This post explores the potential for screening tests to predict sports injury and how they fit into an injury risk minimisation framework.
In our athlete testing YouTube series so far, we’ve explored various tests commonly associated with injury, such as hamstring strength, hip and groin strength, calf capacity, and shoulder assessments. These tests evaluate factors like strength, asymmetries, and rate of force development.
While such measures are often linked to injury prevention, an influential paper by Professor Roald Bahr, published in the British Journal of Sports Medicine, challenges this notion. The paper, titled Why Screening Tests to Predict Injury Do Not Work, and Probably Never Will, raises critical questions about the predictive power of these tests. In this post, we delve into these findings and discuss if we're not using them for injury prediction, how do we apply them?
Screening Tests and Injury Prediction
In the BJSM paper, Professor Bahr outlines three key steps required to determine if a screening test can predict outcomes:
Conducting a Prospective Cohort Study – Identifying risk factors and cut-off values to separate high-risk and low-risk groups.
Validating Test Results – Confirming the accuracy of these risk groups across multiple cohorts.
Randomised Control Trials – Testing interventions based on the established risk groups.
These steps highlight the rigorous process required to establish predictive validity, which is rarely achieved in sports contexts. These are more commonly applied when trying to predict a disease outcome.
Sensitivity and Specificity: The Basics
A foundational concept in evaluating predictive tests is understanding sensitivity and specificity:
Sensitivity refers to a test’s ability to correctly identify individuals with a condition (minimising false negatives).
Specificity is the ability to correctly identify individuals without a condition (minimising false positives).
These two measures are often inversely related, making it challenging to achieve a test that is both highly sensitive and specific. For example, a COVID-19 rapid test with 80% sensitivity may miss two out of ten infected individuals, while a 90% specificity means one out of ten uninfected individuals might falsely test positive. This example is illustrated below by mdd.gov.hk.
Translating this principle to sports, screening tests often fail to distinguish clearly between those who will sustain an injury and those who won’t, due to overlapping data between injured and uninjured populations.
Case Study: Knee Abduction Moment and ACL Injuries
In Professor Bahr’s discussion, knee abduction moment data from female athletes demonstrates the limitations of screening in relation to injury prediction. Taking data from the study by Hewett and colleagues (2005), of 205 athletes screened, nine suffered ACL injuries during the season. Returning to the prospective dataset, where would we have set the threshold for high vs low risk?
In the video below I discuss how different cut-off values for the "high-risk" group influence the sensitivity and specificity of external knee abduction moment to predict ACL injury in this cohort.
Screening Test Purposes Beyond Prediction
The examples discussed underscore the complexity of injury causation, as factors interact dynamically rather than linearly. So, if screening tests of single risk factors cannot reliably predict injuries, how can they be used? A paper by Nicol van Dyk and colleagues in the Aspetar Journal highlights three key purposes:
Detecting Current Conditions – Identifying existing musculoskeletal issues that may not have been reported.
Establishing Performance Baselines – Monitoring athletes’ physical capacities over time.
Building Practitioner-Athlete Relationships – Strengthening rapport and encouraging athlete buy-in for interventions.
It's worth noting that Professor Bahr is also an author on that Aspetar article, entitled There Are Many Good Reasons to Screen Your Athletes, highlighted that he is not anti-screening but providing a critical narrative of what screening tests should, and should not, be used for!
Group-Level Interventions
Van Dyk’s work emphasises a group-level approach to injury risk management. For instance, reduced hamstring strength in footballers increases group-level injury risk. Interventions, such as targeted strength programmes, benefit the entire team, even if individual outcomes vary. This approach acknowledges the probabilistic nature of risk factors. They explain this with a fictional athlete, Johnny, on a soccer team as follows:
"Although Johnny has decreased eccentric hamstring strength that is associated with an increased risk of injury, he may not have an injury (and someone with normal hamstring strength may have an injury). But since Johnny is part of the group and the risk of the whole group is increased by having decreased eccentric hamstring strength, it is worthwhile improving the hamstring strength of the whole squad as some hamstring injuries will also occur in those with normal hamstring strength. Johnny might still have an injury, but we can improve his odds."
Framework for Injury Risk Management
I've previously mentioned the six stage operational framework for individualising injury risk management in sport from Mark Roe and colleagues (2017). In this paper, the authors advise approaching injury risk management in team sports according to three tiers:
Group-Level Interventions – Universal strategies like warm-ups or general strength programmes.
Cluster-Level Interventions – Tailored to subgroups based on position, screening results, or shared risk factors.
Individualised Interventions – Specific exercises targeting unique needs, such as pre-training isometric routines or balance circuits.
Our screening tests can guide our approach to these cluster and individualised interventions.
Balancing Metrics with Practical Application
While no single test or metric can definitively predict injury (at least, not right now!), screening tests remain valuable for guiding interventions. By combining group-level strategies with cluster and individualised approaches, practitioners can optimise athlete care and reduce injury risk, even without absolute predictive certainty.
This underscores that the complexity of injury causation demands a nuanced approach to athlete testing. Screening tests should not be discarded but used as tools for identifying trends, informing interventions, and fostering athlete engagement. As our understanding of injury mechanisms evolves, integrating data-driven strategies with practical experience will remain key to effective injury prevention and performance enhancement.
FAQs on Screening Tests and Injury Prediction
Can screening tests predict injuries with certainty?
No, screening tests cannot (yet) reliably predict individual injuries. They are more effective for identifying risk factors and guiding group- or individual-level interventions.
What is the primary benefit of screening tests?
Screening tests are valuable for detecting existing issues, establishing baselines, and fostering athlete-practitioner relationships. They also inform injury prevention strategies on a broader scale.
What do sensitivity and specificity mean in the context of injury prediction?
Sensitivity: The test’s ability to identify those at risk (minimising false negatives).
Specificity: The test’s ability to identify those not at risk (minimising false positives).
How should practitioners use screening results?
Practitioners should apply screening results to inform group interventions, cluster interventions for subgroups, and individualised strategies, focusing on reducing overall injury risk.
Comments