How does the situation before a tackle influence a tackler’s head placement in rugby union?: application of the decision tree analysis

Concussion data

The present study was a case–control study. Injury surveillance of a Japanese collegiate rugby union team was conducted between 2008 and 2015 with informed consent from the players.8 There was no patient or public involvement.

In all playing seasons, medically trained personnel from teams recorded injuries according to the consensus statement for injury surveillance.31 Furthermore, all players who sustained a head impact or who sustained a suspected concussion during matches were screened by a team doctor in accordance with the Concussion Guidance published by the World Rugby.32 All suspected or confirmed concussions were analysed in the present study.33

Tackle data

A tackle has been defined as any event in which one or more tacklers attempt to stop or impede the ball-carrier, whether or not the ball-carrier was brought to the ground.3 We used data from our previous study that collected tackles leading and not leading to concussions.8 These data included tackles from 34 collegiate rugby union team matches recorded between 2008 and 2015, in which the tacklers sustained concussions. We selected 538 tackles, 34 of which led to concussions (experimental group) and 504 did not result in any type of injury (control group). An a priori power analysis indicated that an estimated total sample size of 208 events, including both concussion and non-injury events, would be required to achieve a power (1−β) of 0.80 with a confidence level of 95% (α=0.05), while an OR of 1.50 is required to produce the smallest worthwhile effect.

Data analysis

We defined an incorrect head placement as satisfying either (1) when the tackler’s head was in front of the ball-carrier (figure 1A–C)21 or (2) when any tackles gave an impact force to the tackler’s head contacting from the diagonal, lateral or posterior direction (figure 1D–F). The first author (KS) coded all tackles described previously from seven factors, which configured the aspect of both the tackler and ball-carrier during the pre-contact phase based on both consensus statement on video analysis in rugby union and previous studies (table 1).8 19 34–36 Tackles were excluded when the placement of the tackler’s head at the time of contact could not be identified. According to our previous research, we obtained high inter-rater reliability between the two raters.37 The κ statistic result for inter-tester reliability was 0.85 (pre-contact), indicating excellent agreement.38

Statistical analysis

When we used the tackler’s head placement as the main outcome of the decision tree analysis, we needed to clarify whether it is also related to concussion in this cohort. Therefore, we calculated OR and 95% CIs from the results of the binomial logistic regression analysis. If the OR was >1.0, concussion was more likely to occur compared with correct head placement. We also recorded situations in pre-contact features that led to incorrect head placement using the decision tree within the rpart package39 using R V.3.5.3 (R Foundation for Statistical Computing, Vienna, Austria). The rpart package uses the classification and regression tree (CART) algorithm to build a tree model. In CART, some variables merge such that there are always two branches based on the Gini indexes.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Effects of incorrect head placement on the risk of concussion

First, we confirmed whether the tackler’s head placement was related to concussion by logistic regression analysis in this cohort, in order to use that binary variable as the main outcome of the decision tree analysis. The risk of concussion was significantly higher in tackles with incorrect head placement (OR 8.21; 95% CI 3.73 to 18.11; p<0.01; table 2). In addition, we assigned tackles with incorrect head placements to groups depending on when the tackler’s head was in front of the ball-carrier or when any tackles gave an impact force to the tackler’s head contacting from the diagonal, lateral or posterior direction. The risk of concussion increased when the tackler’s head was in front (OR 14.78; 95% CI 6.27 to 34.84; p<0.01; table 2), and when any tackles gave impact force to the tackler’s head contacting from the diagonal, lateral or posterior direction (OR 3.83; 95% CI 1.38 to 10.64; p=0.01; table 2) with the ball-carrier compared with the correct head placement.

Situations before contact leading to tackles with incorrect head placement

Figure 2 shows the features associated with a higher probability of tackling an incorrect head placement at the time of contact. The frequency of tackles with an incorrect head placement was 60.0% when a ball-carrier side-stepped, ran an arc or was tackled by others, and when the tackler’s head was angled downward.

The frequency of an incorrect head placement was also 60.0% when ball-carriers ran straight or diagonally towards the defence line, when they did not move, or moved at high or low speed; and when the tackler moved laterally, stood with both feet flat on the ground or was located obliquely to the ball-carrier.

Request permission

Evasive manoeuvres and speed of the ball-carrier were the same as the second set of features. Furthermore, the frequency of tackles with an incorrect head placement was 61.9% when the tacklers moved forward, moved at high or low speed, and were located in front of, next to or behind the ball-carrier.

