Critical Fluctuations as an Early Warning Signal of Sports Injuries? Applying the Complex Dynamic Systems Toolbox to Football Monitoring Data

Critical Fluctuations as an Early Warning Signal of Sports Injuries? Applying the Complex Dynamic Systems Toolbox to Football Monitoring Data | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Critical Fluctuations as an Early Warning Signal of Sports Injuries? Applying the Complex Dynamic Systems Toolbox to Football Monitoring Data Niklas D. Neumann, Jur J. Brauers, Nico W. Van Yperen, Mees Van der Linde, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4429464/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 16 Dec, 2024 Read the published version in Sports Medicine-Open → Version 1 posted 5 You are reading this latest preprint version Abstract Background There has been an increasing interest in the development and prevention of sports injuries from a complex dynamic systems perspective. From this perspective, injuries may occur following critical fluctuations in the psychophysiological state of an athlete. Our objective was to quantify these so-called Early Warning Signals (EWS) to determine their predictive validity for injuries. The sample consisted of 23 professional youth football (soccer) players. Self-reports of psychological and physiological factors as well as data from GPS sensors were gathered on every training and match day over two competitive seasons, which resulted in an average of 339 observations per player (range = 155–430). We calculated the Dynamic Complexity (DC) index of these data, representing a metric of critical fluctuations. Next, we used this EWS to predict injuries based on different mechanisms (traumatic and overuse) and duration. Results Results showed a significant peak of DC in 31% of the incurred injuries, regardless of mechanism and duration, in the seven data points (roughly one and a half weeks) before the injury. The warning signal exhibited a specificity of 94%, that is, correctly classifying non-injury instances. We followed up on this promising result with additional calculations to account for the naturally imbalanced data (fewer injuries than non-injuries). The relatively low F 1 we obtained (0.08) suggests that the model's overall ability to discriminate between injuries and non-injuries is rather poor, due to the high false positive rate. Conclusion By detecting critical fluctuations preceding one-third of the injuries, this study provided support for the complex systems theory of injuries. Furthermore, it suggests that increasing critical fluctuations may be seen as an EWS on which practitioners can intervene. Yet, the relatively high false positive rate on the entire data set, including periods without injuries, suggests critical fluctuations may also precede transitions to other (e.g., stronger) states. Future research should therefore dig deeper into the meaning of critical fluctuations in the psychophysiological states of athletes. Football Complex Dynamic Systems nonlinear time series analysis injury prediction dynamic complexity process monitoring multidisciplinarity personalized approach early warning signals critical fluctuations Figures Figure 1 Figure 2 Key points Complex Systems Theory suggests that sports injuries may be preceded by a warning signal characterized by a short window of increased critical fluctuations. Results of the current study showed such increased critical fluctuations before 31% of the injuries, which supports previous theoretical notions. Across the entire data set, we also found a considerable number of critical fluctuations that was not followed by an injury, suggesting that the warning signal may also precede transitions to other (e.g., healthier) states. Increased critical fluctuations may be interpreted as a window of opportunity for the practitioner to launch timely and targeted interventions, and researchers should dig deeper into the meaning of such fluctuations. 1. Background Hardly any athlete finishes their career unscathed, as traumatic and overuse sports injuries are a ubiquitous and emotionally and physically disturbing aspect of the athletic trajectory [ 1 ]. Those who are affected often experience performance decrements and suffer financially [ 2 , 3 ]. Given the substantial incidence of injuries [ 4 ] and the harm they cause, the sports field would highly benefit from any kind of anticipation or prediction. In that way, practitioners may be supported in their decision-making on whether, for instance, the training plan needs to be adjusted for an athlete. While the prediction of sports injuries is a subject of significant interest and value, it continues to pose a challenge even after many years of dedicated research. Obviously, injuries can never be predicted perfectly, but recent research suggests that strides can be made by accounting for the dynamic complexity through which injuries come about [ 5 – 7 ]. In the present empirical research, we will specifically target this dynamic complexity. To be more specific, we will apply and validate an analytic strategy from the complex dynamic systems toolbox to detect warning signals, which can be an important avenue in the field of sports science and medicine. 1.1 Past Approaches and Future Directions of Sports Injury Prediction Previous research on sports injury prediction measured isolated or monodisciplinary risk factors at one or a few points in time, and analysed data with linear methods at the group level [ 8 – 12 ]. Prediction models performed rather low in forecasting injuries, they were poorly developed (i.e., not validated, no code provided, high risk of bias), and not applicable in practice [ 8 , 13 , 14 ]. Further, much focus has been put on why injuries occur (i.e., revealing relevant factors) but the question of how those factors lead to injuries and when athletes are at increased risk is generally lacking. As a response, researchers have suggested to study sports injuries from a complex dynamic systems perspective [ 5 – 7 , 15 – 20 ]. Complex systems consist of factors that interact over time in a non-linear manner [ 5 , 6 ]. Those factors coordinate their behaviour while generating patterns that are adapted to their environment. When a pattern is maintained by a system, the pattern can be called an attractor state, that is a state to which the system is “attracted” [ 21 , 22 ]. Complex systems typically have certain tipping points in which abrupt changes (phase transitions) from one attractor state to another can occur [ 5 , 6 , 19 , 22 – 26 ]. An injury would be the result of this so-called phase transition, a system-wide reorganisation, from a healthy state to an injured state [ 6 ]. For such a transition to take place, the stability of the prevailing attractor state has become weakened [ 27 ]. Theory and empirical findings in the fields of ecology, financial markets, and psychology show that, in such a phase of instability, critical fluctuations can be observed as a relatively short window of increased variability and turbulence (for an illustration see Fig. 1 ) [ 6 , 24 , 27 – 32 ]. Such fluctuations are therefore called “Early Warning Signals” (EWS) and might be an explanation of how injuries occur. To be more specific, instead of analysing absolute values of screening tests or daily measures at the group level, which will likely not work in predicting injuries [ 12 , 33 ], theory suggests that strides can be made by investigating the complex dynamics, and more specifically, critical fluctuations, at the individual level [ 6 , 34 ]. Given the parallel between injury development and complex system dynamics, an interesting question is therefore whether such critical fluctuations can also be found before injuries. 1.2 Non-linear Time Series Analysis To detect critical fluctuations, time series analysis methodologies are needed [e.g., 27–30,35,36]. A promising methodological tool in this regard is the Dynamic Complexity (DC) algorithm, which can detect heavy and irregular variability in a system’s behaviour [ 24 , 28 , 34 , 37 ]. DC was developed for the real-time monitoring of human change processes and is therefore specifically suited for non-linear, non-stationary, and short and coarse-grained time series that are typical in psychology, sports science and medicine. So far, DC has already successfully been applied to investigate changes in mood of psychological disorders and physical activity. For instance, a study by Olthof and colleagues [ 28 ] showed that increased critical fluctuations predicted sudden shifts in mood in the previous four days for patients receiving psychotherapy. In the same line, research on walking behaviour found a positive and significant association between increased critical fluctuations and a loss of step counts in the next few days [ 37 ]. Recently, Schiepek et al. [ 34 ] provided a first exploration of the potential merits of DC as an EWS of sports injuries. In their case study, one professional football (soccer) player filled out the sports process questionnaire across 77 days using an app-based system. During that time, he suffered a twisted cruciate ligament. Results showed that DC reached a peak of critical instability just before the injury. This is a promising finding, but on the basis of this case study, we cannot conclude that such a signal is specific for identifying injuries. Hence, more research is needed to confirm the potential and robustness (validation) of these results across athletes and injuries (e.g., injury mechanism and duration). Further, the study did not determine the time window for which the warning signal appears before the injury, and it is unclear how a peak was defined. In the current in-depth study of DC on sports injuries we address these limitations in Schiepek et al.’s [ 34 ] study and account for the complex multidisciplinarity of traumatic and overuse sports injuries. 1.3 The Role of Injury Mechanism (traumatic vs overuse) and Duration Previous research has dichotomized injuries into two fundamental mechanisms: traumatic injuries, stemming from specific, identifiable events, and overuse injuries, resulting from repeated micro-traumas without a single, discernible event precipitating the injury [ 38 ]. Scholars have posited that the prediction of overuse injuries may be more feasible compared to traumatic injuries, given the opportunity to capture the repeated micro-traumas which lead to the injury [e.g., 39]. However, previous investigations propose a more nuanced perspective, contending that, for instance, athletes experiencing high physical load may be more susceptible to traumatic injuries [ 40 ], but that generally, causal pathways between load and injury are poorly understood [ 41 , 42 ]. Furthermore, no study has yet examined the effect of injury duration. Clarifying whether a prediction model exhibits differential efficacy across injury mechanism (traumatic vs overuse) and duration would suggest that the model may be more sensitive to specific patterns. For instance, the model’s validity in predicting all injuries of a given sample may be rather low. However, overuse injuries or injuries that last longer, for instance, may be better predicted than traumatic injuries or injuries that result in a short time-loss, respectively. This information can be useful for targeted prevention strategies and interventions and may prevent huge costs, as well as physical and emotional suffering. 1.4 The Present Study The present study aims to predict injury occurrence of youth football players of a professional academy, based on EWS in the form of critical fluctuations. To meet this aim, it is important to monitor psychological and physiological variables of individual players on a daily basis [ 9 , 43 – 46 ]. Such monitoring can be realized via an online application and sensors, for instance [ 34 , 43 , 47 ].[1] There is a myriad of factors that can be measured, yet, the decision on what to measure should be based on what prior scholarly work suggests may be precursors of sports injuries and what is realistic and feasible to collect from the sample. For instance, on the psychological level, self-efficacy, motivation, and mood are key factors in sport performers and injuries [ 34 , 48 – 53 ]. In addition, athletes may regularly be perturbed by unenjoyable training sessions and bad performances [ 43 ]. On the physiological level, internal load factors such as the heart rate and the session rating of perceived exertion (sRPE), as well as external load factors such as sprints and total distance covered are typically related to injuries [ 17 , 39 , 54 – 58 ]. Further, the recovery status may be monitored before every training session or competition [ 59 – 61 ]. In line with the complex dynamic systems theory, we hypothesized that increased critical fluctuations (i.e., cumulative peaks of DC) precede the occurrence of an injury. Next to the detection of critical fluctuations at the individual level, this research seeks to validate this EWS by testing its predictive validity at the group level. Last, the study intends to explore whether injury mechanism (traumatic vs overuse) and duration affect the predictive validity and which measured factor(s) (i.e., psychological, physiological, self-reports, sensors) reveal the most peaks in destabilizations. 2. Methods 2.1 Subjects We collected data from 55 youth male players of the U-18 and U-21 teams (16 to 20 years old) of a premier league (Eredivisie) football club in the Netherlands. 23 of these players were included in the analysis because they fulfilled the inclusion criteria (more details under 2.3.1. Data Pre-Processing). Once a player started playing at the club, he was informed about the data collection process. By signing an informed consent, he could then decide whether he wanted to permit the use of his data for research purposes or not (for the present study, all players provided consent). The players competed in the highest national league of their age category. They had between six and eight training sessions per week, composed of two strength sessions and four to six field sessions of 60 to 75 minutes and 75 to 90 minutes, respectively, and matches on the weekend. Due to personal data protection, further potentially identifiable information (e.g., height, weight, position, team, specific type of injury) is not reported. 2.2 Design, Measures, and Procedure For every training and match up to two competitive seasons, we collected psychological and physiological data. To be more specific, players answered self-report questions on self-efficacy, motivation, mood, performance self-evaluation, enjoyment, session rating of perceived exertion (sRPE) and total quality of recovery (TQR) on a tablet computer near the locker room without staff or team members being present. All measures were part of the normal, daily team monitoring routine at the club. Polar TeamPro sensors (Polar Electro Oy, Kempele, Finland) were used to collect data on the duration, total distance covered, sprints, and heart rate for every training session and match (see Table 1 ). The self-report questions were extracted from validated questionnaires and adjusted to the present context (see Table 1 for references). The RPE consists of one item (“How hard was the training/competition?”) and is a subjective estimate of the psychological and physiological stress imposed on the athlete [ 62 ]. Later, we multiplied the RPE score by the duration of the training session to obtain a measure for the internal training load for the analysis, the session RPE (sRPE) [ 63 ]. If there were two load scores a day, which occasionally happened after two different types of training, we added them up to capture the daily load [ 64 ]. Finally, next to the questions explained in the table, the team physician recorded injury occurrence (yes vs. no), and injury-related time-loss (in days), including the mechanism (i.e., traumatic or overuse) [ 56 ]. Table 1 Data Collection Time of the day Measured factor Self-report question Measurement scale References T1: In the morning up to 30 minutes before the first training session or match Recovery How good is your recovery? CRS from 6 ( very, very poor recovery ) to 20 ( very, very good recovery ) [e.g., 65,66] Self-efficacy How confident are you that you can perform maximally today? VAS from 0 ( not at all confident ) to 100 ( very confident ) [e.g., 52,67] Motivation How motivated are you to perform maximally today? VAS from 0 ( not at all motivated ) to 100 ( maximally motivated ) [e.g., 52,68] Mood How much are you in the mood to train/play the match today? VAS from 0 ( not at all in the mood ) to 100 ( very much in the mood ) [e.g., 51,69] T2: During the training session or match Distance Meter [e.g., 17,54] Sprint Number of sprints > 25km/h (for the U-21) and > 19.8km/h (for the U-18) per session [e.g., 54,57] Duration Minutes [e.g., 55,63] Heart rate in zone 5 Beats per minute; output is the number of seconds spent in zone 5 (i.e., 92–100% of an individual’s max heart rate) [e.g., 54,56] T3: At the end of the day up to 30 minutes after the last training session or the match Exertion How hard was the training/match? CRS from 6 ( very, very light ) to 20 ( very, very hard ) [e.g., 56,70] Perceived performance How well did you perform today? VAS from 0 ( very bad (far below my capabilities )) to 100 (maximally ( to the best of my capabilities )) [e.g., 43,71] Enjoyment How much did you enjoy the training session(s)/the match today? VAS from 0 ( not at all ) to 100 ( very much ) [e.g., 43,72] CRS, Category-Ratio Scale; T, time point; VAS, Visual Analogue Scale. 2.3 Data Set and Statistical Analysis The analysis was performed with R and RStudio [ 73 ]. The R code has been made publicly available (see Code Availability). 