The Mental Health of Elite Athletes: A Narrative Systematic Review

1High-quality, systematic studies on the nature and impacts of physical injuries in elite athletes—most notably, head injuries/concussion and limb injuries—have led to advances in how these injuries are optimally managed or, ideally, prevented. There is comparatively less research on, but growing interest in, the mental health and psychological wellbeing of elite-level athletes. The prevalence of diagnosable psychiatric disorders in this population remains a matter of debate; however, notions that elite athletes are devoid of mental health problems have been increasingly scrutinised by sports medicine practitioners.

2The intense mental and physical demands placed on elite athletes are a unique aspect of a sporting career, and these may increase their susceptibility to certain mental health problems and risk-taking behaviours. Furthermore, the peak competitive years for elite athletes tend to overlap with the peak age for the risk of onset of mental disorders. In addition to physical and competition stress, elite athletes face a unique array of ‘workplace’ stressors, including the pressures of increased public scrutiny through mainstream and social media, limited support networks due to relocation, group dynamics in team sports and the potential for injuries to end careers prematurely. The ways by which athletes appraise and cope with these stressors can be a powerful determinant of the impact the stressors have on both their mental health and their sporting success.

3Athletes tend not to seek support for mental health problems, for reasons such as stigma, lack of understanding about mental health and its potential influence on performance, and the perception of help seeking as a sign of weakness. While there have been efforts to disseminate sport-related mental health findings in order to advance the prevention, identification and early treatment of psychopathology in elite athletes, there are suggestions that some sporting governing bodies continue to minimise the significance of mental ill-health in this population. This has sobering implications if elite athletes within such organisations are not provided with access to timely or adequate mental health care, or do not feel that the culture of the sporting organisation is such that they can even raise their mental health concerns. While it is well established that physical activity has a positive effect on mental health, a review has found that intense physical activity performed at the elite athlete level might instead compromise mental wellbeing, increasing symptoms of anxiety and depression through overtraining, injury and burnout. Some, though not all, research suggests that this population has an increased risk of mental health problems, including eating disorders and suicide. A recent national survey of elite athletes in Australia found that almost half acknowledged symptoms of at least one of the mental health problems that were assessed, with prevalence rates similar to those reported in the community. Emerging research suggests that retired elite athletes may be at particularly elevated risk of mental ill-health, corresponding to both low rates of formal athlete mental health screening processes and player perceptions of inadequate availability of mental health support.

4Given the early-stage state of sports psychiatry and its research base, the current delivery of mental health care for elite athletes might not take into account sport-related factors that potentially influence vulnerability to mental health problems, nor diagnostic or treatment issues that may be unique to this population. Developing a comprehensive understanding of the mental health and psychological wellbeing specific to elite athletes has the potential to advance models of care and management of this population, which may, in turn, facilitate performance gains. Such an understanding is required to provide guidance for sport practitioners—including coaches, medical staff and sport psychologists—in developing the coping abilities of elite athletes and, in turn, improving their emotional wellbeing.

5The utility of systematic reviews to synthesise research on discrete topics and identify gaps in knowledge is well established; however, to date, there have been no such reviews of the mental health and psychological wellbeing of elite athletes. The objective of this review was to synthesise the growing evidence base regarding the incidence and nature of mental ill-health (including substance use) and psychological wellbeing among elite-level athletes in order to identify gaps that future research should prioritise, and inform strategies or guidelines to advance the detection and management of mental ill-health in this population.

6A systematic search of five electronic databases (PubMed, EMBASE, SPORTDiscus, PsycINFO, Cochrane) was conducted between January and February 2015, using the relevant database search engines. A subsequent search was conducted using the Google Scholar database in May 2015 to ensure that all recently published articles meeting the inclusion criteria were identified. The search strategy for each database, MeSH descriptors and corresponding number of hits per database are presented in Electronic Supplementary Material Tables S1 and S2.

