Self-tracking via smartphone app: Potential tool for athletes’ recovery self-management?

Performance and recovery management increasingly comprises monitoring external and internal training load and the athlete’s response. Monitoring may help to maintain performance capacity and health, including injury prevention and optimizing recovery. From a holistic point of view, the main goal is to achieve balance between effective training impulses and sufficient recovery and rest periods (Kellmann et al., 2018). Monitoring systems have evolved over the past years and a number of approaches are available to athletes, coaches, and sport scientists. While training-related assessments need to be tailored to the specific characteristics of each type of sport, global approaches such as subjective self-report measures of well-being and recovery-stress states are available on a broad level (Kölling & Kellmann, 2020).

However, monitoring is only half the battle. Recovery self-management and self-regulation have to be considered as an important task of successful athletes—on the elite as well as on the leisure level. Recognising one’s current recovery need and acting accordingly, i.e. initiating and applying appropriate measures, is crucial (Balk & Englert, 2020). Since sleep is considered one of the most efficient and important recovery strategies, it seems an obvious target when it comes to optimizing regeneration (Kölling, Duffield, Erlacher, Venter, & Halson, 2019; Walsh et al., 2021). At the same time, sleep is highly vulnerable to external and internal stressors, which are prevalent among athletic as well as university populations (Doherty, Madigan, Nevill, Warrington, & Ellis, 2021; Wang & Bíró, 2021). While training/competition and schedule-related stress may be one reason for unrestful or insufficient sleep, dysfunctional cognitions and behaviours may be another influencing factor (Hiller, Johnston, Dohnt, Lovato, & Gradisar, 2015; Kölling et al., 2019; Kroese, Evers, Adriaanse, & de Ridder, 2016).

As the sleeping process is inherently beyond the sleeper’s consciousness, objective assessments that provide information about sleep-related events may be quite appealing. While standardised measures such as polysomnography and actigraphy are the method of choice in research and sleep medicine, these may be too costly and complex for the average athlete (Halson, 2019; Shelgikar, Anderson, & Stephens, 2016). Self-tracking wearable technologies and smartphone applications have made the approach to one’s sleep behaviour more easily accessible, although their validity and reliability are considered highly doubtful (Khosla et al., 2018). Possible benefits are that they enhance the user’s awareness of their sleeping patterns (Watson, Lawlor, & Raymann, 2019). However, several drawbacks are discussed among experts and even negative effects of the usage of smartphone applications may be observed (Baron, Abbott, Jao, Manalo, & Mullen, 2017; Shelgikar et al., 2016; Van den Bulck, 2015). For instance, users may become too obsessed with self-optimisation and overemphasise the dubious feedback. Especially people with subjective sleep complaints and those concerned about their health are assumed to be prone to misguidance via self-tracking consumer technologies (Baron et al., 2017). It is hypothesised that ambitious and performance-oriented athletes represent a target group that is attracted by those technologies. This assumption needs yet to be confirmed.

Moreover, the 2020 confinements due to the coronavirus disease 2019 (COVID-19) significantly affected sleep and mental health of Australian athletes (Facer-Childs, Hoffman, Tran, Drummond, & Rajaratnam, 2021) as well as the psychobiosocial state of Italian athletes (di Fronso et al., 2022). While physical activity of healthy adults generally declined during lockdown in the USA, this decline was buffered by the use of fitness apps (Yang & Koenigstorfer, 2020). It is possible that the lockdown regulations in Germany raised athletes’ self-awareness and openness to explore their physical activity and sleep. The restricted access to sport facilities and organised activities as well as the promoted social distancing and staying-at-home measures may have contributed to sensitise the individuals to focus more on their psychophysiological state. As training and exercise opportunities were limited, more attention could be paid to assess and optimise recovery processes. In this context, athletes may consider inexpensive and easy-to-use tools as a convenient opportunity to track and analyse physical activity, sleep, and nutrition. Therefore, it is hypothesised that the majority of athletes will report to use either smartphone applications or wearable technologies to support their training and recovery management.

