BJSM profiles three-part AI project on athletics injuries

Pierre-Eddy Dandrieux outlined a three-study PhD on machine learning and athletics injury prevention in a British Journal of Sports Medicine article published Oct. 9.

Evidence grade: Claim (no study cited) for any conclusion that artificial intelligence is associated with fewer injuries. The article describes research aims, including injury burden measured as days with injury per 1,000 hours of athletics exposure during one season.

Dandrieux, affiliated with Université Jean-Monnet Saint-Etienne, Mines Saint-Etienne and MSH Medical School Hamburg, focused the project on athletics, meaning track and field. Machine learning refers to computer methods that learn patterns from data to make classifications or predictions.

“I explored the use of artificial intelligence and, in particular, machine learning in sports injury prevention, and specifically in athletics (track and field),” Dandrieux wrote in the BJSM article.

The first part of the PhD aimed to describe machine-learning methods used to build sports-injury prediction models. The second examined how elite athletics athletes, coaches and health professionals viewed injury prediction as a prevention strategy.

The third part analysed whether self-reported use of injury risk estimation feedback, abbreviated I-REF in the article, was associated with injury burden during a competitive athletics season. The excerpt does not provide sample size, participant demographics, model performance, or absolute results.

That absence matters for readers because prediction tools can sound more actionable than the evidence supports. A model that estimates risk is not the same as an intervention that has been shown in this excerpt to reduce injuries, missed training days or competition loss.

The article says injury prevention in athletics still needs more individualised strategies because injuries are complex and multifactorial. It does not establish that AI-based prediction changes decision-making, changes behaviour, or is associated with fewer injury days.

The PhD was funded by a scholarship from the University of Jean Monnet Saint-Etienne, with partial funding from Deutscher Akademischer Austauschdienst e.V. The article says software for data collection was developed jointly by Mines Saint-Etienne and the University of Jean-Monnet Saint-Etienne.

The funding sources did not influence the design, conduct, analysis, interpretation or presentation decision, according to the article. Dandrieux declared no competing interests, and BJSM listed the piece as externally peer reviewed and not commissioned.

For athletes, coaches and clinicians, the practical takeaway is limited: the article documents a research programme, not a validated tool for training decisions. The full paper is behind access controls in the source material provided.

Source: peer-reviewed journal article, British Journal of Sports Medicine, Oct. 9.

Atlas360 reports health research and does not provide medical advice.