7Three researchers independently assessed the eligibility of each retrieved record on the basis of the title and abstract. If the information was unclear, the full-text article was screened. All included studies were subsequently re-screened (i.e. double screened) by a fourth researcher. The included studies were required to meet the following inclusion criteria: (1) participants were currently competing at the elite level, as able-bodied athletes, where the elite level was defined a priori to be competitive at either the Olympic, international, national or professional level; (2) the study reported quantitative data on a mental health, wellbeing or coping outcome; and (3) the study was published in English. Studies were excluded from the review on the basis of the following criteria: (1) the mean age of the participants was <18 years; (2) participants were competing at a school or collegiate level; (3) the study was undertaken with a heterogeneous sample (i.e. a mixed sample of elite and non-elite athletes) without reporting group findings separately; (4) the study assessed only physiological wellbeing or stress responses without assessing or reporting psychological wellbeing; (5) the study was available in abstract form only (i.e. conference presentations), precluding full quality assessment; and (6) the study described substance use focused on performance enhancement (i.e. doping) as opposed to personal use. The systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (see Fig. 1 for flow diagram).

8A standardised data extraction template was designed. One researcher sourced the required information from the included studies, using this template, including the study type and design, sport population, study aim, sex ratio and key outcomes mapped to study measures and main findings.

9Given the heterogeneity of the included study designs and the lack of randomised, controlled trials identified in the search process (see Sect. 3.2), it was not possible to conduct a standard risk-of-bias assessment. In place of this, all studies were appraised for reporting quality based on the established standards outlined below.

10The literature search yielded a total of 2279 records. After screening of the titles and abstracts, 103 studies were identified as likely meeting the inclusion criteria. After double screening, exclusion of qualitative studies and a final manual search of the literature (i.e. screening of references lists), a total of 60 studies were included in the quantitative synthesis (see Fig. 1 for the study selection diagram).

11Sixty quantitative studies were included in the review. The study designs varied, though most were either cross-sectional observational studies (N = 38; 63.3 %), longitudinal studies (N = 11; 18.3 %) or of a mixed-method design (N = 8; 13.3 %). In addition, there were one randomised, controlled trial (1.6 %), one meta-analysis and one intervention case study report. Given the heterogeneity of both the study designs and the outcome variables that were assessed, it was not possible to conduct a meta-analysis as part of this review.

12The methodological rigour of the included studies was assessed according to relevant published criteria. Quality appraisal of the 60 studies is presented in Electronic Supplementary Material Table S3. Only two studies met all methodological criteria, with one quarter (N = 15; 25 %) assessed to be of good reporting quality (i.e. scoring ≥4 out of 5). The mean quality rating was 2.88 (standard deviation 0.87). Over one third of the remaining studies (N = 21; 45 %) were assessed to be of moderate quality (scoring 3 out of 5), while 24 studies (40 %) were assessed to be of low quality (scoring 2 out of 5). Almost all included studies (N = 59) defined their participants well and either reported use of standardised, validated questionnaires or clearly described the outcomes measured. Fewer than half of the included studies clearly reported ethical review (N = 27) and either reported a participant rate of more than 80 % or, in the absence of that, provided a description comparing responders with non-responders (N = 25). Over one quarter of the included studies did not report on the participant rate (N = 16), and very few used random sampling (N = 3).

13Tables 1, 2, 3, 4, 5, 6, 7 provide a summary of the key characteristics and main outcomes of the 60 included studies. For the purposes of reporting and analysis, studies were grouped according to the following major mental health constructs: anger and aggression (N = 2), anxiety (N = 4), eating disorder and body image (N = 10), general-prevalence studies (N = 10), help seeking (N = 1), sleep (N = 1), stress and coping (N = 22), substance use (N = 9) and wellbeing (N = 1). The included studies examined athletes from a broad range of individual sports (e.g. swimming, tennis, wrestling) and team-based sports (e.g. soccer, football, rugby), with some studies including elite athletes from a range of sporting disciplines.

14Of the two included studies that focused on anger and aggression (see Table 1), one was conducted with rugby players, while the other was a case study of an elite table tennis player. Anger tended to be experienced relatively frequently by the rugby players, was viewed as facilitative as opposed to debilitative and was positively associated with anxiety. In both studies, cognitive aspects (self-confidence or problem solving) were associated with less expression of anger in competition. No studies were found that evaluated off-field expressions of anger or aggression.

