The extraordinary situation in the context of the COVID-19 pandemic—known as the “first lockdown”—was decreed by the Swiss Federal Council lasting from 17 March to 10 May 2020. This included the prohibition of all events and the closure of many public places including sport facilities. Private training was therefore still possible, but competitions and team training were not.

In the middle of this period, a REDCap-based online survey (REDCap 9.10.0 - © 2020 Vanderbilt University) was sent to different elite athletes in Switzerland who performed in Olympic sports or IOC-approved disciplines on a national or international level. This online questionnaire was shared by sports clubs for team sports, whereas for individual sports, it was distributed by the Swiss national sports federations. So that athletes could participate in their mother tongue, questionnaires were offered in German and French. Open questions were primarily designed in German. The French translation was validated by two people proficient in both languages and differences in translation were resolved by bilateral discussion. Regarding the standardized questionnaires, validated translations for both languages were used.

The survey evaluated the 4 weeks before the lockdown as well as the first 4 weeks of the lockdown. For both periods, athletes reported the frequency of video gaming, gaming time, and parameters related to mental health. As demographic parameters, the survey queried age, gender, type of sport and time spent on educational and/or part-time occupational duties. Video game playing time was quantified by frequency (days per month [d/m]) and mean playing time (hours per month [h/m]). We also asked whether the participants were able to earn enough income from their athletic performance to make a living. Finally, we evaluated the performance in athletes (before and during the lockdown) with three different variables: the objective activity (training time hours per day [h/d]), the subjective training intensity and the subjective maximum of athletic performance (both in [%] of the respective z-value). Although athletic performance is usually quantified by athletic success, this was not possible in the current study due to canceled tournaments and events.

With respect to COVID-19 itself, participants were asked whether they had been infected and about their personal quarantine requirements. Furthermore, the questionnaire evaluated to what extent the athletes worried about their athletic career due to the COVID-19 pandemic and how they were personally affected by the lockdown measures.

To understand the psychological effects of the situation, depressive symptoms, sleep behavior, and anxiety characteristics were evaluated using the Patient Health Questionnaire (PHQ9), the Insomnia-Severity-Index (ISI), and certain aspects resp. questions extracted from the Pittsburgh Sleep Quality Index, as well as the Spielberger State-Trait-Anxiety-Index (STAI) (Spielberger et al., 1983; Buysse et al., 1989; Bastien et al., 2001; Kroenke et al., 2001). Also, substance use was evaluated, including alcohol and cannabis (frequency of consumption).

The aspects of fear and anxiety were evaluated in three different questions and standardized questionnaires using the STAI survey. The athletes were asked whether they experienced existential fears (4 weeks before and/or during the lockdown), whether they worried about their careers due to the pandemic, and how well they were able to cope with the restrictions established by the Swiss government. For each of these questions, participants had to choose a value on a scale from 0 to 100, where zero corresponded to “I have no existential fears,” “I have no worries at all about my career due to COVID-19,” and “I cannot cope at all with the measures due to the pandemic.” At the other extreme, the value 100 expressed “I experience very strong existential fears,” “I have major worries about my career due to COVID-19,” and “I cope very well with the measures due to the pandemic.” Additionally, all participants filled out the STAI questionnaire regarding their anxiety traits and states (Spielberger et al., 1983; Spielberger, 2010).

To gain an understanding of the physical complaints of the participants, current health problems like traumatic/overuse injuries or illnesses were assessed using the Oslo Sports Trauma Research Centre (OSTRC) questionnaire on health problems (Clarsen et al., 2013). For the classification of self-reported injuries or illness (yes/no), the first question of the OSTRC was used (“Did you have any difficulties participating in normal training and competition due to injury, illness or other health problems during the past 4 weeks?”).

Inclusion criteria were training volume (before the COVID-19 lockdown) of at least 1 h/d and a minimum age of 18. Participants with incomplete data and those who did not participate in Olympic disciplines or in sports recognized by the IOC were excluded. There was no compensation for participation in this study.

