Diminishing Returns from Added Sensing in Athlete Injury Forecasting: A Two-Cohort Evaluation of Monitoring Tiers with Data-Integrity Audits

Zhuopeng Li

doi:10.54691/jxxb3k93

Authors

Zhuopeng Li The Experimental High School Attached to Beijing Normal University, Beijing, China

DOI:

https://doi.org/10.54691/jxxb3k93

Keywords:

Injury prediction, athlete monitoring, wearable sensors, data leakage, ascertainment bias, clinical prediction models, value of information.

Abstract

Background. Multimodal “digital athlete” systems assume that adding sensor streams improves injury forecasting, but the marginal value and data integrity of each layer are rarely tested. Objective. To quantify the predictive gain from successive monitoring tiers and audit failure modes that inflate performance. Methods. This retrospective prognostic-modelling study analysed SoccerMon (50 elite women footballers; 25,529 monitored player-days; 63 injury episodes) and a seven-year runner cohort (74 athletes; 42,766 athlete-days; 583 injury days). Five nested tiers (T0–T4) were ordered by instrumentation and reporting burden and compared within cohort. Penalised logistic regression and gradient boosting were assessed with athlete-grouped cross-validation, athlete-level bootstrap intervals, natural-rate calibration, and explicit alert budgets. Pre-modelling audits tested injury-report ascertainment and historical coverage. Results. SoccerMon injury incidence fell 11.9-fold at a September 2020 changepoint despite stable or rising monitoring. In runners, injury rows contained 10.8 recoverable prior days versus 25.8 for non-injury rows; coverage alone achieved AUROC 0.961, and reconstructed-history features yielded AUROC 0.988 versus 0.660 within the released window. Cohort A showed no generalisable forecasting (AUROC 0.32–0.48). In Cohort B, AUROC increased from 0.641 at T0 to 0.663 at T3; the full stack added ΔAUROC 0.043 for logistic regression and 0.019 for gradient boosting over exposure alone. Self-report carried 38.2% of attribution. At a 10% alert budget, the best model detected 19.6% of injuries with 38 alerts per detection. Conclusions. Added sensing showed diminishing returns, hardware-free self-report carried the most signal, and operational utility remained low. Ascertainment-stability and coverage-balance audits should precede performance claims.

Downloads

Download data is not yet available.

References

[1] Bourdon, P. C., Cardinale, M., Murray, A., Gastin, P., Kellmann, M., Varley, M. C., et al. (2017). Monitoring athlete training loads: Consensus statement. International Journal of Sports Physiology and Performance, 12, S2-161–S2-170. https://doi.org/10.1123/ijspp.2017-0208

[2] Halson, S. L. (2014). Monitoring training load to understand fatigue in athletes. Sports Medicine, 44, 139–147. https://doi.org/10.1007/s40279-014-0253-z

[3] Seshadri, D. R., Li, R. T., Voos, J. E., Rowbottom, J. R., Alfes, C. M., Zorman, C. A., et al. (2019). Wearable sensors for monitoring the physiological and biochemical profile of the athlete. npj Digital Medicine, 2. https://doi.org/10.1038/s41746-019-0150-9

[4] Corral-Acero, J., Margara, F., Marciniak, M., Rodero, C., Loncaric, F., Feng, Y., et al. (2020). The ‘digital twin’ to enable the vision of precision cardiology. European Heart Journal, 41, 4556–4564. https://doi.org/10.1093/eurheartj/ehaa159

[5] Zadeh, A., Taylor, D., Bertsos, M., Tillman, T., Nosoudi, N., & Bruce, S. (2021). Predicting sports injuries with wearable technology and data analysis. Information Systems Frontiers, 23, 1023–1037. https://doi.org/10.1007/s10796-020-10018-3

[6] Rossi, A., Pappalardo, L., Cintia, P., Iaia, F. M., Fernàndez, J., & Medina, D. (2018). Effective injury forecasting in soccer with GPS training data and machine learning. PLOS ONE, 13, e0201264. https://doi.org/10.1371/journal.pone.0201264