In summary, athletes on either performance level in the pursuit of optimising performance and recovery constitute a vulnerable group that might be attracted to self-tracking technologies. If usage was highly prevalent, the characteristics of users need to be investigated. For instance, it is conceivable that users either have poor sleeping patterns and rely on the self-tracking technologies to deal with these issues or, on the contrary, that users show better sleeping patterns than non-users because they are able to manage the necessary requirements with the help of these technologies. The former scenario would lead to increased concerns of researchers and practitioners, as the above-mentioned negative effects need to be addressed. If athletes place too much confidence into customer technologies, they may become less accessible for more scientific approaches. The latter scenario, on the other hand, would indicate the potential of using self-tracking technologies to increase athletes’ self-awareness and responsibility. There is currently a lack of data on user behaviour and current sleeping patterns. Therefore, this manuscript evaluates the experience of German athletes with self-tracking technologies, especially smartphone apps and wearable technologies. The second aim of this study was to analyse the young athletes’ behavioural sleeping patterns and how these differ among users and non-users of sleep self-trackers. Furthermore, correlations of sleep parameters with participants’ characteristics will be examined to shed more light on the state of athletes in a period of time that was characterised by gradual reversed restrictions of exercise and training opportunities based on coronavirus disease 2019 (COVID-19) pandemic regulations in Germany.

Data were collected between May and July 2020 via online surveys which were distributed via personal contacts and social media channels. After receiving information about the purpose of the study and reading the confidentiality statement of data processing, respondents provided informed consent by ticking a specific box to start participation. The survey consisted of a demographic questionnaire and standardised questionnaires to assess usage of technologies, sleep quality, daytime sleepiness, attitudes about sleep, bedtime procrastination and self-control. Ethical clearance was secured prior to the study from the local ethical committee (EKS-V-10/2020). On average, it took participants 14.7 ± 4.7 min to complete the survey.

The sample consisted of 217 participants, with 29% (n = 63) female and 71% (n = 154) male respondents. Mean age was 26.9 ± 7 years and average training volume was 10.7 ± 5.4 h per week. The majority declared regular competition participation (n = 150; 69%), with 22.7% (n = 49) competing on the national or international level. The majority (58.5%) participated in individual sports and 41.4% were assigned to team sports. Only 10% (n = 22) stated to be in the regular season at the time of answering the survey, while 22% (n = 48) were in the preseason and 27% (n = 59) in the off-season phase. Only 5 participants were professional athletes, while just over half (51.6%) reported to be employed and 40.6% were school or university students.

Data processing and analysis was performed with the software SPSS 26 (IBM, Armonk, NY, USA). To analyse the prevalence of the use of self-tracking smartphone apps and wearables, descriptive statistics will be presented in absolute and relative numbers according to the stage model described above (König et al., 2018). Chi-square tests were performed to analyse the relationship between the five stages regarding sleep apps and (a) sex, (b) regular competition participation, (c) type of sport, (d) fitness app use, (e) nutrition app use, (f) use of wearable sleep trackers, and (g) use of wearable fitness trackers. In case of significant results (p ≤ 0.05), Cramer’s V will be reported to indicate the size of correlation. To address the second aim of the study, differences between the sleep app use stages and sleep-related variables were analysed via one-way analyses of variance (ANOVA) with Bonferroni corrected post hoc tests in case of significant main effects (p ≤ 0.05). Levene tests were conducted to test for homogeneity of variances and this precondition was fulfilled by each variable. In particular, the habitual time spent in bed, sleep duration, and sleep efficiency (expressed as the percentage of sleep duration relative to time in bed) as well as sleep quality (PSQI), daytime sleepiness (ESS), sleep-related cognitions (DBAS), bedtime procrastination (BPS), and self-control (BSCS) were examined. Finally, correlations among the sleep-related constructs as well as with the participants’ age and weekly training volume were investigated by Pearson’s product–moment correlation coefficients.

The aim of the present study was to analyse the user behaviour of smartphone and wearable technologies in the context of recovery self-management and self-tracking among German athletes. The overall prevalence of sleep app users was with less than 20% surprisingly low. Considering different aspects such as sex, type of sport, competition participation, and training volume, no remarkable characteristics among users versus non-users were identified. In terms of the different types of apps, it seems that fitness apps were more popular than sleep apps followed by nutrition apps. The correlation between the sleep apps and the other two types of apps indicate that non-users of sleep apps are probably also non-users of fitness or nutrition apps. This applies also to the correlation between the use of sleep apps and wearable technologies. Comparing the survey results with those reported by König et al. (2018) for 1215 adults with a mean age of 41 ± 18 years and 64% female respondents, several differences can be observed regarding the adoption process of fitness apps and nutrition apps. While fitness apps were also more frequently used than nutrition apps in that study, only one quarter was identified as acting compared to almost one half of the current sample.

