Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
Trends in match injury risk in professional male rugby
union: a 16-­season review of 10 851 match injuries in
the English Premiership (2002–2019): the rofessional
ugby njury urveillance roject
Stephen W West ‍ ‍,1 Lindsay Starling,1 Simon Kemp ‍ ‍,2 Sean Williams,1
Matthew Cross,3 Aileen Taylor,4 John H M Brooks ‍ ‍,5 Keith A Stokes1,2

►► Additional material is ABSTRACT when breaking injuries down into smaller injury
published online only. To view Objectives The Professional Rugby Injury Surveillance categories). We aimed to examine match injuries in
please visit the journal online
(http://d​ x.​doi.o​ rg/​10.​1136/​ Project is the largest and longest running rugby union professional rugby over the period 2002–2019 and
bjsports-​2020-​102529). injury surveillance project globally and focuses on the to describe trends in injuries over this period.
highest level of rugby in England.
1
Department for Health, Methods We examined match injuries in professional METHODS
University of Bath, Bath, UK
2 men’s rugby over the period 2002/2003 to 2018/2019 Participants
Medical Services, Rugby
Football Union, London, UK and described trends in injuries over this time. In the 2002/2003–2018/2019 seasons (August to
3
Premiership Rugby, London, UK Results Over the period 2002/2003–2018/2019, June), a mean of 576 (SD:98, range: 413–763)
4
Chartered Physiotherapist, 10 851 injuries occurred in 1 24 952 hours of match players consented to participate in the study per
Nairobi, Kenya play, equating to a mean of 57 injuries per club per
5
Connect Health, Merton, UK season, with a total of 9213 player-­seasons (3006
season and one injury per team per match. The mean unique players). Data were collected from the 16
incidence, severity (days absence) and burden (days clubs (12 per season) in the top tier of English
Correspondence to
Professor Keith A Stokes, absence/1000 hours) of injury were 87/1000 hours (95% rugby with a mean squad size for the period of 48
Department for Health, CI 82 to 92), 25 days (95% CI 22 to 28) and 2178 players per season (range: 34 (2002/2003 season)
University of Bath, Bath BA2 days/1000 hours (95% CI 1872 to 2484), respectively. to 64 (2018/2019 season)). Match exposure and
7AY, UK; ​k.​stokes@b​ ath.​ac.u​ k The tackle accounted for 43% injuries with running injury data were recorded daily by club condi-

copyright.
Accepted 27 August 2020 the second most common activity during injury (12%). tioning and medical staff as part of the Profes-
The most common injury location was the head/ sional Rugby Injury Surveillance Project (which is
face with an incidence of 11.3/1000 hours, while the mandated for all teams playing in the competition)
location with the highest overall burden was the knee and includes all match injuries from the Premier-
(11.1 days/1000 hours). Long-­term trends demonstrated ship, the National Cup and European Cups. Each
stable injury incidence and proportion of injured players, team played 22 Premiership games each year with
but an increase in the mean and median severity of four teams involved in semi-­finals and two in the
injuries. Concussion incidence, severity and burden final, while European and National Cup exposure
increased from the 2009/2010 season onwards and from was based on the success of English teams in those
2011 to 2019 concussion was the most common injury. competitions. Individual informed consent was
Conclusion The rise in overall injury severity and obtained from first-­team eligible players on a yearly
concussion incidence are the most significant findings basis.
from this work and demonstrate the need for continued
efforts to reduce concussion risk as well as a greater Procedures
understanding of changes in injury severity over time. From 2002/2003 to 2012/2013, data were collected
using a paper-­ based format and then entered
into a database at the host institution.7 From the
2013/2014 season, injury data were captured
INTRODUCTION directly from the injury surveillance section of an
Rugby Union (herein referred to as ‘rugby’) is a online clinical electronic medical record keeping
field-­based collision team sport comprising both system, ‘Rugby Squad’ (The Sports Office, UK).7
low-­ intensity and high-­intensity periods of exer- For each injury reported, information including
cise.1 In comparison to other team sports, the inci- injury type, site, activity causing injury and severity
© Author(s) (or their dence and severity of injury in professional rugby was collected.7 An injury was defined as ‘any injury
employer(s)) 2020. No is relatively high (83/1000 hours2 and 37 days per that resulted in a player being unable to take a full
commercial re-­use. See rights
and permissions. Published injury,3 respectively). The majority of previous part in future rugby training or match play for
by BMJ. injury surveillance studies in professional club more than 24 hours from midnight at the end of the
rugby have been limited to one or two seasons3–5; day the injury was sustained’.7 Injury severity was
To cite: West SW,
Schwellnus et al6 reported on five seasons of data. defined as the number of days lost from match play
Starling L, Kemp S, et al.
Br J Sports Med Epub ahead Multiple season, league-­wide data are of substan- or training, with the injury return date set as the day
of print: [please include Day tially greater benefit, allowing injury trends to be the player was deemed fully fit to play by the club
Month Year]. doi:10.1136/ examined as well as providing a larger sample size medical staff, irrespective of whether a match or
bjsports-2020-102529 to give greater confidence in the data (especially training was planned for that day.7 Match exposure