[7] Ruddy, J. D., Shield, A. J., Maniar, N., Williams, M. D., Duhig, S., Timmins, R. G., et al. (2018). Predictive modeling of hamstring strain injuries in elite Australian footballers. Medicine & Science in Sports & Exercise, 50, 906–914. https://doi.org/10.1249/mss.0000000000001527

[8] Luu, B. C., Wright, A. L., Haeberle, H. S., Karnuta, J. M., Schickendantz, M. S., Makhni, E. C., et al. (2020). Machine learning outperforms logistic regression analysis to predict next-season NHL player injury: An analysis of 2322 players from 2007 to 2017. Orthopaedic Journal of Sports Medicine, 8. https://doi.org/10.1177/2325967120953404

[9] Van Eetvelde, H., Mendonça, L. D., Ley, C., Seil, R., & Tischer, T. (2021). Machine learning methods in sport injury prediction and prevention: A systematic review. Journal of Experimental Orthopaedics, 8. https://doi.org/10.1186/s40634-021-00346-x

[10] Claudino, J. G., Capanema, D. d. O., de Souza, T. V., Serrão, J. C., Machado Pereira, A. C., & Nassis, G. P. (2019). Current approaches to the use of artificial intelligence for injury risk assessment and performance prediction in team sports: A systematic review. Sports Medicine – Open, 5. https://doi.org/10.1186/s40798-019-0202-3

[11] Wolff, R. F., Moons, K. G., Riley, R. D., Whiting, P. F., Westwood, M., Collins, G. S., et al. (2019). PROBAST: A tool to assess the risk of bias and applicability of prediction model studies. Annals of Internal Medicine, 170, 51–58. https://doi.org/10.7326/m18-1376

[12] Saito, T., & Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLOS ONE, 10, e0118432. https://doi.org/10.1371/journal.pone.0118432

[13] Van Calster, B., McLernon, D. J., van Smeden, M., Wynants, L., & Steyerberg, E. W. (2019). Calibration: The Achilles heel of predictive analytics. BMC Medicine, 17. https://doi.org/10.1186/s12916-019-1466-7

[14] Vickers, A. J., & Elkin, E. B. (2006). Decision curve analysis: A novel method for evaluating prediction models. Medical Decision Making, 26, 565–574. https://doi.org/10.1177/0272989x06295361

[15] Bahr, R. (2016). Why screening tests to predict injury do not work—and probably never will…: A critical review. British Journal of Sports Medicine, 50, 776–780. https://doi.org/10.1136/bjsports-2016-096256

[16] Bittencourt, N. F. N., Meeuwisse, W. H., Mendonça, L. D., Nettel-Aguirre, A., Ocarino, J. M., & Fonseca, S. T. (2016). 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, 50, 1309–1314. https://doi.org/10.1136/bjsports-2015-095850

[17] Hulme, A., & Finch, C. F. (2015). From monocausality to systems thinking: A complementary and alternative conceptual approach for better understanding the development and prevention of sports injury. Injury Epidemiology, 2. https://doi.org/10.1186/s40621-015-0064-1

[18] Gabbett, T. J. (2016). The training—injury prevention paradox: Should athletes be training smarter and harder? British Journal of Sports Medicine, 50, 273–280. https://doi.org/10.1136/bjsports-2015-095788

[19] Impellizzeri, F. M., Tenan, M. S., Kempton, T., Novak, A., & Coutts, A. J. (2020). Acute:Chronic workload ratio: Conceptual issues and fundamental pitfalls. International Journal of Sports Physiology and Performance, 15, 907–913. https://doi.org/10.1123/ijspp.2019-0864

[20] Kalkhoven, J. T., Watsford, M. L., Coutts, A. J., Edwards, W. B., & Impellizzeri, F. M. (2021). Training load and injury: Causal pathways and future directions. Sports Medicine, 51, 1137–1150. https://doi.org/10.1007/s40279-020-01413-6

