Diverse paradigms and stories: mapping ‘mental illness’ in athletes through meta-study

ABSTRACT

This meta-study systematically appraises and synthesizes research into athletes’ experiences of mental illness. Our critical review of 37 studies conformed to the meta-study structure of meta-theory, meta-method, and meta-data analysis. We also produced a meta-synthesis of findings to deliver new insights into athlete mental illness. Athlete accounts of mental illness pertained to experience of the following: depression, eating disorders, gambling addiction and substance-related disorders (alcohol and drugs). Following a critical interrogation of original articles’ theory, method, and findings, we noted a general lack of methodological coherence (congruence between philosophical stance, theoretical position, and methodology). Through the process of a thematic synthesis, we developed 4 new themes: origins of certainty and ambiguity, a gradual sense of decline, mental illness as a threat to identity, and constructing recovery stories. Athletes drew upon dominant illness discourses to construct mental illness and recovery experiences. Our results provide us with an understanding of how mental illness and recovery were experienced within an elite sport involvement. We recommend future research embraces more diverse methodologies and authors ensure a strong alignment between guiding philosophies and methodological approach.

KEYWORDS:

• Mental health in sport
• meta-synthesis
• research philosophy
• qualitative methodology
• athlete well-being

Introduction

Mental health in sport research is currently amidst a growth boom that shows little sign of relenting (Poucher et al., 2019). Whereas naïve notions of ‘mental toughness’ once rendered athletes protected from mental health issues (Bauman, 2016), an increasing body of evidence suggests otherwise. For example, numerous high-profile elite athletes have disclosed experiences of mental illness across a range of media outlets (see Parrott et al., 2019). As well as personal accounts in the popular press, empirical studies show that athletes can, and do, experience mental illness. Conditions such as depression, anxiety, and eating disorders are well documented among elite athlete samples (Rice et al., 2016). Elite sport offers unique challenges which may heighten the risk for the development and maintenance of such conditions in athletes. For example, contextual factors such as injury (Gulliver et al., 2015), performance failure (Hammond et al., 2013), and career termination (Gouttebarge et al., 2015) have been associated with poor mental health in elite athletes. Although prevalence evidence is conflicting, some studies show up to 46% of elite athletes experience at least one mental health problem (Gulliver et al., 2015).

Several narrative systematic reviews on the mental health of elite athletes have been published in the past decade which have advanced knowledge in important ways. Reardon and Factor (2010) take a diagnostic perspective, concluding that major life events, inside and outside of sport, place athletes at similar risk of mental illness to the general population. Rice et al. (2016) reviewed primarily prevalence and risk factor studies on mental health in athletes, with the major finding, again, being elite athletes experience a comparable risk of mental illness to general populations. Rice et al. (2016) noted a lack of methodologically rigorous studies calling for more medically informed intervention studies. Continuing the trend, Souter et al. (2018) produced a narrative review on mental health symptoms and risk-factors specific to male athletes. These reviews make important contributions to understanding the extent of mental illness within sport and the presentation of symptoms across various diagnosed conditions. Despite these valuable insights, the exclusivity of such a medical approach has often been critiqued for its understanding of mental illness as a detached concern with observable symptoms, illness categorizations, and associated treatments; at the expense of understanding individual experience (Smith, 2014).

The dominance of medicalized conceptions of athlete mental illness has filtered through to various position statements addressing best practices for managing mental health in sport (for a review and meta-synthesis of these, see Vella et al., 2021). These statements focus heavily on clinical presentation of symptoms and associated interventions, without acknowledgment of how athletes understand their mental illness experiences. Vella et al. (2021) acknowledged this gap and argued for the inclusion of athlete voices in the process of developing guidelines and recommendations for best practice. The absence of more meaning centered insights, the kind typical in qualitative research, is troubling. The embedded nuances and complexities of mental illness in athletes, the very issues that inform therapy, are missing from these management guides. As such, the abundance of (partial) position statements narrowly prescribe what athlete mental illness is and what can be done about it.

An alternative and complementary approach to medical constructions of athlete mental health might be found within sport psychology’s shift towards more interpretive ways of knowing (Ryba et al., 2010). Qualitative knowledge is essential for understanding human experience as it allows for an in-depth exploration into the social reality of a person, groups, and cultures, as well as the behaviors, perspectives, and experiences of people’s lives (Sparkes & Smith, 2014). In contrast to survey-based methods, qualitative approaches are more fitting with the highly personal journey of mental illness and its recovery, however, defined. This journey has been described as a personal subjective process rather than a diagnostic outcome (Drake & Whitley, 2014). Unlike the neat, objective categorizations of medical psychology, the mental illness experience is subjective and shaped by personal meaning making (Watson, 2012). Nevertheless, medical and interpretive approaches can work together in a mutually helpful way to increase knowledge and support insights within the field. More interpretive work can work with and build upon existing medical approaches.

Recognizing the value of interpretive work, some scholars have explored how athletes socially and culturally construct mental illness experiences. For example, qualitative accounts have revealed gendered constructions of mental illness (Busanich et al., 2014) which highlight how female athletes construct their disordered eating through feminized understandings of an eating disorder, whereas male athletes use masculinized terms to discuss food and the body. Narrative and autobiographical accounts have also highlighted how regulatory practices imbedded within elite swimming culture impact how athletes make sense of mental illness (Jones et al., 2005; McMahon & Dinan-Thompson, 2008), and how these meanings persist into retirement (Cooper & Winter, 2017). Conducted across various sports and underpinned by various epistemologies, these studies move beyond the identification of risk factors and prevalence rates to demonstrate the complexity of mental illness in elite sport.

Despite these rich and varied contributions to knowledge, there is yet to be any form of rigorous review of this qualitative research. This study will address this gap and synthesize existing qualitative research on athlete mental illness. We will adopt meta-study and meta-synthesis, popular approaches in both sport psychology (e.g. Tamminen & Holt, 2010; Williams et al., 2014) and nursing (e.g. Kaite et al., 2015). Previous meta-studies demonstrate how the use of qualitative synthesizes can consolidate experiences within a given field to create a comprehensive picture of a given phenomenon. With regards to qualitative insights into mental illness in sport, a critical review is both timely and necessary if such work is to inform practice recommendations. We sought to answer the following research questions: (1) What is known about athlete mental illness from qualitative research, and how is this knowledge constructed? (2) How have the researchers’ chosen theoretical and methodological frameworks guided interpretations of findings? (3) Based on an original synthesis of existing qualitative findings, what are athletes’ experiences of mental illness in sport?

Methodology

Given that this is a meta-study of mental illness in sport, it is important to identify our own understanding on what mental illness means according to our philosophical positioning. In line with the ontological position of relativism, we understand mental illness and related concepts (e.g. ‘mental health concern’; ‘mental distress’; ‘disorder’) to be socially constructed. We are aware that this dominant scientific discourse to describe human conditions relies heavily on the core assumptions of the medical model and an objective view of reality (Rhodes & Conti, 2016). However, we have not used these terms with the intention of reducing participant experiences to a diagnosis, nor are we suggesting that these individuals possess a human deficit, as the medical perspective can imply. Due to the challenges in defining ‘mental illness’ (see Wakefield 1992 for a critique of mental illness definitions), we want to emphasize that the use of these medicalized terms within this article are a means to broadly group people with experience of perceived disruption to psychosocial health. That experience is constructed by various actors such as study participants or the study researchers.

