Introduction

Sports injuries (SP-INJ) remain one of the most pressing challenges in professional football, with serious consequences for athletes’ health, team performance, and the financial sustainability of clubs and organisations1,2. Injuries impose significant medical and rehabilitation costs while also reducing player availability and career longevity3. Beyond the physical burden, they negatively affect psychological well-being and long-term quality of life4. Although decades of surveillance have consistently documented high SP-INJ rates in elite football5,6, the central issue is not simply recognising prevalence but developing predictive models that can transform monitoring data into evidence-based prevention strategies. Multiple factors contribute to SP-INJ, including body composition, training load, prior injury, biomechanics, playing surface, footwear, and quality of recovery7,8. Despite this breadth of knowledge, most studies evaluate these variables in isolation, limiting the ability to capture the complex interactions that shape SP-INJ outcomes. Wearable monitoring technologies, such as GPS and physiological telemetry devices, provide detailed workload data9,10; however, these are rarely analysed in conjunction with contextual or self-reported measures. As a result, existing approaches remain primarily descriptive and offer limited practical guidance for coaches, practitioners, and medical staff11. Three critical gaps emerge from the literature. First, most predictive studies rely on simple workload ratios such as the acute-to-chronic workload ratio (ACWR)12 or focus on descriptive associations without applying structural models capable of testing both direct and mediated pathways13. Second, the integration of ML with contextual and questionnaire-based data into a unified analytical model has received limited attention14. Third, the potential of data warehouse–supported monitoring systems, which centralize, curate, and analyze multi-source data for real-time decision-making, has not been systematically evaluated in football SP-INJ research15. Addressing these shortcomings requires methodological innovation and frameworks that can better inform applied practice.

This paper develops and evaluates a structural equation model using PLS-SEM to examine how multiple risk factors (RF) interact to influence SP-INJ outcomes in professional football. The analysis focuses on body mass index (BMI), mechanical load (ML), training-program structure (TP), and contextual RF. These were prioritised due to their measurability using wearable telemetry and their consistent association with musculoskeletal stress in prior work16,17. At the same time, it is acknowledged that other determinants, such as nutrition, sleep, recovery, biomechanics, and previous SP-INJ history, are equally important18. These factors were not included in the present dataset but are identified as essential considerations for future research. By focusing on BMI and ML while recognising broader influences, the study demonstrates how ML metrics can be integrated within a structural modelling framework to enhance SP-INJ assessment. The following questions guide the research:

  1. 1.

    To what extent do mechanical load (ML) and body mass index (BMI) predict sports-injury incidence (SP-INJ) in professional football?

  2. 2.

    Are training-program structure (TP), context-specific risk factors (RF), and warehouse-technology effectiveness (EAIS-supported WT) directly associated with SP-INJ, and in what directions?

  3. 3.

    Do musculoskeletal development (MD) and performance mediate the relationships from ML, BMI, TP, RF, and WT to SP-INJ?

  4. 4.

    What is the overall explanatory power of an integrated wearables- and warehouse-supported PLS-SEM model for SP-INJ, and does it enable actionable team-level risk stratification?

In line with these questions, the objectives of this study are:

  1. 1.

    Quantify the direct effects of ML and BMI on SP-INJ.

  2. 2.

    Estimate the directions and magnitudes of the direct effects of TP, RF, and WT (EAIS-supported) on SP-INJ.

  3. 3.

    Test the mediating roles of MD and performance in pathways from ML, BMI, TP, RF, and WT to SP-INJ.

  4. 4.

    Build and validate an integrated PLS-SEM model combining wearable-derived and contextual variables, and assess variance explained in SP-INJ for practical decision-making.

By addressing these objectives, the research contributes in three ways. First, it demonstrates how wearable telemetry and contextual measures can be integrated into a single framework for SP-INJ modeling. Second, it highlights the potential of WT-supported monitoring pipelines for scalable decision-making in professional sport. Third, it provides practical insights for designing evidence-based training and prevention strategies. The contribution lies not in reiterating the prevalence of injuries but in presenting an integrative modelling approach that advances both research and practice in professional football. Figure 1 illustrates the distribution of SP-INJ occurrence across different BMI categories among the professional football players included in the study, further highlighting the association between BMI and the risk of SP-INJ.

Fig. 1
figure 1

BMI vs SP-INJ occurrence. Source(s): figure by authors.

Full size image

Literature review and hypothesis development

Multiple studies have demonstrated a strong link between SP-INJ and BMI. High BMI athletes who participate in high-impact sports, such as football, are more susceptible to injuries because the strain they experience on their musculoskeletal system is significantly higher. Excessive BMI increases the incidence of joint injuries, with the knees and ankles being the areas most affected, as these body parts support the majority of the body’s weight during movement. Logerstedt, et al.8 stressed that in such individuals, the injuries are more severe due to the extra weight placed on their joints and muscles. Research conducted by Zumwalt19 also demonstrated that excess body weight intensifies the load on joints, thereby increasing the risk of both acute and chronic injuries. Accordingly, we hypothesize that individuals with higher BMI are more susceptible to SP-INJ.

