The tackle is a frequently executed body contact skill in rugby union football and is the component of the game associated with the highest incidence of injury across all levels of competition (4,5,7,8,11,15,16). For several decades, high and spear tackles, which are illegal in the game, have also been observed to cause catastrophic cervical spine and spinal cord injury (12,17-19). In rugby, a spear tackle occurs when the ball carrier is inverted and driven head first into the ground. The aims of the project were to measure tackle characteristics from video using a qualitative protocol, to assess whether tackle characteristics differed by level of play, and to measure the associations between tackle characteristics and injury.
Wilson et al. (21) provided an early example of the use of video to study tackle characteristics and injury. The authors observed that tackles from the front and body contact were two important injury risk factors. In 1983, Scher (18) postulated an association between the number and timing of tacklers and cervical spine injury to the ball carrier. Garraway et al. (8) used postmatch interviews to establish the characteristics of injury and noninjury tackles in professional rugby. They observed that player speed and/or speed differential contributed to injury, with the player with the lower momentum being injured (8). Video-based studies of the tackle in professional rugby union have been reported recently. They identified the following injury risk factors in the tackle: player speed and/or speed differential, contact with the head/neck, and front-on and side-on tackles (7,11,15). All the studies indicated that player speed, mass, and body contact, factors contributing to momentum and energy transfer, were associated with injury. In contrast, tackles from behind and when the tackler was in the ball carrier's peripheral vision were observed to be associated with injury by Garraway et al. (8). Explanations for the differences between studies from the 1990s and those undertaken more recently include the evolving nature of the game, regional differences, level of play, and research methods.
In contrast to recent research on the tackle, this study will report on a cross section of rugby participants from five levels of play: schoolboys to provincial/national representative (elite) (7,15). The ability to generalize the findings from research on professional rugby to other levels will be limited if the tackle characteristics and injury risks differ between each level. As players mature physically and develop rugby skills and experience, differences in tackle characteristics may also develop. Studying the tackle in schoolboy and community rugby player cohorts is important because they comprise the largest proportion of all rugby players. Furthermore, injuries in youth rugby are not trivial and may result in hospitalization (6).
The main study was a cohort study conducted on a cross section of rugby players to investigate the following aims: whether tackle characteristics were the same across all main levels of play and whether specific tackle characteristics were associated with injury. The study was approved by a human research ethics committee.
Video and injury data sources.
At the time this study commenced, the incidence of injuries resulting in a missed game had been recorded prospectively by the Rugby Union Injury Surveillance Study from 2000 to 2006 in the following five levels of competition: younger than 15 yr schoolboy, younger than 18 yr schoolboy, colts (younger than 20 yr men), Sydney Grade Competition (open-age men), Super 12/14, and Wallaby (Australian national) teams (both elite professional men). Videos of games in each level in the Rugby Union Injury Surveillance Study cohort were obtained for the period 2002 to 2006. Videos were obtained from the Australian Rugby Union or Sydney rugby clubs or were recorded on behalf of the investigators in younger than 20 yr and schoolboy rugby. A sample of 400 games across these five levels was compiled, each with video and injury data. This represented approximately 15% of all the games within these cohorts and seasons. From this sample, a subsample of 100 games was randomly selected for analysis with stratification by level of play. This was to ensure no sample bias.
Protocols and procedures.
Qualitative skill and technique analysis has been conducted to examine human movement in sport, the performing arts, and the workplace for many years as part of skill acquisition, game analysis, and injury prevention programs. There has been an increasing interest in its use to study sports injury causation because of the availability of video in professional sport and improving computer technology for coding events from video (1,7,15). Ideally, the qualitative skill analysis protocol should identify and characterize skills in a suitable framework. For injury analysis, this is typically, but not solely, biomechanical because a mechanistic relationship must exist among the injury, load, and action generating the load (the skill) (13). In addition to biomechanical methods, frameworks that consider injury etiology, the context, contributing factors, and their sequence, for example, Haddon's (9) approach, can be implemented (3). Finally, in sports, there is a need to tally skill or technique descriptors that are shorthand for a body posture and/or movement, for example, in basketball, shooting a basket is a skill, and a "slam dunk" describes a shooting technique. In rugby, a "shoulder tackle" is a technique within the tackle skill, and it refers to a tackle in which the tackler engages the ball carrier with the shoulder.