2.3.1 Data Pre-Processing The original data set consisted of 55 players from two youth teams. The following criteria were established to include players for the analysis: At least one injury occurrence was reported of a player between measurement seven and the last seven measurements (see the Note in Fig. 2 for more background on this inclusion criteria) and a player did not miss more than 20% of the values per measured factor.[2] This resulted in a final sample of 23 players, yielding an average of 339 observations per player (range = 155–430). First, we imputed missing values with the R package mice [ 74 ]. As a next step, we normalized the data in order to be able to analyse and compare the results. To be more specific, self-report data from the Visual Analogue Scale was divided by 100 (i.e., the maximum), self-report data from the Category-Ratio Scale was divided by 20 (i.e., the maximum), and data from sensors was divided by the maximum value of each factor of every individual. 2.3.2 Early Warning Signals (EWS) Calculation We calculated critical fluctuations of the self-report and sensor data with the R package casnet [ 75 ], using the Dynamic Complexity (DC) algorithm. Mathematically, DC is a multiplication of the degree of fluctuation (F) of a time series and the distribution (D) of values between the theoretical minimum and maximum of a scale [for validation and details see 24]. F is sensitive to the amplitude and frequencies of a time series and D to the scattering of values. Both measures were calculated within a moving window of seven data points and a window step of one, in order to identify non-stationary changes.[3] Next, we checked for cumulative complexity peaks (CCPs), which indicate whether the number of simultaneous peaks in DC of the time series was significant. A CCP is determined by conducting a one-sided z-test (one side, because we are looking for increased critical fluctuations) to check for significantly increased scores, which results in a new time series that mirrors the number of significant peaks on a specific day. Another one-sided z-test is performed on this new time series to determine whether this number is significant. Since critical fluctuations are expected to be present just before a transition, we tested which time window between one and ten data points prior to the injury would be optimal for predicting injures. We increased the window size in a step-by-step manner starting with one. When increasing the window size did not provide better results, the previous window was considered as the optimum. 2.3.3 Validation Analysis We validated the predictive validity of the warning signal in four steps. First, we calculated the sensitivity of the DC analysis for every individual. The sensitivity evaluates the warning signal's effectiveness in correctly identifying injury cases, highlighting its ability to accurately detect true positive cases and emphasizing its validity in identifying injuries [ 76 ]. It is the ratio of the true positive rate (i.e., a significant CCP and the athlete was injured) to the true positive rate plus the false negative rate (i.e., no significant CCP and the athlete was injured). Second, we calculated the specificity of the DC analysis for every individual. The specificity assesses the accuracy of the warning signal to correctly identify true negative cases [ 76 ]. It is the ratio of the true negative rate (i.e., no significant CCP and the athlete was not injured) to the true negative rate plus the false positive rate (i.e., a significant CCP and the athlete was not injured). Third, we calculated the accuracy of the warning signal. Accuracy is simply the percentage of correct predictions (injury and non-injury) made by the model [ 76 ]. It is calculated by dividing the number of correct predictions by the total number of predictions made. The result is multiplied by 100. The inherently imbalanced nature of the dataset in this research, with a strong bias towards the majority class (non-injuries), may result in less stable and reliable estimates of sensitivity, specificity, and accuracy. For this reason, we also calculated an F 1 score (see formula 1) as the fourth and last step, by using the harmonic mean of precision and recall [ 77 ]: (1) $${F}_{1}=\frac{2\text{*} \text{p}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}\text{*} \text{r}\text{e}\text{c}\text{a}\text{l}\text{l}}{\text{p}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}+\text{r}\text{e}\text{c}\text{a}\text{l}\text{l}}$$ Precision is the ratio of the true positive rate (i.e., a significant CCP and the athlete was injured) to the true positive rate plus the false positive rate (i.e., a significant CCP and the athlete was not injured). Recall is another term for sensitivity. The F 1 score falls between 0 and 1, with 0 indicating no predictive power and 1 indicating perfect prediction. Important to note is, however, that the analysis takes place at the group level, since all individual observations must be combined and treated as one large observation. We applied the following rules for the analysis: If at least one warning signal occurred in the seven days prior to an injury, the remaining six days were left out for the analysis (they were not interesting anymore, since a warning signal appeared already). Likewise, significant peaks during an injury period were not taken into account for the analysis, because the injury already happened, and turbulence can be expected in this period. 2.3.4 Analysis of Injury Mechanism (Traumatic vs Overuse) and Duration We calculated the percentage of correctly predicted traumatic and overuse injuries and we checked with a chi-squared test whether the difference was statistically significant. Then, we calculated a bi-serial correlation to test if the duration of an injury (i.e., time-loss in days) is associated with the predictive validity index of the analysis tool. A post-hoc power analysis determined that a sample size of 20 is needed for calculating a bi-serial correlation with a desired power of 0.8, an alpha level of p = 0.05, and an effect size of d = 0.9. In our final step, we counted how often each measured variable occurred at least once in the seven days prior to an injury in the CCPs for correct predictions. 3. Results 3.1 Validation Analysis Players were injured 2.8 times on average across the measurement period (range = 1–6; total = 64). To illustrate, Fig. 2 shows the multivariate raw data, the complexity resonance diagram, and the critical instability plot of one representative player (Player 9; for the plots of the other players, see online resource). Note that for one of the two injuries, a significant CCP appeared in the seven data points before – an EWS. CCP = Cumulative Complexity Peak, DC = Dynamic Complexity, sRPE = session Rating of Perceived Exertion. Multivariate raw time series (top figure), complexity resonance diagram (middle figure) and critical instability plot (bottom figure) of Player 9 suffering two injuries (injury period marked in orange/vertically shaded from top to bottom). The x-axis displays the number of data points. The y-axis shows each measured factor with the raw time series (top), the Dynamic Complexity values (middle), and the significant Dynamic Complexity values (bottom). Note that the time scale starts at seven and stops seven data points before the end of the measurement period, because we used a seven-data point overlapping moving window to calculate the Dynamic Complexity. This means that the first available Dynamic Complexity value is at data point seven. The higher the Dynamic Complexity, the whiter and redder the plot becomes (middle figure). Significant Dynamic Complexity levels are colored grey, whereas significant Cumulative Complexity Peaks are colored black (bottom figure). The sensitivity analysis across all players showed that cumulative complexity peaks appeared for 31% (20 out of 64) of the injuries within seven data points before (range: 0-100%).[4] The analysis further demonstrates that the warning signal exhibits a specificity of 94% (range = 89–100%). The accuracy reached a value of 93%. All three metrics were calculated per individual and then averaged across all individuals (individual-level analysis) [ 78 ]. Table 2 shows the confusion matrix from which these metrics were calculated and Table 3 provides more detailed information on injuries and the DC analysis for all individual players. Finally, the calculation of the F 1 score, conducted at the group level, revealed an estimate of 0.08. This indicates poor performance in terms of both precision and recall, suggesting that the model's ability to correctly identify positive instances and avoid false positives is very low. Table 2 Confusion Matrix Showing the Predictions and Actual Outcomes Prediction Injured Not Injured Actual Outcome Injured 20 (TP) 44 (FN or Type II Error) Not Injured 442 (FP or Type I Error) 5987 (TN) FN, False Negative; FP, False Positive; TN, True Negative; TP, True Positive Table 3 Individual Results of all Players Player Number of Measurements Number of Injuries Time-loss in Days Mechanism Sensitivity Specificity 1 176 1 12 T 0% 93% 2 427 2 22 T 50% 92% 8 T 3 248 4 3 T 50% 94% 38 T 6 T 6 O 4 351 2 5 T 50% 90% 286 NA 5 419 2 8 T 50% 95% 12 T 6 413 3 3 T 33% 93% 13 O 35 T 7 241 2 25 T 50% 94% 37 T 8 398 4 12 T 0% 97% 30 T 17 T 25 O 9 419 2 20 O 100% 94% 26 T 10 421 6 11 O 33% 95% 9 O 22 O 28 T 8 T 51 T 11 413 1 16 T 0% 95% 12 229 2 7 T 50% 92% 5 T 13 408 2 6 O 50% 89% 22 T 14 288 5 14 T 60% 93% 27 T 12 O 10 T 271 T 15 422 3 8 T 33% 93% 21 O 24 T 16 155 4 9 T 25% 100% 15 T 6 T 62 T 17 418 2 14 O 100% 90% 23 O 18 415 4 14 T 50% 95% 23 O 22 T 4 T 19 258 4 32 T 50% 99% 5 T 13 T 20 T 20 430 4 19 T 0% 92% 64 O 26 NA 41 O 21 423 3 29 T 33% 93% 3 T 5 T 22 168 1 2 T 0% 90% 23 247 1 4 T 0% 93% NA, Not Available; O, Overuse; T, Traumatic. Injuries that occurred before data point seven and in the last seven data points are left out (n = 3), because the window size of seven data points only allows to capture Dynamic Complexity changes in between. Time-loss in days refers to the lost time in actual days, not days of measurement 3.2 Analysis of Injury Mechanism (Traumatic vs Overuse) and Duration We tested whether our prediction model exhibits differential efficacy across injury mechanisms and duration. Differences would suggest that the model may be more sensitive to specific injuries and could help prevent those instances. The analysis showed no significant difference (χ 2 (3, N = 20) = 2.86, p = .41) in the prediction of traumatic and overuse injuries on average (calculated per individual and then averaged across all individuals). To be more specific, 27% of overuse injuries (14 out of 47) and 30% of traumatic injuries (4 out of 15) could be predicted (two injuries were not classified in the dataset). Next, we tested whether injury duration (in days) is related to the predictive validity index of the analysis tool (0 for not predicted and 1 for predicted). The correlation revealed a coefficient of .31 between injury duration and the predictive validity index with p = .11. Finally, we visually inspected all critical instability plots (see online resource) to determine which measured factor(s) (i.e., psychological, physiological, self-reports, sensors) reveal the most peaks in destabilization: The factors sprint and total distance appear most often in the cumulative complexity peaks that correctly predicted an injury, that is, ten times. Two injuries could be predicted with psychological factors only, another two injuries with physiological factors only, three injuries with only self-reports (i.e., psychological and physiological factors), and two injuries with sensor data only. In the 11 remaining injuries, a combination of the factors was found. 4. Discussion The present study tested if critical fluctuations can serve as an EWS of sports injuries by applying the complex dynamic systems toolbox – specifically the DC algorithm – to sports monitoring data. In doing so, we collected longitudinal data on psychological and physiological factors of 23 football players across two competitive seasons. The outcomes of the study are threefold. As hypothesized, the results show that critical fluctuations appear quite often before injuries. Theoretically speaking, this supports the notion that injuries may come about through a non-linear dynamic process and that the methodological toolbox of complex dynamic systems could provide such insights [ 5 , 34 ]. Also, the findings are in line with DC research on mood and physical activity [ 28 , 37 ], and specifically, the potential promising DC application to injuries as presented by Schiepek et al. [ 34 ]. Second, we took additional measures to check the robustness and practical usefulness of our results. More specifically, the warning signal exhibits a sensitivity of 31% and a specificity of 94%. This means that the model can correctly predict 31% of injury (20 of 64) and 94% of non-injury instances, respectively. The accuracy revealed an estimate of 93%, meaning that it correctly predicts injuries and non-injuries in 93% of the time. These results suggest good predictive validity of the warning signal. However, we also controlled for the imbalance in the data by calculating an F 1 score (at the group level). The result revealed an estimate of 0.08, which suggests that the model's overall ability to discriminate between injuries and non-injuries is rather poor, especially given the high false positive rate. In the section Practical Implications , we put these results into perspective by explaining what this would mean for the practitioner. Given the discrepancy between the results based on two different analytic strategies, we strongly recommend including additional checks for future research as well. So far, studies on EWS before critical transitions have not validated warning signals extensively. More specifically, studies outside the sports context often only study periods before transitions instead of the whole time series, which does not capture the warning signals’ actual predictive validity. Furthermore, previous work tended to provide only one or two metrics (e.g., sensitivity and/or specificity) and did not consider the imbalance in the dataset. This could have led to an overly optimistic report of the predictive validity of EWS metrics and their relevance for practice. With regard to the current study, while critical fluctuations can signal an increased likelihood of injuries, such a warning signal did not always lead to an injury, which can be due to different reasons. For instance, the system can also return to the previous attractor after destabilization, meaning that it does not result in a phase transition. Given that the data in this study is collected as part of the club's everyday routine, where practitioners utilize the data to adjust training processes, it is possible that injuries were prevented without our or the practitioner’s awareness. Destabilization, therefore, could better be considered a window of opportunity, meaning that it provides information that the system is currently out of equilibrium and re-organizes itself [ 28 ]. In the same line, critical fluctuations can be seen as a general indicator of instability and not a predictor of a specific kind of order transition. This means that critical fluctuations may also precede other kinds of transitions (e.g., a performance loss or gain, a mental dip). Third, although previous theoretical research proposed that overuse injuries hold more information prior to their occurrence [ 39 ], we did not find effects of mechanism (traumatic vs overuse). Similarly, the duration of injuries was no significant factor for the predictive validity of the EWS, that is, injuries that resulted in a large time-loss could not be better predicted than injuries of short duration, and vice versa. These results should be treated with caution, however, because of the small sample size and rather homogenous study population. An interesting additional result is that the physiological sensor factors sprint and total distance appear most often in the CCPs, suggesting that they might play an important role in predicting injuries. However, the results also show that for two injuries, only psychological factors appeared, whereas for most other injuries, a combination of the factors (psychological and physiological, self-report and sensor data) is decisive for the prediction. These outcomes highlight the need for a multidisciplinary approach, tailored to individual athletes [e.g., 43,79]. In the next sections, we discuss future directions for research and practical implications of the results for the sports field. 