Hypothesis 1

There is a significant relationship between BMI and the occurrence of sports injuries (SP-INJ).

ML is the forces exerted on the human body during physical activities. Athletes experience mechanical loading that varies from speed and strength to the impact stress they sustain during training and competition. Research indicates that athletes often undergo mechanical loading at excessive levels without sufficient rest, which can lead to injuries, especially in sports with a high physical exertion component, such as football. Orejel Bustos et al.20 Also reported the connection between overuse injuries that arise from mechanical overload and how muscles and bones could be stressed over time, causing the weakening of the body and consequently increasing the risk of bodily injuries. Also, similar studies by Afonso et al.21 Concluded that there is indeed a clear-cut connection between human biomechanics and musculoskeletal strain, which is usually a significant predictor of SP-INJ. Furthermore, the ML in SP-INJ plays a crucial role, as correct load management ensures an appropriate balance between performance and SP-INJ. This signifies that ML must be adequately balanced so that it can provide athletes with the speed needed in sports while also protecting them from SP-INJ.

Hypothesis 2

There is a significant relationship between mechanical loads (ML) and the occurrence of sports injuries (SP-INJ).

Latest technologies, such as the Zephyr BioHarness wearable, are revolutionising SP-INJ prevention by providing instantaneous data on athletes’ movements and physiological responses. This technology, in turn, enables the continuous monitoring of several critical variables, including ML, motion patterns, and overall physical performance, which contributes to the SP-INJ. Kute et al.22 showed that WT help to understand SP-INJ patterns by giving precise data to/from the field of sports analysis, by presenting symptoms at an earlier phase, which is helpful in terms of early treatment of the SP-INJ. Furthermore, Hamlin et al.23 discovered how fantasy WT help to have a deeper understanding of athletes’ performance, which can be used to adjust training loads so that injuries will not happen due to overexertion. Data collection plays a crucial role in developing long-term strategies for preventing injuries, not just in making real-time adjustments. On the other hand, the Electronic Athlete Information System (EAIS) would be powering these WT by efficiently managing resources related to SP-INJ prevention. EAIS can monitor investments made in WT, control expenses incurred for training facilities, and recommend the amount of financing needed for SP-INJ prevention programs. This coupling, alongside the real-time data subpoenaed by WT, is imperative as it lays the foundation for proper financial planning and resource management, thus giving rise to a comprehensive SP-INJ prevention system that is both data-driven and resource-efficient.

Hypothesis 3

Warehouse-technology effectiveness (EAIS-supported WT) has an adverse direct effect on sports-injury incidence (SP-INJ).

Training methods are highly critical in preparing athletes to withstand the physical challenges of their sport without incurring injuries. A proper TP that gradually provides mechanical leeway, with adequate rest intervals, is critical for preventing SP-INJ. González-Ravé et al.24 Assert that periodisation involves the systematic variation of exercise intensity to mitigate the risk of injuries resulting from excessive physical strain. Periodisation addresses the encumbrances of overloading athletes too soon to ensure they adapt through training progression. Reflexively, Den Hartigh et al.25 Recommended person-specific TP can help reduce the number of injuries suffered by each athlete. Given the importance of personalised and progressive TP in reducing the likelihood of SP-INJ, it is assumed that the comprehensive TP plays a crucial role in this regard.

Hypothesis 4

There is a significant relationship between well-structured TP and the reduction of sports injuries (SP-INJ).

It is not just BMI and biomechanical forces on the body; many other factors and conditions can also contribute to SP-INJ. Factors such as technique, surface types, shoes, and weather elements are part of this context26. It was found that faulty technique or inappropriately fitted shoes can be predictors of SP-INJ with sudden directional changes, such as American Football, and those with frequent high-impact landings, like basketball. The study also indicated that playing on loose or inappropriate surfaces would likely increase the risk of knee and muscle injuries. Similarly, Jauhiainen et al.27 Note that the number of RF varies by sport. However, they all play a role in the athletes’ predisposition to injuries. Walking through the effects of the physiological similarities of several RF with higher BMI and load factor will reveal a well-developed accident prevention strategy.

Hypothesis 5

Risk factors (RF) such as playing techniques, surfaces, and footwear significantly contribute to the occurrence of sports injuries (SP-INJ).