Study design: meta-study

A meta-study is a systematic method that appraises and synthesizes the qualitative data on a given topic through the analysis of its theory, methods, and findings (Paterson et al., 2001). This method seeks to construct new understandings and identify research gaps regarding a phenomenon of conceptual interest. Whereas other forms of qualitative synthesis focus exclusively on the results of included articles, a meta-study also critically appraises their underpinning research processes (Paterson et al., 2001). A rigorously executed meta-study is not only concerned with what qualitative findings are, but how they were produced. Critical interrogation precedes meta-synthesis, equipping the reader with the knowledge to form quality judgements on existing findings and on those of the meta-study itself. In the following sections, we describe the multiple methodological steps involved, including the retrieval of research papers; data extraction; meta-theory; meta-method; meta-data analysis; and meta-synthesis.

Retrieval of research papers

The first author conducted a scoping search which consisted of a brief literature search on Google Scholar to determine whether sufficient qualitative literature existed on athlete mental illness (Armstrong et al., 2011) (see Table 1 in the Supplemental files). In line with recommendations (Soilemezi & Linceviciute, 2018), the first author collaborated with an academic librarian to carefully refine the search strategy and terms used, as well as select appropriate databases (MEDLINE; PsychARTICLES; PsychINFO; SPORTDiscuss; and Scopus). The search protocol in each database was set to search titles, abstracts and keywords. No parameters were set on publication date with articles gathered from ‘earliest’ to the date of the first search: ‘November 2019’. The search was then repeated in January 2021 to provide an up-to-date overview of the literature, and also acted as a means to reduce the likelihood that potentially relevant articles had not been mistakenly eliminated. Three categories of search terms were used (a) mental illness; (b) sport; (c) qualitative (see Table 2 in the supplemental file for a full list of terms).

Screening and selection of research papers

The search yielded 24,134 results. Following the screening of titles, 196 articles remained with potential relevance. Two members of the research team screened abstracts, eliminating a further 155 articles. Full text reading of 42 articles resulted in a total of 30 articles included in the synthesis. Seven studies were added upon a bibliographic screening of identified articles. This resulted in 37 articles being included (see Figure 1 in the supplemental file).

Inclusion and exclusion criteria

Included papers had to be peer reviewed and an English language publication. We also developed the following questions as criteria to include/exclude papers:

1. Did the study include sufficient qualitative data? The inclusion of studies with limited qualitative data may ‘threaten the integrity of the final meta-study’ (e.g. a lack of participant quotes) (Paterson et al., 2001, p. 42).

2. Did the article include first-hand athlete accounts of mental illness? We excluded perspectives of coaches, parents, and others operating within an athletes’ social sphere.

3. Did the research involve competitive level athletes? Solely competitive athletes (see Swann et al., 2015, for definition) were included as these athletes have different pressures and demands associated with competing in sport at a higher level (which could potentially influence mental ill health) rather than those participating in sport at a recreational level (e.g. coach and teammate pressures, retirement, etc.). Therefore, we excluded articles focused on mental illness experiences of those involved in leisure or recreational sport. We also excluded studies that explored sport and exercise as a mental illness therapy.

4. Did the study focus on mental illness experiences? Although personal experience of mental illness was essential, we adopted a broad conceptualization to avoid overlooking potentially important insights. Specifically, mental illness could be clinical or sub-clinical, and either self-identified or clinically diagnosed.

Data extraction

After reading and annotating articles to gain initial familiarity, we developed a customized data extraction spreadsheet to organize our formal analysis. Within this spreadsheet, we noted each study’s aims; location; sample characteristics (i.e. gender, age, sport, mental illness); methodology (i.e. philosophical assumptions, data collection and analysis methods, interpretive framework); and any other relevant information.

Data Analysis

Meta-theory analysis

Meta-theory analysis involves critical inspection of how chosen theoretical frameworks influence research findings. We appraised how chosen paradigmatic and/or theoretical frameworks (declared or assumed) affected insights into athlete mental illness. The first and second authors engaged in a critical meta-theory analysis which included answering questions with the intention of (a) identifying the paradigms that underlie theory; (b) identifying the assumptions that underlie theory; (c) examining the historical evolution of theory; (d) determining how sociocultural, disciplinary, and political contexts influenced theory; (e) evaluating the quality of chosen theory (Massey & Williams, 2020; Paterson et al., 2001).

Meta-method analysis

Meta-method analysis involves critical evaluation of methodology, with particular attention on how methodological decisions influenced findings (Paterson et al., 2001). The first and second authors interrogated each study’s methodology according to (a) consistency between research question and underpinning assumptions; (b) data collection and analysis; (c) sampling; (d) how methodological characteristics influenced research findings (Massey & Williams, 2020).

Meta-data analysis

Meta-data analysis involves the critical examination of findings and the ways in which authors interpreted them (Paterson et al., 2001). A critical meta-data analysis was conducted with the intention of (a) describing key findings; (b) identifying whether interpretations were adequately supported by data; (c) summarizing acknowledged limitations; and (d) examining how theoretical and methodological perspectives shaped results (Massey & Williams, 2020).

Meta-synthesis

We conducted the final meta-synthesis using a thematic synthesis (Paterson et al., 2001). Firstly, we extracted data relevant to our research questions from the 37 articles. Following this, the synthesis involved three sequential steps: line-by-line coding; development of descriptive themes; and the development of analytical themes (Thomas & Harden, 2008). The research team reflected on how themes informed the meta-study research aims. Specifically, for each theme we asked: what does this tell us about athlete experiences of mental illness?

Results

Meta-theory results

We include articles’ theoretical and methodological characteristics in Table 1. Within 10 articles, authors explicitly stated specific theories or frameworks which guided studies. Papathomas and Lavallee (2006) used Schlossberg’s (1981) model of human adaptation to transition, Jones (2014) used Flanagan's (2011) addiction framework, Kühnle (2020) used Luhmann's (1995) sociological systems theory, de Bruin and Oudejans (2018) used a contextual body image framework, two articles (Brownrigg et al., 2018; Papathomas & Lavalle, 2010) used a phenomenological framework, and McMahon et al. (2012, 2017) were guided by Foucault’s theory of disciplinary power, with the latter article also guided by feminist theory. Of those articles that did not explicitly state a guiding theory, the research team interpreted underlying theories or frameworks which guided these articles. A framework was identified if it denoted a particular structure consisting of various concepts, constructs or variables that were related to a phenomenon, whilst a theory was identified if it provided an explanation of how and why specific variables led to specific events to explain a phenomenon (Nilsen, 2015). In total we identified seven articles guided by sociological theories and seven by narrative theory. The remaining 13 articles had an overt medical model presence. These articles presented clinical understandings of mental illness symptoms to frame their research and guide their research questions. The investigation of depression (five articles); clinically diagnosed eating disorders (seven articles); gambling addiction (one article); and substance abuse (one articles) was mostly guided by the medical model.

Table 1. Key features of included meta-study articles.

