Methodology in football injury epidemiology

K. Kristenson, J. Bjørneboe, M. Waldén, T. E. Andersen, J. Ekstrand, M. Hägglund

Methodological considerations of football injury epidemiology have only scarcely been described. The aim of this study was to evaluate the inter-rater agreement in injury capture rate and injury categorization for data registered in two different prospective injury surveillance audits studying the same two Norwegian male professional football clubs for two consecutive seasons, 2008-2009. One audit used team-based exposure (TBE) recording and the other individual-based exposure (IBE). The number of injuries recorded and corresponding injury rates (injuries/1000 h exposure) were compared between audits. Cohen's kappa and prevalence-adjusted bias-adjusted kappa (PABAK) coefficients were calculated for injury variables. Of 323 injuries included, the IBE audit captured 318 (overall capture rate 98.5%, training 98.9%, match 97.8%) and the TBE audit 303 injuries (overall capture rate 93.8%, training 91.4%, match 97.1%). Agreement analysis showed kappa and PABAK coefficients regarded as almost perfect (> 0.81) for 8 of 9 injury variables, and substantial (ƙ 0.75) for the variable injury severity. In conclusion, the capture rate for training injuries was slightly higher with IBE recording, and inter-agreement in injury categorization was very high.