MD entails working out many different types of joints, veins, and bones to help prevent SP-INJ. Athletes undergoing a strength and conditioning program designed to develop musculoskeletal fitness are at little risk of sustained Gravitational discharge injuries. Orejel Bustos et al.20 suggest that strengthening the musculoskeletal system makes it less vulnerable to injuries, particularly those resulting from repetitive stresses or high ML. In line with John et al.28, athletes with stronger muscles and joints are better equipped to handle the physical demands of their sport, thereby reducing their risk of SP-INJ. As MD is the mediator through BMI, ML, TP, will affect the SP-INJ rate, the following hypothesis will be proposed:

Hypothesis 6

MD mediates the relationship between BMI, ML, TP, and SP-INJ.

MD mediates the relationship between BMI, ML, TP, and SP-INJ.Athletic performance is also an intricate juggling act of a player’s skill, strength, speed, and overall fitness, which explains the occurrence of SP-INJ. In a study conducted by Tirla et al.29. They explored how athletes with a higher level of physical performance are less likely to suffer from injuries, as they possess the physical fitness that enables their bodies to withstand the physical stresses exerted during performance. This study emphasised the importance of maintaining sufficient fitness levels, agility, and endurance to prevent injuries, particularly in high-intensity sports such as football. Additionally, research by Martin et al.30. Stated that athletes with higher performance metrics are exposed to a lower risk of overuse injuries as they possess a greater capacity for performance that is associated with less predisposition to repetitive trauma when they compete, which leads us to propose the mediator role of performance in SP-INJ as the critical role of athletic performance in mediating SP-INJ. Then, it is hypothesised that:

Hypothesis 7

Performance mediates the relationship between BMI, ML, TP, and SP-INJ.

SP-INJ represents the primary outcome of this study, defined as injuries sustained during participation in sports and physical activity. As the dependent variable, they provide the central measure against which all other relationships are evaluated. Several independent variables were included in the model. The BMI, calculated as body mass divided by height squared, is considered an essential predictor because an elevated BMI may increase musculoskeletal stress, which in turn increases the likelihood of SP-INJ. ML, another independent variable, refers to the external forces acting on athletes during training and competition. These loads encompass movement speed, power output, and impact forces, which can either heighten SP-INJ when excessive or strengthen resilience when appropriately managed. In this study, wearable monitoring devices, such as the Zephyr BioHarness, were employed to collect real-time data on these forces, providing a precise means of quantifying load exposure. TP was also included as an independent variable, representing structured exercise regimens designed to regulate intensity, duration, and frequency of exercise. The design of these programmes directly affects conditioning and SP-INJ susceptibility. Additionally, contextual RF such as technical execution, surface conditions, and footwear type were examined, as these elements have been consistently linked to the occurrence of injuries. Two mediating variables were proposed in the conceptual framework. Musculoskeletal development (MD) reflects the adaptation and strength of the musculoskeletal system and is hypothesised to mediate the relationship between the independent variables and SP-INJ. Similarly, athletic performance, which encompasses skills, endurance, and agility, is treated as a mediator because it represents the functional outcome of training and physical conditioning that influences SP-INJ risk. Together, MD and performance provide pathways through which BMI, ML, TP, and RF exert their effects on SP-INJ incidence. The conceptual framework integrates these variables to illustrate the direct and indirect relationships among them. It suggests that BMI, ML, TP, and RF influence SP-INJ both independently and through their impact on MD and performance.

The conceptual framework in Fig. 2 below illustrates the key variables and their relationships in the study. It provides a theoretical basis and guides research investigation. In the study context, the conceptual framework examines the factors that influence SP-INJ and their interconnections.

Fig. 2
figure 2

Conceptual framework linking BMI, ML, TP, RF, and WT (EAIS-supported) to MD, performance and, ultimately, to SP-INJ. Source(s): Figure by authors.

Full size image

Methodology

Ethical considerations

Ethical approval for this research was not required under the Measures for the Ethical Review of Biomedical Research Involving Humans (National Health Commission of the People’s Republic of China, 2023), as the procedures involved were non-invasive, anonymised, and carried out with informed consent. All methods complied with relevant international guidelines and regulations. Although the national exemption was sufficient for this project, future studies that involve multiple countries should obtain clearance from institutional review boards to ensure consistency with global publication standards31.