Author/ date of publication	Participants	Mental illness	Sport	Competition level/status	Methods	Theoretical framework
1. Arthur-Cameselle and Baltzell (2012)^b, c	Gender: F (16) Age (mean years): 20.7 Country: USA	Clinically diagnosed ED	Track/ cross country (8); Swimming (2); Tennis (2); Crew (2); Golf (1); Diving (1)	Collegiate. NCAA sports: Division 1 (14); Division 3 (2)	Data Collection: Structured questions via phone/email Data Analysis: Content Analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
2. Arthur-Cameselle et al. (2018a)^b, c	Gender: F (53) Age: (mean years): 21.07 Country: USA	Clinically diagnosed ED	Track/ cross country (13) Track and field (7); Cross country (2); Crew (8); Soccer (5); Lacrosse (4); Tennis (3); Cheerleading/ dance (3); Volleyball (1); Swimming (1); Field hockey (1); Gymnastics (1); Basketball (1); Cycling (1); Squash (1); Softball (1); Equestrian (1); Bowling (1)	Collegiate NCAA sports: Division 1 (42); Division 2 (2); Division 3 (11)	Data Collection: Structured online question Data Analysis: Consensual Qualitative Research (CQR)	Theory: Medical model Paradigm: None stated; post-positivist identified
3. Arthur-Cameselle and Curcio (2018)^b, c	Gender: F (12) Age (years): 18–24 Country: USA	Clinically diagnosed ED	Track & field/cross-country (4); Track & field (4); Basketball (1); Crew (1); Soccer (1); Tennis (1)	Collegiate. NCAA sports: Division 1 (10); Division 3 (2)	Data Collection: Semi-structured interview Data Analysis: CQR	Theory: Medical model Paradigm: None stated; post-positivist identified
4. Arthur-Cameselle and Quatromoni (2011)^b, c	Gender: F (17) Age (mean years): 20.7 Country: USA	Clinically diagnosed ED	Track/cross country (9); Numbers for the following sports not stated: swimming, tennis, crew, golf, diving.	Collegiate. NCAA sports: Division 1 (15); Division 3 (2)	Data Collection: Semi-structured interview Data Analysis: Thematic content Analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
5. Arthur-Cameselle and Quatromoni (2014a)^c	Gender: F (16) Age (mean years): 20.7 Country: USA	Clinically diagnosed ED	Track/cross country (8); Swimming (2); Tennis (2); Crew (2); Golf (1); Diving (1)	Collegiate. NCAA sports: Division 1 (14); Division 3 (2)	Data Collection: Semi-structured interview Data Analysis: Inductive content Analysis	Theory: Medical model Paradigm: Critical realist (postpositivist)
6. Arthur-Cameselle et al. (2017)^c	Gender: F (12) Age (years): 18–24 Country: USA	Clinically diagnosed ED	Track and field/cross-country (4); Only track and field (4); Basketball (1); Crew (1); Soccer (1); Tennis (1)	Collegiate. NCAA sports: Division 1 (10); Division 3 (2)	Data Collection: Semi-structured interview Data Analysis: Inductive content Analysis	Theory: Medical model Paradigm: Critical realist (postpositivist)
7. Arthur-Cameselle & Quatromoni (2014b)^b, c	Gender: F (47) Age (mean years): 19.9 Country: USA	Clinically diagnosed ED	Track (16); Cross country (12); Crew (12); Swimming (3); Gymnastics (3); Field hockey (3); Volleyball (2); Tennis (2); Lacrosse (2); Golf (1); Diving (1); Ice hockey (1); Soccer (1); Softball (1); NS (1)	Collegiate. NCAA sports: Division 1 (43); Division 3 (4)	Data Collection: 1 structured online question Data Analysis: Content Analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
8. Brownrigg et al. (2018)^c	Gender: M (4) Age (years): 24, 27, 33, 43. Country: UK	Alcoholism/ depression (1); Gambling (1); Gambling/ depression (1); Gambling/ alcoholism (1)	Soccer	Retired professional athletes.	Data Collection: Semi-structured interviews Data Analysis: Thematic Analysis	Theory: Phenomenological framework Paradigm: Interpretivist
9. Busanich et al. (2014)^c	Gender: F (1); M (1) Age (years): 19 and 34 Country: USA	Disordered eating	Cross-country (1); Cross-country/track (1)	Former athletes (1); Active athletes (1)	Data Collection: Semi-structured interview Data Analysis: Narrative analysis (Structural and performance)	Theory: Narrative theory Paradigm: Social constructionist
10. Busanich et al. (2016)^c	Gender: M (4) Age (years): 19–27 Country: USA	Disordered eating	Distance runners (4)	All active athletes.	Data Collection: Visual narrative method and narrative interview Data Analysis: Narrative analysis (Thematic/ dialogic/ performance)	Theory: Narrative theory Paradigm: Social constructionist
11. Cooper and Winter (2017)^a	Gender: F (4); M (2) Age (mean years): 26.4 Country: UK	Disordered eating	Swimming (6)	All former athletes.	Data Collection: Semi-structured interview Data Analysis: IPA	Theory: Social theory Paradigm: Interpretivist stated; post-positivist identified.
12. de Bruin and Oudejans (2018)^b,	Gender: F (8) Age (years): 18–33 Country: The Netherlands	Clinically diagnosed ED	Gymnastics (1); Dance (1); Track and field (1); Cycling (1); Rowing (2); Judo (2)	National or international level.	Data Collection: Unstructured interview Data Analysis: Content and narrative analysis	Theory: Contextual body image framework Paradigm: None stated; post-positivist identified
13. de Grace et al. (2017)^c	Gender: F (4); M (8) Ages (years): 25-61. Country: Canada/UK	Drug addiction/ substance abuse	Hockey (3); Basketball (1); Dance (1); Dirt Bike (1); Hockey/ Soccer (2); Rowing (1); Soccer (1); Gymnastics (1); Karate (1); Fastball (1)	Professional (2); Varsity (2); AAA (1); Competitive (2); International (1); College (1); Provincial (1); AA (1); Black belt (1)	Data Collection: Semi-structured interviews Data Analysis: Realistic Evaluation	Theory: Medical model Paradigm: Post-positivist (Critical realism)
14 Doherty et al. (2016)^b	Gender: M (8) Age (years): 22 – 65 Country: Ireland	Depression: Clinically diagnosed (7) Undiagnosed (1)	Sports not disclosed for confidentiality purposes. Team sport (5); Individual (3)	Competitive elite (3); Successful elite (4); World class elite (1)	Data Collection: Semi-structured interview Data Analysis: Descriptive and interpretative thematic analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
15 Jones (2014)^c	Gender: M (1) Age: NS Country: UK	Alcoholism	Soccer	Former athlete	Data Collection: Unstructured, open ended interview Data Analysis: Life history	Theory: Flanagan (2011) addiction framework Paradigm: Interpretivist
16. Jones et al. (2005)^b, c	Gender: F (1) Age (years): 28 Country: UK/ New Zealand	Eating disorder	Swimming/ water polo	Former athlete.	Data Collection: Unstructured interview Data Analysis: Interpretive biography	Theory: Social theory Paradigm: None stated; Interpretivist (social constructionist) identified
17. Kühnle (2020)^c	Gender: M (4) Age: NS Country: Germany	Depression (3); Depression and atypical anorexia (1)	Soccer (3) Skiing (1)	Retired professional athletes.	Data Collection: Autobiography data Data Analysis: Narrative structural analysis	Theory: Sociological systems theory Paradigm: Interpretivist (constructivist)
