The purpose of the investigation was to develop a set of epidemiological and statistical models to aid in the interpretation of risk patterns in sports. The data used in the study were drawn from the National Athletic Injury/1llness Reporting System (NAIRS). From the thirty different sports for which NAIRS collects data, college football was chosen for the investigation because it has the largest team enrollment. There were an average of 49 college teams over the period 1975-1978 which represented over 17,000 athlete-years and 195 team seasons.

The data within football were limited to 592 significant game-related knee injuries which occurred over the four-year study period. The Null llypotheses to be tested were that the variables of player position, activity (blocking and tackling) and situation (rushing and passing) had no effect on the occurrance of these game-related significant knee injuries.

It was found that the knee injuries that occur in college football has the highest time-loss ratio of all types of traumatic lesions. Within the significant game-related knee injuries a relatively consistent frequency of occcurrence was revealed year by year. As these knee injuries were examined relative to position, activity and situation using the log linear model, interactions among the variables were established. Those which achieved statistical significance (p.05) were identified as areas for future study. Within this approach, the observed cell frequencies were adjusted to reflect exposure ratios for each variable. This technique provided estimates of risk for each cell in the multidemensional frequency table. The addition of the log linear approach coupled with the use of epidemiological techniques serves as an excellent methodology for sports data interpretation.

This study generally revealed that the position, activity and situation at the time of significant game-related knee injury are an Integral aspect of the occurrence. The circumstances which seemed associated with the highest risk were the offensive and defensive 11nemen Involved in the black and rushing plays outside the tackle. For the running back, the highest risk was being involved in the tackle on rushing plays outside tackle. The frequency of injury to the quarterback during the study period was relatively low. For the 26 quarterback injuries, all occurred while being involved in the tackle on rushing plays outside the tackle.

Finally it was noted that for these data, the severity of the injury, as indicated by both the need for surgery and time-loss, and the exact diagnosis of the significant injury could not be predicted from position, activity and situation.