4.1 Recommendations for Future Research To the best of our knowledge, the present study is the first to test the predictive validity of the DC algorithm as a potential EWS of sports injuries in a sample of football players. From a complex dynamic systems perspective, the DC algorithm can be considered a strong method since it does not have any specific statistical or parametric assumptions about the data, it mirrors increased complexity and noise of system dynamics before a phase transition, it can model linear and non-linear phenomena, and it can be used in real-time and not only for post data collection. The study itself has high ecological validity because measuring took place in the field as a daily routine and not in an experimental setting. In accordance with the high false positive rate on the full time series data, previous research has shown that destabilization is not necessarily related to negative outcomes, such as injuries or increases in symptom severity. Indeed, it may also precede decreases in symptom severity, that is, the transition to a healthier state [ 21 , 28 ]. Hence, future research should test whether destabilization in athletes’ attractors can also predict, for instance, performance gains or decreases in physical complaints. It must be noted that complex systems theory does not suggest all injuries to take place via abrupt transitions preceded by critical fluctuations, they may also develop more gradually and within the same state of attraction. Future studies may test this assumption, as demonstrated by Schiepek and colleagues [ 34 ]. Forthcoming studies may further exploit the complex dynamic systems toolbox to detect dissimilar dynamics preceding critical transitions (e.g., through recurrence quantification analysis and recurrence plots) [ 27 , 34 ]. Last, in our study, we merely relied on intensive monitoring data collected daily across a season. However, as Hecksteden et al. [ 80 ] proposed, the integration of screening tests (i.e., risk factors and protective factors) with monitoring data may be a promising avenue for sports injury forecasting (see online resource for more recommendations for future research). 4.2 Practical Implications In a perfect world, practitioners would always be correctly informed by an EWS about an upcoming injury. However, the world is too complex to capture all the information needed to make such reliable forecasts. The findings from our study indicate that, on average, an injury is quite often preceded by critical fluctuations, but at the same time false positive predictions regularly occur. In practice, and based on the data we analysed, this would mean that staff needs to intervene 21 times per player over the course of two seasons, while one of these signals actually leads to an injury which can be prevented (and 1.9 injuries on average per player would be missed due to false negatives). While a high false negative rate may be more harmful than a high false positive rate, intervening 21 times may still appear extensive. However, it is important to recognize that interventions do not always need to be resource-intensive in terms of time and cost. They may simply be short verbal interactions between the (multidisciplinary) staff and the player. For such verbal interactions, practitioners may specifically address those variables that showed increased critical fluctuations in the DC analysis. For Player 9 in Fig. 2 this concerns his motivation, the seconds of heart rate he spent in zone 5, and the total distance covered. In that way, practitioners are provided with a conversation starter and information on which knob to turn for which athlete. However, if implemented, such a warning system should be considered a decision-support tool for practitioners to adapt training processes and implement interventions instead of a replacement of the coach [ 81 , 82 ]. In a recent study by Hecksteden et al. [ 83 ], it was emphasized that a blend of practitioner experience, judgment, and scientific evidence should be used for decision-making in sports. 4.3 Conclusion The results of this study suggest that critical fluctuations often precede sports injuries, thereby supporting the complex dynamic systems perspective on injuries. At the same time, it cannot (yet) be considered a valid EWS for the real-time prediction of injuries in sports practice. Critical fluctuations might be interpreted as a window of opportunity, because turbulence indicates re-organization of the system which is currently out of equilibrium. Practitioners may make use of this window by planning timely and targeted interventions to guide athletes toward desirable conditions. Future research should dig deeper into the meaning of critical fluctuations in the psychophysiological states of athletes. Abbreviations DC = Dynamic Complexity EWS = Early Warning Signal sRPE = session Rating of Perceived Exertion TQR = Total Quality of Recovery CCP = Cumulative Complexity Peak Declarations Ethics approval and consent to participate: The present study was conducted according to the requirements of the Declaration of Helsinki and was approved by the ethics committee of the Faculty of Behavioral and Social Sciences of the University of Groningen (research code: PSY-1819-S-0308). Consent of participants was obtained to participate in the data collection. Consent for publication : Consent of participants was obtained to publish their data. Availability of data and material: The data contains sensitive information about the subjects and is therefore restricted from openly sharing it. Requests can be made by researchers affiliated with universities or independent, non-commercial research institutes via DataVerseNL at https://dataverse.nl/privateurl.xhtml?token=e0f11c98-9377-4522-8c81-a80cda00c750. The R code is openly accessible via the same link. Competing interests : We have no competing interests to declare. Funding: This work was supported by The Netherlands Organization for Health Research and Development (ZonMw, grant number 546003004). Author Contributions NDN took the lead in writing the manuscript. MvdL collected the data. RJRdH is the leader of the project. NDN, JJB, and FH analysed the data and wrote the R code. All authors designed the study, edited this article critically, and read and approved the final manuscript. Acknowledgements The authors thank Dr. Merlijn Olthoff for helpful and insightful comments on the analysis. References Green SJ, Weinberg RS. Relationships among Athletic Identity , Coping Skills , Social Support , and the Psychological Impact of Injury in Recreational Participants. Journal of Applied Sport Psychology. 2001;13(1):40–59. Cumps E, Cumps E, Verhagen E, Annemans L, Meeusen R. Injury rate and socioeconomic costs resulting from sports injuries in Flanders : data derived from sports insurance statistics 2003. British Journal of Sports Medicine. 2007;42:767–72. Bolling C, Delfino Barboza S, van Mechelen W, Pasman HR. How elite athletes, coaches, and physiotherapists perceive a sports injury. Translational Sports Medicine. 2019;2(1):17–23. Ekstrand J, Hägglund M, Waldén M, Ekstrand J, Hägglund M, Waldén M. Injury incidence and injury patterns in professional football - the UEFA injury study. British Journal of Sport. 2011;45(7):553–8. Bittencourt NFN, Meeuwisse WH, Mendonça LD, Ocarino JM, Fonseca ST. Complex systems approach for sports injuries: moving from risk factor identification to injury pattern recognition — narrative review and new concept. British Journal of Sports Medicine. 2016;50(21):1309–14. Fonseca ST, Souza TR, Verhagen E, van Emmerik R, Bittencourt NFN, Mendonça LDM, et al. Sports Injury Forecasting and Complexity: A Synergetic Approach. Sports Medicine. 2020;50:1757–70. Pol R, Hristovski R, Medina D, Balague N. From microscopic to macroscopic sports injuries. Applying the complex dynamic systems approach to sports medicine: a narrative review. British journal of sports medicine. 2019;53(19):1214–20. Bullock GS, Mylott J, Hughes T, Nicholson KF, Riley RD, Collins GS. Just How Confident Can We Be in Predicting Sports Injuries ? A Systematic Review of the Methodological Conduct and Performance of Existing Musculoskeletal Injury Prediction Models in Sport. Sports Medicine. 2022;52(10):2469–82. Johnson U, Tranaeus U, Ivarsson A. Current status and future challenges in psychological research of sport injury prediction and prevention: A methodological perspective. Revista de Psicología del Deporte. 2014;23(2):401–9. Bache-Mathiesen LK, Andersen TE, Dalen-Lorentsen T, Clarsen B, Fagerland MW. Not straightforward: Modelling non-linearity in training load and injury research. BMJ Open Sport and Exercise Medicine. 2021;7(3):1–9. Stern BD, Hegedus EJ, Lai Y cheng. Injury prediction as a non-linear system. Physical Therapy in Sport. 2020;41:43–8. Bahr R. Why screening tests to predict injury do not work-and probably never will.: A critical review. British Journal of Sports Medicine. 2016;50(13):776–80. Seow D, Graham I, Massey A. Prediction models for musculoskeletal injuries in professional sporting activities: A systematic review. Translational Sports Medicine. 2020;3(6):505–17. Van Eetvelde H, Mendonça LD, Ley C, Seil R, Tischer T. Machine learning methods in sport injury prediction and prevention : a systematic review. Journal of Experimental Orthopaedics. 2021;8(27):1–15. Cook C. Predicting future physical injury in sports: it ’ s a complicated dynamic system. British Journal of Sports Medicine. 2016;50(22):1356–7. Bekker S, Clark AM. Bringing complexity to sports injury prevention research: From simplification to explanation. British Journal of Sports Medicine. 2016;50(24):1489–90. Hulme A, Thompson J, Nielsen RO, Read GJM, Salmon PM. Towards a complex systems approach in sports injury research: Simulating running-related injury development with agent-based modelling. British Journal of Sports Medicine. 2018;53(9):560–9. Tee JC, McLaren SJ, Jones B. Sports Injury Prevention is Complex: We Need to Invest in Better Processes, Not Singular Solutions. Sports Medicine. 2020;50(4):689–702. Yung KK, Ardern CL, Serpiello FR, Robertson S. Characteristics of Complex Systems in Sports Injury Rehabilitation: Examples and Implications for Practice. Sports Medicine - Open. 2022;8(1). Lövdal SS, Den Hartigh RJR, Azzopardi G. Injury Prediction in Competitive Runners With Machine Learning. International Journal of Sports Physiology and Performance. 2021;16(10):1522–31. Olthof M, Hasselman F, Strunk G, Aas B, Schiepek G, Lichtwarck-Aschoff A. Destabilization in self-ratings of the psychotherapeutic process is associated with better treatment outcome in patients with mood disorders. Psychotherapy Research. 2020;30(4):520–31. Haken H. Synergetics: An Introduction. Nonequilibrium Phase-Transitions and Self-Organization in Physics, Chemistry and Biology. Springer Series in Synergetics 1. Berlin: Springer; 1983. Stern BD, Hegedus EJ, Lai YC. State dependence: Does a prior injury predict a future injury? Physical Therapy in Sport. 2021;49:8–14. Schiepek G, Strunk G. The identification of critical fluctuations and phase transitions in short term and coarse-grained time series — a method for the real-time monitoring of human change processes. Biological Cybernetics. 2010;102:197–207. Scheffer M, Carpenter SR, Lenton TM, Bascompte J, Brock W, Dakos V, et al. Anticipating critical transitions. Science. 2012;338(6105):344–8. Scheffer M, Bolhuis JE, Borsboom D, Buchman TG, Gijzel SMW, Goulson D, et al. Quantifying resilience of humans and other animals. Proceedings of the National Academy of Sciences of the United States of America. 2018;115(47):11883–90. Hasselman F. Early Warning Signals in Phase Space: Geometric Resilience Loss Indicators from Multiplex Cumulative Recurrence Networks. Frontiers in Physiology. 2022;13:859127. Olthof M, Hasselman F, Strunk G, van Rooij M, Aas B, Helmich MA, et al. Critical Fluctuations as an Early-Warning Signal for Sudden Gains and Losses in Patients Receiving Psychotherapy for Mood Disorders. Clinical Psychological Science. 2020;8(1):25–35. Scheffer M, Bascompte J, Brock WA, Brovkin V, Carpenter SR, Dakos V, et al. Early-warning signals for critical transitions. Nature. 2009;461(7260):53–9. Helmich MA, Olthof M, Oldehinkel AJ, Wichers M, Bringmann LF, Smit AC. Early warning signals and critical transitions in psychopathology: challenges and recommendations. Current Opinion in Psychology. 2021;41:51–8. Kelso JAS, Scholz JP, Schöner G. Nonequilibrium phase transitions in coordinated biological motion: critical fluctuations. Physics Letters A. 1986;118(6):279–84. Hill Y, Den Hartigh RJR, Cox RFA, De Jonge P, Van Yperen NW. Predicting resilience losses in dyadic team performance. Nonlinear Dynamics, Psychology, and Life Sciences. 2020;24(3):327–51. Impellizzeri FM, Menaspa P, Coutts AJ, Kalkhoven J, Menasp MJ. Training Load and Its Role in Injury Prevention , Part I : Back to the Future. Journal of Athletic Training. 2020;55(9):885–92. Schiepek G, Schorb A, Schöller H, Aichhorn W. Prediction of sports injuries by psychological process monitoring. Sports Psychiatry. 2023;2(4):135–42. Dakos V, Carpenter SR, Brock WA, Ellison AM, Guttal V, Livina V, et al. Methods for Detecting Early Warnings of Critical Transitions in Time Series Illustrated Using Simulated Ecological Data. PLoS ONE. 2012;7(7). Wichers M, Smit AC, Snippe E. Early warning signals based on momentary affect dynamics can expose nearby transitions in depression: A confirmatory single-subject time-series study. Journal for Person-Oriented Research. 2020;6(1):1–15. Chevance G, Baretta D, Heino M, Perski O, Olthof M, Klasnja P, et al. Characterizing and predicting person-specific, day-to-day, fluctuations in walking behavior. PLoS ONE. 2021;16(5):e0251659. Fuller CW, Ekstrand J, Junge A, Andersen TE, Bahr R, Dvorak J, et al. Consensus statement on injury definitions and data collection procedures in studies of football (soccer) injuries. Scandinavian Journal of Medicine and Science in Sports. 2006;16(2):83–92. Martínez-Silván D, Díaz-Ocejo J, Murray A. Predictive indicators of overuse injuries in adolescent endurance athletes. International Journal of Sports Physiology and Performance. 2017;12:153–6. Soligard T, Schwellnus M, Alonso JM. How much is too much? (Part 1) nternational Olympic Committee consensus statement on load in sport and risk of injury. British Journal of Sports Medicine. 2016;50(17):1042. Kalkhoven JT, Watsford ML, Coutts AJ, Edwards WB, Impellizzeri FM. Training Load and Injury : Causal Pathways and Future Directions. Sports Medicine. 2021;51(6):1137–50. Halson SL. Monitoring Training Load to Understand Fatigue in Athletes. Sports Medicine. 2014;44(Suppl 2):139–47. Den Hartigh RJR, Meerhoff LRA, Van Yperen NW, Neumann ND, Brauers JJ, Frencken WGP, et al. Resilience in Sports: A Multidisciplinary, Dynamic, and Personalized Perspective. International Review of Sport and Exercise Psychology. 2022; Ivarsson A, Johnson U, Podlog L. Psychological Predictors of Injury Occurrence : A Prospective Investigation of Psychological Predictors of Injury Occurrence : A Prospective Investigation of Professional Swedish Soccer Players. Journal of Sport Rehabilitation. 2013;22:19–26. de Leeuw AW, van der Zwaard S, van Baar R, Knobbe A. Personalized machine learning approach to injury monitoring in elite volleyball players. European Journal of Sport Science. 2022;22(4):511–20. Jones CM, Griffiths PC, Mellalieu SD. Training Load and Fatigue Marker Associations with Injury and Illness: A Systematic Review of Longitudinal Studies. Sports Medicine. 2017;47(5):943–74. Zadeh A, Taylor D, Bertsos M, Tillman T, Nosoudi N, Bruce S. Predicting Sports Injuries with Wearable Technology and Data Analysis. Information Systems Frontiers. 2021;23:1023–37. Lathlean TJH, Gastin PB, Newstead S V., Finch CF. Player wellness (soreness and stress) and injury in elite junior Australian football players over 1 season. International Journal of Sports Physiology and Performance. 2020;15(10):1422–9. Fletcher D, Sarkar M. A grounded theory of psychological resilience in Olympic champions. Psychology of Sport and Exercise. 2012;13(5):669–78. Galli N, Vealey RS. “Bouncing back” from adversity: Athletes’ experiences of resilience. Sport Psychologist. 2008;22(3):316–35. Kleinert J. Mood states and perceived physical states as short term predictors of sport injuries: Two prospective studies. International Journal of Sport and Exercise Psychology. 2007;5(4):340–51. Wiese-Bjornstal DM. Psychological predictors and consequences of injuries in sport settings. In: APA handbook of sport and exercise psychology. 2019. p. 699–725. Galambos SA, Terry PC, Moyle GM, Locke SA. Psychological predictors of injury among elite athletes. British Journal of Sports Medicine. 2005;39(6):351–4. Ehrmann FE, Duncan CS, Sindhusake D, Franzsen WN, Greene DA. GPS And Injury Prevention In Professional Soccer. Journal of Strenght and Conditioning Research. 2016;2:360–7. Cross MJ, Williams S, Trewartha G, Kemp SPT, Stokes KA. The influence of in-season training loads on injury risk in professional rugby union. International Journal of Sports Physiology and Performance. 2016;11(3):350–5. Bourdon PC, Cardinale M, Murray A, Gastin P, Kellmann M, Varley MC, et al. Monitoring athlete training loads: Consensus statement. International Journal of Sports Physiology and Performance. 2017;12(s2):161–70. Aiello F, Di C, Fanchini M, Impellizzeri FM, Mccall A, Sharp C, et al. Do non-contact injuries occur during high-speed running in elite football ? Preliminary results from a novel GPS and video-based method. Journal of Science and Medicine in Sport. 2023;26(9):465–70. Gabbett T. The development and application of an injury prediction model for noncontact, soft-tissue injuries in elite collision sport athletes. Journal of Strength and Conditioning Research. 2010;24(10):2593–603. Van Der Does HTD, Brink MS, Otter RTA, Visscher C, Marie Lemmink KAP. Injury risk is increased by changes in perceived recovery of team sport players. Clinical Journal of Sport Medicine. 2017;27(1):46–51. Johnson U, Ivarsson A. Psychosocial Factors and Sport Injuries: Prediction, Prevention and Future Research Directions. Current Opinion in Psychology. 2017;(16):89–92. Kellmann M, Bertollo M, Bosquet L, Brink M, Coutts AJ, Duffield R, et al. Recovery and performance in sport: Consensus statement. International Journal of Sports Physiology and Performance. 2018;13(2):240–5. Impellizzeri FM, Rampinini E, Coutts AJ, Sassi A, Marcora SM. Use of RPE-based training load in soccer. Medicine and Science in Sports and Exercise. 2004;36(6):1042–7. Foster C, Florhaug JA, Franklin J, Gottschall L, Hrovatin LA, Parker S, et al. A new approach to monitoring exercise training. Journal of Strenght and Conditioning Research. 