This study aims to present a new approach towards the analysis of intervention time-series studies in the context of sports-related injury data. We used Victoria-wide hospital admission injury data associated with the sport of Australian football during the period 2006 to 2013. To estimate the state-wide effect of an implemented exercise training intervention that aimed to reduce the number of football-related injuries, time-series analysis was performed using a generalised least square (GLS) method. We show how the GLS method can be used to evaluate the impact of the intervention. Trend and seasonal patterns time series were also assessed using the 'Seasonal and Trend decomposition using Loess' nonparametric seasonal decomposition procedure. The model identified a decreasing trend in the seasonally adjusted number of injuries after the implementation of the intervention in the hospital admission data. The seasonal decomposition plots also indicate strong seasonal patterns in the injury time series.
BACKGROUND: It is known that some people can, and do, sustain >1 injury over a playing season. However, there is currently little high-quality epidemiological evidence about the risk of, and relationships between, multiple and subsequent injuries. PURPOSE: To describe the subsequent injuries sustained by Australian Football League (AFL) players over 1 season, including their most common injury diagnoses. STUDY DESIGN: Cohort study; Level of evidence, 3. METHODS: Within-player linked injury data on all date-ordered match-loss injuries sustained by AFL players during 1 full season were obtained. The total number of injuries per player was determined, and in those with >1 injury, the Subsequent Injury Classification (SIC) model was used to code all subsequent injuries based on their Orchard Sports Injury Classification System (OSICS) codes and the dates of injury. RESULTS: There were 860 newly recorded injuries in 543 players; 247 players (45.5%) sustained >/=1 subsequent injuries after an earlier injury, with 317 subsequent injuries (36.9% of all injuries) recorded overall. A subsequent injury generally occurred to a different body region and was therefore superficially unrelated to an index injury. However, 32.2% of all subsequent injuries were related to a previous injury in the same season. Hamstring injuries were the most common subsequent injury. The mean time between injuries decreased with an increasing number of subsequent injuries. CONCLUSION: When relationships between injuries are taken into account, there is a high level of subsequent (and multiple) injuries leading to missed games in an elite athlete group.