The video analysis protocol was developed at the University of New South Wales (UNSW) in consultation with the Australian Rugby Union and local rugby experts (14,20). The authors collaborated with the Rugby Football Union (RFU) tackle study team through a series of formal meetings in Sydney and London in 2005 and 2006 (7). In mid-2006, an interrater reliability study was conducted with the RFU group and the authors using a draft protocol. Nine coders were asked to review videos that contained a maximum of 31 tackle events. Agreement ranged between 48% for the tackler's body region struck in the tackle to 91% for the event type (i.e., whether injury occurred). The median agreement for the nine fields reported in this article was 75%, which was considered good. A final protocol template and definitions document was produced after a review meeting involving the UNSW and RFU teams. At that meeting, the need to define the time limits within which observations were made, for example, body region struck, was identified as a factor that had contributed to low agreement in specific fields during the interrater reliability study. This was addressed in the final protocol, and initial tackle event coding was undertaken in this study by one analyst. Each research group derived their video analysis protocol from this template, typically in an abbreviated format (7).
A tackle event occurred when an opposition team member attempted to tackle a ball carrier. The intention of the tackle is for the ball carrier to be held and brought to the ground by an opponent (the tackler) as per the laws of the game (10). A more common perception is that any attempt by one or more opponents to stop and constrain a ball carrier is a tackle. A tackle event could involve more than one tackler and, thus, tackles. One tackle event with two tacklers would create the potential to code tackle characteristics of two ball carrier-tackler pairs. The video analysis consisted of 102 fields describing the tackle event and salient features of each tackle as well as player information. All videos were first reviewed by an author (T.N.S.), and each tackle event was identified. Coding of each tackle event was performed, and then a team of trained analysts was assigned specific tasks to complete the full analysis of the tackle event and incorporated tackles. The tackle event analysis was undertaken using the sports coding software Snapper™ (Webbsoft Technologies, Australia). The tackle event was described by the phase of play before and after tackle, the number of tacklers, the sequence of tackler contact with ball carrier, the tackle technique (Fig. 1), the completeness of tackle, and the time in game (Appendix 1). The analysis of each tackle was undertaken using a Microsoft Access™ database prepopulated with fields from the tackle event analysis. For each tackle, the main fields coded were body region of contact, tackle direction, and ball carrier awareness (Appendix 1). Owing to the varying quality of the video and tackle characteristics being obscured, it was not possible to code all fields.
One game half was randomly selected from each of the 100 games and subjected to a full tackle event analysis. In the other game half, only tackle events associated with an injury were examined to facilitate an ancillary case-control analysis.
Injury was defined in two ways: (1) game injury and (2) missed-game injury, that is, an injury that led to a player missing a subsequent game. From the video, game injuries (and their associated tackle events) were recorded as player "knocked down," "received medical attention," or "left the field." This is analogous to a combination of Quarrie and Hopkins' "injury assessments" and "injury replacements" tackle events (15). A "knockdown" was recorded when a player remained on the ground in apparent discomfort for more than 5 s after participating in a tackle event. Missed-game injuries had been recorded prospectively after obtaining written consent from each player or his parent/guardian. Players who sustained a missed-game injury were tracked using their player position or jersey number during the game in which they were injured.
Tables of frequencies were produced, and χ2 analyses were conducted to examine differences in tackle characteristics between levels, assuming independence of each tackle event. The risk factors level of play, tackle technique, number of tacklers, and sequence of tacklers were analyzed by logistic regression with respect to the outcome "injury/no injury" from each tackle event or tackle, as appropriate. All of these analyses assumed independence of tackle events but used robust SE to allow for correlation among constituent tackles of a tackle event (20). The regressions were undertaken with and without adjustment for level of play. The same method of statistical analysis was undertaken in an ancillary case-control sample of tackle events resulting in missed-game injury observed in both halves of the games together with the noninjury tackle events observed in the halves fully analyzed.