There were also no remarkable differences among sleep indices between sleep app users and non-users. However, self-control was found to be highest among sleep app users compared to non-users (i.e. stage 1: unengaged). As the application of self-tracking devices requires certain engagement by the users, higher self-control seems to support this behaviour. It may also be speculated that athletes with higher self-regulatory abilities are more willing to investigate further resources into their recovery self-management. The negative correlation between self-control and bedtime procrastination supports this assumption, as delaying bedtime becomes less likely with higher self-control ratings (Kroese et al., 2016). According to Choi et al. (2018), smartphone applications have the potential to raise awareness and promote healthy sleep habits and by this means may support sleep self-management. Unfortunately, the present survey does not provide further information about the type of sleep apps among users as well as those of the disengaged non-users and why they decided to disengage. Only few apps incorporate behavioural constructs to encourage healthy sleep hygiene (Grigsby-Toussant et al., 2017) and it seems worthwhile to examine the apps’ functionality and appeal to this group. Moreover, further investigation on the user’s motives would provide deeper insights. For instance, Roomkham, Lovell, Cheung, and Perrin (2018) identified five styles of personal tracking based on the users’ needs. Specifically, these (potentially overlapping) styles were classified as (a) directive tracking, (b) documentary tracking, (c) diagnostic tracking, (d) collection rewards tracking and (e) fetishized tracking. The concern that certain people may develop an unhealthy obsession with healthy sleeping that could be enforced by the risk of false-positive diagnoses by sleep apps raised by Van den Bulck (2015) apparently does not apply to the current athletic population. Thus, this phenomenon that Baron et al. (2017) described as “orthosomnia” probably is rather an issue for patients with diagnosed sleep disorders who indeed seek relief and optimisation of their sleep insufficiency. Thus, based on the current findings, consumer sleep tracking technology apparently cannot be considered a threat to sport science research and practice. On the contrary, it may comprise some potential which still has to be further explored. For instance, Reichert et al. (2021) discuss the potential of ambulatory assessment for precision psychiatry which can also be considered for applied sport science and sport psychological interventions. One advantage is that data can be assessed in real-time and real-life conditions over longer periods of time. However, users of commercial devices have to take possible lack of data security into account. On the other hand, large-scale data are available to evaluate physical activity and sleep behaviour on the population level. By this means, Rezaei and Grandner (2021) as well as Capodilupo and Miller (2021) were able to examine trends in sleep and physical activity before and during COVID-19 pandemic developments retrospectively by using consumer wearable technology data from Fitbit (Fitbit Inc., San Francisco, CA, USA) and WHOOP (WHOOP, Boston, MA, USA), respectively. Nevertheless, most of the available commercial activity and sleep technologies lack sufficient data on reliability and validity in terms of detecting sleep parameters (Khosla et al., 2018), why it is recommended to rely on empirically evaluated devices (Halson, 2019). There is quite a wide range of ambulatory assessment methods for sleep monitoring (in athletes) so that the most convenient approach can be chosen according to the purpose. Portable polysomnography (Hof zum Berge et al., 2020) and actigraphy (Sadeh, 2011) evolved as promising methods to investigate sleep behaviour in an ecologically valid way.

Regarding the sleep quality of the participants, the average PSQI score was just around the cut-off (M = 5.3), with the majority categorised as good sleepers (59%). The score was higher than that of a German sample of adults below 40 years (M = 4.04 ± 2.73, n = 509), and prevalence of good sleep quality was lower than that of a general community sample (64.1%) according to Hinz et al. (2017). However, comparing the present findings to another sub-elite athletic population (n = 146), prevalence of good sleep quality was higher than the 35% reported by Doherty et al. (2021). It needs to be acknowledged, though, that Doherty et al. (2021) set the cut-off at PSQI ≥ 5, while the current study applied the more conservative cut-off with PSQI ≥ 6 (Buysse et al., 1989; Hinz et al., 2017). Comparing the findings to those of Bender, Van Dongen, and Samuels (2019), similar PSQI scores for the athletic sample (n = 63) were found (M = 5.0 ± 2.6).