18. Lebrun et al. (2018)^b	Gender: F (1); M (4) Age (mean years): 33 Country: UK	Clinically diagnosed depression	Sports not disclosed for confidentiality purposes. Team sport (2); Individual (2)	Former athletes (2); Active athletes (2)	Data Collection: Semi-structured interview Data Analysis: IPA	Theory: Medical model Paradigm: None stated; post-positivist identified
19. Lim et al. (2017)^b	Gender: M (11) Age (years): 18–45 Country: UK	Gambling addiction	Soccer	Active professional (6); Active semi-professional (2); Retired (3)	Data Collection: Semi-structured interview Data Analysis: Thematic analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
20. McMahon and Barker-Ruchti (2017)^b, c	Gender: F (3) Age: NS Country: Australia	Disordered eating	Swimming	All former athletes.	Data Collection: Unstructured interviews Data Analysis: Narrative autoethnography	Theory: Foucault’s theory of disciplinary power; Feminist theory Paradigm: None stated; Interpretivist (social constructionist) identified
21. McMahon and Dinan-Thompson (2008)^b, c	Gender: F (1) Age: NS Country: Australia	Disordered eating	Swimming	International elite level (Olympic athletes) Former athlete.	Data Collection: Autoethnography Data Analysis Autoethnography	Theory: Social theory Paradigm: None stated; Interpretivist (social constructionist) identified
22. McMahon and Dinan-Thompson (2011)^b, c	Gender: F (1) Age: NS Country: Australia	Disordered eating	Swimming	International elite level Former athlete.	Data Collection: Autoethnography Data Analysis Autoethnography	Theory: Social theory Paradigm: None stated; social constructionist identified
23. McMahon et al. (2012)^b, c	Gender: F (3) Age: NS Country: Australia	Disordered eating	Swimming	All former athletes.	Data Collection: Unstructured interviews Data Analysis: Narrative autoethnography	Theory: Foucault’s theory of disciplinary power Paradigm: None stated; Interpretivist (social constructionist) identified
24. McMahon and Penney (2013a)^b, c	Gender: F (3) Age: NS Country: Australia	Disordered eating	Swimming	All former athletes.	Data Collection: Semi-structured interviews Data Analysis: Narrative ethnography and autoethnography	Theory: Foucault’s theory of disciplinary power Paradigm: None stated; Interpretivist (social constructionist) identified
25. McMahon and Penney (2013b)^b, c	Gender: F (3) Age: NS Country: Australia/New Zealand	Disordered eating	Swimming	All former athletes.	Data Collection: Semi-structured interviews Data Analysis: Narrative ethnography and autoethnography	Theory: Foucault’s theory of disciplinary power Paradigm: None stated; Interpretivist (social constructionist) identified
26. McGannon et al. (2020)^c	Gender: F (3) Age: NS Country: Canada/Australia	Alcohol/drug addiction (1); Drug addiction/ED (1)	Ultrarunning	All former athletes.	Data Collection: Autobiography data Data Analysis: Thematic narrative analysis	Theory: Narrative theory Paradigm: Relativist/Interpretivist
27. McGannon and McMahon (2019)^b, c	Gender: F (2) Age: NS Country: Australia/USA	Disordered eating	Swimming	International elite level (Olympic athletes) Both former athletes.	Data Collection: Autobiography data Data Analysis: Thematic narrative analysis	Theory: Narrative theory Paradigm: None stated; Interpretive (social constructionist) identified
28. Newman et al. (2016)^b, c	Gender: F (3); M (9) Age: NS Country: UK	Depression (clinical and subclinical symptoms)	Cycling (2); Cricket (1); Tennis (2); Snooker (2); Rugby union (1); Soccer (2); Swimming (2); Boxing (1)	All former athletes.	Data Collection: Autobiography data Data Analysis: Thematic structural analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
29 Papathomas and Lavallee (2006)^b, c	Gender: M (1) Age: NS Country: UK	Clinically diagnosed ED	Soccer	University level. Former athlete.	Data Collection: Semi-structured interview Data Analysis: Life history analysis	Theory: Schlossberg’s model of human adaptation to transition Paradigm: None stated; Interpretivist (social constructionist) identified
30. Papathomas and Lavallee (2010)^c	Gender: F (4) Age (years): 18 – 24 Country: UK	Disordered eating	Ice skating (1); Figure skating (1); Triathlon (1); Distance running (1)	National (1); National/ international (1); International (1); School (1)	Data Collection: Semi-structured interview Data Analysis: IPA	Theory: Phenomenological framework Paradigm: Interpretivist (social constructionist)
31. Papathomas and Lavallee (2012)^b, c	Gender: F (1) Age (years): 24 Country: UK	Clinically diagnosed ED	Tennis	Former athlete.	Data Collection: Unstructured life history interview Data Analysis: Life history analysis	Theory: Narrative theory Paradigm: None stated; Interpretivist (social constructionist) identified
32 Papathomas and Lavallee (2014)^c	Gender: F (1) Age (years): 20 Country: UK	Disordered eating	Basketball	Active athlete.	Data Collection: Unstructured life history interview Data Analysis: Narrative analysis (Thematic and structural)	Theory: Narrative theory Paradigm: Interpretivist (social constructionist)
33. Papathomas et al. (2015)^b, c	Gender: F (1) Age (years): 21 Country: UK	Clinically diagnosed ED	Triathlon	Active athlete.	Data Collection: Unstructured interview Data Analysis: Narrative analysis	Theory: Narrative theory Paradigm: None stated; Interpretivist (social constructionist) identified
34. Pereira Vargas and Winter (2021)^c	Gender: F (17) Age: (mean years): 22.8 Country: UK	Disordered eating	Powerlifting	All active athletes. International and national level.	Data Collection: Semi-structured interviews. Data Analysis: Reflexive thematic analysis	Theory: Social theory Paradigm: Interpretivist (social constructionist)
35. Plateau et al. (2017)^b	Gender: F (13) Age (mean years): 23.9 Country: UK	Clinically diagnosed ED	Specific sports not disclosed. Lean sports (7); Non-lean sports (6)	Active athletes (6); Former athletes (1); Not training (6). National/ international (8); Regional/ university (3); Club level (1)	Data Collection: Semi-structured interview Data Analysis: Thematic analysis	Theory: Medical model Paradigm: None stated; post-positivist identified
36. Stirling and Kerr (2012)^a	Gender: F (17) Age (years): 18–25 Country: Canada	Disordered eating	Athletics (5); Artistic gymnastics (1); Basketball (1); Dance (1); Figure skating (3); Soccer (3); Swimming (1); Volleyball (2)	All active athletes. National (3); Junior national (3); Collegiate (4); Provincial (2); International (5)	Data Collection: Semi-structured interview Data Analysis: Not specified; content analysis identified	Theory: Social theory Paradigm: Interpretivist stated; post-positivist identified
37. Wood et al. (2017)^b	Gender: M (7) Age (years): 32–41. Country: UK	Clinically diagnosed ‘mental health difficulty’	Soccer	Active part-time athletes (1); Retired (6)	Data Collection: Semi-structured interviews Data Analysis: IPA	Theory: Social theory Paradigm: None stated; interpretivist identified

