Youth Tackle Football Head-Impact Estimation by Players and Parents: Is the Perception the Reality?

Participants

Male youth tackle football players and parents from 4 teams in a single recreation department in the rural southern United States were invited to participate over the course of 2 football seasons. Descriptive information is presented in the Table. This study was approved by the Institutional Review Board of the University of Georgia. All players assented to be in the study, and at least 1 parent or guardian provided written informed consent.

Table Descriptive Information for Youth Tackle Football Players and Parents

View Fullscreen

Head-Impact Estimation

At the beginning of the football season, players and parents were invited to complete a head-impact estimation tool developed by our research team to capture player and parent head-impact estimations. The tool contained 10 items, as follows: 3 related to frequency (Q1–Q3), 3 related to severity (Q4–Q6), and 4 related to location (Q7–Q10). All items and the response options as viewed by the participants are shown in Figures 1 through 3. Items Q2 and Q3, Q5 and Q6, Q7 and Q8, and Q9 and Q10 were identical questions but split between practices and games. The estimation tool was generated by the lead author (J.D.S.) and last author (R.C.L.). Four external researchers with expertise in youth tackle football head-impact exposure reviewed the tool for content validity.

View Figure

• Download as Powerpoint
• Download Figure

Participants were provided the following operational definition of a head impact: “Getting hit in the head can either be when you hit someone else or when someone else hits you.” Paper forms were distributed at practices, but participants were allowed to complete the forms at home and return them on a subsequent day. Recruits were excluded if they did not complete at least 30% of the responses. Otherwise, missing responses remained blank and were excluded from the corresponding analyses. Final samples are reported in Figures 1 through 3.

The primary focus of this study was measuring head-impact estimation; however, we did include 2 questions regarding concussion-risk estimation. We asked players and parents: “Do you think [you] [your child] will get a concussion this season?” and “How many players on your [team] [child's team] do you think will be concussed this season?”

Head-Impact Outcomes

Smart Impact Monitors (SIM-G; Triax Technologies) were inserted in headbands or skullcaps and used to collect head-impact biomechanics data for all players. All parents had 1 child instrumented. The SIM-G sensor consists of a triaxial accelerometer and triaxial gyroscope for measuring linear and rotational acceleration, respectively. Previous researchers^12,13 have used the SIM-G sensor to record on-field head-impact biomechanics. The SIM-G validation studies have demonstrated mixed results regarding impact location, linear acceleration, and rotational acceleration validity.^14–16 However, the validity of SIM-G is similar to that of other on-field head-impact biomechanics systems,^17,18 and we chose to bin our outcomes into categories to account for some of the measurement variability. We selected the SIM-G system because our study included both helmeted and nonhelmeted participants and was part of a larger study in which head-impact exposures in youth tackle and flag football were examined.¹⁹

Transformation of the SIM-G linear acceleration data to the head center of gravity was described in an earlier investigation.^19,20

Procedures

Tackle football players were fitted with a specific Triax SIM-G sensor and headband or skullcap unique to each participant. All participants wore their assigned SIM-G throughout the season (full-padded practices, scrimmages, and games), with the device placed below the external occipital protuberance. A research team member was present to distribute and collect SIM-Gs before and after each session; ensure sustained proper placement and fit; and record each event's start and end times, water breaks, injury time-outs, any miscellaneous pauses in play, and if athletes removed or tampered with their devices.¹⁹

Data Reduction

Head-impact data went through a multistep cleaning process before being analyzed.¹⁹ We first removed all impacts that were deemed transient or invalid by the Triax proprietary filters. Researcher-recorded time stamps for start and end times, water breaks, injury time-outs, and any miscellaneous pauses in play were used to remove false-positive head impacts or impacts unrelated to youth tackle football. Head impacts were not verified on video. We also applied a high-frequency threshold (the sum of median impacts per session and third-quartile impacts per session) to remove likely fraudulent impacts due to frequent SIM-G removal or improper wearing. We flagged players whose impacts in a single session exceeded the high-frequency threshold and removed these data from the analysis dataset. Sessions were classified as described in a previous study.¹⁹ Players were excluded from practice and game analyses if they did not participate in at least 30% of all events.

Our goal was to compare the on-field head-impact outcomes with the head-impact estimations of players and parents. Thus, each player's on-field head-impact outcomes were converted to match a head-impact estimation response category as follows.

Frequency

Players and parents overestimated the head-impact frequency in practices but underestimated the frequency in games (Figure 1). We asked players and parents to compare (Q1) and quantify head-impact frequency during practices (Q2) and games (Q3). Most parents (66.7%) said that head impacts would be more frequent during practices, whereas only 33.3% of players actually sustained most of their head impacts in practices.²⁷ Interestingly, players more closely estimated their true practice head-impact frequency than parents did (Figure 1A and B). Through their actual on-field experiences, players may gain a better appreciation for the frequency of head impacts. Parents' estimations may be less accurate because they are often on the sideline observing the activity. It was outside the scope of our study to determine how players and parents formed their head-impact exposure estimations, but researchers have suggested that parents may base concussion estimation risks on their observations, their own previous experiences with sport, or media influences.²⁸ Future researchers should determine which sources inform layperson estimations of head-impact exposure. Twelve players reported participation in at least 1 previous season of tackle football, which may have served as a reference point when estimating head-impact exposure. Earlier authors²⁹ proposed that the Triax SIM-G system may overestimate head-impact frequency; however, we sought to compensate for this by using more stringent timestamp filtering than prior investigators have used, and our results were similar to those of other estimates^21,30 of head-impact frequency in youth tackle football. Furthermore, we have no reason to believe this would affect game-to-practice referenced comparisons (Q1).