Relationship between incorrect head placement and concussion

The risk of tackler-related concussion increased 8.2-fold with tackles with incorrect head placement, which concurred with some previous findings.7–9 21 In addition to the traditional definition of incorrect head placement, when the tackler’s head was in front of the ball-carrier, we also analysed an incorrect head placement when a tackler’s head was in a position of contact with other players without in front of the ball-carrier. As a result, the risk of concussion was significantly increased (OR 3.83) even after excluding tackles with the head in front of the ball-carrier. Rugby Ready, which is a website intended to raise awareness of good practice and help stakeholders manage the inherent risks of contact by the World Rugby, advises positioning the head behind or to one side, and never in front of the ball-carrier in all tackle situations.40 Our results provide support for this recommendation by demonstrating the combined effects of head placement of the tackler and direction of the tackle.

Situations leading to the placement of the tackler’s head at the time of contact

Evasive manoeuvres implemented by the ball-carrier significantly contributed to the head placement at the time of contact. The probability of tackles with incorrect head placement also increases when a ball-carrier tries to avoid the tackle by side-stepping and running an arc, or when the tackler’s head faces downward. Conversely, even if the evasive manoeuvre of the ball-carrier was the same, the frequency of the tackle with a correct head placement increased when the tackler faced toward the ball-carrier or tackler’s head was moving in an attempt to follow the ball-carrier. At a professional level, the characteristics ‘head up and forward/face up’ and ‘identify/track ball-carrier onto shoulder’ had a lower propensity to result in a head injury assessment for tacklers.9 Further, lowered tackler’s head increased the frequency of tackler’s primary anterior shoulder dislocation.16 Moreover, in terms of tackle performance, having the head up, looking forward and facing the ball-carrier before contact affected successful outcomes.19 20 Therefore, our results suggest that tacklers should keep their heads up to identify the movements of the ball-carrier, which might lead to tackling the head on the correct side at the moment of tackling and decrease the likelihood of tackler-related injuries and tackle failure.

In addition, when the ball-carrier ran straight to the defence line at high speed, the likelihood of tackling an incorrect head placement increased whether a tackler moved at high or low speeds. When both the tackler and the ball-carrier sprinted, the estimated contact time was shortened, which would interfere with the ability of the tacklers to control their own movements at the time of contact. Conversely, the risk of tackles with incorrect head placements decreased in the same situation mentioned previously, even when a tackler moved at a moderate speed. Therefore, these results suggest that controlling the speed of tacklers to match the movements of ball-carriers is essential. For example, shortening steps to control speed might help to prevent tacklers from placing incorrect head positions and improve tackle performance. ‘Shortening steps’ was significant in enabling successful tackles20 and decreasing the likelihood tackle-related injuries.2 14

In the present study, we suggested that the two aforementioned characteristics, ‘head up’ and ‘controlling the tackler’s speed’, decrease the probability of incorrect head placement during the tackle. We applied the decision tree analysis for clarifying the situations that led to tackler’s head placement. As a result, we identified the situations that led to both tackling correct head placement and incorrect head placement. We believe that analysis of the situations that decreased the risk of concussion might lead to better recommendations for the prevention of concussion.

Applying decision tree analysis

In the present study, we clarified the relationship between a tackler’s head placement (the main outcome) and the pre-contact situations of a tackler and the ball-carrier by decision tree analysis. We believe that the decision tree analysis is an approach that allows players, coaches and trainers to visually understand the relationship between the main outcome and multiple factors. Our results showed that the probability of tackles with correct head placement increased regardless of the direction of the tackle when a ball-carrier tries to avoid the tackle by side-stepping, running an arc, and when a tackler faced toward the ball-carrier or tackler’s head was moving in an attempt to follow the ball-carrier. When players, especially beginners, improve their interpersonal tackle technique by restricting the ball-carrier’s movement and keeping the tackler’s head up to check the ball-carrier’s movements, they will contribute to training under safe conditions with a lower probability of concussions.

Moreover, this is the first study to adapt decision tree analysis to video analysis of tackling scenes in rugby union. The results showed that a combination of pre-contact characteristics influenced the head placement of the tackler. Therefore, we believe that new insights can be identified by adapting decision tree analysis to other tackle-related injury mechanisms and tackle performance.

Limitations

The present study generated important information, and several limitations should be addressed. At first, our study analysed isolated characteristics in the pre-contact phase, which were not involved in any other players or the ground area occurred a tackle. As concussions occur after contact, our results do not directly influence the occurrence of concussions. Therefore, future studies need to determine the effects of concussion by combining the phases in which concussions occur with some of the situations identified in this study. Second, since our results were based on a single collegiate rugby union team, caution should be exercised when generalising our findings, even at the same playing level. Other contextual factors, such as match time and match status at the time of tackles, were not analysed. Tierney et al reported that most direct impacts to the heads of tacklers occur during the second half of matches.18 Therefore, the characteristics of tackles with incorrect head placements might differ significantly according to the amount of time spent in matches.