2001;15(1):109–15. Foster C. Monitoring training in athletes with reference to overtraining syndrome. Medicine & Science in Sports & Exercise. 1998;30(7):1164–8. Kenttä G, Hassmén P. Overtraining and recovery. Sports Medicine. 1998;26(1):1–16. Brink MS, Visscher C, Arends S, Zwerver J, Post WJ, Lemmink KAPM. Monitoring stress and recovery: new insights for the prevention of injuries and illnesses in elite youth soccer players. British Journal of Sports Medicine. 2010;44(11):809–15. Bandura A. Guide for constructing self-efficacy scales. In: Urdan T, Pajares F, editors. Self-Efficacy Beliefs of Adolescents. Greenwich: Information Age Publishing; 2006. p. 307–37. Barte JCM, Nieuwenhuys A, Geurts SAE, Kompier MAJ. Motivation counteracts fatigue-induced performance decrements in soccer passing performance. Journal of Sports Sciences. 2019;37(10):1189–96. Cohen AB, Tenenbaum G, English RW. Emotions and golf performance: An IZOF-based applied sport psychology case study. Behavior Modification. 2006;30(3):259–80. Borg GAV. Psychophysical bases of perceived exertion. Medicine & Science in Sports & Exercise. 1982;14(5):377–81. Brink MS, Nederhof E, Visscher C, Schmikili SL, Lemmink KAPM. Monitoring Load, Recovery, and Performance in Young Elite Soccer Players. Journal of Strenght and Conditioning Research. 2010;24(3):597–603. Van Yperen NW, Jonker L, Verbeek J. Predicting Dropout From Organized Football: A Prospective 4-Year Study Among Adolescent and Young Adult Football Players. Frontiers in Sports and Active Living. 2022;3:752884. R Core Team. R: A language and environment for statistical computing (Version 4.3.1) [Software] [Internet]. Vienna, Austria: R Foundation for Statistical Computing; 2023. Available from: https://www.r-project.org/. van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained equations in R. Journal of Statistical Software. 2011;45(3):1–67. Hasselman F. casnet: A toolbox for studying Complex Adaptive Systems and NETworks. R package version 01. 2018;3. Ruddy JD, Cormack SJ, Whiteley R, Williams MD, Timmins RG, Opar DA, et al. Modeling the Risk of Team Sport Injuries : A Narrative Review of Different Statistical Approaches. Frontiers in Physiology. 2019;10(413742). Flach PA, Kull M. Precision-Recall-Gain Curves : PR Analysis Done Right. In: Advances in neural information processing systems. 2015. Neumann ND, Van Yperen NW, Brauers JJ, Frencken W, Brink MS, Lemmink KAMP, et al. Nonergodicity in Load and Recovery: Group Results Do Not Generalize to Individuals. International Journal of Sports Physiology and Performance. 2021;17(3):391–9. Neumann ND, Van Yperen NW, Arens CR, Brauers JJ, Lemmink KA, Emerencia AC, et al. How Do Psychological and Physiological Performance Determinants Interact Within Individual Athletes? An Analytical Network Approach. International Journal of Sport and Exercise Psychology. 2024;1–22. Hecksteden A, Schmartz GP, Egyptien Y, Aus der Fünten K, Keller A, Meyer T. Forecasting football injuries by combining screening, monitoring and machine learning. Science and Medicine in Football. 2023;7(3):214–28. Robertson S, Bartlett JD, Gastin PB. Red, Amber, or Green ? Athlete Monitoring in Team Sport: The Need for Decision-Support Systems. International Journal of Sports Physiology and Performance. 2017;12(s2):73–9. Thornton HR, Delaney JA, Duthie GM, Dascombe BJ. Developing Athlete Monitoring Systems in Team Steports : Data Analysis and Visualization. International Journal of Sports Physiology and Performance. 2019;14(6):698–705. Hecksteden A, Keller N, Zhang G, Meyer T, Hauser T. Why Humble Farmers May in Fact Grow Bigger Potatoes : A Call for Street ‑ Smart Decision ‑ Making in Sport. Sports Medicine - Open. 2023;9(1). Olthof M, Hasselman F, Lichtwarck-Aschoff A. Complexity In Psychological Self-Ratings: Implications for research and practice. BMC Medicine. 2020;1–16. Tenan MS. Missing Data in Sport Science : A Didactic Example Using Wearables in American Football. Sports Medicine. 2023;53(6):1109–16. Footnotes Previous research has demonstrated that self-reports and sensor data reveal complex dynamics [ 37 , 84 ]. Therefore, we can assume to actually capture and study them under these methods. Missing data is inevitable in sports monitoring. In consultation with the football club, we decided that a maximum of 20% missingness per variable is still acceptable for the analysis (this refers to roughly one day per week on average). Data was mainly missing, because players forgot to fill out the questionnaire or the sensors did not work. However, the analysis showed that the data was not missing completely at random, which is why we performed multiple imputation (R package mice ) on the data. Recently, it was shown that imputing missing data (preferably using multiple imputation methods) is important to avoid biases and to make valid data-driven decisions in sports [ 85 ]. The window size of seven data points refers to roughly one and a half weeks in the present context (note that the football club did not measure every day of the week but for every training or match day). Previous studies used the same (relatively short) window size [ 27 , 28 , 37 ]. A validation study has shown that short window sizes still report sufficiently good outcomes [ 24 ]. The analysis indicated that a seven-data point window before injuries was optimal since extending the window by an additional data point did not improve the predictive validity. Supplementary Files SupplementaryInformation.pdf Cite Share Download PDF Status: Published Journal Publication published 16 Dec, 2024 Read the published version in Sports Medicine-Open → Version 1 posted Reviewers agreed at journal 23 May, 2024 Reviewers invited by journal 22 May, 2024 Editor invited by journal 20 May, 2024 Editor assigned by journal 19 May, 2024 First submitted to journal 17 May, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4429464","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":305375959,"identity":"06bf5b8c-1882-4493-aeb3-caaaa661f6fa","order_by":0,"name":"Niklas D. Neumann","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYHACAxiD8QGYYidBCzOExUyCFjYJorSYMzBvYC6oOBzNP7v3WHVlWy2DOSEtlg1sBcwzzhzOnXHnXNrNs23HGSybCbnqAI8BM2/b4dyGGzlmNxvbjjEYHCZKy7/DufOBWgpJ0NJwOHcDUAtjY1sNEVoOsxUc5jmWnrvxRo6xZMM5oAkEtRxv3viYp8Y6d96NHMOPDWV1cgbHGwjoAQbpASTuYR4C6jFBHck6RsEoGAWjYPgDAES6QU/AbtWVAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-2062-1958","institution":"University of Groningen: Rijksuniversiteit Groningen","correspondingAuthor":true,"prefix":"","firstName":"Niklas","middleName":"D.","lastName":"Neumann","suffix":""},{"id":305375960,"identity":"eb104ecc-e1c4-4737-96dc-672b1faa14a5","order_by":1,"name":"Jur J. Brauers","email":"","orcid":"","institution":"University Medical Centre Groningen: Universitair Medisch Centrum Groningen","correspondingAuthor":false,"prefix":"","firstName":"Jur","middleName":"J.","lastName":"Brauers","suffix":""},{"id":305375961,"identity":"4ab3ec0a-cada-4187-9195-04b16484bb58","order_by":2,"name":"Nico W. Van Yperen","email":"","orcid":"","institution":"University of Groningen: Rijksuniversiteit Groningen","correspondingAuthor":false,"prefix":"","firstName":"Nico","middleName":"W. Van","lastName":"Yperen","suffix":""},{"id":305375962,"identity":"54580527-abbd-4dad-94c0-ed6cff2aacea","order_by":3,"name":"Mees Van der Linde","email":"","orcid":"","institution":"Football club Groningen","correspondingAuthor":false,"prefix":"","firstName":"Mees","middleName":"Van der","lastName":"Linde","suffix":""},{"id":305375963,"identity":"e3b04cd3-c9b8-4d4c-92dc-77b8ba7d4be3","order_by":4,"name":"Koen A.P.M. Lemmink","email":"","orcid":"","institution":"University Medical Centre Groningen: Universitair Medisch Centrum Groningen","correspondingAuthor":false,"prefix":"","firstName":"Koen","middleName":"A.P.M.","lastName":"Lemmink","suffix":""},{"id":305375964,"identity":"f5d3261e-b745-43f6-8b74-22de6d48b38c","order_by":5,"name":"Michel S. Brink","email":"","orcid":"","institution":"University Medical Centre Groningen: Universitair Medisch Centrum Groningen","correspondingAuthor":false,"prefix":"","firstName":"Michel","middleName":"S.","lastName":"Brink","suffix":""},{"id":305375965,"identity":"99c5c8ef-d3a4-4a3a-8009-f0e707968ef6","order_by":6,"name":"Fred Hasselman","email":"","orcid":"","institution":"Radboud University Nijmegen: Radboud Universiteit","correspondingAuthor":false,"prefix":"","firstName":"Fred","middleName":"","lastName":"Hasselman","suffix":""},{"id":305375966,"identity":"0b3e977d-1459-4639-b314-b01b3c7fee5b","order_by":7,"name":"Ruud J.R. Den Hartigh","email":"","orcid":"","institution":"University of Groningen: Rijksuniversiteit Groningen","correspondingAuthor":false,"prefix":"","firstName":"Ruud","middleName":"J.R. Den","lastName":"Ha","suffix":"J.R."}],"badges":[],"createdAt":"2024-05-16 08:08:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4429464/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4429464/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40798-024-00787-5","type":"published","date":"2024-12-16T15:58:09+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":57625901,"identity":"4a1be959-61a2-4979-8339-462fec0683d8","added_by":"auto","created_at":"2024-06-03 14:00:31","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":85712,"visible":true,"origin":"","legend":"\u003cp\u003eTheoretical illustration of critical fluctuations before an injury\u003cstrong\u003e. \u003c/strong\u003eThe transition from one stable state (e.g., non-injured) to another stable state (e.g., injured) is often characterized by a period of instability, like critical fluctuations. This period can be considered an EWS.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4429464/v1/1e39bec49db3fdb3fe863152.png"},{"id":57625902,"identity":"dc05779a-eae5-4681-907d-b094d0794b1c","added_by":"auto","created_at":"2024-06-03 14:00:31","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":488395,"visible":true,"origin":"","legend":"\u003cp\u003eResults of Player 9\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4429464/v1/6c7cf7e8931eb038eda3b6f5.png"},{"id":72202528,"identity":"0238884a-fe31-444d-892b-6549f14d5364","added_by":"auto","created_at":"2024-12-23 16:14:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1605333,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4429464/v1/234c83ac-2f79-4841-97c8-085d1e3e6e27.pdf"},{"id":57625903,"identity":"35a2e1f8-a14e-4304-ad91-e513d1cbd085","added_by":"auto","created_at":"2024-06-03 14:00:31","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":5569276,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4429464/v1/918d981d6823133d6a93e905.pdf"}],"financialInterests":"","formattedTitle":"Critical Fluctuations as an Early Warning Signal of Sports Injuries? Applying the Complex Dynamic Systems Toolbox to Football Monitoring Data","fulltext":[{"header":"Key points","content":"\u003cul\u003e\n \u003cli\u003eComplex Systems Theory suggests that sports injuries may be preceded by a warning signal characterized by a short window of increased critical fluctuations.\u003c/li\u003e\n \u003cli\u003eResults of the current study showed such increased critical fluctuations before 31% of the injuries, which supports previous theoretical notions. Across the entire data set, we also found a considerable number of critical fluctuations that was not followed by an injury, suggesting that the warning signal may also precede transitions to other (e.g., \u0026nbsp;healthier) states.\u003c/li\u003e\n \u003cli\u003eIncreased critical fluctuations may be interpreted as a window of opportunity for the practitioner to launch timely and targeted interventions, and researchers should dig deeper into the meaning of such fluctuations. \u003cem\u003e\u003cbr\u003e\u0026nbsp;\u003c/em\u003e\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"1. Background","content":"\u003cp\u003eHardly any athlete finishes their career unscathed, as traumatic and overuse sports injuries are a ubiquitous and emotionally and physically disturbing aspect of the athletic trajectory [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Those who are affected often experience performance decrements and suffer financially [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Given the substantial incidence of injuries [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] and the harm they cause, the sports field would highly benefit from any kind of anticipation or prediction. In that way, practitioners may be supported in their decision-making on whether, for instance, the training plan needs to be adjusted for an athlete.\u003c/p\u003e \u003cp\u003eWhile the prediction of sports injuries is a subject of significant interest and value, it continues to pose a challenge even after many years of dedicated research. Obviously, injuries can never be predicted perfectly, but recent research suggests that strides can be made by accounting for the dynamic complexity through which injuries come about [\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. In the present empirical research, we will specifically target this dynamic complexity. To be more specific, we will apply and validate an analytic strategy from the complex dynamic systems toolbox to detect warning signals, which can be an important avenue in the field of sports science and medicine.\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003e1.1 Past Approaches and Future Directions of Sports Injury Prediction\u003c/h2\u003e \u003cp\u003ePrevious research on sports injury prediction measured isolated or monodisciplinary risk factors at one or a few points in time, and analysed data with linear methods at the group level [\u003cspan additionalcitationids=\"CR9 CR10 CR11\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Prediction models performed rather low in forecasting injuries, they were poorly developed (i.e., not validated, no code provided, high risk of bias), and not applicable in practice [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Further, much focus has been put on why injuries occur (i.e., revealing relevant factors) but the question of how those factors lead to injuries and when athletes are at increased risk is generally lacking. As a response, researchers have suggested to study sports injuries from a complex dynamic systems perspective [\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan additionalcitationids=\"CR16 CR17 CR18 CR19\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Complex systems consist of factors that interact over time in a non-linear manner [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Those factors coordinate their behaviour while generating patterns that are adapted to their environment. When a pattern is maintained by a system, the pattern can be called an attractor state, that is a state to which the system is \u0026ldquo;attracted\u0026rdquo; [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eComplex systems typically have certain tipping points in which abrupt changes (phase transitions) from one attractor state to another can occur [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan additionalcitationids=\"CR23 CR24 CR25\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. An injury would be the result of this so-called phase transition, a system-wide reorganisation, from a healthy state to an injured state [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. For such a transition to take place, the stability of the prevailing attractor state has become weakened [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Theory and empirical findings in the fields of ecology, financial markets, and psychology show that, in such a phase of instability, critical fluctuations can be observed as a relatively short window of increased variability and turbulence (for an illustration see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan additionalcitationids=\"CR28 CR29 CR30 CR31\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Such fluctuations are therefore called \u0026ldquo;Early Warning Signals\u0026rdquo; (EWS) and might be an explanation of how injuries occur. To be more specific, instead of analysing absolute values of screening tests or daily measures at the group level, which will likely not work in predicting injuries [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e], theory suggests that strides can be made by investigating the complex dynamics, and more specifically, critical fluctuations, at the individual level [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Given the parallel between injury development and complex system dynamics, an interesting question is therefore whether such critical fluctuations can also be found before injuries.