Article Text

Download PDFPDF

Depressive symptoms in high-performance athletes and non-athletes: a comparative meta-analysis
  1. Paul Filip Gorczynski1,
  2. Melissa Coyle2,
  3. Kass Gibson2
  1. 1 Department of Sport and Exercise Science, University of Portsmouth, Portsmouth, UK
  2. 2 Physical and Coach Education Department, University of St Mark and St John, Plymouth, UK
  1. Correspondence to Dr Paul Filip Gorczynski, Department of Sport and Exercise Science, University of Portsmouth, Spinnaker Building, Cambridge Road, Portsmouth PO12ER, UK; paul.gorczynski{at}port.ac.uk

Abstract

Objective To assess whether a difference exists in the prevalence of mild or more severe depressive symptoms between high-performance athletes and non-athletes.

Design Comparative OR meta-analysis.

Data sources We searched PsycINFO, PubMed, MEDLINE, CINAHL, SPORTDiscus and Google Scholar, as well as the reference lists of reviews of mental health issues in high-performance athletes.

Eligibility We included studies that compared high-performance athletes and non-athletes, included a validated measure of depressive symptoms and included the prevalence of individuals who indicated at least mild depressive symptoms.

Results Five articles reporting data from 1545 high-performance athletes and 1811 non-athletes were examined. A comparative OR meta-analysis found high-performance athletes were no more likely than non-athletes to report mild or more severe depressive symptoms (OR=1.15, 95% CI=0.954 to 1.383, p=0.145). Male high-performance athletes (n=940) were no more likely than male non-athletes (n=605) to report mild or more severe depressive symptoms (OR=1.17, 95% CI=0.839 to 1.616, p=0.362). For females, high-performance athletes (n=948) were no more likely than non-athletes (n=605) to report mild or more severe depressive symptoms (OR=1.11, 95% CI=0.846 to 1.442, p=0.464). Overall, male high-performance athletes (n=874) were 52% less likely to report mild or more severe depressive symptoms than female high-performance athletes (n=705) (OR=0.48, 95% CI=0.369 to 0.621, p<0.001).

Summary/conclusions High-performance athletes were just as likely as non-athletes to report depressive symptoms. Researchers need to move beyond self-report measures of depressive symptoms and examine the prevalence of clinically diagnosed depressive disorders in athletes.

  • mental health
  • depressive symptoms
  • elite athletes
  • meta-analysis

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Recent research has indicated that high-performance athletes are at risk for mental health problems.1 Researchers have pointed to several unique stressors and challenges that high-performance athletes face that may partially explain mental health problems, such as depression and anxiety, in this population. Such unique challenges and stressors include demands of competition and training,2 3 as well as dealing with injury and recovery.4 Additionally, some high-performance athletes are also concerned with body image, especially among those who compete in aesthetic sports such as diving and gymnastics, which may result in eating disorders and substance use.4 Research has also shown that concerns with retirement, especially when brought on suddenly and unexpectedly, leave high-performance athletes feeling vulnerable and depressed.5 6

There are several reasons why high-performance athletes may wish to not disclose their mental health status and seek support. As found in the general population, athletes fear dealing with the stigma associated with mental illness and the consequences that stem from it.4 7 8 Perhaps most pertinently, athletes do not wish to disrupt their athletic careers.9 As a result, many athletes may decide to stay silent about their mental health and not receive the support and treatment that may be necessary. That being said, some high-profile athletes have disclosed mental health challenges and illnesses as part of consciousness raising exercises, although these are predominately retired athletes. Ultimately, non-disclosure may mean that mental health problems, especially depression, among high-performance athletes may be far more prevalent than what current research findings show.8 However, given mental illness non-disclosure among the general population, it is not possible to say that non-disclosure affects high-performance athletes more than those in the general population and that high-performance athletes would be less likely to report symptoms that may be indicative of mental illness. In fact, few direct comparisons between high-performance athletes and the general population have been made. Although several recent narrative reviews have examined depression in high-performance athletes,10–12 no systematic approaches to examine the literature have been taken. Furthermore, no meta-analytic approaches have been made to directly compare prevalence rates of self-reported depressive symptoms among high-performance athletes and non-athletes. As a result, the purpose of this meta-analysis was to directly compare high-performance athletes and non-athletes with respect to prevalence of reporting mild or more severe depressive symptoms as measured by validated measures. Findings will help illustrate if high-performance athletes are more likely to report depressive symptoms and if additional support strategies are necessary with this population.

