in improving biopsychosocial outcomes for people with cancer?

What factors are associated with treatment outcomes?

2
Methods
This systematic review was developed in accordance with the PICO (Population, Intervention, Control, Outcomes) framework, PRISMA guidelines [34], and was pre‐registered with PROSPERO (ID:CRD42024557824).
2.1
Search Strategy
A comprehensive search strategy was developed in consultation with the research team and an academic librarian to identify studies reporting the efficacy or effectiveness of NT for people affected by cancer (see Table 1 for key terms). The following databases were included: Ovid MEDLINE, Ovid PsycINFO, Ovid Embase, and Web of Science. To identify unpublished studies and counter publication bias, the first 100 search results generated through searching for key terms in Google Scholar were reviewed for relevance. Although this does not completely offset publication bias, this method was chosen to increase search comprehensiveness while limiting the number of non‐relevant results that could reduce search precision, specificity and feasibility [35]. The reference lists of included studies were also screened for additional relevant studies.
Inclusion was not limited by types or presence of comparison groups because it may be unethical to employ random allocation or use comparison groups when studying a population affected by cancer. To acknowledge risk of bias, quality of study design was considered in the evaluation of outcomes of included studies using JBI critical appraisal tools [36]. Additionally, no specific search terms were entered for outcomes to capture all evaluations of NT; relevance in terms of biopsychosocial outcomes was determined at the screening stage.

2.2
Eligibility Criteria
Studies were included in the review if they investigated an NT intervention in a sample of adults (over 18 years of age) with cancer, reported biopsychosocial outcomes for people with cancer, and reported primary, quantitative data.
Initially, studies were excluded if the intervention was not implemented by a qualified psychotherapist or counselor, however, early in the screening process, it was determined that this criterion was too stringent. The criterion was changed to exclude interventions which were not delivered by an individual sufficiently trained in NT. To ensure that no relevant studies were missed due to this change, any study that had been excluded based on intervention implementation was rechecked prior to moving on to full text screening.
All cancer types and stages were included due to evidence of biopsychosocial risk throughout the cancer trajectory for all cancer types [5, 37]. Studies including participants with other illnesses were included if they reported outcomes for people affected by cancer separately. An intervention was considered an NT intervention if it included the following foundational elements: the sharing of client stories, externalised descriptions of the problem, and the construction of alternative meaning for events [11]. Interventions could be delivered in an individual or group setting and either face‐to‐face or online.
In terms of methodology, studies needed to report primary quantitative data to meet eligibility criteria. Considering the ethical complexity associated with randomised allocation to control conditions in cancer research, studies were not excluded based on research design; however, a critical appraisal of methodological rigor was conducted using the JBI scale, with greater weight being placed on high quality studies [36]. All biopsychosocial outcomes for individuals with cancer were included to capture all evaluations of NT's potential efficacy.
Studies were limited to those published in English to ensure accurate interpretation of data and implications within this review. White and Epston's first publication on NT was published in 1990. Therefore, this study did not include research prior to this date due to likely differences in psychotherapeutic approach.
In summary, studies were included in this review if they (1) evaluated an NT intervention, (2) included adults (over 18 years of age) with cancer, (3) reported primary quantitative data for biopsychosocial outcomes, and (4) were published in English from 1990 onwards. Studies were excluded if they (1) did not evaluate an NT intervention, (2) participants did not include adults with cancer, (3) primary quantitative data were not reported for biopsychosocial outcomes, (4) were published prior to 1990 or (5) were not published in English.

2.3
Study Selection Process
The study selection process is represented in the PRISMA flowchart (Figure 1). Search results were uploaded into Endnote software to collate results and remove duplicates. Independent screening was conducted in Covidence by two reviewers, first based on title and abstract, then full text articles, to determine whether studies met the inclusion criteria. Conflicts were moderated by a third independent reviewer in discussion with the research team.

2.4
Data Extraction
Data were extracted from each included study using a standardised data extraction form, after consultation with an academic librarian. Study author, year of publication, study design, country in which the research was conducted, sample characteristics, intervention characteristics, outcomes, outcome measures and limitations were extracted by the first author, and the second author was consulted where information or data were unclear. The JBI Critical Appraisal Checklist for Randomised Control Trials and the JBI Critical Appraisal Checklist for Quasi‐experimental Studies were used to assess the quality of each study respective to its design [36]. The JBI tools provide a comprehensive procedure by which to assess risk of bias in each study and this was consulted throughout the process. Any uncertainty regarding the JBI criterion or the risk of bias in a study was resolved through consultation with the research team.