Measures and instrumentation

The research employed two complementary data sources. Physiological and workload variables were collected using the Zephyr BioHarness system, which recorded metrics such as heart rate, respiratory rate, and ML indices. These data were aggregated to create latent indicators of external workload, reflecting the physical demands placed on players32. In parallel, a structured three-part questionnaire was used. The first section captured demographic details, including age, playing experience, and competition level. The second section addressed anthropometric and training-related variables, including BMI, ML perception, and contextual RF. BMI was derived from self-reported height and weight. Although BMI is recognised as a limited indicator of body composition in professional athletes, as it does not distinguish between muscle and fat mass, it remains a widely used and standardised measure for large, international samples where advanced methods, such as DEXA or bioimpedance analysis, may not be feasible33. Its inclusion provided comparative value and complemented the telemetry data. The third section of the survey examined perceptions of SP-INJ-prevention strategies and the role of monitoring systems in managing player health. To capture the constructs of interest, a 5-point Likert scale was employed for items related to ML and SP-INJ. In contrast, a 7-point scale was used to evaluate the effectiveness of prevention strategies. These items were adapted from previously validated instruments34. Expert consultation and pilot testing with a subset of players ensured clarity and contextual relevance. Feedback led to the simplification of technical terminology, particularly concerning ML, and the separation of training- and competition-related items to minimise confusion, which is consistent with recommendations in survey methodology35. Reliability analysis produced coefficients ranging from 0.76 to 0.89, and convergent validity was supported by inter-item correlations above 0.70, confirming robust measurement quality36. Both reflective constructs, such as BMI and ML, and formative constructs, such as prevention strategies, were incorporated into the measurement model to capture the multidimensional nature of SP-INJ37.

Procedures to reduce bias

Several measures were adopted to reduce potential bias. Recall bias was addressed by restricting the recall windows to two weeks and by cross-referencing questionnaire responses with available training logs. The administration of questionnaires in person allowed respondents to clarify ambiguous items, thereby reducing misinterpretation, a common concern in SP-INJ surveys38. The inclusion of telemetry-based ML data further complemented self-reported measures, thereby enhancing the validity of the dataset.

Sample size and participants

Participants consisted of professional football players recruited from multiple international leagues. Out of 247 distributed questionnaires, 153 complete and usable responses were obtained. Sample adequacy was evaluated using the gamma-exponential method Álvarez Chaves, et al.39, which indicated a minimum threshold of 146, confirming that the final dataset exceeded the recommended requirement. Although modest in size, the sample was sufficient for PLS-SEM, which is well-suited for small to medium-sized datasets and models of moderate complexity40,41.

Data collection and analysis

Data collection took place between April 6, 2022, and November 11, 2023. The telemetry data provided continuous ML and physiological indicators, which were merged with questionnaire responses at the construct level. This integration enabled both objective and subjective predictors of SP-INJ to be analysed together. Data analysis was performed using PLS-SEM in WarpPLS 8.0. This method was selected due to its ability to evaluate complex relationships involving both reflective and formative constructs, its robustness under conditions of non-normal data distribution, and its emphasis on explaining variance in dependent variables42,43. A two-stage approach was employed, beginning with the assessment of measurement models to confirm reliability and validity, followed by the estimation of structural relationships. This enabled the evaluation of both direct and mediated pathways linking ML, BMI, MD, performance, and SP-INJ outcomes, thereby advancing a comprehensive understanding of SP-INJ mechanisms44. Table 1 shows the study constructs, indicators, and sources.

Table 1 Study constructs, indicators, and source Source(s): table by authors.
Full size table

Figure 3 illustrates how SP-INJ reduces as performance level increases. It displays the association between athletes’ performance levels and their corresponding SP-INJ risks, highlighting the protective role of higher physical performance in reducing the susceptibility to SP-INJ.

Fig. 3
figure 3

Performance level vs. SP-INJ risk. Source(s): Figure by authors.

Full size image

Figure 4 presents the methodological framework, illustrating the various stages of the study, including questionnaire design, sample size determination, data collection, and the analysis technique employed59.

Fig. 4
figure 4

Research-methodology flowchart with the WT (EAIS-supported) data pipeline. Source(s): Figure by authors.

Full size image

Table 2 provides an overview of the demographic characteristics of the professional football players who participated in the study. The sample represents a diverse group in age, playing position, and preferred foot, offering a broader perspective on the athlete population under investigation. Where available, information on previous SP-INJ history was also recorded to enrich the dataset. Including these demographic details enhances the contextual understanding of the participants and contributes to the transparency and generalizability of the study’s findings.

Table 2 Demographic characteristics of the soccer players. Source(s): table by authors.
Full size table

Results

The PLS-SEM analysis provided detailed insights into the relationships among body mass index (BMI), mechanical load (ML), training-program structure (TP), warehouse-technology effectiveness (EAIS-supported WT), risk factors (RF), musculoskeletal development (MD), performance (PERFOR), and sports-injury (SP-INJ) incidence in professional football players.

