An international team of researchers led by the Barça Innovation Hub (BIHUB), the FC Barcelona Medical Department, the Barcelona Institute for Global Health (ISGlobal), the University of Bonn, and Made of Genes has developed a new computational framework that represents a significant advance in injury prediction in elite women's football. Published in npj Digital Medicine, part of the Nature Portfolio, the study combines artificial intelligence, survival analysis, statistical calibration and decision theory to transform risk prediction into a useful tool for clinical and sporting practice.

The study addresses one of the main challenges in sports medicine: anticipating injuries before they occur. Although numerous artificial intelligence-based models have been developed in recent years, most have important limitations that make them difficult to apply in professional clubs. Rather than simply building a more accurate model, this work proposes a new methodological framework that addresses several of the main shortcomings identified in the scientific literature.

A new way to understand injury risk

The main innovation is to treat injury prediction as a survival analysis problem. This makes it possible to represent mathematically something practitioners already know: risk increases progressively as training or competition minutes accumulate. Unlike traditional classifiers, this behaviour is built into the model by design.

This approach is complemented by three further innovations:

•  Beta Calibration, which adjusts probabilities so that they accurately reflect the observed real-world risk;

•  the incorporation of decision theory, which adapts risk thresholds according to the sporting importance of each situation;

•  the introduction of Rest Benefit Certainty (RBC), which allows coaching staff to define the level of certainty they want before deciding whether a player should rest.

Together, these elements turn an abstract probability into a contextualized recommendation that helps professionals assess when it is more beneficial for a player to participate or rest.

Four seasons of data from the women's first team

The study was developed using one of the most comprehensive longitudinal datasets available in elite women's football: four consecutive seasons from the FC Barcelona women's first team (2019-2023), covering 34 players, nearly 14,000 daily observations and 83 non-contact musculoskeletal injuries. The data integrate GPS information, training load, minutes played, competitions, international duty and injury duration.

A much more reliable injury-risk detection system

The results show that survival analysis-based models consistently outperform the traditional machine-learning algorithms used to date for injury prediction. In particular, these models achieved better discrimination than conventional classifiers, substantially improving the identification of high-risk situations.

Statistical calibration also corrected the tendency of the models to underestimate actual risk, making the resulting probabilities much more reliable and interpretable for the medical team.

In temporal validation on a completely new season, the system produced positive gains in player availability, suggesting that this type of model can help reduce the impact of injuries by keeping more players available throughout the season.

Cumulative fatigue, the strongest predictor

The analysis shows that the most important predictor is the distance accumulated over the previous 21 days, followed by high-speed running intensity and same-day accelerations and decelerations. This result reinforces the idea that cumulative fatigue is the main mechanism associated with non-contact injury risk.

"Unlike previous approaches, this framework makes it possible to identify, rank and calibrate the factors associated with injury risk with statistical rigour, and to quantify the contribution of cumulative fatigue. Above all, it offers a new scientific perspective on a complex phenomenon," says Juan R. González, a researcher at ISGlobal.

"The methodological challenge was twofold: to represent how risk accumulates with exposure while also contextualising the resulting probabilities so that they can be interpreted according to the relevance of each situation. Incorporating decision theory makes it possible to adapt risk thresholds to each sporting context and turn predictions into a useful tool to support professional decision-making," explains Manuel Huth, first author, from the University of Bonn.

From research to the pitch and beyond

This research provides unique insight into the factors that explain injury risk, with external load as the starting point. The athlete's digital twin integrates this with internal load - how the body responds to effort, through biomarkers, physiological indicators and molecular data - and each athlete's genetic profile: an essential tool to support the technical and medical staff, personalise each athlete's monitoring and anticipate injuries. This is the line that Barça Innovation Hub is driving together with Made of Genes.

"This kind of research brings scientific rigour to factors that, until now, we often interpreted intuitively. The real value is to better understand how load and each athlete's biology relate to injury risk," explains Gil Rodas, physician at FC Barcelona.

"Advancing in this field requires research and methodology of the highest scientific rigour, like this work. As a data and artificial intelligence platform partner, at Made of Genes we bring research onto the pitch through anticipation and performance-enhancement models that integrate, within a digital twin, each athlete's external load, internal load and genetic profile," adds Berta Canal, data scientist at Made of Genes.

The potential reaches well beyond elite football: as wearable sensors are now widely used not only by professional athletes but also in amateur sports and physically demanding occupations, the same approach could support injury prevention far beyond the football pitch.

About the study

Title: Injury prediction in elite women's football: an integrative machine learning-based decision-support framework
Journal: npj Digital Medicine (Nature Portfolio)
Publication: Online, 8 July 2026
DOI: 10.1038/s41746-026-02937-3
Data: FC Barcelona women's first team · 4 seasons (2019/20-2022/23) · 34 players · 83 non-contact injuries
Authors: Manuel Huth, Berta Canal-Simón, Eva Ferrer, Gil Rodas, Xavier Yanguas, Jan Hasenauer and Juan R. González
Institutions: University of Bonn · ISGlobal · Made of Genes · FC Barcelona Medical Department and Barça Innovation Hub · Hospital Clínic-Sant Joan de Déu · Leitat · CIBERESP · Universitat Autònoma de Barcelona

More news here