Owing to time and resource restrictions, only 77 of the planned 100 games were analyzed. A total of 6618 tackle events were observed and coded in the 77 game halves fully analyzed (Tables 1 and 2). In total, 45.6 h of rugby was coded. An average rate of 145 tackle events per hour of rugby was observed. There were 81 tackle events resulting in game injury (knocked down, received medical attention, or left the field) observed in the 77 game halves, equating to 12 game injuries per 1000 tackle events. A total of 96 tackle events resulting in game injury were identified in the other half of those games for which one half had been fully analyzed. The value "fully obscured" denoted a noninjury tackle event in which detailed analysis of the incorporated tackles was not possible because the video was of too poor quality to make observations (Table 2). It was possible to code to some extent fields in 5029 tackles (ball carrier-tackler pair) from the 6618 tackle events.
A total of 66 missed-game injuries had been recorded in the 77 games prospectively (both halves included). Injuries were to the following regions: thigh (18.2%), knee (15.2%), shoulder (13.6%), head/face/neck (12.1%), ankle (9.1%), trunk (7.6%) lower leg (6.1%), wrist/hand/fingers (6.0%), elbow (4.5%), and unspecified (7.6%). Forty injuries (61%) were sprain/strain-type injuries. There were two fractures, three dislocations, and three concussions. Of the 66 missed-game injuries, 22 had been recorded prospectively as having occurred in a tackle event. Six missed-game injury tackle events were identified on video in the 77 half games fully analyzed and a further 10 in the other halves. The overall success rate for identifying the tackle event that caused the missed-game injury in the 77 full games was 73% (16/22).
Tackle event characteristics and level of play.
The tackle descriptor for the primary tackle in each tackle event, the number of tacklers, tackler sequence, and tackle outcome were recorded. Approximately 59.5% of the primary tackles met the International Rugby Board's (IRB) definition of a tackle, that is, ball carrier brought to ground. In the remaining cases, the following occurred: tackler missed tackle (3.0%), ball carrier broke tackle (14.3%), ball carrier held up/maul (6.8%), ball carrier offload while on feet (13.4%), and unclear (3.6%). The distribution of tackle characteristics by level of play is presented in Table 3. In 15% of cases, it was not possible to identify the tackle characteristics from the video because the tackle event was obscured. Differences in tackle characteristics were observed between each level of play. An increase in the proportion of active shoulder tackles was observed from younger than 15 yr (13%) to elite (31%), as well as smother tackles. Younger players engaged in more passive tackles and tended to stay on their feet more than more experienced players. There was a significantly greater proportion of IRB-compliant primary tackles in the elite and grade levels compared with the younger age groups (Table 3).
Approximately 90% of tackle events had either one or two tacklers (Table 3). There was a significantly higher proportion of tackle events with two tacklers in the elite and grade levels. There was a greater proportion of two or more tacklers contacting the ball carrier simultaneously at the elite level compared with all other levels (Table 3).
Tackle event characteristics and game injury.
The following factors-level of play, tackle technique, number of tacklers, and sequence of tacklers-were analyzed with respect to observed game injury incidence. The analyses were run with and without adjustment for level of play. Younger than 15 yr rugby players had a significantly lower risk of game injury compared with elite. No specific tackle technique was observed to be associated with a significantly increased risk of game injury. An "ankle tap," where a player dives and grabs or pushes the ankle of a ball carrier, typically running at pace, was associated with the greatest risk of game injury. There was a greater risk of game injury associated with two or more tacklers involved in the tackle event and the greatest risk associated with simultaneous contact by tacklers. These findings were slightly changed by taking into account (adjusting for) the differing injury risks by level of play (Table 4).
