Information about project titled 'Modelling relative training load and injury risk'
Modelling relative training load and injury risk
| Details about the project - category | Details about the project - value |
|---|---|
| Project status: | Published |
| Project manager: | Lena Kristin Bache-Mathiesen |
| Supervisor(s): | Morten Wang Fagerland, Thor Einar Andersen |
| Coworker(s): | Ben Clarsen, Karim Chamari, Torstein Dalen-Lorentsen |
Description
Background: Researchers strive to determine causes of sports injury, especially causes we can intervene on. One potential cause of injury is training load: the short- and long-term exposure to physical activity. Experts theorize that short-term, recent training load relative to long-term, past training load—known as relative training load—contributes to injury risk. Such theories have, due to statistical limitations, been difficult to test. Traditionally, analysts calculate a ratio between recent and past training loads, which cannot handle training loads of 0. Unless we improve the statistical tools, how relative training load affects injury risk will continue to be difficult to determine, and therefore, the potential of training load interventions will not be developed to full capability.
Aim: First,to demonstrate how modelling recent and past training loads separately, without calculating a ratio, is a potential approach to modelling relative training load and injury risk. Second, to see if recent and past training loads interact in their association with injury risk in football.
Methods: We ran a mixed model on two football datasets: Qatar Stars League (QSL, 1 465 players, 1 977 injuries), and Norwegian U-19 football (81 players, 60 injuries). The outcome was injury (yes/no). The independent variables were the recent and past training load, with an interaction between them. Recent training load was defined as the current day, while past training load was a cumulative effect of the previous 27 days (4 weeks). For the QSL model, training load was measured with the minutes in activity; for the Norwegian model, it was the minutes in activity multiplied by the player’s perceived intensity of the activity on a scale from 1 to 10 (sRPE).
Results: Both models demonstrated the lowest injury risk at medium levels of long-term, past training load, intermediate risk at high levels of past training load, and highest risk at low levels of past training load. This pattern changed for each level of short-term, recent training load, also demonstrating an interaction between recent and past training load.
Conclusion: Recent and past training load can be modelled separately to determine the relationship between relative training load and injury risk. Researchers should consider the potential for an interaction between recent and past training load.