Measurement model evaluation

The measurement model demonstrated strong reliability and validity. Cronbach’s alpha values ranged from 0.78 to 0.91, exceeding the recommended threshold of 0.70. Composite reliability values exceeded 0.80, and average variance extracted values surpassed 0.50, confirming convergent validity36. Discriminant validity was established using the Fornell–Larcker criterion, as each construct shared more variance with its own items than with the items of other constructs, supporting the adequacy of the measurement model60. As shown in Fig. 5, the path coefficients for all constructs were above the recommended threshold of 0.70, with values for BMI (0.836), ML (0.821), EAIS-supported WT (0.784), TP (0.752), and RF (0.754). These results provide strong evidence of satisfactory convergent validity. To further confirm model robustness, variance inflation factor (VIF) values were examined. Following36 VIF values below 3.3 indicate no concern with multicollinearity; all constructs in this study satisfied this criterion. The significance of indicator weights was evaluated using t-values, with a cut-off of 1.96 representing the 95% confidence level. While a few indicators (e.g., BMI1, ML3, and WT-EAIS 2) reported t-values slightly below 1.96, they were retained based on theoretical justification, consistent with prior studies employing SEM-PLS in sports SP-INJ prediction61. Their outer loadings exceeded the 0.70 threshold, and previous literature has highlighted their relevance62. Retaining these items ensured appropriate representation of the underlying constructs. Figure 5 presents the measurement model, illustrating the tested relationships among the constructs.

Fig. 5
figure 5

Structural model relationships. BMI body mass index, ML mechanical load, WT EAIS-supported Warehouse Technology effectiveness, TP training-program structure, RF risk factors, MD musculoskeletal development, PERFOR performance, SP-INJ sports-injury incidence. Source(s): Figure by authors

Full size image

Before presenting the results in Table 3, the rationale for assessing the measurement model is briefly outlined. The analysis evaluated convergent validity, outer loadings, indicator weights, p-values, and VIF scores to ensure construct reliability and validity. All constructs, BMI, ML, WT, TP, and RF, demonstrated satisfactory convergent validity, with indicator loadings above the 0.70 threshold. VIF values were well below acceptable cutoffs. VIF values were well below the recommended cutoff of 3.3, confirming the absence of multicollinearity. A small number of indicators with borderline significance were retained due to their theoretical relevance and support from prior literature. Collectively, these results confirm the robustness of the measurement model and provide a strong foundation for the subsequent structural analysis. Table 3 presents the detailed findings.

Table 3 Measurement model results. Source(s): table by authors.
Full size table

The SEM-PLS analysis has been effective in capturing the link between BMI and SP-INJ in professional footballers. In Fig. 6 below, it is evident that a strong correlation exists between high BMI and the risk of trauma in the knee and ankle areas that are under mechanical stress. Individuals with a greater than average BMI also had more impaired joints and joint and muscular systems, making them more likely to experience acute or chronic injuries. On the other hand, these documents imply that professional sports coaching regarding BMI monitoring and management should be included in the intervention programs, considering that overweight athletes not only suffer performance failure but also increase the risk and susceptibility to injuries. Imposed measures aimed at caloric restriction and SP-INJ prevention could serve as a reason for athletes with a high BMI to suffer fewer injuries of this kind. The structural equation modelling results also emphasise the significant impact of BMI on SP-INJ among professional football players. This relationship is graphically depicted in Fig. 6.

Fig. 6
figure 6

Relationship between BMI and SP-INJ among professional football players. Source(s): Figure by authors

Full size image

Additionally, Fig. 7 illustrates the influence of ML levels on the frequency of injuries, reinforcing the importance of load management in SP-INJ prevention strategies.

Fig. 7
figure 7

Impact of ML on SP-INJ occurrence in football players: A SEM-PLS analysis. Source(s): Figure by authors.

Full size image

As shown in Fig. 7, the effect of ML on SP-INJ incidence is apparent: higher levels of mechanical stress are associated with an increased frequency of musculoskeletal strains and overuse injuries, particularly in high-impact areas of the body. These findings underscore the importance of load management in preventing injuries. They further suggest that progressively TP, tailored to the needs of individual athletes, can help mitigate the risk of overload. When combined with WT that enable real-time monitoring of mechanical stress, such strategies can reduce SP-INJ rates and support long-term performance and health.

Structural model outcomes

The structural model revealed that ML was the strongest positive predictor of SP-INJ incidence (β = 0.33, p < 0.001, f² = 0.17), indicating that higher external loads, as captured by wearable telemetry, substantially increase the risk of SP-INJ. This finding is consistent with previous studies linking sudden increases in training load to musculoskeletal injuries in elite athletes63. BMI was also positively associated with SP-INJ (β = 0.29, p = 0.003, f² = 0.09). Although BMI does not fully capture body composition in athletic populations, it reflects an additional mechanical burden on joints that contributes to SP-INJ risk, and remains a relevant comparative marker across diverse samples33. Protective effects were observed for EAIS-supported WT and TP, both of which were negatively associated with SP-INJ (WT: β = − 0.21, p = 0.010, f² = 0.08; TP: β = − 0.18, p = 0.014, f² = 0.06). These findings indicate that effective technology use and structured programming mitigate SP-INJ risk by supporting recovery and optimising load distribution5. In addition, RF was independently predictive of SP-INJ (β = 0.23, p = 0.007, f² = 0.10), underscoring the influence of surfaces, footwear, and technique on SP-INJ incidence. Together, the independent variables explained 61% of the variance in SP-INJ (R² = 0.61), representing a strong explanatory power for the model. Effect sizes were small to medium, consistent with64. Table 4 summarises these structural paths.