The following game injury risk factors for each tackle were analyzed within the cohort design framework: body region struck for tackler and ball carrier, tackle direction, apparent awareness of ball carrier, and speed into the tackle. Impacts to the tackler's head and the ball carrier's legs and the tackler being unaware were observed to have the highest incidence of game injury. However, none of these tackle characteristics could be concluded to have been associated with game injury risk in general (Table 5).
Missed-game injury ancillary analysis.
The ancillary case-control analysis comprised 6588 tackle events and 16 missed-game injuries. No significant differences were observed for the risk factors: level, tackle technique, number of tacklers, and sequence of tacklers. The findings were generally consistent with the analysis of all tackle events. For example, after adjustment for level of play, tackle events with simultaneous tackles by two or more players had a greater risk of resulting in a missed-game injury than tackle events with only one tackler (OR = 2.57, P = 0.13). The ankle tap technique had the highest injury risk compared with the active shoulder tackle (OR = 2.92, P = 0.33). As indicated by the P values, the differing injury risks observed could have occurred by chance alone.
This is the first study to describe tackle technique and analyze associated injury risks in a cross section of rugby union football. Injury events related to the tackle in 77 games of rugby union were analyzed, and all tackles in one-half of each of these games were analyzed using a standard protocol. In total, 6618 tackle events were analyzed in younger than 15 yr, younger than 18 yr, younger than 20 yr, open grade, and Super 14/Wallaby rugby.
Significant differences in tackle characteristics were observed between levels of play. The more skilled and older players (elite and grade) were engaged in a higher proportion of active shoulder tackles and smother tackles compared with the younger players (younger than 15 yr to younger than 20 yr). The elite and grade players executed complete IRB tackles more frequently than other levels, whereas missed or broken tackles were more frequent in the school-aged players. Elite- and grade-level rugby was observed to have more multiplayer tackles than the other levels. Level was observed to be a general injury risk factor, with the younger than 15 yr having an OR of 0.28 compared with elite. This is consistent with injury surveillance studies. This confirms that observations made on professional rugby are not necessarily valid for community rugby and may not form the basis for successful interventions for the majority of rugby players.
Whereas the relationships between missed-game injury and tackle event descriptors in the 77 halves fully analyzed were not significant, they showed similar characteristics to tackle events resulting in a player being knocked down, receiving medical attention, or leaving the field. This observation suggests that there is some similarity between events leading to a game injury and a missed-game injury. The severity of the contact forces, the tolerance of the structure to the load, the resultant incapacity, and injury management protocols will determine the duration of absence from the game.
The main findings from recent studies of the tackle in professional rugby were that the direction of the tackle, the tackle speed, and the height of the tackle were injury risk factors (7,15). In addition to level of play, we observed that having at least two tacklers per tackle event and simultaneous contact were associated with the greatest incidence of injury to either tackler/s or ball carrier, even after adjustment for level of play. This was consistent with Scher's (18) observations regarding cervical spine injury. These tackles were more frequent in elite rugby than the other levels. No specific tackle technique was identified as a risk factor, although ankle taps were over twice as likely to result in an injury as an active shoulder tackle. Ankle tap techniques are used when a ball carrier is running at full pace and the tackler must dive to stop them. The tackler dives to the ground, reaches out to tap the ball carriers foot or ankle, and, if successful, the ball carrier falls at high speed. No significant effects were observed based on tackle direction, awareness of the ball carrier, or region of the body involved in the initial contact for either ball carrier or tackler. It was observed that contact with the ball carrier's lower leg and with the tackler's head had the greatest risk of injury. The latter is consistent with other studies (7,15).
Study limitations will now be discussed. The selection process for the cohort games minimized any bias, although resource limitations prevented the completion of the target 100 games. One limitation was that few missed-game injury events related specifically to the tackle occurred in the cohort. In the context of the number of tackle events and tackle injury incidence rates, increasing the number of tackle events studied represents a resource challenge. In studies of professional rugby that have piggybacked an analysis of the tackle and other skills resulting in injury on an existing game video analysis structure, those resource challenges are slightly reduced in comparison with community rugby.
