(D-123) Evaluating the relationship between youth football games and game-simulating practice drills

Youth football is popular sport for young athletes with over 2.8 million participants each year.[1] There is a growing concern for football safety [2] as participants have a greater risk of injury compared to other sports such as basketball, soccer, wrestling, and gymnastics.[3] Practice structure plays a direct role in the safety of players. The practice drills selected influence the volume and severity of head acceleration events (HAEs) experienced. [4] Certain practice drills, such as team scrimmage, skelly, and inside run, are intended to simulate game-like contact to prepare players for live game play. The objective of the following study was to compare the HAEs during live game play and game simulated practice drills.

Two seasons of game and practice data were collected from three youth football organizations (one 12U, two 13U) consisting of N=35 athletes across 44 player-seasons. Athletes were instrumented with a mouthpiece-based head acceleration sensor with a recording threshold set at 5g for the football season. [5] Peak resultant linear and rotational acceleration, and rotational velocity were quantified for HAEs. A camera was time synchronized to the mouthpiece sensors to collect video data of all sessions. The research team analyzed the film to verify the HAEs were real using a custom MATLAB R2022B app. All game HAEs from the following practice drills, intended to simulate game play, were used for data analysis: skelly (i.e., game pass play drill), inside run (i.e., game run play drill), and team scrimmage (i.e., full team game play drill). The number of impacts, median, and 95^th percentile linear acceleration and rotational acceleration were calculated in JMP Pro 15 and compared across drill categories and game play.

Across two seasons, 5,983 HAEs were classified as game (run/pass plays), special teams’ plays, inside run, skelly, and team scrimmage. 21 of these HAEs were excluded due to review of rotational frequency content. [6] 6,375 minutes of practice were conducted with 3,412 minutes in team scrimmage (53.5%), 296 minutes in inside run (4.6%,), and 243 minutes in skelly (3.8%). Games resulted in a total of 2,493 minutes. The median (95^th percentile) linear acceleration and rotational velocity and acceleration for game-simulated practice drills was 8.9 (23.2) g, 7.7 (15.5) rad/s, and 631 (1963) rad/s², respectfully. The median (95^th) for games was 9.9 (30.3) g, 8.5 (18.8) rad/s, and 697 (2424) rad/ s², respectfully. Figure 1 shows the linear accelerations for direct (head contact) and indirect (non-head contact plays) for each game-simulating practice drill and play type. Percentages of direct head impacts were 37.1% (game), 40.8% (special teams), 37.7% (inside run), 13.8% (skelly), and 33.7% (team scrimmage). Game HAEs were greater than those in game-simulated practice drills but comparable HAEs were measured in game-simulated practice drills. Special teams resulted in the greatest mean direct linear acceleration compared to game-simulated drills, and non-special teams’ game play. Among practice drills, skelly and inside had the highest and lowest mean linear acceleration for direct HAEs, respectively. All practice drill mean linear accelerations were similar to games for indirect HAEs.

The results demonstrate that game-simulating practice drills collect similar magnitude indirect HAEs to indirect HAEs collected in games. However, direct HAEs collected in games are higher on average than direct HAEs in practices. The lower mean linear acceleration in inside run may be due to the close contact involved in this drill, with the majority of the HAEs occurring on line (not allowing for maximal acceleration). The distribution of HAEs trending toward higher magnitudes in the special teams’ category compared to all other drills and game plays is consistent with previous literature. High event HAEs can be reduced in practice by running practice drills at slower speeds, incorporating thud drills (limiting contact to the ground), or by decreasing the closing distance.

This project was supported in part by the NSF REU Site (Award #1950281) in the Department of Biomedical Engineering at Wake Forest University School of Medicine and the NIH NICHD K25HD101686 grant.

[1] Powell JW et al., J Am Med Assoc., 282(10), 1999.

[2] Brenner JS et al., The American Academy of Pediatrics, 136(5), 2015.

[3] Peterson AR et al., Orthopedic Journal of Sports Medicine, 5(2) 2017.

[4] Holcomb TD et al., Journal of Applied Biomechanics, 39(4), 2023.

[5] Rich AM et al., Ann Biomed Eng, 47(10), 2019

[6] Marks ME, et al., Ann Biomed Eng, 50(11), 2022.