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e1.2 Non-linear Time Series Analysis\u003c/h2\u003e \u003cp\u003eTo detect critical fluctuations, time series analysis methodologies are needed [e.g., 27\u0026ndash;30,35,36]. A promising methodological tool in this regard is the \u003cem\u003eDynamic Complexity\u003c/em\u003e (DC) algorithm, which can detect heavy and irregular variability in a system\u0026rsquo;s behaviour [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. DC was developed for the real-time monitoring of human change processes and is therefore specifically suited for non-linear, non-stationary, and short and coarse-grained time series that are typical in psychology, sports science and medicine. So far, DC has already successfully been applied to investigate changes in mood of psychological disorders and physical activity. For instance, a study by Olthof and colleagues [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] showed that increased critical fluctuations predicted sudden shifts in mood in the previous four days for patients receiving psychotherapy. In the same line, research on walking behaviour found a positive and significant association between increased critical fluctuations and a loss of step counts in the next few days [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRecently, Schiepek et al. [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] provided a first exploration of the potential merits of DC as an EWS of sports injuries. In their case study, one professional football (soccer) player filled out the sports process questionnaire across 77 days using an app-based system. During that time, he suffered a twisted cruciate ligament. Results showed that DC reached a peak of critical instability just before the injury. This is a promising finding, but on the basis of this case study, we cannot conclude that such a signal is specific for identifying injuries. Hence, more research is needed to confirm the potential and robustness (validation) of these results across athletes and injuries (e.g., injury mechanism and duration). Further, the study did not determine the time window for which the warning signal appears before the injury, and it is unclear how a peak was defined. In the current in-depth study of DC on sports injuries we address these limitations in Schiepek et al.\u0026rsquo;s [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] study and account for the complex multidisciplinarity of traumatic and overuse sports injuries.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e1.3 The Role of Injury Mechanism (traumatic vs overuse) and Duration\u003c/h2\u003e \u003cp\u003ePrevious research has dichotomized injuries into two fundamental mechanisms: traumatic injuries, stemming from specific, identifiable events, and overuse injuries, resulting from repeated micro-traumas without a single, discernible event precipitating the injury [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Scholars have posited that the prediction of overuse injuries may be more feasible compared to traumatic injuries, given the opportunity to capture the repeated micro-traumas which lead to the injury [e.g., 39]. However, previous investigations propose a more nuanced perspective, contending that, for instance, athletes experiencing high physical load may be more susceptible to traumatic injuries [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e], but that generally, causal pathways between load and injury are poorly understood [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. Furthermore, no study has yet examined the effect of injury duration. Clarifying whether a prediction model exhibits differential efficacy across injury mechanism (traumatic vs overuse) and duration would suggest that the model may be more sensitive to specific patterns. For instance, the model\u0026rsquo;s validity in predicting all injuries of a given sample may be rather low. However, overuse injuries or injuries that last longer, for instance, may be better predicted than traumatic injuries or injuries that result in a short time-loss, respectively. This information can be useful for targeted prevention strategies and interventions and may prevent huge costs, as well as physical and emotional suffering.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e1.4 The Present Study\u003c/h2\u003e \u003cp\u003eThe present study aims to predict injury occurrence of youth football players of a professional academy, based on EWS in the form of critical fluctuations. To meet this aim, it is important to monitor psychological and physiological variables of individual players on a daily basis [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan additionalcitationids=\"CR44 CR45\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Such monitoring can be realized via an online application and sensors, for instance [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e].[1]\u003ca class=\"FNLink\" href=\"#Fn1\" id=\"#FNLinkFn1\"\u003e\u003c/a\u003e There is a myriad of factors that can be measured, yet, the decision on what to measure should be based on what prior scholarly work suggests may be precursors of sports injuries and what is realistic and feasible to collect from the sample. For instance, on the psychological level, self-efficacy, motivation, and mood are key factors in sport performers and injuries [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan additionalcitationids=\"CR49 CR50 CR51 CR52\" citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e]. In addition, athletes may regularly be perturbed by unenjoyable training sessions and bad performances [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. On the physiological level, internal load factors such as the heart rate and the session rating of perceived exertion (sRPE), as well as external load factors such as sprints and total distance covered are typically related to injuries [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan additionalcitationids=\"CR55 CR56 CR57\" citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. Further, the recovery status may be monitored before every training session or competition [\u003cspan additionalcitationids=\"CR60\" citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn line with the complex dynamic systems theory, we hypothesized that increased critical fluctuations (i.e., cumulative peaks of DC) precede the occurrence of an injury. Next to the detection of critical fluctuations at the individual level, this research seeks to validate this EWS by testing its predictive validity at the group level. Last, the study intends to explore whether injury mechanism (traumatic vs overuse) and duration affect the predictive validity and which measured factor(s) (i.e., psychological, physiological, self-reports, sensors) reveal the most peaks in destabilizations.\u003c/p\u003e \u003c/div\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Subjects\u003c/h2\u003e \u003cp\u003eWe collected data from 55 youth male players of the U-18 and U-21 teams (16 to 20 years old) of a premier league (Eredivisie) football club in the Netherlands. 23 of these players were included in the analysis because they fulfilled the inclusion criteria (more details under 2.3.1. Data Pre-Processing). Once a player started playing at the club, he was informed about the data collection process. By signing an informed consent, he could then decide whether he wanted to permit the use of his data for research purposes or not (for the present study, all players provided consent). The players competed in the highest national league of their age category. They had between six and eight training sessions per week, composed of two strength sessions and four to six field sessions of 60 to 75 minutes and 75 to 90 minutes, respectively, and matches on the weekend. Due to personal data protection, further potentially identifiable information (e.g., height, weight, position, team, specific type of injury) is not reported.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Design, Measures, and Procedure\u003c/h2\u003e \u003cp\u003eFor every training and match up to two competitive seasons, we collected psychological and physiological data. To be more specific, players answered self-report questions on self-efficacy, motivation, mood, performance self-evaluation, enjoyment, session rating of perceived exertion (sRPE) and total quality of recovery (TQR) on a tablet computer near the locker room without staff or team members being present. All measures were part of the normal, daily team monitoring routine at the club. Polar TeamPro sensors (Polar Electro Oy, Kempele, Finland) were used to collect data on the duration, total distance covered, sprints, and heart rate for every training session and match (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The self-report questions were extracted from validated questionnaires and adjusted to the present context (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e for references). The RPE consists of one item (\u0026ldquo;How hard was the training/competition?\u0026rdquo;) and is a subjective estimate of the psychological and physiological stress imposed on the athlete [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. Later, we multiplied the RPE score by the duration of the training session to obtain a measure for the internal training load for the analysis, the session RPE (sRPE) [\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. If there were two load scores a day, which occasionally happened after two different types of training, we added them up to capture the daily load [\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. Finally, next to the questions explained in the table, the team physician recorded injury occurrence (yes vs. no), and injury-related time-loss (in days), including the mechanism (i.e., traumatic or overuse) [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eData Collection\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTime of the day\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMeasured factor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelf-report question\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMeasurement scale\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eReferences\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eT1: In the morning up to 30 minutes before the first training session or match\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRecovery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow good is your recovery?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCRS from 6 (\u003cem\u003every, very poor recovery\u003c/em\u003e) to 20 (\u003cem\u003every, very good recovery\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 65,66]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSelf-efficacy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow confident are you that you can perform maximally today?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVAS from 0 (\u003cem\u003enot at all confident\u003c/em\u003e) to 100 (\u003cem\u003every confident\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 52,67]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMotivation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow motivated are you to perform maximally today?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVAS from 0 (\u003cem\u003enot at all motivated\u003c/em\u003e) to 100 (\u003cem\u003emaximally motivated\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 52,68]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow much are you in the mood to train/play the match today?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVAS from 0 (\u003cem\u003enot at all in the mood\u003c/em\u003e) to 100 (\u003cem\u003every much in the mood\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 51,69]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eT2: During the training session or match\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDistance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMeter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 17,54]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSprint\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNumber of sprints\u0026thinsp;\u0026gt;\u0026thinsp;25km/h (for the U-21) and \u0026gt;\u0026thinsp;19.8km/h (for the U-18) per session\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 54,57]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDuration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMinutes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 55,63]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHeart rate in zone 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eBeats per minute; output is the number of seconds spent in zone 5 (i.e., 92\u0026ndash;100% of an individual\u0026rsquo;s max heart rate)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 54,56]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eT3: At the end of the day up to 30 minutes after the last training session or the match\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eExertion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow hard was the training/match?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCRS from 6 (\u003cem\u003every, very light\u003c/em\u003e) to 20 (\u003cem\u003every, very hard\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 56,70]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerceived performance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow well did you perform today?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVAS from 0 (\u003cem\u003every bad (far below my capabilities\u003c/em\u003e)) to 100 (maximally (\u003cem\u003eto the best of my capabilities\u003c/em\u003e))\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 43,71]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnjoyment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHow much did you enjoy the training session(s)/the match today?\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVAS from 0 (\u003cem\u003enot at all\u003c/em\u003e) to 100 (\u003cem\u003every much\u003c/em\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[e.g., 43,72]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eCRS, Category-Ratio Scale; T, time point; VAS, Visual Analogue Scale.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Data Set and Statistical Analysis\u003c/h2\u003e \u003cp\u003eThe analysis was performed with R and RStudio [\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e]. The R code has been made publicly available (see Code Availability).\u003c/p\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1 Data Pre-Processing\u003c/h2\u003e \u003cp\u003eThe original data set consisted of 55 players from two youth teams. The following criteria were established to include players for the analysis: At least one injury occurrence was reported of a player between measurement seven and the last seven measurements (see the Note in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e for more background on this inclusion criteria) and a player did not miss more than 20% of the values per measured factor.[2]\u003ca class=\"FNLink\" href=\"#Fn2\" id=\"#FNLinkFn2\"\u003e\u003c/a\u003e This resulted in a final sample of 23 players, yielding an average of 339 observations per player (range\u0026thinsp;=\u0026thinsp;155\u0026ndash;430). First, we imputed missing values with the R package \u003cem\u003emice\u003c/em\u003e [\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e]. As a next step, we normalized the data in order to be able to analyse and compare the results. To be more specific, self-report data from the Visual Analogue Scale was divided by 100 (i.e., the maximum), self-report data from the Category-Ratio Scale was divided by 20 (i.e., the maximum), and data from sensors was divided by the maximum value of each factor of every individual.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e2.3.2 Early Warning Signals (EWS) Calculation\u003c/h2\u003e \u003cp\u003eWe calculated critical fluctuations of the self-report and sensor data with the R package \u003cem\u003ecasnet\u003c/em\u003e [\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e], using the Dynamic Complexity (DC) algorithm. Mathematically, DC is a multiplication of the degree of fluctuation (F) of a time series and the distribution (D) of values between the theoretical minimum and maximum of a scale [for validation and details see 24]. F is sensitive to the amplitude and frequencies of a time series and D to the scattering of values. Both measures were calculated within a moving window of seven data points and a window step of one, in order to identify non-stationary changes.[3]\u003ca class=\"FNLink\" href=\"#Fn3\" id=\"#FNLinkFn3\"\u003e\u003c/a\u003e Next, we checked for cumulative complexity peaks (CCPs), which indicate whether the number of simultaneous peaks in DC of the time series was significant. A CCP is determined by conducting a one-sided z-test (one side, because we are looking for \u003cem\u003eincreased\u003c/em\u003e critical fluctuations) to check for significantly increased scores, which results in a new time series that mirrors the number of significant peaks on a specific day. Another one-sided z-test is performed on this new time series to determine whether this number is significant. Since critical fluctuations are expected to be present just before a transition, we tested which time window between one and ten data points prior to the injury would be optimal for predicting injures. We increased the window size in a step-by-step manner starting with one. When increasing the window size did not provide better results, the previous window was considered as the optimum.