The issue of interrater reliability was addressed in the development of the tackle protocol. The final protocol had evolved in an iterative manner with input from a formal interrater reliability study and expert review. An important outcome of the interrater reliability study was the need to identify clearly the timeframe for determining each tackle descriptor and group body region descriptors. In practice, a single person (author T.N.S.) identified all tackles and undertook the initial tackle event coding. Analysts were trained, and their work was reviewed regularly. The analysts had participated in the interrater reliability study. When uncertainties arose, the event was reviewed by multiple analysts with the default outcome being "unknown." Using a purely descriptive analysis does not permit the estimation of forces or moments applied to the rugby players, or energy transfer, which are typically measured in biomechanical studies of injury. Because of this limitation, it remains possible that the cause of injury in a tackle could not be observed. We successfully identified 73% of the tackle events resulting in a missed-game injury. In comparison with other studies, Quarrie (15) identified 738 injury assessments and injury replacements (equivalent to our "medical aid sought" and "leave field" events) from 434 matches and matched 38% of these to injury data. In video analysis studies of injury mechanisms in football, Andersen et al. identified 57% of ankle injury events (2) and 43% of all injury events (1).
On the basis of the findings of this study, the following aspects of the tackle require consideration from an injury prevention perspective: the number of tacklers, the sequence of tackler contact with the ball carrier, and positioning of the tackler's head when executing a tackle. However, it is recognized that the removal of multiplayer tacklers may lead to unintended consequences. One message that can be derived from the findings is that school-age rugby at the level of younger than 15 yr is significantly different to and safer in the tackle than adult rugby.
This project was funded by the International Rugby Board (IRB). No investigator's salary was funded by the sponsor. R.W. was paid to undertake the statistical analyses from the research grant. The investigators, not the IRB, were responsible for the following: decisions regarding the design and conduct of the study as well as collection, management, analysis, and interpretation of the data. The authors prepared this article, and it was sent to the IRB for approval. All authors and researchers were independent of the IRB, and the project was conducted independently. Statistical analyses were conducted in part by Dr. Rory Wolfe from Monash University. The decision to publish was controlled by the authors and not the sponsor. There are no conflicts of interest.
Contributors: A.S.M. was the principal investigator and is the guarantor. He contributed to the study design, project management, data analysis, interpretation, drafting, and writing of the article. T.N.S., P.M., B.O.F., and R.W. contributed to the study design and data interpretation as well as in the drafting and writing of the article. The authors thank the following people and their contributions: Dean McNamara, Maciej Swistak, Arthur Kapsimalis, Megan Stringer, and Jennifer Coogan who undertook video analyses; John Searl, John Muggleton, and Brian O'Shea from the Australian Rugby Union; Dr. Simon Kemp from the Rugby Football Union; and Ken Quarrie from the New Zealand Rugby Union, who contributed to the development of the tackle protocol.
This study was approved by the UNSW Human Research Ethics Committee: approvals HREC 99085 and HREC 02146.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Andersen TE, Larsen O, Tenga A, Engebretsen L, Bahr R. Football incident analysis: a new video based method to describe injury mechanisms in professional football. Br J Sports Med
2. Andersen TE, Floerenes TW, Arnason A, Bahr R. Video analysis of the mechanisms for ankle injuries in football. Am J Sports Med
. 2004;32(1 suppl):69S-79S.
3. Bahr R, Krosshaug T. Understanding injury mechanisms: a key component of preventing injuries in sport. Br J Sports Med
4. Bathgate A, Best JP, Craig G, Jamieson M. A prospective study of injuries to elite Australian rugby union players. Br J Sports Med
5. Best JP, McIntosh AS, Savage TN. Rugby World Cup 2003 injury surveillance project. Br J Sports Med
6. Boufous S, Dennis R, Finch CF. A Profile of Hospitalisations and Deaths Due to Sport and Leisure Injuries in New South Wales, 2000-2004: NSW Injury Risk Management Report
. Sydney (Australia): The University of New South Wales; 2006. p. 11-20.