3
Results
Our searches identified 2823 records. After the removal of duplicates, 1939 records remained. Screening of titles and abstracts resulted in 113 studies to be assessed for eligibility via full text screening. 105 of these records were excluded, including 31 which contained some elements of narrative therapy but were insufficient to meet the inclusion criteria (see Figure 1). Interventions described as ‘narrative interview’ were typically excluded because highlighted strengths were not integrated into a new story due to the limited session duration and frequency. Similarly, ‘narrative medicine’ and ‘storytelling’ interventions were typically excluded due to the lack of re‐authoring. The reference lists of the eight remaining studies were examined for additional potentially relevant studies based on titles. All titles that may have been relevant were identified and the abstracts and full text screened according to the inclusion and exclusion criteria. One additional study met the inclusion criteria. Therefore, nine studies were included in the review.
3.1
Data Synthesis
The included studies were not homogeneous in research design and presentation of results. Therefore, a meta‐analysis was not possible. A descriptive synthesis of the results of included studies to inform psychotherapists about the current evidence for using NT with people affected by cancer follows with effect sizes reported where available (see Table 2).

3.2
Characteristics of Included Studies
Nine studies published between 2006 and 2024 met the inclusion criteria. Two studies were conducted in China (n = 2), and the Netherlands (n = 2), with one study conducted in Iran, the USA, Spain, Japan, and Canada. Eight studies were peer reviewed journal articles, and one was a published doctoral dissertation [6].
The sample sizes ranged from 10 to 110 (M = 59.33, median = 66). Breast cancer was the most common cancer type (29.03%) followed by gynecologic/cervical (17.79%), oral (17.42%), colon (9.6%), and skin cancer (5.62%). Other cancers comprised 13.67% of the data with 1.87% of cancer type unknown. Included studies reported a wide range of cancer stages with one study including 7 participants who were in remission [6]. Participants were predominantly female (73%), and their age ranged from 18 to 82.

3.3
Intervention
While all included studies contained sufficient elements to be considered narrative therapy interventions, the way in which the client's problem was treated once it had been externalised via therapist prompts differed. Some interventions stated a clear objective that the client be separated from their problem(s) to establish an identity outside of their diagnosis (n = 5). Other interventions stated an objective of healthily re‐integrating the problem(s) (n = 4). For the purposes of this review the former studies are categorised as ‘separation’ interventions and the latter ‘re‐integration’ interventions.
3.3.1
Separation Interventions
Four of the five studies categorised as separation interventions were defined as ‘narrative therapy’ and one as ‘narrative nursing’ (NN). Where all other interventions were delivered or supervised by registered psychologists or counsellors, the NN intervention was delivered by nursing graduate students trained in the theory and practice of NT [38].
In terms of modality, three interventions were delivered individually and two were delivered in group settings [6, 39]. Both studies examining group interventions adhered to a quasi‐experimental design and the studies of individual interventions were RCTs. One of the RCTs provided a video chatting alternative to increase accessibility [29]. Interventions ranged from five to 12 sessions (M = 7.8) with approximately 30–90 min per session. Follow up data were collected by Rodríguez Vega et al. [40] and Sun et al. [29] at 24 and 4 weeks respectively, with the other separation studies reporting only pre and post intervention data. Four studies compared results with participants receiving usual care or on a waiting list. The remaining study compared the intervention group with participants receiving CBT, however this group was terminated early due to poor retention, limiting comparability [6]. Overall adherence rates for intervention and control groups of separation studies were approximately 90%.

3.3.2
Re‐Integration Interventions
Of the four studies categorised as re‐integration interventions three were defined as structured, or spiritual life review and one was defined as a meaning making intervention [41]. In terms of delivery, the meaning making intervention and telephone‐based life‐review were delivered by PhD candidates, with the other interventions delivered by clinical psychologists.
In terms of study design, two studies were RCTs [41, 42]. Lee et al. utilised a usual care control group and Wise et al. provided access to a control care website which omitted the life‐review components. Ando et al. utilised a pre/post design without a comparison group [43] and Post et al. utilised a quasi‐experimental, mixed methods, within groups design [44].
The study by Wise et al. [42] consisted of a 45–124 min interview prior to giving independent access to a ‘miLivingStory’ website with tips on revising/sharing narratives, a social network, health information and support groups. The remaining re‐integration interventions had four to eight face‐to‐face sessions of varying duration. Two interventions included follow‐up measures at nine‐ and 4‐months respectively [42, 44]. Overall adherence rates for intervention and control groups of re‐integration studies were 77% and 89% respectively.