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e2.3.3 Validation Analysis\u003c/h2\u003e \u003cp\u003eWe validated the predictive validity of the warning signal in four steps. First, we calculated the sensitivity of the DC analysis for every individual. The sensitivity evaluates the warning signal's effectiveness in correctly identifying injury cases, highlighting its ability to accurately detect true positive cases and emphasizing its validity in identifying injuries [\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e]. It is the ratio of the true positive rate (i.e., a significant CCP and the athlete was injured) to the true positive rate plus the false negative rate (i.e., no significant CCP and the athlete was injured). Second, we calculated the specificity of the DC analysis for every individual. The specificity assesses the accuracy of the warning signal to correctly identify true negative cases [\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e]. It is the ratio of the true negative rate (i.e., no significant CCP and the athlete was not injured) to the true negative rate plus the false positive rate (i.e., a significant CCP and the athlete was not injured). Third, we calculated the accuracy of the warning signal. Accuracy is simply the percentage of correct predictions (injury and non-injury) made by the model [\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e]. It is calculated by dividing the number of correct predictions by the total number of predictions made. The result is multiplied by 100.\u003c/p\u003e \u003cp\u003eThe inherently imbalanced nature of the dataset in this research, with a strong bias towards the majority class (non-injuries), may result in less stable and reliable estimates of sensitivity, specificity, and accuracy. For this reason, we also calculated an F\u003csub\u003e1\u003c/sub\u003e score (see formula 1) as the fourth and last step, by using the harmonic mean of precision and recall [\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e]: (1)\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$${F}_{1}=\\frac{2\\text{*} \\text{p}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}\\text{*} \\text{r}\\text{e}\\text{c}\\text{a}\\text{l}\\text{l}}{\\text{p}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}+\\text{r}\\text{e}\\text{c}\\text{a}\\text{l}\\text{l}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ePrecision is the ratio of the true positive rate (i.e., a significant CCP and the athlete was injured) to the true positive rate plus the false positive rate (i.e., a significant CCP and the athlete was not injured). Recall is another term for sensitivity. The F\u003csub\u003e1\u003c/sub\u003e score falls between 0 and 1, with 0 indicating no predictive power and 1 indicating perfect prediction. Important to note is, however, that the analysis takes place at the group level, since all individual observations must be combined and treated as one large observation.\u003c/p\u003e \u003cp\u003eWe applied the following rules for the analysis: If at least one warning signal occurred in the seven days prior to an injury, the remaining six days were left out for the analysis (they were not interesting anymore, since a warning signal appeared already). Likewise, significant peaks during an injury period were not taken into account for the analysis, because the injury already happened, and turbulence can be expected in this period.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e \u003ch2\u003e2.3.4 Analysis of Injury Mechanism (Traumatic vs Overuse) and Duration\u003c/h2\u003e \u003cp\u003eWe calculated the percentage of correctly predicted traumatic and overuse injuries and we checked with a chi-squared test whether the difference was statistically significant. Then, we calculated a bi-serial correlation to test if the duration of an injury (i.e., time-loss in days) is associated with the predictive validity index of the analysis tool. A post-hoc power analysis determined that a sample size of 20 is needed for calculating a bi-serial correlation with a desired power of 0.8, an alpha level of \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.05, and an effect size of \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.9. In our final step, we counted how often each measured variable occurred at least once in the seven days prior to an injury in the CCPs for correct predictions.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Validation Analysis\u003c/h2\u003e \u003cp\u003ePlayers were injured 2.8 times on average across the measurement period (range\u0026thinsp;=\u0026thinsp;1\u0026ndash;6; total\u0026thinsp;=\u0026thinsp;64). To illustrate, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the multivariate raw data, the complexity resonance diagram, and the critical instability plot of one representative player (Player 9; for the plots of the other players, see online resource). Note that for one of the two injuries, a significant CCP appeared in the seven data points before \u0026ndash; an EWS.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCCP\u0026thinsp;=\u0026thinsp;Cumulative Complexity Peak, DC\u0026thinsp;=\u0026thinsp;Dynamic Complexity, sRPE\u0026thinsp;=\u0026thinsp;session Rating of Perceived Exertion. Multivariate raw time series (top figure), complexity resonance diagram (middle figure) and critical instability plot (bottom figure) of Player 9 suffering two injuries (injury period marked in orange/vertically shaded from top to bottom). The x-axis displays the number of data points. The y-axis shows each measured factor with the raw time series (top), the Dynamic Complexity values (middle), and the significant Dynamic Complexity values (bottom). Note that the time scale starts at seven and stops seven data points before the end of the measurement period, because we used a seven-data point overlapping moving window to calculate the Dynamic Complexity. This means that the first available Dynamic Complexity value is at data point seven. The higher the Dynamic Complexity, the whiter and redder the plot becomes (middle figure). Significant Dynamic Complexity levels are colored grey, whereas significant Cumulative Complexity Peaks are colored black (bottom figure).\u003c/p\u003e \u003cp\u003eThe sensitivity analysis across all players showed that cumulative complexity peaks appeared for 31% (20 out of 64) of the injuries within seven data points before (range: 0-100%).[4]\u003ca class=\"FNLink\" href=\"#Fn4\" id=\"#FNLinkFn4\"\u003e\u003c/a\u003e The analysis further demonstrates that the warning signal exhibits a specificity of 94% (range\u0026thinsp;=\u0026thinsp;89\u0026ndash;100%). The accuracy reached a value of 93%. All three metrics were calculated per individual and then averaged across all individuals (individual-level analysis) [\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e]. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the confusion matrix from which these metrics were calculated and Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e provides more detailed information on injuries and the DC analysis for all individual players. Finally, the calculation of the F\u003csub\u003e1\u003c/sub\u003e score, conducted at the group level, revealed an estimate of 0.08. This indicates poor performance in terms of both precision and recall, suggesting that the model's ability to correctly identify positive instances and avoid false positives is very low.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eConfusion Matrix Showing the Predictions and Actual Outcomes\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003ePrediction\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInjured\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNot Injured\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eActual Outcome\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eInjured\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (TP)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44 (FN or Type II Error)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eNot Injured\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e442 (FP or Type I Error)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5987 (TN)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFN, False Negative; FP, False Positive; TN, True Negative; TP, True Positive\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eIndividual Results of all Players\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlayer\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eNumber of Measurements\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNumber of Injuries\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTime-loss in Days\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMechanism\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" morerows=\"1\" nameend=\"c2\" namest=\"c1\" rowspan=\"2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e427\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e92%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e248\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e94%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e351\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e90%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e286\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e419\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e95%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e413\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e33%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e241\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e94%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e398\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e97%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e419\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e94%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e421\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e33%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e95%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e413\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e95%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e229\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e92%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e408\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e89%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e288\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e60%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e422\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e33%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e155\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e25%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e418\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e100%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e90%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e415\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e95%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e258\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e99%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e430\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e92%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eO\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e423\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e33%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e168\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e90%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e247\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e93%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"8\"\u003eNA, Not Available; O, Overuse; T, Traumatic. Injuries that occurred before data point seven and in the last seven data points are left out (n\u0026thinsp;=\u0026thinsp;3), because the window size of seven data points only allows to capture Dynamic Complexity changes in between. Time-loss in days refers to the lost time in actual days, not days of measurement\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Analysis of Injury Mechanism (Traumatic vs Overuse) and Duration\u003c/h2\u003e \u003cp\u003eWe tested whether our prediction model exhibits differential efficacy across injury mechanisms and duration. Differences would suggest that the model may be more sensitive to specific injuries and could help prevent those instances. The analysis showed no significant difference (χ\u003csup\u003e2\u003c/sup\u003e (3, N\u0026thinsp;=\u0026thinsp;20)\u0026thinsp;=\u0026thinsp;2.86, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.41) in the prediction of traumatic and overuse injuries on average (calculated per individual and then averaged across all individuals). To be more specific, 27% of overuse injuries (14 out of 47) and 30% of traumatic injuries (4 out of 15) could be predicted (two injuries were not classified in the dataset).\u003c/p\u003e \u003cp\u003eNext, we tested whether injury duration (in days) is related to the predictive validity index of the analysis tool (0 for not predicted and 1 for predicted). The correlation revealed a coefficient of .31 between injury duration and the predictive validity index with \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.11. Finally, we visually inspected all critical instability plots (see online resource) to determine which measured factor(s) (i.e., psychological, physiological, self-reports, sensors) reveal the most peaks in destabilization: The factors \u003cem\u003esprint\u003c/em\u003e and \u003cem\u003etotal distance\u003c/em\u003e appear most often in the cumulative complexity peaks that correctly predicted an injury, that is, ten times. Two injuries could be predicted with psychological factors only, another two injuries with physiological factors only, three injuries with only self-reports (i.e., psychological and physiological factors), and two injuries with sensor data only. In the 11 remaining injuries, a combination of the factors was found.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe present study tested if critical fluctuations can serve as an EWS of sports injuries by applying the complex dynamic systems toolbox \u0026ndash; specifically the DC algorithm \u0026ndash; to sports monitoring data. In doing so, we collected longitudinal data on psychological and physiological factors of 23 football players across two competitive seasons. The outcomes of the study are threefold. As hypothesized, the results show that critical fluctuations appear quite often before injuries. Theoretically speaking, this supports the notion that injuries may come about through a non-linear dynamic process and that the methodological toolbox of complex dynamic systems could provide such insights [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Also, the findings are in line with DC research on mood and physical activity [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e], and specifically, the potential promising DC application to injuries as presented by Schiepek et al. [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSecond, we took additional measures to check the robustness and practical usefulness of our results. More specifically, the warning signal exhibits a sensitivity of 31% and a specificity of 94%. This means that the model can correctly predict 31% of injury (20 of 64) and 94% of non-injury instances, respectively. The accuracy revealed an estimate of 93%, meaning that it correctly predicts injuries and non-injuries in 93% of the time. These results suggest good predictive validity of the warning signal. However, we also controlled for the imbalance in the data by calculating an F\u003csub\u003e1\u003c/sub\u003e score (at the group level). The result revealed an estimate of 0.08, which suggests that the model's overall ability to discriminate between injuries and non-injuries is rather poor, especially given the high false positive rate. In the section \u003cspan refid=\"Sec19\" class=\"InternalRef\"\u003ePractical Implications\u003c/span\u003e, we put these results into perspective by explaining what this would mean for the practitioner.\u003c/p\u003e \u003cp\u003eGiven the discrepancy between the results based on two different analytic strategies, we strongly recommend including additional checks for future research as well. So far, studies on EWS before critical transitions have not validated warning signals extensively. More specifically, studies outside the sports context often only study periods before transitions instead of the whole time series, which does not capture the warning signals\u0026rsquo; actual predictive validity. Furthermore, previous work tended to provide only one or two metrics (e.g., sensitivity and/or specificity) and did not consider the imbalance in the dataset. This could have led to an overly optimistic report of the predictive validity of EWS metrics and their relevance for practice. With regard to the current study, while critical fluctuations can signal an increased likelihood of injuries, such a warning signal did not always lead to an injury, which can be due to different reasons. For instance, the system can also return to the previous attractor after destabilization, meaning that it does not result in a phase transition. Given that the data in this study is collected as part of the club's everyday routine, where practitioners utilize the data to adjust training processes, it is possible that injuries were prevented without our or the practitioner\u0026rsquo;s awareness. Destabilization, therefore, could better be considered a window of opportunity, meaning that it provides information that the system is currently out of equilibrium and re-organizes itself [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. In the same line, critical fluctuations can be seen as a general indicator of instability and not a predictor of a specific kind of order transition. This means that critical fluctuations may also precede other kinds of transitions (e.g., a performance loss or gain, a mental dip).