West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529    1

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
was calculated as the number of matches per club multiplied RESULTS
by the number of players exposed (15), multiplied by the time Over the period 2002/2003–2018/2019, 124 953 player-­hours
exposed (1.33 hours (80 min)). Data capture was overseen by of match exposure were recorded across 11 079 matches with
a lead researcher at the host institution, who implemented a 10 851 time loss match injuries captured. In total, 268 343 days
quality control process to ensure all injury details were captured were missed by players as a result of match injury, with 43% of
on a regular basis, as well as a validation of all reported match players returning to full participation in 7 days or less, 63% in
injuries using match report cards completed by match officials 14 days or less, 71% in 21 days or less and 77% in 28 days or
(as of the 2016/2017 season). This process conforms with the less. Ninety-­three per cent of players had returned within 84
International Olympic Committee (IOC) consensus statement days. On average, 54% of players sustained at least one injury
for injury epidemiology studies.8 each season (range: 39%–66%). The mean incidence of injury
was 87/1000 hours (95% CI 82 to 92, range: 62–103). The mean
and median severity of injury was 25 and 9 days, respectively
Data analysis (95% CI 22 to 28 and 8 to 10). The mean burden of injury was
Injury incidence was calculated as the count of injuries per 2178 days per 1000 hours (95% CI 1872 to 2484).
1000 player-­match-­hours (approximately 25 matches).4 9 Mean
severity was calculated as the total sum of days absence divided
by the total count of injuries, while median severity was calcu- Position
lated as the midpoint of the range of injury severities within the There was no significant difference in the incidence of injury in
dataset. Both the mean and median were calculated to account forwards compared with backs (89/1000 hours, 95% CI 86 to 92
for the potential skew in mean severity caused by a small number compared with 83/1000 hours, 95% CI 80 to 86). The mean and
of long-­ term injuries. Injury burden was calculated as the median severity of injury in both positional groups was 25 days
product of mean severity and incidence to give the number of (95% CI 24 to 25) and 9 days (95% CI 8 to 10), respectively. The
days absence per 1000 player-­match hours.4 9 Incidence, severity burden of match injuries was not different between forwards and
and burden were calculated each season to identify trends over backs (2186 days/1000 hours, 95% CI 1286 to 2547 compared
time. Injuries which led to a player retiring from the sport were with 2042 days/1000 hours, 95% CI 1744 to 2340).
included in the calculation of incidence but not severity, there
were 118 such cases over the 16-­season period. Injuries were Activity at the time of injury
subdivided into four categories; 2–7 day injuries, 8–28 day Injuries associated with the tackle accounted for 43% of all inju-
injuries, 29–84 day injuries and injuries greater than 84 days.7 ries, with the next most common activity during injury being
Linear regression was used to identify meaningful trends over running (12%). Of the tackle-­related injuries, 59% were to the
time. Significance was set at p≤‍ ‍0.05, accompanied by 95% CI. player being tackled (21.6/1000 hours, 95% CI 20.8 to 22.4)

copyright.
To account for the multiple tests being undertaken, p values were and 41% to the tackling player (15.0/1000 hour, 95% CI 14.3
adjusted using the false discovery rate method.10 All analyses to 5.7)). The greatest burden of injury was to the player being
were undertaken using IBM SPSS Statistics for Apple (V.24.0.0). tackled, followed by the tackling player (figure 1).