3.4
Outcomes
Outcomes were primarily quality of life, psychological (e.g., mood, self‐esteem) or social (i.e., spirituality, stigma, social support). Quality of life measures were defined as biopsychosocial as these measures typically included physical, psychological, and social domains.
3.4.1
Biopsychosocial Outcomes
Quality of life was evaluated using validated measures in three separation studies [6, 38, 40]. Overall quality of life significantly improved in all studies. In addition, Rodriguez Vega et al. 40found significant improvements in pain, and all dimensions of functioning, and Perry [6] reported moderate effects in overall QOL. Additionally, one re‐integration study reported significant improvement in overall QOL [43]. These studies included participants with a range of cancer types, suggesting that improvements in QoL were not limited to specific cancer types. Collectively, these findings indicate significant improvement in quality of life and related outcomes for people with cancer who receive NT, with separation studies reporting more robust effects than re‐integration studies.
Other biopsychosocial outcomes included sexual performance and couple burnout, which incorporated both psychological and social elements, and improved following group based NT [39].

3.4.2
Psychological Outcomes
Mood, broadly defined, was assessed by measuring angry, anxious, and depressed mood, depressive symptomatology, despair, or optimism in the included studies. A study evaluating a telephone‐based life review found non‐significant results for angry and anxious mood, although a trend toward significance (p = 0.097) was reported for improvement in depressed mood at 2‐ and 4‐months [42]. When measuring depressive symptomatology, one study found non‐significant outcomes at all timepoints post‐intervention [40] while another reported a small significant improvement in despair, as well as a moderate improvement in ego‐integrity sustained more than 9‐months post‐intervention [44]. In terms of positive affect, optimism was shown to have a small significant increase in one study [41] and hope to re‐enter life increased in another [6]. All of these studies included participants with a range of cancer types prohibiting comparisons between cancer types. Collectively, these findings indicate mixed findings with some indications of small improvements in mood when people with cancer receive NT with re‐integration elements. In summary, negative moods such as anger, anxiety or depressive symptomology did not display significant improvement, however small significant improvement in positive mood and meaning was found.
In terms of psychological wellbeing, Post et al. [44] reported significant findings for 4/8 subscales, reporting small, significant improvements in purpose in life, self‐acceptance, inner‐strength, and relationship with a higher power, 3‐month post‐intervention. A narrative nursing intervention reported significant improvements in post‐traumatic growth and reduced reproductive concerns, which included psychological constructs such as worry, fear, anxiety and acceptance [38].
Self‐esteem and self‐efficacy were also evaluated. Lee et al. [41] reported small, significant improvements in self‐esteem and self‐efficacy via a re‐integration intervention. Using the same self‐esteem measure, the Rosenberg Self‐Esteem Scale, Sun et al. [29] also found significant improvement in self‐esteem over time with a moderate‐large effect (d = 0.739) resulting from a separation intervention. Therefore, both studies indicated increased self‐esteem with the separation intervention reporting larger effects.

3.4.3
Social Outcomes
Key social outcomes reported by individual studies included significant improvements in social support, perception of stigma, and spirituality [29, 42, 44]. Interestingly, three re‐integration studies found significant outcomes in terms of spirituality. They highlighted post‐intervention improvements in peace, orientation, spiritual growth, and mood [42, 44]. Sun et al. [29] reported significant improvements in stigma and social support following NT in people with oral squamous cell carcinoma who had undergone surgery. Overall, these findings indicate significant improvement in social wellbeing, in personal and community relationships, for people with cancer who have received NT.

3.4.4
Efficacy and Effectiveness
This review examined the results of five RCTs and four quasi‐experimental studies. The RCTs reported that NT led to significant improvement in multiple biopsychosocial outcomes for people with cancer under ideal, experimental conditions, however, efficacy could not be established due to the lack of high‐quality studies currently available (see critical appraisal) [29, 38, 40, 41, 42]. In terms of effectiveness, the findings of the quasi‐experimental studies extended these improvements to relate to applied clinical settings for many biopsychosocial outcomes [6, 39, 43, 44].

