Introduction: Sports injury surveillance studies relying on information technology sometimes report poorly on data quality control (2). The aim of this study was to present a data quality algorithm for sports injury epidemiology and to apply it in an analysis of running-related injury (RRI) incidence in novice and experienced runners. Method: This study consists in a 24-week follow-up of 301 recreational runners. They were asked to run at least once a week and to upload their training data for running (duration, surface, intensity, mileage) plus any other sport activity on an internet-based diary. They were also asked to self-report all injuries (RRI or not). To ascertain for high-quality data, three inclusion criteria were applied before statistical treatment: uploading compliance was conformed when the athletes status (training, at rest, injured, sick) was known for .75% of the weeks. Uploading delay was conformed when .75% of the sessions were uploaded within 15 days after the session was achieved. Self-assessed compliance evaluated monthly was conformed when the participant reported to have uploaded .75% of the actually performed sport sessions. In addition, two criteria assessing for the regularity of the running training were applied: the minimum average running training, evaluated throughout the observation period or until RRI occurrence, was set to 1 session/week, and there could be no period of 4 consecutive weeks with no running. Results: From the initial 301 logged-on runners, 100 did not comply with the 3 data quality criteria. Another 24 participants were excluded from the final analysis after applying the two running training criteria. The demographics were not different between excluded and included participants. From the 177 runners retained, 43 (27%) sustained a total of 60 RRI; 75% were due to overuse, and the most frequently injured location was the knee (27%). RRI incidence was 14.2 and 4.7 RRI/1000 running hours in novice and experienced runners, respectively. Novice runners were 3.0 times (95% CI: 1.75-5.18) more at risk of sustaining a RRI than experienced (Chi2 = 17.3, p < 0.001). Discussion: Applying the present quality algorithm reduces the bias caused by lack of reporting and recall in the presented results. However, a selection bias due to this algorithm cannot be excluded. The fact that novice runners are more at risk than experienced is an important finding considering the high prevalence of RRI (1).
© Copyright 2012 17th Annual Congress of the European College of Sport Science (ECSS), Bruges, 4. -7. July 2012. Published by Vrije Universiteit Brussel. All rights reserved.