Note: ED: eating disorder; M: male; F: female; NS: not specified; CQR: consensual qualitative research; IPA: interpretive phenomenological analysis.

^a Stated philosophical positioning was incongruent with what researchers produced.

^b These articles did not state ontology/epistemology.

^c These articles demonstrated methodological coherence.

Only 14 articles acknowledged ontology and epistemology. Of those articles that stated philosophical underpinnings, we deemed two misaligned with what the researcher(s) produced. For example, an article claiming an interpretivist approach still offered numerous examples of causal language and interpretations that might be better described as post-positivist. Post-positivist interpretations are often reductionist, cause-and-effect oriented, and determined based on previous theory (Creswell, 2007), whereas interpretive articles adhere to a relativist position that assumes multiple and equally valid realities and emphasize how these are socially constructed (Ponterroto, 2005). In total, we identified 15 articles underpinned by post-positivist assumptions. Finally, we identified 22 articles which we considered broadly informed by interpretivism: a constructionist (or constructivist) epistemology and relativist ontology. Of these studies, 9 articles explicitly articulated a philosophical position.

Meta-method results

A total of 281 participants (78% females and 22% males) were recruited from 32 sports, with track-and-field/athletics (18%) and soccer (13%) the most represented. Male athletes made up 93% of participants with experience of depression, whereas females constituted 96% of participants with eating disorders. A gambling addiction was experienced by 4% of participants, 1.7% experienced alcoholism, 2.8% experienced drug addiction, whilst 2.4% of participants experienced a combination of these disorders.

We identified 29 articles which demonstrated methodological coherence – congruence between philosophical stance, theoretical position, and methodology (Mayan, 2009) – and eight which did not. For example, within Lebrun et al. (2018), we identified realist overtones (e.g. focus on prevalence, use of causal language, focus on between group differences) despite their use of IPA which aligns with a phenomenological and interpretive underpinning (Shinebourne, 2011).

The one-to-one interview, used across 27 studies, dominated data collection (19 semi-structured, eight were unstructured), whilst three articles used structured online questions to collect data. Alternative approaches to data gathering included autobiography (four articles), auto-ethnography (two articles), and visual narrative (one article). There was greater diversity in data analysis as authors used the following methods of analysis; content (six articles); auto-ethnographical (six articles); thematic, dialogic and/or structural narrative (eight articles); thematic (five articles); IPA (four articles); life history (three articles); consensual qualitative analysis (two articles); interpretive biography (one article); and realist evaluation (one article). Drawing on principles of analytical pluralism, one study combined content analysis with narrative analysis – de Bruin and Oudejans (2018). Although the benefits of analyzing data from multiple vantage points are widely acknowledged, it can be problematic when the conceptual and philosophical framing of these is misaligned (Clarke et al., 2014).

Meta-data analysis results

All articles provided adequate data to support proposed findings. Post-positivist studies presented broad themes describing ‘factors’ or ‘perceptions’ of conceptual interest. For example, perceived ‘triggers’ for mental illness (Arthur-Cameselle et al., 2011, 2017; Doherty et al., 2016; Plateau et al., 2017; Stirling & Kerr, 2012), mental health support recommendations (Arthur-Cameselle & Baltzell, 2012), and perceived factors which initiated and hindered recovery (Arthur-Cameselle et al., 2014a, 2014b, 2018a, 2018b; Plateau et al., 2017).

Articles involving a clinical population often focus heavily on the clinical presentation of symptoms guided primarily by the medical model. These articles provide an understanding of mental illness risk, onset, and recovery using a post-positivist framework. Where interpretive articles included clinically diagnosed participants, the focus was often critical; how medical narratives of mental illness influence athletes’ experiences. Examples of medical narratives include a restitution narrative, described as an ill individual’s belief that full health restoration or recovery is possible (Frank, 1995) (e.g. Papathomas et al., 2015). In considering gender, three articles presented athlete accounts grounded in gendered constructions of their disordered eating (Busanich et al., 2014, 2016; Pereira Vargas & Winter, 2021). For example, Busanich et al. (2014) highlighted how a female athlete’s narrative aligned with dominant medical understandings of an eating disorder. Therefore, alignment to a medical narrative normalized an eating disorder for the female athlete. On the other hand, the male athlete’s account did not align with dominant medical understandings of an eating disorder as he instead felt his masculinity threatened by the diagnosis and refuted it.

Most articles exemplified what Clarke and Braun (2018) term ‘small q’ qualitative. ‘Small q’ studies use qualitative tools under a broadly positivist paradigm. ‘Bucket themes’, which are lists of descriptive themes derived from literature driven semi-structured interview guides, are favored over more interpretive approaches to presenting results. The ‘postpositivising’ of qualitative data, whereby authors calculate numbers and percentages of participants represented in a theme, was also a feature of several articles (Ponterotto, 2005). Post-positivist articles were mostly guided by the medical model and rarely offered experiential insight. In contrast, studies guided by interpretivism aimed to understand athlete meaning making. Within these articles, mental illness experiences were understood as socially constructed based on the narrative resources available to the participant. As such, these articles present examples of ‘big Q’ qualitative work. Despite these distinctions, in no way do we wish to infer that one paradigmatic position has its place over another. In line with Poucher et al. (2020), through drawing on methodological coherence to assess the body of literature, we steer away from arguments reinforcing a ‘paradigm war’ and are cautious about acting as gatekeepers for what paradigmatic position constitutes as best for qualitative research.

Where studies offered limitations, these tended towards positivist judgement criteria and an allegiance to a single reality. For example, several studies cited the possibility of recall bias and an inability to verify the ‘accuracy’ of participants accounts (Arthur-Cameselle & Quatromoni, 2014b; Doherty et al., 2006; Lebrun et al., 2018; Plateau et al., 2017; Wood et al. 2017). Similarly, poor generalizability due to small sample size (Arthur-Cameselle & Curcio, 2018b), purposeful sampling, and limited diversity in represented sports (Lebrun et al., 2018). Although these articles did not explicitly state ontological and epistemological assumptions, these limitations align with a positivist philosophy. In contrast, interpretivist articles offered limitations and quality measures which were compatible with relativist assumptions. These measures included the use of a critical friend to aid in the construction of discussion points in relation to data (Brownrigg et al., 2018) and member reflections to assist in generating further data and insights (Pereira Vargas & Winter, 2021). Following recommendations by Smith and McGannon (2018), future research must be clear about their ontological and epistemological position and attempt to apply quality and rigor in accordance to these.

Meta-synthesis

Through the process of thematic synthesis, we constructed four broad themes: (1) Origins of certainty and ambiguity; (2) A sense of gradual decline; (3) Mental illness as a threat to identity; and (4) Constructing recovery stories. We now describe each theme and support our interpretations with participant quotes. See Table 2 for a breakdown of contributions made by articles to the thematic synthesis.

Table 2. Contribution made by articles to the thematic synthesis.