15Four studies focused on anxiety in elite athletes (see Table 2): two on swimmers and two on athletes from mixed sporting populations. These studies focused primarily on the performance aspect of symptoms of anxiety (i.e. where athletic performance is evaluated as threatening and is associated with elevated levels of arousal or worry) as opposed to generalised clinical or subclinical experiences of non-competitive anxiety, which are summarised in the general-prevalence studies listed in Table 4. Athlete interpretation of anxiety states was identified as critical to the impact of anxiety. For example, a focus on performance (as opposed to cooperation and effort) predicted athlete worry, while interpretation of anxiety as facilitative was associated with more adaptive anxiety management strategies (i.e. approach-focused coping) and performance levels. Higher levels of athlete anxiety were also found to be related to negative patterns of perfectionism. Recommendations to elite-level coaches included development of athlete skills in appraisal and interpretation of anxiety states and the type of training culture facilitated among athletes (i.e. mastery as opposed to performance).

16Of the ten studies examining eating disorders and body image (see Table 3), six were conducted with mixed populations, two with rowers and the remaining studies with distance runners and figure skaters. With the exception of the three studies that used an interviewer-administered diagnostic interview, all studies on eating disorders and body image used self-report data from standardised measures. Of the ten studies, five evaluated either eating disorder or body image issues in comparison with general community samples. The results from these studies were inconsistent. Three reported a higher incidence of eating disorder or body dissatisfaction in elite athletes relative to controls, especially in sports emphasising leanness or lower body weight. In contrast, one study found no difference between elite athletes and controls when the sample was restricted to females. The included meta-analytic study that examined differences in body image reported no differences between athletes and non-athletes, or by sex or body mass index; however, it must be noted this meta-analysis used homogenous inclusion criteria incorporating a large number of studies reporting data from non-elite athletes. The remaining studies identified athlete-specific risk factors for eating disorders or body image concerns, including sport-specific body-type demands (i.e. leanness), age (higher risk in younger athletes) and sex (higher risk in females). One study recruiting only males identified onset-related characteristics of eating disorders as commencement of training at an earlier age, dieting and experience of traumatic events, such as significant injury.

17The ten general-prevalence studies on elite athlete mental health (see Table 4) reported data from either mixed samples or specific codes, including football, swimming, weight lifting, equestrian, distance running and tennis. Sample sizes for these studies varied on the basis of the population of interest, ranging from N = 2067 for a national-level study to N = 14 for a study of distance runners. Of note, with the exception of one study utilising a structured diagnostic interview and one study utilising clinician diagnoses, outcomes for general-prevalence studies were based on athlete self-report data from standardised measures. In the studies with larger sample sizes (N > 100), combined rates of high-prevalence disorders (i.e. mood or anxiety disorders) were frequently reported. In one large self-report study, up to 46.4 % of Australian athletes (N = 224; a mixed sample) met the clinical cut-off for a diagnosable mental health disorder based on standardised scales, with identified rates of depression (27.2 %), eating disorder (22.8 %) and anxiety disorder (social phobia; 14.7 %). Similar findings were reported for prevalence rates of depression (34 %) based on diagnostic interviews undertaken with swimmers and self-reported depression and anxiety (26 %) in European football players, though rates of self-reported depression were lower (15 %) in a mixed sample of German elite athletes. Markedly lower case rates, based on clinician diagnosis, were reported from a large French study (a mixed sample), with a lifetime prevalence rate for any disorder of 25.1 % and recent diagnosis rates of 8.6 % for anxiety, 4.9 % for eating disorder and 3.6 % for depression (and 0.6 % for suicidal ideation). This study did, however, report relatively frequent sleep problems (delayed onset, frequent waking and daytime drowsiness) in 21.5 % of athletes. Major life events, including injury and chronic stress, were associated with higher rates of distress, anxiety and depression. The self-report general-prevalence studies reporting smaller sample sizes (N < 100) were relatively heterogeneous in outcomes. While male athletes tended to report lower anxiety than their female counterparts, sex differences in depression were inconsistent. Lower ratings of depression and distress were reported in highly achieving athletes and in older versus younger female athletes, with global mood and anxiety predicting athlete performance.

