Background The ability to predict injury is difficult. Prior injury is the only risk factor that has been reported consistently in multiple research studies. Convenient and easy to perform, physical performance tests (PPTs) have great allure as prognostic factors.

Methods 11 PPTs were issued to 359 participants over the course of three seasons of National Collegiate Athletic Association Division I athletic competition. Injuries were monitored and reported in a centralised university tracking system. Exploratory factor analysis was performed in order to group the PPTs into constructs. The relationship between injury and these PPT-based constructs and other known predictors of injury was explored using univariate and multivariate regression.

Results PPTs clustered into five constructs: (1) active motion, (2) power, (3) hip stability, (4) flexibility and (5) motor control. When these five were placed into a multiple regression equation along with known risk factors (age, body mass index (BMI), gender, excessive flexibility and past injury), hip stability and active motion were predicted injury. In addition, motor control predicted non-traumatic injury. Past injury did not predict injury in the multivariate model.

Summary In college athletes, hip stability, active motion and motor control as assessed through PPTs can be useful as part of preseason screening. These PPT-related constructs seem to have a mediating effect on the relationship between past injury and future injury. This study provides the rationale to test targeted interventions to address these limitations.

Clinical trial registration number NCT01804894.