Severity

Players and parents overestimated head-impact severity, particularly in games. Similar to previous studies,^19,31 our on-field data captured from players throughout the season indicated that head-impact severity did not differ between practices and games. Yet most players (78.6%, n = 11/14) and parents (63.2%, n = 12/19) estimated that games would yield more severe head impacts (Figure 2A). Players and parents may expect physicality and risk to be higher during competition, but the actual head-impact exposures were similar across event types. Media attention surrounding football often focuses on more severe collisions, specifically those that occurred during televised games. When penalties are involved, these severe collisions are often replayed several times, with close focus on the head-impact severity. Football games are televised, whereas practices mostly are not, which may lead players and parents to focus more on the risk during games.

We categorized head impacts using a cutoff of 40g, but no perfect operational definition of head-impact severity categorization exists. We chose this threshold because it has been widely used in youth tackle football, allowing for comparisons with other studies, and because it approximates high or very high mean principal strain values using finite element modeling to model a youth brain,^24,25 but other estimates of concussion risk indicate potentially higher cutoffs.^32,33 In recent work, researchers^32,34,35 have suggested that injury thresholds are not as simple as prescribing cutoffs across all athletes but are also based more on each player's head-impact exposure and relative head-impact severity. Our results are similar to those of Savino et al,³⁶ who found that perceived head-impact severity by youth and high school tackle football players was only weakly correlated with on-field head-impact kinematic outcomes. Nonetheless, they did not discern whether players over- or underestimated head-impact severity, only that players did not accurately estimate their head-impact severity in a single season.

Location

Most players and parents underestimated the number of head impacts to the top of the head, particularly during practices (Figure 3A). Among our sample, 68.8% (n = 11/16) of players and 52.2% (n = 12/23) of parents thought that none of the head impacts sustained would be to the top of the head. However, no single youth tackle football player completed the entire season without sustaining at least 1 impact to the top of the head. Most players had only a few (1%–19%) or fewer than half (20%–39%) of their head impacts to the top of the head. Impacts to the top of the head are considered dangerous because spinal loading transmitted through the top of the head while the cervical spine is in a flexed position may pose a catastrophic threat to the spinal cord.³⁷ Cervical forward flexion decreases the natural lordosis of the cervical spine such that the vertebral bodies are stacked on one another in a column formation, the cervical musculature is more relaxed, and the vertebrae more easily collapse and fracture, impinging the spinal cord.³⁷ Top-of-the-head impacts are between 2.4 and 8.5 times more severe than those at other locations³¹ and are more likely to result in loss of consciousness.³⁸

Players and parents estimated that most head impacts would occur to the front of the head (Figure 3C and D). However, our on-field data showed that side impacts were the most common, followed by front, top, and back impacts. In contrast, other authors^38,39 showed that front impacts were the most common but also that head impacts were considerably distributed across locations. Players and parents may think that most collisions occur with 2 opponents approaching face to face, resulting in a front impact. Yet the sport of football is much more dynamic than uniplanar collisions, and not all head impacts occur to the front of the head.

Concussion-Risk Estimation

No players or parents estimated that they or their child would sustain a concussion in the upcoming season. Previous researchers^9,40 demonstrated that approximately 3% to 5% of players on a youth tackle football team will sustain a concussion in a single season. It is plausible that our findings reveal a pattern of optimism bias among youth tackle football players and parents, which has been identified in college-aged football players.⁴¹ Players and parents who elect to participate in youth tackle football may hold a more optimistic view of the concussion risk. Research should be done to compare these estimations among sports with various concussion risks, such as flag football or soccer. It is also plausible that, with rates as low as 3% to 5%, our participants responded unanimously this way because the odds are more likely that players will not sustain a concussion in a single season.

In contrast to their responses regarding their risk or their child's risk, players and parents had less optimistic responses to how many players on their team would sustain a concussion. The studied teams carried rosters of approximately 25 players. Applying published concussion rates for youth tackle football^9,40 to a team of 20 to 25, we should expect 0.60 to 0.75 concussions per team per year (ie, 1.2 to 1.5 per every 2 seasons). Most players and parents estimated that ≥1 player on their team would sustain a concussion, which suggests a slight overestimation of injury risk. Our results are similar to but not quite as pronounced as those reported by Chrisman et al.⁷

Limitations

This study had limitations. No true criterion standard instrumentation exists for capturing on-field head-impact exposures. Some notable concerns with the Triax SIM-G device should be considered. All head-impact systems exhibit considerable variability in their ability to capture head-impact frequency, severity, and location. We sought to account for some of this variability by categorizing both on-field and estimated head-impact outcomes in response options. Earlier investigators²⁹ have also reported variability in head-impact frequency measures. We sought to reduce this limitation by recording and removing data captured during water breaks, injury time-outs, and any miscellaneous pauses. Furthermore, our conservative data-cleaning procedures were meant to exclude spurious impacts.¹⁹ In addition, the Traix SIM-G uses a threshold of 14g, whereas players and parents may have higher or lower thresholds for what they consider a head impact. The Triax SIM-G system relies on adult male anthropometrics to convert raw accelerometry data to the head center of mass and, therefore, may not be as accurate in youth athletes with smaller head sizes and masses. The head-impact estimation tool was administered at the beginning of the season. Player and parent accuracy may improve over the course of a season after participating in or observing a season's worth of head impacts, which would be an interesting focus for future examination. However, most players and parents had experience with tackle football before our study. Finally, our sample size was small, and this research question should be explored in larger samples.