7. Fuller CW, Ashton T, Brooks JHM, Cancea RJ, Hall J, Kemp SPT. Injury risks associated with tackling in rugby union. Br J Sports Med
8. Garraway WM, Lee AJ, Macleod DAD, Telfer JW, Deary IJ, Murray GD. Factors influencing tackle injuries in rugby union football. Br J Sports Med
9. Haddon W. The changing approach to epidemiology, prevention, and amelioration of trauma: the transition to approaches etiologically rather than descriptively based. Am J Public Health
10. International Rugby Board. The Laws of the Game of Rugby Union 2002, as Framed by the International Rugby Board and Australian Rugby Union, Australian Rugby Union Edition
. North Sydney (Australia): Australian Rugby Union; 2002. p. 81-5.
11. McIntosh AS, McCrory P, Finch CF, Wolfe R. Head, face and neck injury in youth rugby: incidence and risk factors. Br J Sports Med
12. McIntosh AS, McCrory P. Preventing head and neck injury. Br J Sports Med
13. McIntosh AS. Risk compensation, motivation, injuries, and biomechanics in competitive sport. Br J Sports Med
14. McIntosh AS, Savage TN, Nicholson A. Investigating the tackle in Rugby Union. J Sci Med Sport
. 2005;8(4 suppl):187.
15. Quarrie KL, Hopkins WG. Tackle injuries in professional Rugby Union. Am J Sports Med
16. Quarrie KL, Alsop JC, Waller AE, Bird YN, Marshall SW, Chalmers DJ. The New Zealand rugby injury and performance project: VI. A prospective cohort study of risk factors for injury in rugby union football. Br J Sports Med
17. Scher AT. The high rugby tackle: an avoidable cause of cervical spinal injury? S Afr Med J
18. Scher AT. The 'double tackle'-another cause of serious cervical spinal injury in rugby players: case reports. S Afr Med J
19. Scher AT. Rugby injuries to the cervical spine and spinal cord: a 10-year review. Clin J Sport Med
20. Savage TN, McIntosh AS, Kemp SP, Frechede BO, McCrory PJ. Development of a method to study injury risk factors associated with the tackle in Rugby Union. J Sci Med Sport
. 2007;10(1 suppl):89.
21. Wilson BD, Quarrie KL, Milburn PD, Chalmers DJ. The nature and circumstances of tackle injuries in rugby union. J Sci Med Sport
Appendix 1: Summary of Tackle Descriptors
- Tackle event type: no injury/knockdown/medical aid/player left field/fully obscured
- Number of tacklers
- Tackle sequence: one-on-one/simultaneous/sequential/unsure
- IRB-compliant tackle: yes/no/unsure
- Estimate of impact force: high/low/unsure
- Referee's decision: legal/free kick/penalty (technical or foul play)/sin bin/sent off
- Phase of play before and after: tackle/scrum/ruck/maul/line-out/open play/kick/try/restart
- Ball carrier team retains ball: yes/no/unsure
- Field location
- Ground dry: yes/no/unsure
- Raining: yes/no/unsure
- Player attempting to score: yes/no/unsure
Tackle characteristics (applied to tackler and ball carrier)
- Tackle type: shoulder active or passive/jersey/ankle tap/smother/arm short-long
- Tackle direction: front/behind/right side/left side/unsure
- Speed of ball carrier: still/slow/fast/unknown
- Injury: knockdown/medical aid/player left field/missed game
- Ball carrier lifted by tackle: yes/no/unsure
- Awareness of tackler: yes/no/unsure
- Tackler aware of ball carrier: yes/no/unsure
- Tackler and ball carrier body regions struck: anatomical description
- Tackler head position: above/behind/in front/beside/unsure
- Ball carrier evasion technique: none/hand fend/arm/fend/shoulder bump/ball bump
- Relative height and body mass: same/larger/smaller (with reference to opponent)
- Player's positions in team
- Posture (back, shoulders, leg drive, feet alignment, preparation)
- Ball carrier team support: yes/no/unsure
- Direction of travel on field: with reference to try line-to-try line
- Ground impact: yes/no/unsure
- With specific characteristics of ground contact