Baseline data was expressed as mean ±SD and frequency tables for categorical data. Demographic data of the gamers and non-gamers were compared using the two-tailed independent sample t-test for continuous data and χ² tests for categorical data. In tables with more than two groups, χ² test was used for tables with more than four fields, one-way analysis of variance for continuous variables, and Kruskal-Wallis rank tests for variables that violated the assumptions of parametric tests. Differences between the groups were considered significant at p < 0.05. For gaming behavior during the lockdown, we fitted a general linear model (GLM) using current gaming time as the dependent variable, previous gaming as the independent variable and gender, team sports, injury, activity, pandemic measures, and occupation percentage as covariates. In an attempt to adjust for psychiatric measures, we added the scores of PHQ9, ISI, TRAIT, and STATE, as well existential fears and worries about career to the model. All explanatory variables that had an association with gaming at P < 0.20 in the univariable analyses were included in the multivariable-adjusted analyses. Using a stepwise backward elimination process, the least significant variables were then removed from the base model. Only variables with P < 0.05 remained in the final parsimonious model.

Dominance analysis was additionally performed as was the calculation of the correlation coefficients between all variables.

The local ethics committee accepted this survey by a declaration of non-responsibility (KEK-ZH-NR: Req-2020-00408).

In total, 203 athletes answered the survey, 193 of whom provided information on their gaming behavior: 87 women (45%) and 106 men (55%), with a mean age of 24.1 years (range 18–38 years). Within this sample, winter (n = 94, 49%) and summer (n = 99, 51%) sports were represented equally, and 59 athletes (31%) were active in team sports. Alongside their activity in sports 116 (60%) of these athletes were employed or studying, with an average workload of 67.5%. From their performance in sports alone, 101 (52%) earned sufficient income, meaning that 8% of the athletes were working or studying even though they earned enough money from their athletic performance alone. Three (2%) had a positive SARS-CoV-2 test during the first period of the lockdown (March and April 2020), and 10 athletes (5%) had to be in quarantine, with an average time of 10.8 days (SD 5.9).

Fifty-two (27%) athletes stated “yes,” when they were asked if they played video games 4 weeks before the pandemic, and 41 (21%) athletes were considered as playing regularly (≥ 1 h/m), with an overall average playing time of 14.6 h/m (SD 17.1) distributed over 9 days per month. Maximum reported playing time was 93 h/m, averaging roughly 3 h/d. Thirty-six gamers (88%) were male and five (12%) were female. The proportion of individual athletes who gamed was 17.2%, whereas 30.5% of team athletes gamed, which represents a significantly higher proportion (P < 0.05). Athletes who gamed regularly tended to work or study more, reporting a higher workload. We found no statistical difference between gamers and non-gamers regarding athletic performance, substance use or self-reported injuries or illness.

Despite spending the same amount of time in bed, gamers reported significant 0.4 h less effective total sleep duration than non-gamers (p = 0.027). In the context of COVID-19 related issues, video gamers coped noticeably better with the measures of the government due to the pandemic (P < 0.05) and worried less about their career due to the COVID-19 pandemic. Overall, athletes who gamed regularly experienced less existential fear in general, and significantly less during the first lockdown (P < 0.05). Six (15%) of the gamers reported an injury or an illness compared to 42 (28%) athletes in the non-gamer group (also see Tables 1, 2).

During the first period of the lockdown, 63 athletes (33%) stated “yes” to the question of whether they had played video games within the last 4 weeks, and 29% of the athletes reported playing video games regularly (>1 h/d) with a mean playing time of 21.8 h/m (SD 22.8). Playing time increased by 165% within the group of athletes who used to play video games before the isolation period.

In our sample, 41 athletes spent more time video gaming, 14 athletes less time, 18 athletes did not change their gaming time and 120 athletes never gamed, neither before nor during the first lockdown. Athletes who gamed more during the lockdown increased their gaming time from 5.5 to 21.8 h/m. and, thus spending on average 16.3 h/m more on video gaming than before. In particular, the group of “occasional” gamers increased their gaming time from 4.6 to 12.6 h/m, therefore, increased the playing time on average by 8.0 h/m [SD 14.3]. Maximum gaming time was 84 h/m.