Methods

To ensure transparency and comprehensive reporting of the methods and results, this meta-analysis adhered to the PRISMA statement.13

Search strategy and selection criteria

To identify appropriate studies, published work and grey literature were examined in August 2016. No limits were imposed on study types or study dates while grey literature included programme evaluations, reports, book chapters and reviews. Grey literature was used primarily to identify additional eligible studies. The following databases were searched: PsycINFO, PubMed, MEDLINE, CINAHL and SPORTDiscus. Search terms for PsycINFO, MEDLINE, CINAHL and SPORTDiscus included ‘elite athlete’ AND ‘dep*’; ‘elite sport’ AND ‘dep*’. For searches in PubMed, we used the following Medical Subject Headings (MeSH) terms: ‘depression’, ‘depressive disorders’, ‘athletes’ and ‘sports’. A hand search of the reference lists of identified literature was conducted, including a hand search of recent reviews of mental illness in high-performance athletes. Google Scholar was also searched using the same keywords to locate additional relevant studies.

For this review we focused specifically on depressive symptoms as measured by a validated measure. Following Swan et al,14 the term ‘high-performance athlete’ was equated to ‘elite athletes’ as defined on a continuum ranging from semi-elite (eg, participating in high-performance youth development programmes, competitive-level college sports, programmes below top-level leagues) to world-class elite (eg, participating and winning consistently at the Olympics or other world-stage events). Other classifications included competitive elite (eg, participating regularly at highest level competitions) and successful elite (eg, participating regularly at highest level competitions and experiencing infrequent success). Non-athletes were defined as individuals who were not engaged in any sports at a semi-elite, competitive elite, successful elite or world-class elite level. Ultimately, studies that met the following inclusion criteria were reported in the review:

  1. published in full in English;

  2. compared high-performance athletes and non-athletes;

  3. included a validated measure of depressive symptoms; and

  4. included the prevalence of individuals who indicated at least mild depressive symptoms.

Studies were excluded if they did not include a control sample of non-athletes, did not report depressive symptoms using a validated measure of depressive symptoms or did not report the number of individuals who indicated symptom severity using a validated measure of depressive symptoms. Where studies did not report fully on depressive symptoms, the corresponding author of the study was contacted to obtain this information. Two reviewers (PG and KG) screened the articles independently to assess their eligibility. The reviewers met to confirm study eligibility, and any discrepancies between the authors were resolved through discussion until an agreement was reached.

Data extraction and analysis

Data extraction was conducted by two of the study authors independently (PG and KG). The authors then met to ensure all appropriate data were extracted. A systematic tool was designed and used to facilitate data extraction of the following information:

  1. Year of publication, country, sample size, recruitment strategy, age, sex, level of high-performance athleticism, non-athlete population, type of sport and study type.

  2. Depressive symptoms: (i) scale used and data collection methods, and (ii) prevalence of mild or more severe depressive symptoms. Identical data were extracted for the non-athlete control samples.

The primary outcome was the overall prevalence of at least mild or more severe depressive symptoms in high-performance athletes and non-athletes, as assessed with validated measures.