Themes	Paper	Athlete quote
Origins of certainty and ambiguity	6, 4, 9, 11, 12, 16, 18, 22, 20, 23, 21, 24, 27, 28, 31, 33, 36	• ‘It is so difficult that I don’t have it all clear to myself. So much has happened, I cannot understand it all yet. Why did things happen? How was that for me?’ (12)
Gradual sense of decline	3, 4, 5, 6, 7, 9, 11, 12, 14, 16, 18, 20, 21, 22, 23, 24, 27, 28, 29, 30, 31, 32, 33, 34, 36, 37	• ‘Say if I was rid of it and I still wasn’t successful, would I think hmmm maybe this isn’t the key to it? Maybe I was better slightly with the eating disorder?’ (33) • ‘I couldn’t be bothered, although I did train, but I’d just turn up. Felt like I were just turning up. I wasn’t there because I wasn’t enjoying it’ (18)
Mental illness as a threat to identity	3, 5, 9, 10, 11, 12, 14, 16, 18, 20, 21, 22, 23, 24, 27, 28, 29, 30, 31, 33, 34, 35, 37	• ‘Models have eating disorders, Mike and sportsmen don’t. Certainly not footballers. Gymnasts, ballerinas, but not footballers’ (29) • ‘I’m not going to be running competitively for the rest of my life’ (3)
Construction of recovery stories	1, 3, 5, 7, 9, 10, 11, 12, 18, 27, 28, 29, 30, 31, 32, 29, 33, 34, 35, 37	• ‘Going through a lot of struggle and adversity and having to really stay at it and be persistent to get over these struggles or roadblocks that impede you from getting where you want to go’ (10)
1- Arthur-Cameselle et al. (2012); 2- Arthur-Camselle et al. (2018a); 3- Arthur-Cameselle et al. (2018b); 4- Arthur-Cameselle et al. (2011); 5- Arthur-Cameselle et al. (2014a); 6- Arthur-Cameselle et al. (2017); 7- Arthur-Cameselle et al. (2014b); 8- Brownrigg et al. (2018); 9- Busanich et al. (2014); 10- Busanich et al. (2016); 11- Cooper et al. 2017; 12- de Bruin et al. (2018); 13- de Grace et al. (2017); 14- Doherty et al. (2016); 15- Jones (2014); 16- Jones et al. (2005); 17- Kühnle (2020); 18- Lebrun et al. (2018); 19- Lim et al. (2017); 20- McMahon et al. (2017); 21- McMahon et al. (2008); 22- McMahon et al. (2011); 23- McMahon et al. (2012); 24- McMahon et al. (2013a); 25- McMahon et al. (2013b); 26- McGannon et al. (2020); 27- McGannon et al. (2019); 28- Newman et al. (2016); 29- Papathomas et al. (2006); 30- Papathomas et al. (2010); 31- Papathomas et al. (2012); 32- Papathomas et al. (2014); 33- Papathomas et al. (2015); 34- Pereira Vargas et al. (2021); 35- Plateau et al. (2017); 36- Stirling et al. (2012); 37- Wood et al. (2017)

Origins of certainty and ambiguity

Origin stories tell of participants’ beliefs about the onset of their mental illness. In the reviewed studies, ‘perceived causes’ are typically considered in a realist sense (i.e. indicative of factors that contribute to mental illness in sport). In our interpretive analysis, we categorized origin stories according to how they were told by participants. Several articles documented athletes’ perceived causes of their mental illness. Our interpretive analysis interrogated the narrative of these causal beliefs and identified two distinct types of construction: narrative certainty (e.g. ‘What caused it [eating disorder] to start? It was the injury’, 6) and narrative ambiguity (e.g. ‘It is difficult that I don’t have it all clear to myself. So much has happened, I cannot understand it all yet. Why did things happen?’, 12).

Narratives of certainty include athletes perceiving the origin of their mental illness to stem from certain characteristics which they possessed. One athlete states: ‘I simply have everything that makes you vulnerable for it’ (12). This account aligns to the deficit model within medical psychology in that it suggests mental illness occurs due to individual weakness (Walker, 2006). Other accounts of certainty included attaching blame to sport and pressure from being expected to produce high level performance:

I got put under so much pressure and I think ethically it was completely wrong. They caused me problem after problem … they caused my deterioration in health … They would devalue what I was doing and then expect me to go and perform. (18)

This athlete expresses certainty toward a singular explanation for the origin of depression which has absolved him of blame for the development of his mental illness. Although such a construction can absolve the athlete of blame, acceptance of this singular explanation can create ineffective recovery outcomes by failing to look more deeply at other personal or sociocultural influences which may have impacted mental illness (Deacon & Baird, 2009).

Although it was sometimes difficult to ascertain the degree of certainty which participant accounts possessed, we determined that narratives of certainty were more aligned to a master medical narrative as athletes recount their experiences through medical understandings of illness. As inferred by McAdams (1993), the alignment to a cultural storyline (i.e. the medical model) is attractive as it can provide a person with a sense of meaning to life, through the telling of a story that ‘makes sense’. However, Crossley (2000) warns, the alignment to stories that do not ring true to individual experience may limit the development of a person’s authentic self. Therefore, individuals are influenced, and in some cases limited, by dominant discourses available to them (Shapiro, 2014). Nevertheless, participant accounts within this theme highlight the power that dominant illness storylines can have.

In contrast, athlete accounts that possessed more of an ambiguous narrative were characterized by a lack of certainty on the origin of their mental illness:

You have already experienced so much as a child, big competitions, all those impressions, feelings, living separated from others … so much goes through your mind. That’s why [the origin] is so unclear to me (12)

When some athletes were presented with a coherent narrative to explain their mental illness, this is questioned, and an alternative understanding is sought. Ochs and Capps (2001) infer that ambiguity narratives are told to try and gain an authentic reconstruction of mental illness, as individuals question the nature and origin of their experiences. In sum, ambiguous and certain origin stories have been associated with poor and effective recovery, respectively (Shohet, 2007).

A sense of gradual decline

This theme highlights athletes’ initial perception that mental illness was manageable, followed by a gradual sense of decline in many aspects of athletes lives as mental illness developed. Although previous reviews have focused on findings around symptom presentation (Reardon & Factor, 2010), this theme goes beyond acknowledging worsening symptom severity, to instead integrating how it is experienced through a performance perspective.

The occurrence of a mental illness initially did not lead to a decline in performance, and symptoms were able to coexist with performance, as an athlete with depression described: ‘I was still doing fairly well, if you look back on the record books, I was still maintaining a top 30 or 40 world ranking and winning an odd tournament’ (14). In fact, some athletes believed there were functional benefits to their mental illness. Particularly for athletes experiencing eating disorders, disordered behaviors (e.g. self-starvation and excessive exercise) were understood as functional and a means to be successful in sport. A belief that disordered behaviors were beneficial led athletes to ponder if recovery was worth it. For example, ‘say if I was rid of it and I still wasn’t successful, would I think hmmm maybe this isn’t the key to it? Maybe I was better slightly with the eating disorder?’ (33). For some athletes, the beliefs held that disorders perhaps aided performance may be what Hughes and Coakley (1991) describe as a manifestation of over conformity to the norms emphasized within sporting cultures.

Although sport initially created a perception that mental illness could coexist with performance, athletes described gradually experiencing a decline in mental health which eventually affected sport: ‘Gambling took priority over training … [I am] physically exhausted after gambling through the night’ (19), and ‘I would compromise my training to have a binge, that’s how unbothered I was about myself as an athlete’ (30). In some accounts, this gradual decline or turning point was characterized by an increase in severity of behaviors (e.g. ‘Binge drinking gradually changed into the more familiar pattern of alcoholic drinking and addiction took hold’ 15). Similarly, in eating disorder accounts, a change in perception surrounding food, exercise, and the body was described. Disordered behaviors were initially described as ‘innocent’ (6) and ‘healthy’ (9), however, over time athletes experienced a compulsion to engage in more extreme eating behaviors.