18The studies on elite athlete substance use focused on either alcohol, or other drugs (see Table 5). Studies related to alcohol use indicated higher rates of consumption in athletes relative to the general community. However, during the playing season, the rates of risky alcohol use may be lower than general community levels. Studies focusing on illicit drugs suggested relatively low use (7–8 %) in the previous year but significantly higher combined lifetime use when athletes were asked if they knew other athletes who used illicit drugs (up to 45 %). These studies reported that knowing another athlete who had used drugs was a significant predictor of their own use. In terms of drug-related attitudes, there was a tendency for athletes to overestimate levels of drug use in those competing in sports other than their own. One intervention study that examined the long-term effects of an illicit drug–testing programme found that rates of positive tests among athletes declined over an extended period, and the authors attributed this to education, harm minimisation and testing frequency. One study examined athlete knowledge regarding the effects of illicit drug use. Despite potential stigma relating to athletes seeking drug-related information (i.e. fellow players or staff assuming that information seeking equates to actual drug use), a substantial proportion of elite athletes expressed a desire to receive additional information regarding the effects of some classes of recreational drugs.

19There were 22 studies examining stress and coping in elite athletes (see Table 6), comprising nine longitudinal studies and 13 cross-sectional observational studies. Injury, errors on the sporting field, fatigue and club/organisational climate were identified as common sources of stress among elite athletes. Four studies emphasised the impact of the coach in setting the organisational climate, noting the negative implications of a performance culture over a mastery culture for athletes’ stress. Longitudinal studies found that stress related to injuries, external distractions and fatigue was highest during training periods, while opponents, officials/umpires and the crowd were the predominant sources of stress for athletes during competition.

20The majority of the retrieved studies examined the different strategies employed by elite athletes to cope with the various stressors that were encountered. Adaptive, active coping strategies (such as problem solving, use of imagery, seeking social support and planning ahead) were frequently reported; however, there was a tendency for athletes to engage in less adaptive (i.e. avoidance-coping) strategies when faced with unexpected stressors. One study found that coping behaviour varied between sporting types, with athletes in team-based sports more likely to seek social support than individual competitors, and female athletes more likely to engage in strategies focused on managing affect. Coping strategies based on problem solving and behavioural change were found to be the most effective in managing stress and developing resilience. Emotional reactivity and stressful life events were associated with poor on-field performance and injury.

21There were single studies identified for the domains of help seeking, sleep and wellbeing (see Table 7). The help-seeking study was the only randomised, controlled trial that met the inclusion criteria. While this small study (N = 59) found no difference between athletes in the intervention conditions and control conditions with regard to attitudes, intentions and behaviours related to mental health help seeking, significant improvements were noted in depression and anxiety mental health literacy scores, as well as stigma, at 3-month follow-up. The one study examining sleep reported on a small sample of Australian Rules footballers (N = 19) and found that match-related interstate air travel exerted relatively minimal effects on athlete sleep quality. Finally, the study examining athlete wellbeing identified distinct profiles, with feeling unappreciated, greater perfectionism and lower self-esteem impacting on athlete general wellbeing.

22Researchers have emphasised the limited peer-reviewed literature regarding the mental health and wellbeing of elite athletes. This narrative systematic review is the first to synthesise data from the existing knowledge base with the goal of identifying the incidence and/or nature of mental ill-health and substance use in elite athletes. Given the paucity of research in the field, the present review took a broad and inclusive approach to study both outcomes and designs. In doing so, it identified the relatively poor overall quality of study reporting to date and the lack of well-designed, intervention-based research in the area of elite athletes’ mental health and wellbeing. Despite the limitations of the extant literature, a number of key observations and tentative conclusions can be drawn from our data synthesis.

