Related Experiment Video
Updated: Aug 6, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Can Field-Based Screening Predict ACL Injury Risk in Women Footballers? External Validation of a Prediction Model
Yuri Lopes Lima1,2, Tyler Collings1,2, Michelle Hall3
1School of Allied Health, Sport and Social Work, Griffith University, Gold Coast, Australia.
Background:
Anterior cruciate ligament (ACL) rupture is a devastating injury that occurs 3 to 7 times more frequently in women footballers than in their male counterparts. Identifying players at elevated risk is critical for targeted injury prevention; however, no ACL injury prediction model has undergone external validation in women. Therefore, this study aimed to externally validate a field-based ACL injury prediction model in women footballers.
Hypothesis:
A previously developed field-based ACL injury prediction model would demonstrate good predictive performance in an independent cohort of women footballers.
Study Design:
Prospective cohort study.
Level Of Evidence:
Level 2.
Methods:
Women footballers (n = 320) completed preseason assessments of single-leg hop kinematics, countermovement jump (CMJ) kinetics, hip adductor/abductor strength, and self-reported injury history. Players were followed prospectively for 18 months for noncontact ACL injuries. Model performance was evaluated via discrimination (area under the curve [AUC]) and calibration (observed versus predicted risk). The model was updated subsequently using the combined development and validation cohorts (n = 642), incorporating the hip adductor/abductor strength ratio and ipsilateral trunk flexion angles. Internal validation was performed via bootstrapping.
Results:
In the external validation cohort, the original model demonstrated reduced discrimination (AUC, 0.67; 95% CI, 0.49-0.82) and poor calibration (intercept, -0.79; slope, 0.53), indicating risk overestimation. The updated 5-variable model improved discrimination (apparent AUC, 0.76; optimism-corrected AUC, 0.72; 95% CI, 0.61-0.82), and calibration intercept (-0.00), but the calibration slope (1.38) and instability across bootstrapped samples indicated imprecise individual risk estimates.
Conclusion:
The original ACL injury prediction model did not generalize well to an independent cohort of women footballers. Although model updating improved overall predictive performance, calibration remained suboptimal, and predictions were unstable. Field-based strength and biomechanical assessments may support group-level ACL injury risk stratification but are not yet suitable for precise individual risk prediction.
Clinical Relevance:
Field-based measures of strength and biomechanics can help identify groups of women footballers at elevated ACL injury risk, achieving 72% classification accuracy. These screening measures are practical and scalable, requiring <10 minutes per player, and assess modifiable factors that may inform targeted injury prevention strategies.