Table 4 Structural model results. Source(s): table by authors.
Full size table

The analysis revealed that BMI and ML were positively correlated with SP-INJ, while WT (EAIS-supported) and TP were negatively correlated; RF showed a positive correlation. These results support H1–H5. In addition, MD and Performance were significant negative predictors of SP-INJ, consistent with the proposed mediating roles and supporting Hypotheses 6 and 7. See Table 4 for the complete set of coefficients and test statistics.

Hypothesis testing

Table 5 presents the results of hypothesis testing. A t-value greater than 1.96 indicates statistical significance at the 95% confidence level. The results confirmed that all five direct-effect hypotheses (H1–H5) were supported: BMI and ML were positively associated with SP-INJ, WT (EAIS) and TP were negatively associated, and RF was positively associated. For the indirect effects, both MD and Performance acted as significant negative predictors of SP-INJ, providing evidence in support of the mediating hypotheses (H6 and H7). These mediators transmit part of the protective influence of training programs and technology, consistent with prior SP-INJ-prevention literature65. Overall, the findings provide robust support for the proposed model. The results highlight that both direct predictors and mediating processes shape SP-INJ risk. At the same time, the complexity of SP-INJ mechanisms suggests additional influences such as nutrition, sleep quality, and prior SP-INJ history, which were not captured in this study, may contribute to overall risk18.

Table 5 Results of hypothesis testing. Source(s): table by authors.
Full size table

Table 5 presents the results of hypothesis testing, including path coefficients (β), t-values, significance levels, confidence intervals (95% CI), effect sizes (f²), and decision outcomes. A t-value greater than 1.96 was considered statistically significant at the 95% confidence level. The results confirmed support for all seven hypotheses. Specifically, Hypothesis 1 was supported, as BMI showed a significant positive relationship with SP-INJ (β = 0.29, t = 3.12, p = 0.003, 95% CI = 0.102–0.478, f² = 0.09). Hypothesis 2 was also supported, with ML demonstrating a significant positive effect on SP-INJ (β = 0.33, t = 4.78, p < 0.001, 95% CI = 0.186–0.468, f² = 0.17) and for Hypothesis 3, WT (EAIS) exhibited a significant negative association with SP-INJ (β = − 0.21, t = 2.88, p = 0.010, 95% CI = − 0.372 to − 0.048, f² = 0.08), providing further support. Similarly, Hypothesis 4 was confirmed, as TP showed a significant negative relationship with SP-INJ (β = − 0.18, t = 2.56, p = 0.014, 95% CI = − 0.329 to − 0.031, f² = 0.06). Hypothesis 5 was also supported, with RF significantly and positively associated with SP-INJ (β = 0.23, t = 3.00, p = 0.007, 95% CI = 0.079–0.381, f² = 0.10). The mediating hypotheses were also validated. Hypothesis 6 (mediation via MD) and Hypothesis 7 (mediation via Performance) were both supported, as bootstrapped confidence intervals indicated significant indirect effects. These findings suggest that MD and performance not only act as protective mechanisms but also mediate the relationship between training factors and SP-INJ risk. Overall, the results provide robust evidence in support of the proposed model. Both direct predictors (BMI, ML, WT, TP, and RF) and mediating processes (MD and Performance) were confirmed as significant, highlighting the multifactorial nature of sports-related SP-INJ risk. The consistency across direct and indirect effects highlights the importance of considering both physiological load and adaptive mechanisms in SP-INJ prevention strategies.

Figure 8 illustrates the distribution of SP-INJ risk across player positions, providing further insights into how positional demands influence susceptibility to SP-INJ among professional football players.

Fig. 8
figure 8

SP-INJ risk by player position. Source(s): figure by authors

Full size image

Explained variance

The model accounted for 39% of the variance in SP-INJ outcomes, 42% of the variance in MD, and 31% of the variance in prevention strategies. These values indicate that the integrated model provides meaningful explanatory power compared with single-variable approaches. At the same time, the unexplained variance confirms the need to include other determinants, such as recovery quality, psychological stress, and historical SP-INJ data, in future research66.