23The data from studies reporting larger samples, although limited in scope, suggest that elite athletes experience a broadly comparable risk of high-prevalence mental disorders (i.e. anxiety, depression) relative to the general population. That said, there may be subgroups of athletes at elevated risk of mental ill-health, including those in the retirement phase of their careers or those experiencing performance failure. As in the general population, major negative life events, including injury, may increase the risk of mental ill-health in elite athletes, though focused quantitative studies with adequate follow-up assessment periods are needed to confirm this. Findings regarding the prevalence of eating disorders and body image concerns relative to the general population were inconsistent. However, there was a tendency for higher vulnerability to these conditions in athletes involved in sports requiring a particularly lean body shape and in female athletes—the latter being consistent with the findings of general population studies. Objective data, based on the results of medical review and tests, would likely assist in the assessment of eating pathologies and help counter the limitations of self-reporting (i.e. underreporting). Low social support was noted as a key risk factor for general mental ill-health, highlighting the importance of both formal and informal support networks for athletes. All of the included general-prevalence studies were cross-sectional in nature. A natural advance for the field will be to assess athletes prospectively and better identify factors within the competitive spheres (i.e. performance or team success) and non-competitive spheres (i.e. approach coping, social support) for managing symptoms of mental ill-health. Given the significant overlap between the competitive years for elite athletes and the peak onset of mental disorders, future work should also assess low-prevalence disorders, such as psychosis or mania, in order to detect and direct at-risk athletes to early-intervention programmes or services.

24Contextual factors also appear important for athlete substance use, though more rigorous studies are needed. For example, no research has examined illicit substance dependence in elite athletes. Nonetheless, higher rates of alcohol use may occur in elite athletes relative to the general population, though this may be largely due to a binge pattern of consumption during non-competitive or vacation periods. Rates of self-reported illicit drug use were relatively low in the previous 12 months (i.e. 8 %), which may be due to rigorous drug-testing procedures. Further targeted research in the domain of athlete substance use is warranted, given the frequent harmful effects (e.g. fighting) reported and the possibility of patterns of misuse developing in the transition to retirement. Given the possible stigma (i.e. assumed use) associated with elite athletes seeking drug information, improvements to specific, targeted and accessible (i.e. internet-based) information may be warranted.

25The literature related to athlete coping is more established than that for mental health outcomes. Most studies evaluated coping strategies employed by athletes to manage performance-related and non-performance-related stressors. In this way, coping was general in nature relative to psychosocial stressors (i.e. managing poor performance, injury or content on social media) rather than specific strategies for coping with a diagnosable disorder. Adaptive and maladaptive coping strategies were reported, though there was a lack of studies that sought to improve athlete coping. While a small-scale (N = 59), internet-based intervention failed to boost help-seeking-related attitudes, intentions or behaviours, increases were noted in mental health literacy—an essential component of the help-seeking process. Given that stigma, poor mental health literacy and negative past experiences of mental health help seeking are key barriers for elite athletes, more well-designed studies, drawing on larger samples, are needed. Common athlete-specific stressors noted across studies included injury, poor performance, fatigue and organisational factors, such as the coaching environment and coaching expectations. The consistency of findings related to athlete stressors highlights these areas as potential avenues for targeted skills-based intervention programmes, including problem solving and resilience training.

26Management of athlete-specific stressors was also highlighted in the included studies focusing on athlete performance-related anxiety. Improved coping may be enabled by coaching staff emphasising a supportive training culture whereby athletes can interpret performance-related anxiety as facilitative, developing approach (as opposed to avoidance) strategies. Indeed, coaching staff themselves were identified as critical to setting the organisational climate—in turn, impacting on the level of stress experienced by athletes. Given the positive associations between coaches emphasising a mastery climate relative to a performance climate, future mental health intervention-based research should ensure the involvement and support of key coaching staff.