Statistical analyses

All statistical analyses were conducted in OpenMetaAnalyst.15 Random-effects models were applied to all meta-analyses to account for heterogeneity.16

The proportions of high-performance athletes and non-athletes indicating at least mild depressive symptoms were pooled across all studies reporting this variable in order to calculate a weighted estimate with 95% CIs. The prevalence of at least mild depressive symptoms in high-performance athletes and non-athletes was reported using OR meta-analysis, with non-athletes used as the reference group. A subgroup analysis of sex differences was also conducted, with females used as the reference group. Specifically, a comparative OR meta-analysis examined depressive symptoms between males and females in both high-performance athletes and non-athletes. Variance between studies was assessed using Cochran’s Q and reported as I2. Visual inspection of funnel plots assessed the degree of potential publication bias.

Risk of bias

A 10-item risk of bias in prevalence studies tool was used to assess the internal and external validity of each study.17 Specifically, the risk of bias assessment provided an overview of the main methodological characteristics of the reported studies. Two study authors (PG and KG) independently assessed each study for risk of bias. The overall agreement between the study authors was considered excellent (kappa statistic=0.90).

Results

The PRISMA search process is presented in figure 1. The initial database search returned 855 articles and a hand search identified an additional 17 articles. Of the 872 articles, 193 duplicates were removed and 679 articles were screened. In total, 25 full-text articles were assessed for eligibility and 20 were excluded for the following reasons: having no non-athlete comparison group (n=15),6 18–31 no reported prevalence of depressive symptoms (n=4)32–35 and no measure of depressive symptoms (n=1).36 A list of full-text excluded articles is presented in table 1.

Table 1

Full-text articles excluded with reasons

In total, five articles were included in this review, reporting data from 1545 high-performance athletes and 1811 non-athletes.37–41 Ages for high-performance athletes ranged from 12 to 41 years and 54.0% of the athletic participants were female. Studies were conducted in Iran, Germany, Switzerland and USA (n=2). All studies were cross-sectional and used validated measures of depressive symptoms. Validated measures included the Beck Depression Inventory-II,42 Center for Epidemiological Studies Depression Scale,43 Composite International Diagnostic-Screener44 and the depression subscale of the scales of the Personality Assessment Inventory.45 Study characteristics are detailed in table 2.

Table 2

Summary of included studies

The included studies showed to be of low risk for bias for prevalence studies, ranking highly in both overall measures of internal and external validity. The risk of bias results can be seen in table 3. A visual inspection of a generated funnel plot revealed no publication bias.46 The publication bias funnel plot can be seen in figure 2.

Figure 2

Publication bias funnel plot.

Table 3

Risk of bias assessment

Depressive symptoms’ prevalence rates

Prevalence rates for mild or more severe depressive symptoms were reported by all studies. The rates of mild or more severe depressive symptoms ranged from 3.7%41 to 26.7%38 for high-performance athlete males and from 9.8%41 to 36.5%37 for high-performance athlete females. For non-athletes, the rates ranged from 7.6%41 to 34.4%38 for males and from 6.1%41 to 42.5%38 for females. A comparative OR meta-analysis found that high-performance athletes (n=1545) were no more likely than non-athletes (n=1811) to report mild or more severe depressive symptoms (OR=1.15, 95% CI=0.954 to 1.383, p=0.145). There was no between-study heterogeneity (p=0.453, I2=0%) (figure 3). When sex differences were examined, a comparative OR meta-analysis found that male high-performance athletes (n=940) were no more likely than male non-athletes (n=863) to report mild or more severe depressive symptoms (OR=1.17, 95% CI=0.839 to 1.616, p=0.362). There was negligible heterogeneity between studies (p=0.276, I2=21.71%) (figure 4). For females, a comparative OR meta-analysis found that high-performance athletes (n=948) were no more likely than non-athletes (n=605) to report mild or more severe depressive symptoms (OR=1.11, 95% CI=0.846 to 1.442, p=0.464). There was no between-study heterogeneity (p=0.518, I2=0%) (figure 5). Overall, male high-performance athletes (n=874) were 52% less likely to report mild or more severe depressive symptoms than female high-performance athletes (n=605) (OR=0.48, 95% CI=0.369 to 0.621, p<0.001). There was no between-study heterogeneity (p=0.547, I2=0%) (figure 6). For non-athletes, a comparative OR meta-analysis found that males (n=812) were 42% less likely to report mild or more severe depressive symptoms than females (n=948) (OR=0.58, 95% CI=0.350 to 0.977, p=0.040). There was substantial heterogeneity (p=0.024, I2=68.14%) (figure 7).