Compared to all other athletes, those who gamed more tended to be male and used to already play video games before the lockdown (for all applies P < 0.05). Again, summer and winter sports were distributed evenly between the groups, and groups did not differ in their outside employment status or in earning sufficient income from sport. However, team athletes tended to spend significantly more time video gaming during the lockdown than those involved in individual sports (P < 0.05). While 40% (23) of the team athletes were gaming regularly, only 23% (32) of the individual athletes did so. Those who gamed “more” also reported a significantly higher workload when employed or studying (79.2% [SD 25.9]) in comparison to the groups that gamed “less” (67.8% [SD 22.8]), the “same” (74.0% [SD 28.0]) or “never” (63.5% [SD 29.1]) (P < 0.05). Twenty-one athletes started playing video games during the first lockdown and 10 athletes stopped playing video games.

Regarding performance levels, fears, depressive symptoms, sleep, and substance use, the groups “same,” “never,” “more,” and “less” showed no statistical difference within the context of the first lockdown (also see Table 3).

Except for male gender and the fact that the athlete used to play before the COVID-19 lockdown, we found that video gaming time during the lockdown was not affected at a statistically significant level by any of the other variables we evaluated (type of sport, training, occupation, aspects of anxiety, and fears, depressive symptoms and disordered sleeping). In this multivariable regression of gaming time during the lockdown, the influence of male gender was significant, with p < 0.05, and athletes who already regularly played video games before the lockdown were very likely to also play during the lockdown (p < 0.01) (also see Table 4). The statistical power of the final parsimonious regression model is >0.9. Dominance analysis was additionally performed and resulted in the same model as the backwards elimination of the GLM. Results of the dominance analysis are shown in Supplementary Table S5. Due to possible unstable results of stepwise regression, we provide the correlation matrix in order to show all relations between the used variables, presented in Supplementary Table S6.

Internal validity is judged to be according to current standards and reported according to STROBE (Cuschieri, 2019). The generalisability however is only possible with professional athletes in a similar setting.

The gaming time of elite athletes is considerably low in the context of studies concerning gaming time and problematic gaming. Rehbein et al. found that 36% of the general adult population in Germany played video games for 44 h/m on average. Of these, 3.5% played occasionally (at least once in 12 months), 28.2% were “normal” players (daily gaming time up to 2.5 h) and 4.2% were classified as “frequent” (daily gaming time >2.5 h) (Rehbein et al., 2010). Other authors have defined “dedicated gamers” as those who played for approximately 30 h per week and “very excessive gaming” as up to 80 h per week, i.e., four times the maximum playing time within our sample (Chappell et al., 2006; Griffiths, 2010; Ballabio et al., 2017). Therefore, active elite athletes seem to spend less time video gaming, and fewer of them consume video games than among the general population.

Also, the gender gap is notable: only 12% of gamers in our sample were female, which does not follow the current trend of the general population, where the proportion of women is 45% (ESA, 2021).

To date, it has often been hypothesized that elite athletes may be prone to problem gaming (Håkansson et al., 2018; Reardon et al., 2019; Bitar et al., 2022) and it can in fact be challenging to distinguish between “enthusiastic” gaming and “problematic” gaming. There is as yet no clear differentiation between these behaviors, as this is still a clinical issue requiring more research (APA, 2013; Király et al., 2015). The existing orienting criteria for defining problematic gaming are dichotomous, and defined by the circumstance, that the patient is severely impaired in his performance or functionality over a time period of at least 12 months (APA, 2013). It is therefore not likely to find this level of gaming behavior among functioning elite athletes. Also, in this survey, no standardized screening tool for problematic gaming was used, so we are not able to make definite statements regarding problematic gaming. However, in our sample, we detected no excessive gaming time, severe performance impairment, or higher level of mental symptoms within the gamer group.