Discussion

ML management remains central to both SP-INJ prevention and performance optimisation in professional sport. Internal workloads, assessed through physiological measures such as heart rate and ratings of perceived exertion, together with external workloads, captured through distance covered, sprint frequency, and exposure to ML, collectively define the demands placed on athletes. Evidence consistently shows that the acute-to-chronic workload ratio (ACWR) is a critical predictor of SP-INJ risk, as sharp increases in acute training load relative to an athlete’s established baseline are strongly associated with an elevated incidence of SP-INJ67. Incorporating systematic ML monitoring into training periodisation allows for a more precise balance between stress and recovery, promoting musculoskeletal adaptation while reducing SP-INJ risk68. The findings of this study reinforce this principle by highlighting the importance of tailoring load progression to individual physiological characteristics. The role of recovery is equally important in sustaining an athlete’s health. Previous research indicates that incomplete rehabilitation substantially increases the probability of re-SP-INJ and often limits post-injury performance capacity69. Comprehensive rehabilitation protocols, social support, and staged return-to-play strategies are therefore essential to secure safe reintegration. Nonetheless, even with structured protocols, some athletes do not fully regain pre-injury performance levels, highlighting the need for preventive measures that address long-term functional outcomes in addition to immediate recovery. Mechanisms such as perception–action coupling, which integrate sensory input with motor response, may be disrupted following SP-INJ and create heightened vulnerability to recurrence. Developing targeted interventions that enhance perceptual-motor control may help reduce the likelihood of repeated injuries while improving performance stability70.

The results also emphasise the contribution of wearable monitoring technologies in advancing SP-INJ prevention. These systems enable practitioners to track load and movement patterns in real-time, supporting the early identification of deviations from safe thresholds and facilitating timely adjustments to training71. Previous studies have demonstrated that predictive models based on telemetry data can forecast SP-INJ risk with increasing accuracy72. The incorporation of PLS-SEM further enhances these models by capturing both direct and mediated relationships among multiple variables. Within this research, BMI and ML emerged as the most significant predictors of SP-INJ risk, which aligns with earlier evidence that excess body mass relative to height and elevated training loads increase mechanical strain and vulnerability to SP-INJ73. Contextual factors, such as footwear and playing surface, did not significantly predict injuries in this dataset, although prior research suggests that these variables may exert effects under particular conditions74. At the same time, the analysis highlights that SP-INJ risk is multifactorial and cannot be attributed solely to BMI and ML. Additional influences, including sleep quality, nutritional adequacy, psychological stress, and prior SP-INJ history, have been widely identified as relevant determinants of athlete health18. The exclusion of these factors from the current dataset represents a limitation, but it also provides a direction for future investigations. Expanding the analytical framework to integrate such dimensions alongside wearable-derived measures may enhance predictive validity and generalisability.

Methodologically, the study confirmed the robustness of the measurement model, as factor loadings, composite reliability values, and average variance extracted all exceeded established thresholds36. Variance inflation factor scores confirmed the absence of multicollinearity, indicating that constructs such as BMI and ML made independent contributions to the prediction of SP-INJ. The structural analysis revealed that progressive load adaptation, combined with structured recovery strategies, is a practical approach to mitigating risk. These findings are consistent with longitudinal studies demonstrating that stable ML progression and carefully managed recovery significantly reduce the incidence of overuse injuries75. The practical implications of these results are clear. Data-driven monitoring frameworks can equip practitioners with actionable insights into athlete risk profiles. Periodized TP that progressively increases load, combined with systematic assessment of body composition and individualised recovery plans, provides a structured pathway to reducing SP-INJ incidence while maintaining performance. The study also holds relevance for the broader development of predictive analytics in sport. Although the analysis focused on professional football players, the integrative framework, which combines wearable monitoring with PLS-SEM modelling, can be adapted to other sports with appropriate validation. Nevertheless, caution is warranted when extending findings beyond football, as biomechanical and physiological demands differ across disciplines76.

In conclusion, the research underscores the combined value of wearable monitoring systems and advanced analytical approaches in predicting and mitigating SP-INJ risk. The integration of ML indicators, body composition metrics, and preventive strategies within a single predictive model enhances the capacity of sports science to transition from descriptive surveillance to actionable, evidence-based interventions. The study also acknowledges several limitations, including the reliance on BMI as a proxy for body composition, a relatively modest sample size, and the omission of important confounding factors. Future investigations should incorporate longitudinal data, broader physiological and behavioural measures, and machine learning approaches capable of modelling complex interactions within larger datasets. By addressing these areas, sports science can continue to develop increasingly precise and transferable models that support athlete well-being and performance longevity.