\u003c/p\u003e \u003cp\u003eThird, although previous theoretical research proposed that overuse injuries hold more information prior to their occurrence [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e], we did not find effects of mechanism (traumatic vs overuse). Similarly, the duration of injuries was no significant factor for the predictive validity of the EWS, that is, injuries that resulted in a large time-loss could not be better predicted than injuries of short duration, and vice versa. These results should be treated with caution, however, because of the small sample size and rather homogenous study population. An interesting additional result is that the physiological sensor factors \u003cem\u003esprint\u003c/em\u003e and \u003cem\u003etotal distance\u003c/em\u003e appear most often in the CCPs, suggesting that they might play an important role in predicting injuries. However, the results also show that for two injuries, only psychological factors appeared, whereas for most other injuries, a combination of the factors (psychological and physiological, self-report and sensor data) is decisive for the prediction. These outcomes highlight the need for a multidisciplinary approach, tailored to individual athletes [e.g., 43,79]. In the next sections, we discuss future directions for research and practical implications of the results for the sports field.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Recommendations for Future Research\u003c/h2\u003e \u003cp\u003eTo the best of our knowledge, the present study is the first to test the predictive validity of the DC algorithm as a potential EWS of sports injuries in a sample of football players. From a complex dynamic systems perspective, the DC algorithm can be considered a strong method since it does not have any specific statistical or parametric assumptions about the data, it mirrors increased complexity and noise of system dynamics before a phase transition, it can model linear and non-linear phenomena, and it can be used in real-time and not only for post data collection. The study itself has high ecological validity because measuring took place in the field as a daily routine and not in an experimental setting.\u003c/p\u003e \u003cp\u003eIn accordance with the high false positive rate on the full time series data, previous research has shown that destabilization is not necessarily related to negative outcomes, such as injuries or increases in symptom severity. Indeed, it may also precede decreases in symptom severity, that is, the transition to a healthier state [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Hence, future research should test whether destabilization in athletes\u0026rsquo; attractors can also predict, for instance, performance gains or decreases in physical complaints. It must be noted that complex systems theory does not suggest all injuries to take place via abrupt transitions preceded by critical fluctuations, they may also develop more gradually and within the same state of attraction. Future studies may test this assumption, as demonstrated by Schiepek and colleagues [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Forthcoming studies may further exploit the complex dynamic systems toolbox to detect dissimilar dynamics preceding critical transitions (e.g., through recurrence quantification analysis and recurrence plots) [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Last, in our study, we merely relied on intensive monitoring data collected daily across a season. However, as Hecksteden et al. [\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e] proposed, the integration of screening tests (i.e., risk factors and protective factors) with monitoring data may be a promising avenue for sports injury forecasting (see online resource for more recommendations for future research).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Practical Implications\u003c/h2\u003e \u003cp\u003eIn a perfect world, practitioners would always be correctly informed by an EWS about an upcoming injury. However, the world is too complex to capture all the information needed to make such reliable forecasts. The findings from our study indicate that, on average, an injury is quite often preceded by critical fluctuations, but at the same time false positive predictions regularly occur. In practice, and based on the data we analysed, this would mean that staff needs to intervene 21 times per player over the course of two seasons, while one of these signals actually leads to an injury which can be prevented (and 1.9 injuries on average per player would be missed due to false negatives). While a high false negative rate may be more harmful than a high false positive rate, intervening 21 times may still appear extensive. However, it is important to recognize that interventions do not always need to be resource-intensive in terms of time and cost. They may simply be short verbal interactions between the (multidisciplinary) staff and the player. For such verbal interactions, practitioners may specifically address those variables that showed increased critical fluctuations in the DC analysis. For Player 9 in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e this concerns his motivation, the seconds of heart rate he spent in zone 5, and the total distance covered. In that way, practitioners are provided with a conversation starter and information on which knob to turn for which athlete.\u003c/p\u003e \u003cp\u003eHowever, if implemented, such a warning system should be considered a decision-support tool for practitioners to adapt training processes and implement interventions instead of a replacement of the coach [\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e, \u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e]. In a recent study by Hecksteden et al. [\u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e], it was emphasized that a blend of practitioner experience, judgment, and scientific evidence should be used for decision-making in sports.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Conclusion\u003c/h2\u003e \u003cp\u003eThe results of this study suggest that critical fluctuations often precede sports injuries, thereby supporting the complex dynamic systems perspective on injuries. At the same time, it cannot (yet) be considered a valid EWS for the real-time prediction of injuries in sports practice. Critical fluctuations might be interpreted as a window of opportunity, because turbulence indicates re-organization of the system which is currently out of equilibrium. Practitioners may make use of this window by planning timely and targeted interventions to guide athletes toward desirable conditions. Future research should dig deeper into the meaning of critical fluctuations in the psychophysiological states of athletes.\u003c/p\u003e \u003c/div\u003e"},{"header":"Abbreviations","content":"\u003cul\u003e\n \u003cli\u003eDC = Dynamic Complexity\u003c/li\u003e\n \u003cli\u003eEWS = Early Warning Signal\u003c/li\u003e\n \u003cli\u003esRPE = session Rating of Perceived Exertion\u003c/li\u003e\n \u003cli\u003eTQR = Total Quality of Recovery\u003c/li\u003e\n \u003cli\u003eCCP = Cumulative Complexity Peak\u003cbr\u003e\u0026nbsp;\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate: \u003c/strong\u003eThe present study was conducted according to the requirements of the Declaration of Helsinki and was approved by the ethics committee of the Faculty of Behavioral and Social Sciences of the University of Groningen (research code: PSY-1819-S-0308). Consent of participants was obtained to participate in the data collection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e: Consent of participants was obtained to publish their data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material: \u003c/strong\u003eThe data contains sensitive information about the subjects and is therefore restricted from openly sharing it. Requests can be made by researchers affiliated with universities or independent, non-commercial research institutes via DataVerseNL at https://dataverse.nl/privateurl.xhtml?token=e0f11c98-9377-4522-8c81-a80cda00c750. The R code is openly accessible via the same link.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e: We have no competing interests to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e This work was supported by The Netherlands Organization for Health Research and Development (ZonMw, grant number 546003004).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNDN took the lead in writing the manuscript. MvdL collected the data. RJRdH is the leader of the project. NDN, JJB, and FH analysed the data and wrote the R code. All authors designed the study, edited this article critically, and read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors thank Dr. Merlijn Olthoff for helpful and insightful comments on the analysis.\u003cstrong\u003e\u003cbr\u003e \u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGreen SJ, Weinberg RS. Relationships among Athletic Identity , Coping Skills , Social Support , and the Psychological Impact of Injury in Recreational Participants. Journal of Applied Sport Psychology. 2001;13(1):40\u0026ndash;59. \u003c/li\u003e\n\u003cli\u003eCumps E, Cumps E, Verhagen E, Annemans L, Meeusen R. Injury rate and socioeconomic costs resulting from sports injuries in Flanders : data derived from sports insurance statistics 2003. British Journal of Sports Medicine. 2007;42:767\u0026ndash;72. \u003c/li\u003e\n\u003cli\u003eBolling C, Delfino Barboza S, van Mechelen W, Pasman HR. How elite athletes, coaches, and physiotherapists perceive a sports injury. Translational Sports Medicine. 2019;2(1):17\u0026ndash;23. \u003c/li\u003e\n\u003cli\u003eEkstrand J, H\u0026auml;gglund M, Wald\u0026eacute;n M, Ekstrand J, H\u0026auml;gglund M, Wald\u0026eacute;n M. Injury incidence and injury patterns in professional football - the UEFA injury study. British Journal of Sport. 2011;45(7):553\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eBittencourt NFN, Meeuwisse WH, Mendon\u0026ccedil;a LD, Ocarino JM, Fonseca ST. Complex systems approach for sports injuries: moving from risk factor identification to injury pattern recognition \u0026mdash; narrative review and new concept. British Journal of Sports Medicine. 2016;50(21):1309\u0026ndash;14. \u003c/li\u003e\n\u003cli\u003eFonseca ST, Souza TR, Verhagen E, van Emmerik R, Bittencourt NFN, Mendon\u0026ccedil;a LDM, et al. Sports Injury Forecasting and Complexity: A Synergetic Approach. Sports Medicine. 2020;50:1757\u0026ndash;70. \u003c/li\u003e\n\u003cli\u003ePol R, Hristovski R, Medina D, Balague N. From microscopic to macroscopic sports injuries. Applying the complex dynamic systems approach to sports medicine: a narrative review. British journal of sports medicine. 2019;53(19):1214\u0026ndash;20. \u003c/li\u003e\n\u003cli\u003eBullock GS, Mylott J, Hughes T, Nicholson KF, Riley RD, Collins GS. Just How Confident Can We Be in Predicting Sports Injuries ? A Systematic Review of the Methodological Conduct and Performance of Existing Musculoskeletal Injury Prediction Models in Sport. Sports Medicine. 2022;52(10):2469\u0026ndash;82. \u003c/li\u003e\n\u003cli\u003eJohnson U, Tranaeus U, Ivarsson A. Current status and future challenges in psychological research of sport injury prediction and prevention: A methodological perspective. Revista de Psicolog\u0026iacute;a del Deporte. 2014;23(2):401\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eBache-Mathiesen LK, Andersen TE, Dalen-Lorentsen T, Clarsen B, Fagerland MW. Not straightforward: Modelling non-linearity in training load and injury research. BMJ Open Sport and Exercise Medicine. 2021;7(3):1\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eStern BD, Hegedus EJ, Lai Y cheng. Injury prediction as a non-linear system. Physical Therapy in Sport. 2020;41:43\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eBahr R. Why screening tests to predict injury do not work-and probably never will.: A critical review. British Journal of Sports Medicine. 2016;50(13):776\u0026ndash;80. \u003c/li\u003e\n\u003cli\u003eSeow D, Graham I, Massey A. Prediction models for musculoskeletal injuries in professional sporting activities: A systematic review. Translational Sports Medicine. 2020;3(6):505\u0026ndash;17. \u003c/li\u003e\n\u003cli\u003eVan Eetvelde H, Mendon\u0026ccedil;a LD, Ley C, Seil R, Tischer T. Machine learning methods in sport injury prediction and prevention : a systematic review. Journal of Experimental Orthopaedics. 2021;8(27):1\u0026ndash;15. \u003c/li\u003e\n\u003cli\u003eCook C. Predicting future physical injury in sports: it \u0026rsquo; s a complicated dynamic system. British Journal of Sports Medicine. 2016;50(22):1356\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eBekker S, Clark AM. Bringing complexity to sports injury prevention research: From simplification to explanation. British Journal of Sports Medicine. 2016;50(24):1489\u0026ndash;90. \u003c/li\u003e\n\u003cli\u003eHulme A, Thompson J, Nielsen RO, Read GJM, Salmon PM. Towards a complex systems approach in sports injury research: Simulating running-related injury development with agent-based modelling. British Journal of Sports Medicine. 2018;53(9):560\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eTee JC, McLaren SJ, Jones B. Sports Injury Prevention is Complex: We Need to Invest in Better Processes, Not Singular Solutions. Sports Medicine. 2020;50(4):689\u0026ndash;702. \u003c/li\u003e\n\u003cli\u003eYung KK, Ardern CL, Serpiello FR, Robertson S. Characteristics of Complex Systems in Sports Injury Rehabilitation: Examples and Implications for Practice. Sports Medicine - Open. 2022;8(1). \u003c/li\u003e\n\u003cli\u003eL\u0026ouml;vdal SS, Den Hartigh RJR, Azzopardi G. Injury Prediction in Competitive Runners With Machine Learning. International Journal of Sports Physiology and Performance. 2021;16(10):1522\u0026ndash;31. \u003c/li\u003e\n\u003cli\u003eOlthof M, Hasselman F, Strunk G, Aas B, Schiepek G, Lichtwarck-Aschoff A. Destabilization in self-ratings of the psychotherapeutic process is associated with better treatment outcome in patients with mood disorders. Psychotherapy Research. 2020;30(4):520\u0026ndash;31. \u003c/li\u003e\n\u003cli\u003eHaken H. Synergetics: An Introduction. Nonequilibrium Phase-Transitions and Self-Organization in Physics, Chemistry and Biology. Springer Series in Synergetics 1. Berlin: Springer; 1983. \u003c/li\u003e\n\u003cli\u003eStern BD, Hegedus EJ, Lai YC. State dependence: Does a prior injury predict a future injury? Physical Therapy in Sport. 2021;49:8\u0026ndash;14. \u003c/li\u003e\n\u003cli\u003eSchiepek G, Strunk G. The identification of critical fluctuations and phase transitions in short term and coarse-grained time series \u0026mdash; a method for the real-time monitoring of human change processes. Biological Cybernetics. 2010;102:197\u0026ndash;207. \u003c/li\u003e\n\u003cli\u003eScheffer M, Carpenter SR, Lenton TM, Bascompte J, Brock W, Dakos V, et al. Anticipating critical transitions. Science. 2012;338(6105):344\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eScheffer M, Bolhuis JE, Borsboom D, Buchman TG, Gijzel SMW, Goulson D, et al. Quantifying resilience of humans and other animals. Proceedings of the National Academy of Sciences of the United States of America. 2018;115(47):11883\u0026ndash;90. \u003c/li\u003e\n\u003cli\u003eHasselman F. Early Warning Signals in Phase Space: Geometric Resilience Loss Indicators from Multiplex Cumulative Recurrence Networks. Frontiers in Physiology. 2022;13:859127. \u003c/li\u003e\n\u003cli\u003eOlthof M, Hasselman F, Strunk G, van Rooij M, Aas B, Helmich MA, et al. Critical Fluctuations as an Early-Warning Signal for Sudden Gains and Losses in Patients Receiving Psychotherapy for Mood Disorders. Clinical Psychological Science. 2020;8(1):25\u0026ndash;35. \u003c/li\u003e\n\u003cli\u003eScheffer M, Bascompte J, Brock WA, Brovkin V, Carpenter SR, Dakos V, et al. Early-warning signals for critical transitions. Nature. 2009;461(7260):53\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eHelmich MA, Olthof M, Oldehinkel AJ, Wichers M, Bringmann LF, Smit AC. Early warning signals and critical transitions in psychopathology: challenges and recommendations. Current Opinion in Psychology. 2021;41:51\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eKelso JAS, Scholz JP, Sch\u0026ouml;ner G. Nonequilibrium phase transitions in coordinated biological motion: critical fluctuations. Physics Letters A. 1986;118(6):279\u0026ndash;84. \u003c/li\u003e\n\u003cli\u003eHill Y, Den Hartigh RJR, Cox RFA, De Jonge P, Van Yperen NW. Predicting resilience losses in dyadic team performance. Nonlinear Dynamics, Psychology, and Life Sciences. 2020;24(3):327\u0026ndash;51. \u003c/li\u003e\n\u003cli\u003eImpellizzeri FM, Menaspa P, Coutts AJ, Kalkhoven J, Menasp MJ. Training Load and Its Role in Injury Prevention , Part I : Back to the Future. Journal of Athletic Training. 2020;55(9):885\u0026ndash;92. \u003c/li\u003e\n\u003cli\u003eSchiepek G, Schorb A, Sch\u0026ouml;ller H, Aichhorn W. Prediction of sports injuries by psychological process monitoring. Sports Psychiatry. 2023;2(4):135\u0026ndash;42. \u003c/li\u003e\n\u003cli\u003eDakos V, Carpenter SR, Brock WA, Ellison AM, Guttal V, Livina V, et al. Methods for Detecting Early Warnings of Critical Transitions in Time Series Illustrated Using Simulated Ecological Data. PLoS ONE. 2012;7(7). \u003c/li\u003e\n\u003cli\u003eWichers M, Smit AC, Snippe E. Early warning signals based on momentary affect dynamics can expose nearby transitions in depression: A confirmatory single-subject time-series study. Journal for Person-Oriented Research. 2020;6(1):1\u0026ndash;15. \u003c/li\u003e\n\u003cli\u003eChevance G, Baretta D, Heino M, Perski O, Olthof M, Klasnja P, et al. Characterizing and predicting person-specific, day-to-day, fluctuations in walking behavior. PLoS ONE. 2021;16(5):e0251659. \u003c/li\u003e\n\u003cli\u003eFuller CW, Ekstrand J, Junge A, Andersen TE, Bahr R, Dvorak J, et al. Consensus statement on injury definitions and data collection procedures in studies of football (soccer) injuries. Scandinavian Journal of Medicine and Science in Sports. 