Taking also the findings of the reported habitual sleep duration (M > 7 h) and the normal amount of daytime sleepiness (79% with ESS ≤ 10) into account, it may be assumed that the current sample was not chronically sleep deprived. It should be considered that the survey was conducted in a timeframe of social- and sport-restricted regulations during the 2020 COVID-19 pandemic activities. As university classes took place online and most of the employees were working at home, individuals were more flexible in terms of scheduling their leisure activities and sleep behaviour. This may have enhanced sleep quality and/or quantity, as Blume, Schmidt, and Cajochen (2020) reported positive effects on sleep–wake patterns in European adults between 26 and 35 years. In addition, Wright Jr. et al. (2020) found increased sleep regularity and sleep quantity among US American university students during COVID-19 lockdown orders. Furthermore, an Australian survey among elite and sub-elite athletes during the lockdown period support the current findings, as the athletes reported spending more time in bed and sleeping longer than before the lockdown (Facer-Childs et al., 2021). Future research should further explore the impact of the pandemic and its consecutive regulations on physical activity patterns, mental health, quality of life and sleep in elite athletes. A systematic review revealed that overall physical fitness and training volume (i.e. number of days, duration) as well as sleep quality decreased, while negative emotions (stress, fatigue, depression) increased (Jurecka, Skucińska, & Gądek, 2021).

Moreover, findings on bedtime procrastination were comparable to scores of Dutch adults (M = 2.7 ± 0.8) and even slightly lower than those of a Polish sample (M = 3.2 ± 0.9) according to Kroese et al. (2016) and Herzog-Krzywoszanska and Krzywoszanski (2019), respectively. Unfortunately, there is currently no cut-off to identify the problematic degree of postponing going to bed and a lack of comparable athletic samples that may help interpret the present findings. It may be generally assumed that getting insufficient sleep is a typical problem in athletes and that sleep is lacking sufficient prioritisation (Halson & Lastella, 2017).

Considering sleep-related cognitions, the average DBAS score (M = 64.4) was slightly higher compared to the baseline scores of two groups in a study with physically active university students (group 1: M = 52.4 ± 18.1, n = 25, group 2: M = 56.4 ± 26.6, n = 33) reported by Kölling and Hof zum Berge (2020). Moreover, the score of the present study corresponds to the cut-off that distinguishes between normal and an unhelpful degree of sleep-related beliefs (Carney et al., 2010). Interestingly, only those categorised as users of sleep apps (i.e. acting) were identified with a normal DBAS score. Nevertheless, the difference between the groups was marginal so that mean values should be considered cautiously. While the effect was small, dysfunctional beliefs were correlated with poor sleep quality and higher daytime sleepiness.

Several limitations need to be addressed. The participants were recruited via a convenient sample and the rather small number does not allow for generalisations. Moreover, the self-ratings may be biased by subjective recall deficiencies, on the one hand, and they constitute only a snapshot, on the other hand. Longitudinal analyses of sleep and recovery behaviour before, during and following lockdown restrictions would provide more detailed insights. The effect of COVID-19 regulations was not the focus of this study, but the results need to be interpreted with the current situation in mind. Follow-up studies that examine recovery self-management activities and sleep patterns during subsequent lockdowns and ‘new normal’ situations will help understand the interplay between recovery and external influencing factors to derive better guidance for athletes and practitioners.

Overall, the results of this study provide interesting insights and allow encouraging conclusions. On the one hand, the present sample seems to be rather unaffected by smartphone and wearable self-tracking technologies so that most of them do not take the risk of the potentially unsubstantiated or even dangerous feedback on their sleep (Ko et al., 2015). This implies that these athletes are not as vulnerable as expected and this may enhance their compliance in the collaboration with sport scientific researchers who are using empirically validated devices. Participants’ attitudes toward sleep monitoring and intervention approaches may probably not be affected by non-validated dubious feedback and dysfunctional self-treatment approaches. Thus, researchers and practitioners will probably not be confronted with contradicting and conflicting occasions regarding monitoring approaches and data interpretation. On the other hand, problematic sleep seems to be an issue for the minority of the sample. Nevertheless, those affected athletes should raise concern as sleep is an essential factor contributing to performance as well as recovery (Walsh et al., 2021). Therefore, sleep monitoring may be a possible approach to raise the athlete’s awareness and to detect dysfunctional sleep indices at an early stage (Halson, 2019). While a perfect sleep assessment method does not exist, several methods should be combined to balance their advantages and disadvantages (Ibáñez, Silva, & Cauli, 2018). Moreover, Halson (2019) highlights that practitioners and athletes should be aware of the limitations of different monitoring methods, especially smartphone applications and consumer-based technologies.