Background A lack of available injury data on non-elite sports participants has hampered the development of informed preventative strategies for the broad base of sports participation. In community sports settings, sports trainers are well-placed to carry out injury surveillance but few studies have evaluated the validity of the data they record.

Objective To investigate the capture rate, completeness and agreement between sports trainers and other sources of injury data in community sport.

Design Validation study.

Setting Community Australian football clubs.

Participants Of the 316 Australian football players eligible from four community clubs (male and aged 18+ years), 44% (n=139) agreed to participate. Seventeen participants were lost to follow up over the season. At each club, an on-site sports trainer agreed to record players' injuries.

Main outcome measurements Throughout the football season (16–18 rounds) players reported their injuries to the club's sports trainer. To validate injury data recorded by sports trainers, we obtained self-reported injury data from players via short message service (SMS) and additionally from treating health-care practitioners. We reported the percentage of all injuries registered by sports trainers only, by players only (via SMS) and by both. For any injuries reported by sports trainers, players and health-care practitioners, we reported agreement between sources on a range of injury variables using the kappa (Κ) co-efficient.

Results In total, 210 discrete injuries were recorded, 59% by players via SMS only, 21% by sports trainers only and 21% by both. Kappa coefficients ranged from 0.32 (0.27–0.37) for date of return to football and 1.00 (1.00–1.00) for activity at time of injury.

Conclusions Injury data collected by sports trainers within community clubs is of adequate quality for providing an understanding of the profile of injuries. However, data are likely to underestimate true injury rates and therefore must be interpreted with caution.