Theoretical contributions

This study advances the theoretical understanding of sports SP-INJ prevention by moving beyond the narrow perspective of earlier work, which typically analyzed isolated RF. By combining BMI, ML, and warehouse technology data within a structural equation modelling framework using SEM-PLS, the research demonstrates the value of a systems-oriented approach to SP-INJ prediction. The findings provide empirical evidence for integrating multiple interacting variables into conceptual models, thereby generating more accurate and holistic insights into SP-INJ risk. From a methodological perspective, the research highlights the capacity of WT to support large-scale, real-time monitoring of athlete ML and movement patterns. At the same time, SEM-PLS enables the examination of complex interrelationships, including both direct and indirect effects, between risk variables. Although this study employed a cross-sectional design, the framework established here provides a foundation for longitudinal research that could clarify temporal dynamics and causal pathways. Taken together, these contributions reinforce the theoretical basis for ML monitoring and body composition in SP-INJ prevention and encourage further exploration of WT within sports science.

Practical contributions

The practical implications of this research are significant. The identification of elevated BMI and ML as primary predictors of SP-INJ provides practitioners with clear targets for intervention. Monitoring body composition and implementing tailored conditioning and ML-management programmes can reduce the risk of SP-INJ associated with excessive mechanical stress. Similarly, structured ML management and recovery planning represent effective strategies to mitigate cumulative training demands. WT plays a central role in operationalising these findings. By providing continuous, objective data on ML and physiological status, these systems allow practitioners to detect early warning signs of elevated SP-INJ risk and implement timely training modifications. The integration of SEM-PLS enhances decision-making by clarifying how RF interact, thereby supporting evidence-based allocation of resources and the design of targeted prevention programmes. Although the research was conducted within the context of professional soccer, the broader applications extend to other sports. Practitioners can adapt the principles identified here to discipline-specific demands, ensuring prevention strategies are contextually appropriate. The translational value of combining WT with advanced statistical modelling lies in its potential to reduce SP-INJ rates, improve rehabilitation outcomes, and sustain performance across sporting domains.

Limitations and suggestions for future studies

Several limitations should be acknowledged. The use of BMI as a proxy for body composition constitutes a methodological constraint. Among professional athletes, BMI does not adequately distinguish between muscle mass and adipose tissue, which may lead to misclassification. Future research should employ more precise tools, such as dual-energy X-ray absorptiometry (DEXA), direct muscle mass measurement, or percentage body fat analysis, to strengthen the validity of SP-INJ prediction. The scope of this study was primarily restricted to ML, BMI, and monitored TP variables. Other important determinants of SP-INJ risk, including sleep quality, nutrition, hydration, psychological stress, genetic predisposition, and biomechanical alignment, were not examined. Additionally, unmonitored TP and the technical limitations of wearable devices may have influenced SP-INJ outcomes but were not captured in the dataset. Expanding future models to incorporate these physiological, behavioural, and environmental factors would provide a more comprehensive understanding of SP-INJ susceptibility. The dataset was confined to professional soccer players, which limits generalisability to other sports and athlete populations. To enhance external validity, future research should include athletes from different sporting disciplines and competitive levels, recognising that distinct physical demands and SP-INJ patterns exist. The cross-sectional nature of the research further restricts the ability to establish causality. Longitudinal designs, which follow athletes across multiple seasons, would allow for stronger conclusions regarding temporal and causal relationships. The reliance on self-reported questionnaires for injury-related data also introduces potential recall bias and subjective distortion. Future studies should incorporate objective measures, including biomechanical assessments, physiological monitoring, and clinical records, to improve both precision and reliability. Although BMI and ML emerged as primary predictors in this study, external influences such as playing surfaces, footwear, and technical execution should also be systematically examined. Incorporating these elements into predictive models would allow for more sport-specific prevention strategies. Long-term investigations are critical for evaluating the sustained effectiveness of WT and integrated prevention programs.

Conclusion

This research highlights the contributions of WT and advanced analytical methods, such as SEM-PLS, to the prediction and prevention of SP-INJ. Analysis of Zephyr BioHarness data confirmed that elevated ML and higher BMI are among the strongest predictors of SP-INJ in professional football players. The findings suggest that progressive increases in ML are essential for supporting MD and reducing SP-INJ risk. WT reinforce this process by enabling real-time monitoring and precise data collection, facilitating timely and individualised interventions to mitigate potential risks. Although these variables were shown to be significant, their influence is likely to vary across sports and among individual athletes. Future studies should therefore explore the long-term applicability of these technologies across multiple sporting contexts and integrate additional variables such as sleep, nutrition, hydration, and recovery practices to strengthen predictive models. Beyond predictive analytics, the application of wearable technologies and advanced modelling supports evidence-based, athlete-specific training and rehabilitation programmes. These approaches promote more effective SP-INJ management and contribute to sustained athletic performance. As training practices and competitive demands continue to evolve, the regular refinement of these technologies will remain essential for safeguarding athlete health. By combining WT with SEM-PLS modelling, this study provides a comprehensive framework that extends beyond earlier work focused on isolated RF. The findings contribute to the advancement of theoretical knowledge and support practitioners in reducing SP-INJ incidence while maintaining performance capacity.