[21] Midoglu, C., Kjæreng Winther, A., Boeker, M., Dahl Pettersen, S., Pedersen, S., Ragab, N., et al. (2024). A large-scale multivariate soccer athlete health, performance, and position monitoring dataset. Scientific Data, 11. https://doi.org/10.1038/s41597-024-03386-x

[22] Lövdal, S. S., Den Hartigh, R. J., & Azzopardi, G. (2021). Injury prediction in competitive runners with machine learning. International Journal of Sports Physiology and Performance, 16, 1522–1531. https://doi.org/10.1123/ijspp.2020-0518

[23] Teramoto, M., Cross, C. L., Cushman, D. M., Maak, T. G., Petron, D. J., & Willick, S. E. (2017). Game injuries in relation to game schedules in the National Basketball Association. Journal of Science and Medicine in Sport, 20, 230–235. https://doi.org/10.1016/j.jsams.2016.08.020

[24] Lewis, M. (2018). It’s a hard-knock life: Game load, fatigue, and injury risk in the National Basketball Association. Journal of Athletic Training, 53, 503–509. https://doi.org/10.4085/1062-6050-243-17

[25] Collins, G. S., Moons, K. G. M., Dhiman, P., Riley, R. D., Beam, A. L., Van Calster, B., et al. (2024). TRIPOD+AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ, 385, e078378. https://doi.org/10.1136/bmj-2023-078378

[26] Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785–794). https://doi.org/10.1145/2939672.2939785

[27] Foster, C. (1998). Monitoring training in athletes with reference to overtraining syndrome. Medicine & Science in Sports & Exercise, 30, 1164–1168. https://doi.org/10.1097/00005768-199807000-00023

[28] Steyerberg, E. W., Vickers, A. J., Cook, N. R., Gerds, T., Gonen, M., Obuchowski, N., et al. (2010). Assessing the performance of prediction models: A framework for traditional and novel measures. Epidemiology, 21, 128–138. https://doi.org/10.1097/ede.0b013e3181c30fb2

[29] Ekstrand, J., Hägglund, M., & Waldén, M. (2011). Injury incidence and injury patterns in professional football: The UEFA injury study. British Journal of Sports Medicine, 45, 553–558. https://doi.org/10.1136/bjsm.2009.060582

[30] Riley, R. D., Ensor, J., Snell, K. I. E., Harrell, F. E., Martin, G. P., Reitsma, J. B., et al. (2020). Calculating the sample size required for developing a clinical prediction model. BMJ, 368, m441. https://doi.org/10.1136/bmj.m441

[31] van Smeden, M., Moons, K. G., de Groot, J. A., Collins, G. S., Altman, D. G., Eijkemans, M. J., et al. (2019). Sample size for binary logistic prediction models: Beyond events per variable criteria. Statistical Methods in Medical Research, 28, 2455–2474. https://doi.org/10.1177/0962280218784726

[32] Saw, A. E., Main, L. C., & Gastin, P. B. (2016). Monitoring the athlete training response: Subjective self-reported measures trump commonly used objective measures: A systematic review. British Journal of Sports Medicine, 50, 281–291. https://doi.org/10.1136/bjsports-2015-094758

[33] Windt, J., & Gabbett, T. J. (2017). How do training and competition workloads relate to injury? The workload-injury aetiology model. British Journal of Sports Medicine, 51, 428–435. https://doi.org/10.1136/bjsports-2016-096040

[34] Fuller, C. W., Ekstrand, J., Junge, A., Andersen, T. E., Bahr, R., Dvorak, J., et al. (2006). Consensus statement on injury definitions and data collection procedures in studies of football (soccer) injuries. British Journal of Sports Medicine, 40, 193–201. https://doi.org/10.1136/bjsm.2005.025270

[35] Drew, M. K., & Finch, C. F. (2016). The relationship between training load and injury, illness and soreness: A systematic and literature review. Sports Medicine, 46, 861–883. https://doi.org/10.1007/s40279-015-0459-8

[36] Harrell, F. E., Lee, K. L., & Mark, D. B. (1996). Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine, 15, 361–387. https://doi.org/10.1002/(sici)1097-0258(19960229)15:4<361::aid-sim168>3.0.co;2-4