Figure 3

Relative odds of mild or more severe depressive symptoms in non-athletes versus high-performance athletes.

Figure 4

Relative odds of mild or more severe depressive symptoms in non-athletes versus high-performance athletes.

Figure 5

Relative odds of mild or more severe depressive symptoms in non-athletes versus high-performance athletes.

Figure 6

Relative odds of mild or more severe depressive symptoms in high-performance athletes.

Figure 7

Relative odds of mild or more severe depressive symptoms in non-athletes.

Discussion

The purpose of this meta-analysis was to compare high-performance athletes and non-athletes with respect to reporting of mild or more severe depressive symptoms. The results from this meta-analysis show that high-performance athletes and non-athletes do not differ with respect to reporting mild or more severe depressive symptoms. Female high-performance athletes were twice as likely to report mild or more severe depressive symptoms than male high-performance athletes. This is the first meta-analysis to aggregate comparative findings for depressive symptoms between high-performance athletes and comparative samples of non-athletes.

Findings from this meta-analysis are in support of previous findings from other reviews and studies that have showed similar levels of prevalence for reporting depressive symptoms between high-performance athletes and non-athletes.10 12 Ultimately, the results show that high-performance athletes were just as likely as non-athletes to report depressive symptoms. Previous research does state that because of the stigma surrounding depression, high-performance athletes may try to ignore or suppress depressive symptoms and may not wish to seek help from sports psychologists and other mental health specialists.47 Given that high-performance athletes may be more skilled at or have more sophisticated training opportunities to develop mental toughness skills, which are defined as a set of mental attributes that allow individuals to cope with stressful situations and more consistently remain determined, focused, confident and in psychological control,48 one may speculate that elite athletes may not report depressive symptoms by either remaining, or seeking to appear, mentally tough. After all, to report depressive symptoms may be perceived as an admission of mental weakness, which may ultimately provide an opponent an advantage. Unfortunately, this perception of mental weakness would only perpetuate the negative attitudes associated with mental illness in sport and prevent athletes from seeking help. Equally, non-athletes may choose to ignore, suppress and not seek help for depressive symptoms for similar stigma-related reasons as well, yet do not have access to sophisticated training opportunities to develop mental toughness skills. As such, to further and better investigate the prevalence of depressive symptoms in both populations, data from clinical diagnoses in the form of structured clinical interviews rather than self-reported measures are necessary.

Although no differences were seen between athletes and non-athletes, findings from this meta-analysis confirm that female athletes were more likely to report depressive symptoms than male athletes.31 41 49 The twofold increased risk for depressive symptoms in high-performance athlete females found in this meta-analysis is similar to what has been found in non-athletes.50 It is believed that this phenomenon is the result of many different and complicated social and psychological factors, including adverse experiences in childhood and upbringing, sociocultural roles, and adverse experiences (eg, greater likelihood of victimisation related to sexual harassment, chronic everyday burdens related to social status), as well as psychological attributes related to stress and coping (eg, negative self-concept, inward focus on feelings).51 52 Strategies should be designed to render more support to female athletes in terms of understanding and acknowledging depressive symptoms, as well as seeking help. Additionally, it must be stressed that this study focused on self-reported measures of depressive symptoms. Therefore, findings indicate that female athletes are more likely than male athletes to report their depressive symptoms rather than have a higher prevalence of depressive symptoms per se. Research that further helps address environmental and cultural factors, particularly gendered attitudes and values, that shape reporting behaviours is urgently required.

