Background International literature recognises that sports people who have sustained an injury have a high likelihood of a subsequent injury, either of the same type again or a different one. Taking into account the dependency of subsequent injuries within athletes is important, otherwise, results have greater precision than is warranted and there is possible biassing of results away from the null. Survival methods for recurrent events data can be applied to take into account the potential statistical relationships between subsequent injuries. Various survival analysis approaches exist and the selection of a correct model is important. The aim of this systematic review was to identify and evaluate the appropriate use of survival analysis methods in sports injury incidence studies.

Methods The review was performed and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement, comprising a traditional systematic search of literature and a hand search of key sports medicine journals. Descriptive data on the studies were extracted (e.g. aim, setting, follow up) along with detailed information on the reported survival analysis methods. The appropriateness of selected models, and their reporting, were examined in-depth.

Results There were 699 publications screened from which 123 were included for further review. Studies were mainly focused on the reported time to first injury (n = 53%) or the time to recover/return to play after injury (19%). Recurrent or subsequent injuries were only considered in 10% of papers. Graphical approaches (Kaplan-Meier curves) were most commonly used (51% of studies), along with simple univariate (30%) and Cox-regression techniques (88%).

Conclusions While a number of researchers have used survival methods for analysis of sports injury data, there is scope for improvement in the statistical approaches used and, specifcally, in the reporting of subsequent injuries.