The situation after the outbreak of the COVID-19 pandemic can be described as an unexpected and significant disruption of the athletes' lives and perspectives. Such situations require behavioral adaptations and high cognitive and emotional flexibility (Pété et al., 2022). Also, several studies showed reduced well-being, sleep quality and physical performance in team athletes during isolation period or during COVID-19 infection, arguing that awareness of the effects of COVID-19 on elite team athletes need to be established (Mon-López et al., 2020; Wagemans et al., 2021).

As expected, during the first weeks of the first lockdown, gaming time increased significantly, especially for those athletes who had only occasionally gamed before. In this particular situation, video gaming appears to be an effective part of the adaptation process; with the restrictions on movement and limitation of training possibilities, athletes could interact with friends and team colleagues through gaming. The proportion of team athletes in the sample population who increased their playing time was significantly higher than in individual sports. This may be due to the more drastic restrictions in team training and fewer limitations on individual training: team athletes had more spare time and may have been missing the social component of team training. This thesis corresponds to a study from Italy, in which athletes reported that being in contact with colleagues and coaches through the internet was helpful when managing the lockdown (di Fronso et al., 2022).

Athletes who had regularly gamed before the pandemic tended to be less anxious, had fewer depressive symptoms, and displayed lower scores regarding sleep disturbances. However, no statistical significance could be found. We did not find significant impairments in athletic performance before the pandemic and in the first period of the lockdown. Also, among athletes with a secondary employment or those who were studying alongside their sports, gamers tended to have a higher workload than non-gamers. Athletes who gamed regularly seemed to cope significantly better with the measures of the Swiss government.

Therefore, video games may have been used as a positive coping mechanism, similar to seeking social support and to active stress coping. Elite athletes are used to deal with stressors due to their intense training and competition routine (Daumiller et al., 2021). Hence, it is likely they were able to use video gaming as a coping strategy under the circumstances of the first lockdown.

As a key limitation of this study, the absence of a control group must be mentioned. The questionnaires were voluntary and based on self-reporting. Therefore, they may suffer from certain biases, such as selective recall due to impaired memory recall and social desirability. Also, the recruited sample of professional athletes in Switzerland may not be representative due to selection bias. Consequently, external validity could be impaired. Furthermore, the differences between the gamer types must be interpreted with caution due to small sample sizes, especially in the group of “frequent gamers.” Another bias, that should not be neglected, is the constellation of the “gamers,” consisting primarily of male team-athletes. Hence, any observed tendency of the “gamers” could also originate from this factor. Another note of caution, when interpreting the results, is the fact, that stepwise regression can throw out unstable results. Finally, it has to be kept in mind that athletes' careers tend to start at an early age and often last less than two decades. In our analysis we evaluated a sample of young and active elite athletes. We cannot estimate the video gaming behavior after the end of a sports career or after a major injury or illness. Also, more qualitative, contextual information would be desirable, e.g., motives and game types. Thus, further follow-up data is needed to evaluate the stability of our results with respect to the psychological symptoms.

Our findings suggest that playing video games may have a positive functional role for active elite athletes. Especially in the context of the COVID-19 pandemic, it appears to help with an effective adaptation to the current circumstances. Pété et al. (2022) evaluated different coping profiles in different athletes during the COVID-19 pandemic in France, pointing out the importance of finding mechanisms to cope with the impact of pandemic on the mental health of elite athletes on a long-term basis. In this regard, video games could open up many helpful possibilities for sports clubs and athletes in terms of athletic training and social interaction. Since neurobiological studies have proven that some types of video games are beneficial to neuroplasticity (e.g., strategy, 3D adventure, puzzles), athletes could also continue training their fine motor skills and reaction time in order to maintain at least some of their athletic requirements (Brilliant T et al., 2019). In respect to further research and transfer of our data, prospective studies with unexpected and sudden events, that are defined by less training, less social interaction, change in social framework and the loss of planned successes, could be promising in order to elaborate video gaming as a coping mechanism. For example, injuries, non-consideration, or end of career show these characteristics for elite athletes.

Ultimately, video gaming may act as a stabilizing factor for elite athletes in the context of the COVID-19 pandemic and similar stressful situations.