27As indicated, the overall study quality in this field is poor, and heterogeneous study outcomes prevented the application of meta-analytic techniques. In addition, the nature of participant self-selection may have reduced the representativeness of the findings. While the included studies focused specifically on elite-level competition at the national or international level, differences between the included sports in terms of training, remuneration, media pressure and other salient variables must be considered. Our study did not include athletes with disabilities—a population in which relatively little is known about mental health outcomes. Furthermore, most studies used self-reporting rather than a diagnostic interview; therefore, the extent of psychiatric disorders, as opposed to mental health symptoms or probable ‘caseness’, in this population remains largely unknown. While some mental health domains were relatively well represented by the included studies, other domains—particularly athlete anger and aggression, help seeking and sleep—had very few studies. In addition, the included studies generally failed to include assessment of athlete psychological strengths, and only one included study assessed wellbeing-related outcomes. Further, there is a lack of research focusing on the transition from playing or competing in elite sport to retirement. Despite this, the current findings are useful for informing the next generation of studies focusing on elite athlete mental health.

28The last two decades have witnessed extraordinary progress in sports medicine, performance coaching and elite athlete nutrition. As comparatively little progress has been made in the area of mental health, there is enormous scope for programmes to boost athlete wellbeing, which would likely flow on to benefits in competitive performance and increase the likelihood of a successful transition to retirement. Although vulnerability to mental ill-health might well be relatively comparable in elite athletes relative to the general community, there is significant scope for coaches, team psychologists and sport administrators to focus on targeted screening and early detection, monitoring and intervention—especially at key risk periods, such as significant injury, transition to retirement and following performance difficulties. Specific mental health help-seeking interventions are being developed for collegiate-level athletes, with a randomised, controlled trial currently underway. The results of this work are likely to be relevant to elite-level athletes. Encouraging progress has been made in the development of mental health guidelines for working with school-aged athletes and collegiate-level athletes. These guidelines show a growing emphasis on the need to provide specific and targeted support for the mental health needs of athletes. They highlight the importance of monitoring changes in specific observable behaviours, appreciating psychological history and the need for a responsive crisis intervention framework specific to athletes. Development of comprehensive, targeted, disorder-specific treatment models are a required next step, and the National Athletic Trainers’ Association statement on preventing, detecting and managing disordered eating provides a useful disorder-specific model. Psychoeducation should also extend to substance use—in particular, alcohol—given the tendency for hazardous use (bingeing) outside competition periods and the stigma related to athlete help seeking in this domain.

29Development of specific models of psychiatric intervention for elite athletes with significant psychopathology and impairment appears to be warranted. Such models should capitalise on an early-intervention framework, ensuring early detection and prompt access to high-quality, evidence-based interventions. This may include implementing mental health screening programmes alongside physical health checks, in addition to dissemination of mental health awareness support to key support people, including partners, friends, family, coaching staff and administrative staff. For this to occur, collaborative efforts would be required between sports medicine practitioners, psychiatrists, psychologists and other mental health professionals, mindful of overcoming treatment barriers and stigma for athletes at the elite level and balancing the need for treatment with the need for ongoing performance in, or commitment to, their chosen career.

30In addition, from a broader public health perspective, better engagement of elite athletes in the domains of positive mental health (and as identifiable role models or ambassadors) may be significant in mental health stigma reduction and in boosting help-seeking behaviours and engagement in services. Importantly, the mental health of hard-to-engage populations, such as young men and older men, could be targeted. For this, the research will need to expand from simple cross-sectional designs and develop innovative strategies to improve athlete help seeking. Technology heralds a unique opportunity for this, especially given that the next generation of elite athletes will be digital natives and highly adept at utilising computer-based or internet-based interventions. Finally, given the paucity of high-quality studies reported to date, we encourage sporting bodies to consider public dissemination of any research (subject to ethical conduct) that is being conducted within the field of elite athlete mental health. Such efforts will enable the field to prosper and develop.

31Elite athletes experience a unique range of stressors that may potentially increase their vulnerability to mental ill-health. Key factors include the psychological impacts of injury, overtraining and burnout; intense public and media scrutiny; and managing ongoing competitive pressures to perform. For the assessment and management of the mental health needs of elite athletes’ to be on a par with their physical needs, more high-quality epidemiological and intervention studies are needed. Ideally, where possible and appropriate, the results of these should be disseminated beyond the organisation or sporting code.

32Below is the link to the electronic supplementary material.