A gradual sense of decline was also highlighted through contrasts made by athletes between the their previous committed, athletic selves, and their current disordered selves, with the latter identity rejected:

I just want to stay in bed with the curtains down and just withdraw … that’s not what I feel like normally, that’s not who I am … normally I can’t wait to get out of bed to see the day, to do training. (30)

Perceiving a change in oneself as the disorder evolved highlights the complexities in being an athlete whilst attempting to perform with mental illness. Stories present a sense of before and after, as athletes perceived themselves as athletes prior to their mental illness which was slowly lost as mental health declined.

The slow gradual sense of decline in mental health created a false impression that symptoms and behaviors associated with mental illness could coexist with successful performance. Athletes adhered to the behaviors and norms reinforced within their sporting cultures to attain performance success. Over time, acceptance of these extreme norms for some gradually influenced mental illness. As such, some accounts within this theme reinforce the notion that there may be dangers in accepting certain norms which are engrained within sporting sub-cultures such as weight-cutting in weight-class sports (Pereira Vargas & Winter, 2021); extreme dieting in weight-sensitive sports (Bratland-Sanda & Sundgot-Borgen, 2013); and a culture of binge-drinking or gambling in male dominated sports (Miller et al., 2001). Therefore, to limit poor athlete mental health, dominant ideas around extreme behaviors to succeed should be challenged. As Glick and Brodwin (2016) argue ‘the rationale underlying sports has to change’ towards measures of success that ‘go above simply winning’ (Glick & Brodwin, 2016, p. 629). Whether it is possible to reconstruct success within elite sport is unclear, however the accounts within this theme suggest that although mental illness can coexist with successful sporting performances, athletes may eventually experience a gradual decline in many aspects of their lives.

Mental illness as a threat to identity

For male athletes, mental illness threatened masculinity. Experiences of illness were incongruent with expected masculinity norms and values accepted within sporting culture such as competitiveness and strength (Steinfeldt & Steinfeldt, 2012). Conformity to these socially constructed ideals was central to male athletes’ identity, which led to feelings of confusion, shame, and embarrassment because ‘[mental illness] was perceived as weakness’ (18). As such, male athletes attempted to downplay their illness by overperforming their masculinity. For example, athletes denied diagnoses, as one athlete referred to their disorder as ‘a mild’ or ‘small case of anorexia’ (9). Another athlete also held the perception that some mental illnesses were exclusive to a particular demographic: ‘Models have eating disorders … sportsmen don’t. Certainly not footballers. Gymnasts, ballerinas, but not footballers’ (29). Given that he did not fit the typical identity of someone experiencing an eating disorder, he perceived his masculine identity threatened and hid his diagnosis. Behaviors that contradict a valued identity can be detrimental to well-being (Kaya et al., 2019). Within accounts of gambling addiction and substance abuse, participants’ involvement with gambling, alcohol and drug use became an expected part of athletic male culture: ‘[drinking] was what we did, it, it went hand in hand with baseball’ (13). Athletes often rejected diagnostic labels as they threatened their athletic identity and masculinity. These accounts highlight the impact which labeling athletes with a mental disorder had on identities. Schwenk (2000) appropriately suggests that these unhelpful labels and diagnoses create stigmatization of mental illness which ultimately ‘deprives the athlete of effective care’ (p. 4).

For some, an exclusive athletic identity was associated with a perception they could not leave sport. For example, ‘I felt like I had to … . There was no way I could turn around to anyone and say ‘I don’t want to do [sport] anymore’ because I felt like it was what I did’ (30). This athlete described sport as something she ‘had to do’ which suggests she may not have been able to present an identity other than an athletic identity. Although this athlete’s athletic identity may have been disrupted by her experience of mental ill health, there remained a perceived pressure to continue performing an athletic self and fulfilling the athlete role.

Within eating disorder accounts, behaviors which were misaligned with athlete identity (e.g. gaining weight) left athletes at risk of identity loss, which promoted extreme dieting behaviors to regain this identity. Thus, athletes perpetuated a process of alignment to disordered behaviors to reinforce identity. However, athletes still feared losing athletic identity later, which allowed disordered behaviors to continue as an identity maintenance strategy. The prospect of gaining weight, not losing enough weight, or not possessing a body that was deemed capable of sporting success, made athletes continuously fear losing their athletic identity. This process was heightened when extreme behaviors were being reinforced by coaches or peers. For example, ‘[coach] was always like ‘you will never make weight eating that stuff’ … [I would] go and make myself sick because I felt so bad’ (34). Within this athlete account, the fear of not making weight led this athlete to align to extreme behaviors to reinforce her identity as an athlete within an idealized weight-class.

As participant identities were rooted in being athletes, they sought to exhibit characteristics representative of successful performers: ‘In sport being tough and being driven are really admired’ (14). Being mentally tough and denying weakness are characteristics which are projected within elite sport, thus internalized by athletes (Bauman, 2016). Although such characteristics were perceived to aid performance: ‘[obsessive drive] made me a star athlete in the water’ (28), the same characteristics heightened depressive symptoms: ‘out of the water it tore me apart’ (28). This athlete expressed tension between the possession of obsessive drive which aided performance, whilst also being a perceived risk-factor for depression. Ultimately embracing this obsessive drive and perceiving it as a marker for successful performance may have heightened depression and created an unwillingness to change or pursue recovery. These accounts highlight that athletic identity may in some cases reinforce mental illness.

Constructing recovery stories

The way in which athletes presented recovery stories differed. Athletes presented experiences of either ‘full’ recovery or being unable to recover from mental illness. Participants’ recovery aligned to either medical discourses or was told through the social model or growth-oriented perspective (e.g. recovery as a journey or process, Jacob, 2015). Recovery experiences aligned to the medical model focused on symptoms, clinical treatments, and interventions. Through such an alignment to medical conceptualizations of recovery, athletes expressed how recovery was characterized by a change from ill person to a healthy person no longer experiencing symptoms. One athlete discussed recovery this way, comparing his disorder to an injury: ‘you know it’s not a flaw, it’s an illness. It can be treated, it can be fixed, it can be sorted out just like any other injury’ (18). Although comparing a mental illness to a physical illness is a narrative used to normalize mental illness and reduce stigma (Whitley & Campbell, 2014), this view can be problematic. The focus on ‘curing’ mental illness in a comparable way to a physical illness presents a reductionist means to ‘treat’ mental illness as it ignores sociocultural influences which may be impacting mental illness and thus may limit treatment options (Albee & Joffe, 2004).

Although some athletes who understood their recovery through an alignment to the medical model perceived they had achieved ‘full’ recovery (i.e. no more symptoms), others who sought professional help for their mental illness described a perceived lack of relevance of treatment: ‘I started to get a bit frustrated actually with NHS stuff because it’s so generic. I don’t feel like I have a typical ED … people didn’t understand me and the athlete side of it all’ (35). Athletes perceived that therapy models took a one size fits all approach not individualized for athletes. This highlights the difficulty that clinical professionals have in assisting athletes’ recovery when using restricted medical criteria (Hughes & Leavey, 2012). Mental illness recovery is a highly complex lived experience which scholars suggest can only be claimed by the person in recovery themselves (Deegan, 1992).