3.5
Critical Appraisal
The appropriate JBI critical appraisal tool for randomised control trials or quasi‐experimental studies was used to assess the risk of bias in each study (Tables 3 and 4) [36, 45]. Overall, all studies were scored as having a high risk of bias in at least two criteria. Although all RCTs demonstrated high or unclear risk of bias in the criteria related to blinding to treatment assignment, blinding in psychotherapy trials is rarely sufficiently reported [46]. Lack of blinding may introduce experimenter effects particularly when allegiance is present, but there is conjecture regarding its importance in assessing bias [46].
The studies by Lee et al. [41] and Wise et al. [42] had high risk of bias through not treating compared participants identically and measuring outcomes unreliably. In the former, unmonitored psychological support was received by 20% of the control group and in the latter, participants' treatment post‐interview was not discussed for either group. In both studies questionnaires were self‐administered without supervision, threatening reliability of measurement. Although, no studies conducted intention to treat analysis, all studies adequately described participants who dropped out of the study and analyzed how their absence may have influenced the results.
Two quasi‐experimental studies lacked control groups [43, 44]. The other quasi‐experimental studies also introduced potential bias by limited reporting on the treatment received by control group participants during the intervention period [6, 39]. Potential bias in various forms mean that the results of these studies should be interpreted with caution.

4
Discussion
Previous studies have noted the potential suitability of NT for people with cancer due to its holistic, culturally sensitive, personalised approach [11, 16, 17, 31]. To inform clinicians and patients about the efficacy/effectiveness of NT for people with cancer, this study reviewed and synthesised the existing evidence. The results suggest that NT leads to improvement in biopsychosocial outcomes for people with cancer. While RCTs provide evidence of NT for supporting people with cancer in multiple domains such as health and QOL, spirituality, appraisal of oneself, and personal and community relationships, firm conclusions cannot be drawn about efficacy due to the limited number of high‐quality studies available, with all 5 RCTs scored as having high risk of bias in multiple JBI criteria. Quasi‐experimental studies supported and extended these findings to demonstrate NT's effectiveness under more applied conditions in improving psychological wellbeing, meaning, spiritual growth and intimate relationships. Negative mood and depressive symptomatology did not significantly improve, however evidence for increased positive affect and mood was found.
This review identified notable factors which may influence the effectiveness of NT for people with cancer. Where many studies predominantly reported small effect sizes, Fallah and Ghodsi reported improvement in sexual performance and couple burnout in a group setting with moderate effects. These findings support Yang et al.’s27 findings that NT in a group setting establishes a shared social world which supports and affirms the ‘new story’. NT theory states that this support is imperative to the success of therapy [11]. Interestingly, another study undertaken in a group setting by Perry [6] found significant improvement in overall QOL with a moderate effect size, but non‐significant findings in the social sub‐category of QOL. Alongside high baseline social‐wellbeing, Perry noted that this study was completed over Zoom due to COVID‐19 restrictions, with the online nature of treatment potentially impeding social connection.
The COVID‐19 pandemic prompted an increase in development and prevalence of tele‐health, with the population being more familiar with this technology since 2020 [47]. A study of an online intervention conducted prior to 2020 reported more non‐significant outcomes compared with face‐to‐face interventions [42]. The authors of this study noted that “narrative interventions might well be more robust through in‐person interviews … [however] paying skilled interviewers/writers poses significant barriers to implementing narrative interventions in healthcare systems”. Therefore, face‐to‐face interventions may be difficult to implement on a broad scale due to cost and limited availability of experts in real‐world settings. Additionally, many cancer patients can experience household confinement due to their weakened immune systems and virtual groups could overcome this social barrier [48]. The findings of this review suggest face‐to‐face NT could have more favourable outcomes, however, future studies conducted in a post‐COVID‐19 climate may report improved outcomes, offering one way to overcome financial and illness barriers to implementation.
Intervention type also influenced NT outcomes. Studies delivering separation interventions reported more robust effects for QOL when compared with re‐integration interventions. Separation studies included participants with a range of cancer stages, though predominantly early stage, whereas re‐integration studies typically reported on participants with later stage and/or terminal cancer. While separation interventions were found to be more efficacious overall, re‐integration interventions may be more suitable for individuals with late‐stage cancer, with available research showing significant improvement for these individuals. These findings support Wang et al.’s49 previous review of life review interventions, which were found to be effective in improving biopsychosocial wellbeing in terminal or advanced cancer patients.
The long‐term effects of NT are currently unclear due to a lack of follow up measures in five out of nine included studies. A similar challenge was reported in Chen et al.’s32 meta‐analysis of psychological interventions targeting fear of cancer recurrence. The authors found NT to be highly effective in alleviating fear in the short term but found no data for long‐term effects. In this review, all studies which included follow up measures (ranging from 4‐weeks to 4‐months post‐intervention) saw sustained or increased improvement in biopsychosocial outcomes [29, 40, 42, 44]. These findings are promising in terms of long‐term effectiveness of NT in psycho‐oncology; however, additional longitudinal research with the inclusion of a well‐controlled comparison condition is needed.
A strength of this review is the diversity of populations found to experience improvement via NT. This supports Bacon's16 proposition that the non‐directive storytelling foundations of NT increase its cross‐cultural applicability. There are multiple barriers to recruiting diverse populations to oncology clinical trials, limiting evidence for intervention in culturally and linguistically diverse populations [50]. However, this review provides evidence for the applicability of NT within different cultural contexts.
Low retention rates and attrition is a noted difficulty in cancer research [51]. This is reflected in this review which found retention rates under 90%. Interestingly, Perry [6] achieved full retention in a NT group and very poor retention (25%) in a CBT comparison group described by participants as having “too much negativity”. Furthermore, Rodriguez Vega and colleagues [40] found poorer adherence in a usual care group (21.9%) compared with a NT group (2.8%) up to 12 weeks post‐treatment. These studies indicate some evidence for lower attrition and greater adherence with NT when compared with CBT or usual care.
4.1
Limitations
This review is limited by a lack of high‐quality studies, with all studies scored as having a high risk of bias in at least two criteria of the JBI [36, 45]. We attempted to mitigate publication bias through searching Google Scholar, however, other sources of gray literature were not searched. Therefore, it is possible that relevant studies were missed. Furthermore, the screening process identified five highly relevant non‐English studies which were unable to be translated due to time and resource limitations. Information retrieved from the translated abstracts supports the effectiveness of NT for people with cancer, reporting improvement in somatic symptomatology, fatigue, post‐surgery complications, post‐traumatic and depressive symptoms, and psychological distress [52, 53, 54]. Two of these studies compared NT to CBT, reporting greater effectiveness of NT in improving resilience, emotional regulation, pain, self‐efficacy, and hope, and greater online delivery efficiency [30, 55]. Professional translation of these studies in future reviews would provide a more complete synthesis of existing evidence regarding the effectiveness of NT.
Only one of the included studies provided comparison to an alternative therapy; CBT [6]. Furthermore, six out of nine studies lacked a control or comparison group. Common factors of therapy such as therapeutic relationship, empathy and collaboration have been documented to significantly contribute to therapeutic change [56]. Therefore, as most of the included studies did not include comparisons to participants who were receiving similar care, it is difficult to determine whether change can be attributed to the NT intervention or to common therapeutic factors. Future research would benefit from including comparison groups of alternative therapies alongside usual care controls to determine if NT is comparable or superior to other approaches for people with cancer. However, it is important to acknowledge the difficulty in recruiting participants to cancer trials and to the ethics of assigning people with cancer to a potentially less efficacious intervention [57].
Another limitation of included studies was that six out of nine studies did not confirm blinding of outcome assessors, potentially leading to detection bias where investigators with allegiance to NT may have scored its outcomes more favourably [49]. Where it is impractical to conceal the intervention, future studies could include independent assessors and compare outcomes with alternative therapies with common factors to avoid detection bias and Hawthorne effects.