Because of the limited number of included studies, mild or more severe depressive symptom prevalence reporting was not examined in different sports. Future research should explore different individual and team sports as stressors, levels of responsibility and rates of injury, such as concussions, vary greatly.25 Although Wolanin and colleagues49 showed differences in depressive symptoms, with athletes involved in track and field and cheerleading carrying the greatest risk among high-performance student athletes, these findings are not consistent with previous research conducted by Schaal et al.28 Additionally, other determinants of depressive symptoms may wish to be further explored, including various demographic variables (eg, age, race, relationship status, sexuality), physical health variables (eg, injuries, surgeries), mental health variables (eg, stress, self-esteem) and sociocultural elements (eg, social connectedness, support structures, mental toughness).

Despite a robust search and evaluation of included studies, a number of limitations with this meta-analysis must be pointed out. First, efforts were made to reach study authors by e-mail for additional information about depressive symptom prevalence. Such information would have expanded the number of eligible studies and provided a more robust understanding of comparing depressive symptom prevalence between high-performance athletes and non-athletes. Ultimately, a small number of studies met the eligibility criteria and were included in the meta-analysis. This potentially limits the generalisability of our results. Second, most included studies examined student high-performance athletes and student non-athletes. Further comparative research is needed with professional and world-class high-performance athletes and non-student populations. Previous research has found that students tend to report higher rates of depressive symptoms than those in the general population, and this may potentially skew results, indicating that high-performance athletes have much higher rates of mental health problems when compared with the non-athletes.53 Third, four different validated and reliable measures of depressive symptoms were used in this meta-analysis. For consistency and easy comparability, researchers may wish to choose to use one validated and reliable measure that has been used previously, such as the Center for Epidemiological Studies Depression Scale.44 Fourth, we excluded unpublished and non-English-language articles, which exposes the current meta-analysis to publication and language biases. Lastly, each of the included studies is based on self-report measures of depressive symptoms. As such only the prevalence rates of depressive symptoms are reported and not clinical diagnoses of depressive disorders. Future research should consider incorporating the use of structured clinical interviews and examine the prevalence rates of clinical diagnoses of depressive disorders. Researchers should also make every effort to follow rigorous standards in reporting in full their results as to minimise publication bias.

Overall, the results of this meta-analysis show that high-performance athletes were just as likely to report depressive symptoms as non-athletes, and female high-performance athletes were twice as likely as male high-performance athletes to report depressive symptoms. Given the limitations of self-reporting depressive symptoms, researchers should aim to use structured clinical interviews in the future to examine the prevalence of depressive symptoms in high-performance athletes.

What is already known?

  • High-performance athletes face unique stressors in training and competition that predispose them to mental health problems, such as depressive symptoms.

  • It is hypothesised that being exposed to stressful environments may increase the likelihood of high-performance athletes to report mild or more severe depressive symptoms than non-athletes.

What are the findings?

  • Female high-performance athletes were just as likely to report mild or more severe depressive symptoms as female non-athletes.

  • Male high-performance athletes were just as likely to report mild or more severe depressive symptoms as male non-athletes.

  • Male high-performance athletes were 52% less likely to report mild or more severe depressive symptoms than female high-performance athletes.

MCQs

1.%20High-performance athletes face unique stressors that predispose them to depression. These stressors include:

a) competition

b) training

c) recovery after injury

*d) all of the above

2. High-performance athletes do not wish to disclose their mental health status _____________.

*a) due to mental health stigma.

b) because they do not recognise the symptoms of mental health problems.

c) because they do not know where to turn to for support.

d) all of the above.

3. Female high-performance athletes are ________ times more likely than male high-performance athletes to indicate mild or more severe depressive symptoms.

*a) 2

b) 3

c) 4

d) 5

4. Male high-performance athletes are ____________ males in the general population to experience mild or depressive symptoms.

*a) as likely as

b) 2 times more likely than

c) 3 times more likely than

d) 4 times more likely than

5. Female high-performance athletes are ____________ females in the general population to experience mild or depressive symptoms.

*a) as likely as

b) 2 times more likely than

c) 3 times more likely than

d) 4 times more likely than

References

Footnotes

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.