Other athletes framed recovery as a journey of growth which focused on personal development rather than symptoms (Jacob, 2015). Athletes understood that recovery was not a linear process and viewed it as an ongoing process: ‘[Recovery] represents a lot of growing as a person. Going through a lot of struggle and adversity and having to really stay at it and be persistent’ (10). Although mental illness was defined as a low period in athletes’ lives, a lack of regret was felt by one athlete: ‘You should never regret something that’s happened but be glad it’s happened and learn from it’ (30). These findings support research into growth following adversity which suggests that positive changes occur to athletes following a period of struggle (e.g. an appreciation of individual skills, and better relationships with others) (Day & Wadey, 2016). These accounts suggest it may have been difficult to experience growth if previous struggle had been dwelled upon.

Ultimately, athletes recognized the transiency of elite sport and therefore the importance of an identity shift:

I definitely didn’t want to be a person with an eating disorder, that’s not part of my identity … There was just this moment of realizing that I’m not going to be running competitively for the rest of my life. To make choices that are really gonna affect me suddenly didn’t make sense anymore. (3)

Through doing so, athletes reconstructed the meanings of success dictated by elite sport and its performance narrative (Douglas & Carless, 2009). The performance narrative has been described as stories of single-minded dedication to sport whereby an athlete’s mental well-being, identity, and self-worth are tied to achieving results and performance success. Alternative narratives to the performance narrative included the engagement in recreational sport for enjoyment: ‘My heart felt happy. I would have so much fun on my runs again, all by myself, doing my own thing’ (9). The identity constructed by this athlete transformed from an elite runner with performance goals, to recreational runner focused on the development of health. Athletes also sought achievement and success in other domains of life, such as academia: ‘I actually did have a brain and that I could use it … I found something else that I was good at’ (30). An understanding that there was more to life than just sport, allowed athletes to broaden their identities through reconstructed meanings of success. As athletes’ athletic identity permeated all other aspects of their lives, the ability to broaden one’s identity whilst still involved in competitive sport is challenging (Warriner & Lavallee, 2008). Therefore, only upon leaving sport did athletes develop an ability to reconstruct and broaden their identity away from performance pressures.

Critical reflections

Through meta-study, we critically appraised the theory, methods, and findings of reviewed articles. In addition, we have shown how a thematic synthesis can organize data to generate new knowledge on athlete mental illness. This approach extends previous reviews of predominately quantitative literature (e.g. Rice et al., 2016). Underpinned by social constructionism, we move beyond the identification of prevalence and risk; already widely discussed (see Küttel & Larsen, 2020), to highlight how athletes construct mental illness. Within this critical reflection we will comment further upon results from this article, acknowledge methodological implications of this study, and finally make recommendations for future research and practice.

From an analytical perspective, the medical model guided many of the reviewed articles. The dominance of a diagnostic perspective was evident within athlete accounts, as many athletes’ mental illness and recovery stories were shaped through medical conceptualizations. Research has indicated that recovery experiences for individuals whose stories align to medical understandings of illness may not be as fulfilling given the promises the medical model makes about ‘curing’ illness (Deacon & Baird, 2009). We encourage future research to explore further narratives which focus on athlete experiences of living with mental illness, which deviate from biomedical understandings of illness. As we have highlighted, particular theoretical choices may enlighten or restrict athlete voices. For example, narrative articles privilege athlete voices more so than articles which conform to post-positivist methodology through more in depth quotes. As such, articles which presented findings as lists of factors with few participant quotes were harder to interpret for the purpose of the present study. We argue that articles which ascribe to interpretivist methodologies revealed more nuanced accounts of mental illness in sport by understanding how athletes’ experiences are constructed as opposed to what these experiences were. As such, an alternate approach to explore athlete accounts of mental illness may be through a social constructionist lens. Through examining how mental illness is socially constructed, it is possible to explore how social and cultural forces shape understandings of illness (Brown, 1995).

In line with Kaite et al.’s (2015) qualitative meta-synthesis, mental illness constructions impacted identities. Within our synthesis, participants contrasted their athletic selves with mentally ill selves leading to self-stigma and lack of disclosure. Rather than separating being mentally ill from being an athlete, studies have advocated for challenging stigma by integrating these identities. Being labeled ‘mentally ill’ and ascribing to such labels has implications for understanding how an individual perceives oneself (Schwenk, 2000). For example, previous research has suggested that individuals who are able to challenge stigma and their illness label may be able to gain autonomy over recovery and reconstruct identity, as they reject the negative connotations associated with being mentally ill (Thoits, 2016). Exploring perceptions toward challenging stigma and the labels associated with having a mental illness may provide a further understanding to the main themes derived from our meta-synthesis, and in particular greater insight into identity and the process of recovery.

Although our results present novel insights into the athlete mental illness literature, it is important to acknowledge methodological considerations. Upon commencement of the research, we noted a large disparity between the types of sports, types of mental illness’, competitive levels, and genders represented within the meta-study. For example, the dominance of female eating disorder accounts within aesthetic sports was reflected in the synthesis due to the large amount of data available for this population. Given the dominance of depression and more so eating disorder accounts within our study, we recommend future research expand to the study of additional mental disorders within athletes. In addition, a limited number of studies offered accounts of the sociocultural context in which athletes operate in. This limited description of the context in which athletes experience mental illness restricted a full understanding of the athlete experience. It is important that future research address this, particularly given the role which sport plays in shaping athletes’ mental illness constructions.

Finally, our findings hold practical implications for those operating within elite sporting environments including athletes, coaches, sporting organizations, and medical professionals including therapists. Providing athletes with education on their illness may allow for them to develop their own discourses outside of the dominant medical storylines reproduced in society. Although individuals’ mental illness stories may all appear different at face value, these follow a common framework. Acceptance of and adherence to biomedical models can restrict understandings of illness which do not possess certain kinds of knowledge (Woods et al., 2019). Therefore, therapists working with athletes may assist athletes with creating their own meanings around experiences which perhaps fall outside of this framework. By privileging stories of uncertainty and unconventional understandings of recovery from mental illness which do not comply to biomedical frameworks, positive recovery experiences may be yielded (e.g. reclaiming a renewed identity) (Woods et al., 2019). As suggested by Davidson and Roe (2007), this recovery-oriented approach offers a more effective means to reclaiming his or her life back. As such, we propose that practitioners may find it useful to consider the social construction of mental illness when working with athletes. Our results emphasize the value of qualitative work, which we recommend be better integrated into future position statements in order to assist applied work within the field. Through the inclusion of athlete voices within position statements, a further knowledge base may be shaped for the target populations of such statements (e.g. stakeholders, coaches, etc.) surrounding the meanings athletes construct around mental illness. As such, it may be possible to expand upon dominant trends within sport on prevalence, treatments and incidence of mental illness.

Conclusion

This study is the first to systematically review and synthesize the qualitative research on athlete mental illness. Through the critical appraisal of authors’ chosen theoretical underpinnings and methodological approaches, we have provided potential research avenues to explore and advance work within the field. We argue that there continues to be good reason to align to medical conceptions of mental illness within sport psychology literature, and to listen to and tell stories which ascribe to these understandings within sport. However, consideration should be given to how this framework can limit knowledge of athlete mental illness and restrict its application to practice. We also encourage open critical dialogue regarding methodological coherence within qualitative work to ensure appropriate alignment of authors’ philosophical perspective, theoretical position, and methodology used, to advance qualitative research exploring athlete mental illness.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article and/or its supplementary materials.