4.2
Clinical Implications
Narrative therapy is increasingly utilised in clinical practice to holistically promote mental health; however, it is not yet widely used for people with cancer [11]. Despite the risk of bias in included studies and lack of active control conditions, the findings of this review present NT as a potentially effective treatment option for people with cancer to improve their biopsychosocial wellbeing. Furthermore, when selecting an appropriate intervention for people with cancer, clinicians should consider that NT may be less burdensome, as reflected in reduced attrition, and is effective across cultural contexts and stages of cancer due to its versatile, person‐centred approach. Current practice of NT in psycho‐oncology should consider providing group therapy options to create a supportive social environment, delivering therapy face‐to‐face where practicable, and selecting a separation or re‐integration approach, dependent upon client needs and stage of cancer, to optimise treatment outcomes.

5
Conclusion
This review indicates that NT is an accessible, culturally sensitive alternative for the diverse population of people with cancer. It is an equitable treatment option which can support various unique needs and improve various biopsychosocial outcomes. It is, however, limited by several factors which should be considered when interpreting its findings. Although the available evidence is not robust, agreement among studies with promising results indicate that it is worthwhile for future research to be undertaken to better inform the use of NT in psycho‐oncology.

Author Contributions
Mitchell A. Jones: conceptualisation, data curation (screening), formal analysis and investigation, writing – original draft, writing – reviewing and editing. Amanda D. Hutchinson: conceptualisation, supervision, writing – reviewing and editing. Liberty Barry: data curation (screening), validation.

Funding
The authors have nothing to report.

Ethics Statement
This study was exempt from ethics committee approval as it did not use primary data. All datasets on which the conclusions of this paper rely are publicly available and contain no material previously published or written by another person except where due reference is made.

Conflicts of Interest
The authors declare no conflicts of interest.