Objectives: The quality of epidemiological injury data depends on the reliability of reporting to an injury surveillance system. Ascertaining whether all physicians/physiotherapists report the same information for the same injury case is of major interest to determine data validity. The aim of this study was therefore to analyse the data collection reliability through the analysis of the interrater reliability. Design: Cross-sectional survey. Methods: During the 2016 European Athletics Advanced Athletics Medicine Course in Amsterdam, all national medical teams were asked to complete seven virtual case reports on a standardised injury report form using the same definitions and classifications of injuries as the international athletics championships injury surveillance protocol. The completeness of data and the Fleiss kappa coefficients for the inter-rater reliability were calculated for: sex, age, event, circumstance, location, type, assumed cause and estimated time-loss. Results: Forty-one team physicians and physiotherapists of national medical teams participated in the study (response rate 89.1%). Data completeness was 96.9%. The Fleiss kappa coefficients were: almost perfect for sex (k = 1), injury location (k = 0.991), event (k = 0.953), circumstance (k = 0,942), and age = 0.870), moderate for type (k = 0.507), fair for assumed cause (k = 0.394), and poor for estimated time loss (k = 0.155). Conclusions: The injury surveillance system used during international athletics championships provided reliable data for "sex", "location", "event", "circumstance", and "age". More caution should be taken for "assumed cause" and "type", and even more for "estimated time-loss". This injury surveillance system displays satisfactory data quality (reliable data and high data completeness), and thus, can be recommended as tool to collect epidemiology information on injuries during international athletics championships. (C) 2018 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.