2006;16(2):83\u0026ndash;92. \u003c/li\u003e\n\u003cli\u003eMart\u0026iacute;nez-Silv\u0026aacute;n D, D\u0026iacute;az-Ocejo J, Murray A. Predictive indicators of overuse injuries in adolescent endurance athletes. International Journal of Sports Physiology and Performance. 2017;12:153\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eSoligard T, Schwellnus M, Alonso JM. How much is too much? (Part 1) nternational Olympic Committee consensus statement on load in sport and risk of injury. British Journal of Sports Medicine. 2016;50(17):1042. \u003c/li\u003e\n\u003cli\u003eKalkhoven JT, Watsford ML, Coutts AJ, Edwards WB, Impellizzeri FM. Training Load and Injury : Causal Pathways and Future Directions. Sports Medicine. 2021;51(6):1137\u0026ndash;50. \u003c/li\u003e\n\u003cli\u003eHalson SL. Monitoring Training Load to Understand Fatigue in Athletes. Sports Medicine. 2014;44(Suppl 2):139\u0026ndash;47. \u003c/li\u003e\n\u003cli\u003eDen Hartigh RJR, Meerhoff LRA, Van Yperen NW, Neumann ND, Brauers JJ, Frencken WGP, et al. Resilience in Sports: A Multidisciplinary, Dynamic, and Personalized Perspective. International Review of Sport and Exercise Psychology. 2022; \u003c/li\u003e\n\u003cli\u003eIvarsson A, Johnson U, Podlog L. Psychological Predictors of Injury Occurrence : A Prospective Investigation of Psychological Predictors of Injury Occurrence : A Prospective Investigation of Professional Swedish Soccer Players. Journal of Sport Rehabilitation. 2013;22:19\u0026ndash;26. \u003c/li\u003e\n\u003cli\u003ede Leeuw AW, van der Zwaard S, van Baar R, Knobbe A. Personalized machine learning approach to injury monitoring in elite volleyball players. European Journal of Sport Science. 2022;22(4):511\u0026ndash;20. \u003c/li\u003e\n\u003cli\u003eJones CM, Griffiths PC, Mellalieu SD. Training Load and Fatigue Marker Associations with Injury and Illness: A Systematic Review of Longitudinal Studies. Sports Medicine. 2017;47(5):943\u0026ndash;74. \u003c/li\u003e\n\u003cli\u003eZadeh A, Taylor D, Bertsos M, Tillman T, Nosoudi N, Bruce S. Predicting Sports Injuries with Wearable Technology and Data Analysis. Information Systems Frontiers. 2021;23:1023\u0026ndash;37. \u003c/li\u003e\n\u003cli\u003eLathlean TJH, Gastin PB, Newstead S V., Finch CF. Player wellness (soreness and stress) and injury in elite junior Australian football players over 1 season. International Journal of Sports Physiology and Performance. 2020;15(10):1422\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eFletcher D, Sarkar M. A grounded theory of psychological resilience in Olympic champions. Psychology of Sport and Exercise. 2012;13(5):669\u0026ndash;78. \u003c/li\u003e\n\u003cli\u003eGalli N, Vealey RS. \u0026ldquo;Bouncing back\u0026rdquo; from adversity: Athletes\u0026rsquo; experiences of resilience. Sport Psychologist. 2008;22(3):316\u0026ndash;35. \u003c/li\u003e\n\u003cli\u003eKleinert J. Mood states and perceived physical states as short term predictors of sport injuries: Two prospective studies. International Journal of Sport and Exercise Psychology. 2007;5(4):340\u0026ndash;51. \u003c/li\u003e\n\u003cli\u003eWiese-Bjornstal DM. Psychological predictors and consequences of injuries in sport settings. In: APA handbook of sport and exercise psychology. 2019. p. 699\u0026ndash;725. \u003c/li\u003e\n\u003cli\u003eGalambos SA, Terry PC, Moyle GM, Locke SA. Psychological predictors of injury among elite athletes. British Journal of Sports Medicine. 2005;39(6):351\u0026ndash;4. \u003c/li\u003e\n\u003cli\u003eEhrmann FE, Duncan CS, Sindhusake D, Franzsen WN, Greene DA. GPS And Injury Prevention In Professional Soccer. Journal of Strenght and Conditioning Research. 2016;2:360\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eCross MJ, Williams S, Trewartha G, Kemp SPT, Stokes KA. The influence of in-season training loads on injury risk in professional rugby union. International Journal of Sports Physiology and Performance. 2016;11(3):350\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eBourdon PC, Cardinale M, Murray A, Gastin P, Kellmann M, Varley MC, et al. Monitoring athlete training loads: Consensus statement. International Journal of Sports Physiology and Performance. 2017;12(s2):161\u0026ndash;70. \u003c/li\u003e\n\u003cli\u003eAiello F, Di C, Fanchini M, Impellizzeri FM, Mccall A, Sharp C, et al. Do non-contact injuries occur during high-speed running in elite football ? Preliminary results from a novel GPS and video-based method. Journal of Science and Medicine in Sport. 2023;26(9):465\u0026ndash;70. \u003c/li\u003e\n\u003cli\u003eGabbett T. The development and application of an injury prediction model for noncontact, soft-tissue injuries in elite collision sport athletes. Journal of Strength and Conditioning Research. 2010;24(10):2593\u0026ndash;603. \u003c/li\u003e\n\u003cli\u003eVan Der Does HTD, Brink MS, Otter RTA, Visscher C, Marie Lemmink KAP. Injury risk is increased by changes in perceived recovery of team sport players. Clinical Journal of Sport Medicine. 2017;27(1):46\u0026ndash;51. \u003c/li\u003e\n\u003cli\u003eJohnson U, Ivarsson A. Psychosocial Factors and Sport Injuries: Prediction, Prevention and Future Research Directions. Current Opinion in Psychology. 2017;(16):89\u0026ndash;92. \u003c/li\u003e\n\u003cli\u003eKellmann M, Bertollo M, Bosquet L, Brink M, Coutts AJ, Duffield R, et al. Recovery and performance in sport: Consensus statement. International Journal of Sports Physiology and Performance. 2018;13(2):240\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eImpellizzeri FM, Rampinini E, Coutts AJ, Sassi A, Marcora SM. Use of RPE-based training load in soccer. Medicine and Science in Sports and Exercise. 2004;36(6):1042\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eFoster C, Florhaug JA, Franklin J, Gottschall L, Hrovatin LA, Parker S, et al. A new approach to monitoring exercise training. Journal of Strenght and Conditioning Research. 2001;15(1):109\u0026ndash;15. \u003c/li\u003e\n\u003cli\u003eFoster C. Monitoring training in athletes with reference to overtraining syndrome. Medicine \u0026amp; Science in Sports \u0026amp; Exercise. 1998;30(7):1164\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eKentt\u0026auml; G, Hassm\u0026eacute;n P. Overtraining and recovery. Sports Medicine. 1998;26(1):1\u0026ndash;16. \u003c/li\u003e\n\u003cli\u003eBrink MS, Visscher C, Arends S, Zwerver J, Post WJ, Lemmink KAPM. Monitoring stress and recovery: new insights for the prevention of injuries and illnesses in elite youth soccer players. British Journal of Sports Medicine. 2010;44(11):809\u0026ndash;15. \u003c/li\u003e\n\u003cli\u003eBandura A. Guide for constructing self-efficacy scales. In: Urdan T, Pajares F, editors. Self-Efficacy Beliefs of Adolescents. Greenwich: Information Age Publishing; 2006. p. 307\u0026ndash;37. \u003c/li\u003e\n\u003cli\u003eBarte JCM, Nieuwenhuys A, Geurts SAE, Kompier MAJ. Motivation counteracts fatigue-induced performance decrements in soccer passing performance. Journal of Sports Sciences. 2019;37(10):1189\u0026ndash;96. \u003c/li\u003e\n\u003cli\u003eCohen AB, Tenenbaum G, English RW. Emotions and golf performance: An IZOF-based applied sport psychology case study. Behavior Modification. 2006;30(3):259\u0026ndash;80. \u003c/li\u003e\n\u003cli\u003eBorg GAV. Psychophysical bases of perceived exertion. Medicine \u0026amp; Science in Sports \u0026amp; Exercise. 1982;14(5):377\u0026ndash;81. \u003c/li\u003e\n\u003cli\u003eBrink MS, Nederhof E, Visscher C, Schmikili SL, Lemmink KAPM. Monitoring Load, Recovery, and Performance in Young Elite Soccer Players. Journal of Strenght and Conditioning Research. 2010;24(3):597\u0026ndash;603. \u003c/li\u003e\n\u003cli\u003eVan Yperen NW, Jonker L, Verbeek J. Predicting Dropout From Organized Football: A Prospective 4-Year Study Among Adolescent and Young Adult Football Players. Frontiers in Sports and Active Living. 2022;3:752884. \u003c/li\u003e\n\u003cli\u003eR Core Team. R: A language and environment for statistical computing (Version 4.3.1) [Software] [Internet]. Vienna, Austria: R Foundation for Statistical Computing; 2023. Available from: https://www.r-project.org/.\u003c/li\u003e\n\u003cli\u003evan Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained equations in R. Journal of Statistical Software. 2011;45(3):1\u0026ndash;67. \u003c/li\u003e\n\u003cli\u003eHasselman F. casnet: A toolbox for studying Complex Adaptive Systems and NETworks. R package version 01. 2018;3. \u003c/li\u003e\n\u003cli\u003eRuddy JD, Cormack SJ, Whiteley R, Williams MD, Timmins RG, Opar DA, et al. Modeling the Risk of Team Sport Injuries : A Narrative Review of Different Statistical Approaches. Frontiers in Physiology. 2019;10(413742). \u003c/li\u003e\n\u003cli\u003eFlach PA, Kull M. Precision-Recall-Gain Curves : PR Analysis Done Right. In: Advances in neural information processing systems. 2015. \u003c/li\u003e\n\u003cli\u003eNeumann ND, Van Yperen NW, Brauers JJ, Frencken W, Brink MS, Lemmink KAMP, et al. Nonergodicity in Load and Recovery: Group Results Do Not Generalize to Individuals. International Journal of Sports Physiology and Performance. 2021;17(3):391\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eNeumann ND, Van Yperen NW, Arens CR, Brauers JJ, Lemmink KA, Emerencia AC, et al. How Do Psychological and Physiological Performance Determinants Interact Within Individual Athletes? An Analytical Network Approach. International Journal of Sport and Exercise Psychology. 2024;1\u0026ndash;22. \u003c/li\u003e\n\u003cli\u003eHecksteden A, Schmartz GP, Egyptien Y, Aus der F\u0026uuml;nten K, Keller A, Meyer T. Forecasting football injuries by combining screening, monitoring and machine learning. Science and Medicine in Football. 2023;7(3):214\u0026ndash;28. \u003c/li\u003e\n\u003cli\u003eRobertson S, Bartlett JD, Gastin PB. Red, Amber, or Green ? Athlete Monitoring in Team Sport: The Need for Decision-Support Systems. International Journal of Sports Physiology and Performance. 2017;12(s2):73\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eThornton HR, Delaney JA, Duthie GM, Dascombe BJ. Developing Athlete Monitoring Systems in Team Steports : Data Analysis and Visualization. International Journal of Sports Physiology and Performance. 2019;14(6):698\u0026ndash;705. \u003c/li\u003e\n\u003cli\u003eHecksteden A, Keller N, Zhang G, Meyer T, Hauser T. Why Humble Farmers May in Fact Grow Bigger Potatoes : A Call for Street ‑ Smart Decision ‑ Making in Sport. Sports Medicine - Open. 2023;9(1). \u003c/li\u003e\n\u003cli\u003eOlthof M, Hasselman F, Lichtwarck-Aschoff A. Complexity In Psychological Self-Ratings: Implications for research and practice. BMC Medicine. 2020;1\u0026ndash;16. \u003c/li\u003e\n\u003cli\u003eTenan MS. Missing Data in Sport Science : A Didactic Example Using Wearables in American Football. Sports Medicine. 2023;53(6):1109\u0026ndash;16. \u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e Previous research has demonstrated that self-reports and sensor data reveal complex dynamics [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e]. Therefore, we can assume to actually capture and study them under these methods.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e Missing data is inevitable in sports monitoring. In consultation with the football club, we decided that a maximum of 20% missingness per variable is still acceptable for the analysis (this refers to roughly one day per week on average). Data was mainly missing, because players forgot to fill out the questionnaire or the sensors did not work. However, the analysis showed that the data was not missing completely at random, which is why we performed multiple imputation (R package \u003cem\u003emice\u003c/em\u003e) on the data. Recently, it was shown that imputing missing data (preferably using multiple imputation methods) is important to avoid biases and to make valid data-driven decisions in sports [\u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e].\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The window size of seven data points refers to roughly one and a half weeks in the present context (note that the football club did not measure every day of the week but for every training or match day). Previous studies used the same (relatively short) window size [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. A validation study has shown that short window sizes still report sufficiently good outcomes [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e The analysis indicated that a seven-data point window before injuries was optimal since extending the window by an additional data point did not improve the predictive validity.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"sports-medicine-open","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"smoa","sideBox":"Learn more about [Sports Medicine-Open](http://sportsmedicine-open.springeropen.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/smoa/default.aspx","title":"Sports Medicine-Open","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Football, Complex Dynamic Systems, nonlinear time series analysis, injury prediction, dynamic complexity, process monitoring, multidisciplinarity, personalized approach, early warning signals, critical fluctuations","lastPublishedDoi":"10.21203/rs.3.rs-4429464/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4429464/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThere has been an increasing interest in the development and prevention of sports injuries from a complex dynamic systems perspective. From this perspective, injuries may occur following critical fluctuations in the psychophysiological state of an athlete. Our objective was to quantify these so-called Early Warning Signals (EWS) to determine their predictive validity for injuries. The sample consisted of 23 professional youth football (soccer) players. Self-reports of psychological and physiological factors as well as data from GPS sensors were gathered on every training and match day over two competitive seasons, which resulted in an average of 339 observations per player (range\u0026thinsp;=\u0026thinsp;155\u0026ndash;430). We calculated the Dynamic Complexity (DC) index of these data, representing a metric of critical fluctuations. Next, we used this EWS to predict injuries based on different mechanisms (traumatic and overuse) and duration.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eResults showed a significant peak of DC in 31% of the incurred injuries, regardless of mechanism and duration, in the seven data points (roughly one and a half weeks) before the injury. The warning signal exhibited a specificity of 94%, that is, correctly classifying non-injury instances. We followed up on this promising result with additional calculations to account for the naturally imbalanced data (fewer injuries than non-injuries). The relatively low F\u003csub\u003e1\u003c/sub\u003e we obtained (0.08) suggests that the model's overall ability to discriminate between injuries and non-injuries is rather poor, due to the high false positive rate.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eBy detecting critical fluctuations preceding one-third of the injuries, this study provided support for the complex systems theory of injuries. Furthermore, it suggests that increasing critical fluctuations may be seen as an EWS on which practitioners can intervene. Yet, the relatively high false positive rate on the entire data set, including periods without injuries, suggests critical fluctuations may also precede transitions to other (e.g., stronger) states. Future research should therefore dig deeper into the meaning of critical fluctuations in the psychophysiological states of athletes.\u003c/p\u003e","manuscriptTitle":"Critical Fluctuations as an Early Warning Signal of Sports Injuries? Applying the Complex Dynamic Systems Toolbox to Football Monitoring Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-03 14:00:26","doi":"10.21203/rs.3.rs-4429464/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2024-05-23T07:17:01+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-05-22T08:23:07+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Sports Medicine-Open","date":"2024-05-21T02:19:45+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-05-19T23:06:18+00:00","index":"","fulltext":""},{"type":"submitted","content":"Sports Medicine-Open","date":"2024-05-17T05:30:27+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"sports-medicine-open","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"smoa","sideBox":"Learn more about [Sports Medicine-Open](http://sportsmedicine-open.springeropen.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/smoa/default.aspx","title":"Sports Medicine-Open","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4f6070f3-c14d-4bf4-8965-1a5c012fd33d","owner":[],"postedDate":"June 3rd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-12-23T16:08:09+00:00","versionOfRecord":{"articleIdentity":"rs-4429464","link":"https://doi.org/10.1186/s40798-024-00787-5","journal":{"identity":"sports-medicine-open","isVorOnly":false,"title":"Sports Medicine-Open"},"publishedOn":"2024-12-16 15:58:09","publishedOnDateReadable":"December 16th, 2024"},"versionCreatedAt":"2024-06-03 14:00:26","video":"","vorDoi":"10.1186/s40798-024-00787-5","vorDoiUrl":"https://doi.org/10.1186/s40798-024-00787-5","workflowStages":[]},"version":"v1","identity":"rs-4429464","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4429464","identity":"rs-4429464","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}