25

Tackled

20
Incidence (Injuries/ 1000 hours)

Tackling
15
Running
Unknown

10
Ruck
Other
Scrum
5 Collision (accidental)

Maul
Collision Collision (not accidental)
Kicking Lineout
0
0 5 10 15 20 25 30 35 40
Severity (mean days absence per injury)

Figure 1 Injury burden as a function of activity causing injury for the seasons 2002/2003–2018/2019. The y-­axis represents incidence (number per
1000 player-­hours) while the x-­axis represents mean severity (days absence).21 Green line: values to the left and below represent the under the 25th
burden percentile, these are low-­risk injuries. Orange line: values to the left and below represent the under the 50th burden percentile, these are low-­
medium risk injuries. Red line: values to the left and below represent the under the 75th burden percentile, these are medium-­high-­risk injuries. Values
to the right and above the red line are the most high-­risk injuries.

2 West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
Ankle
13 Head/Face Shoulder
Knee
12

11

10 Lower Leg
Incidence (injuries/ 1000 hours)

8
Neck/ Cervical Spine
7 Sternum/Chest/
Ribs/ Upper Back Hamstrings
Low Back
6
Hand
5 Foot
Hip/
4 Groin

3 Quadriceps
2 Abdomen Elbow
Other Thigh
Wrist Forearm
1
Pelvis
Buttock Upper Arm
0
0 10 20 30 40 50 60 70 80 90 100
Severity (mean days absence per injury)

Figure 2 Injury burden as a function of body site for the seasons 2002/2003–2018/2019. The y-­axis represents incidence (number per 1000
player-­hours) while the x-­axis represents mean severity (days absence). Green line: values to the left and below represent the under the 25th burden
percentile, these are low-­risk injuries. Orange line: values to the left and below represent the under the 50th burden percentile, these are low-­medium
risk injuries. Red line: values to the left and below represent the under the 75th burden percentile, these are medium-­high-­risk injuries. Values to the
right and above the red line are the most high-­risk injuries.

Injury location and type (table 1). The incidence of injury in the 7–28 days severity cate-
The body region with the highest incidence of injury was the gory was stable with an incidence of 30/1000 hours (95% CI 29
lower limb (45/1000 hours, 95% CI 44 to 47), followed by the to 31, online supplemental table S3). The incidence of injury in

copyright.
head/neck, upper limb and trunk (online supplemental table S4). the 28–84 day severity category was 13/1000 hours (95% CI 13
The body region with the highest injury severity was the upper to 14, online supplemental table S3), while the incidence of the
limb (32 days absence, 95% CI 31 to 34), followed by the lower greater than 84 day injuries was 6/1000 hours (95% CI 5.6 to
limb, head/neck and trunk (online supplmental table S4). When 6.4, online supplemental table S3). Both of these severity catego-
divided into more location specific grouping, injuries to the ries significantly increased over time (table 1).
head/face had the highest incidence (11.3/1000 hours: figure 2). We calculated the incidence, severity and burden of concussion
Injuries to the forearm had the highest mean severity (76 days (online supplemental figure S2). Concussion incidence, mean
absence), but occurred infrequently (0.5/1000 hours, 95% CI severity, median severity and burden all increased significantly
0.4 to 0.7). Knee injuries had the greatest burden (493 days per (table 1), with a steep annual rise in the incidence and burden
1000 hours, 95% CI 413 to 573) as they were among the most from the 2011/2012 season onwards (online supplemental
common (11.1/1000 hours, 95% CI 10.5 to 11.7) and highest figure S2A and D). Mean concussion severity increased annu-
severity (45 days absence, 95% CI 42 to 47) injuries (figure 2). ally from the 2013/2014 season, (online supplemental figure
Of the three highest burden injury types, the tackle was the S2B) while median severity increased annually from 2014/15
cause in 46% of ankle injuries, 45% of knee injuries and 66% of (online supplemental figure S2C). Over the period 2002/2003–
shoulder injuries. 2018/2019, the mean number of concussions per club per season
The injury type with the highest incidence was sprains and liga- was 6, however this ranged from two per club (2005/2006) to 14
ment injuries, which were also the highest burden with an inci- per club (2016/2017).
dence of 22.6/1000 hours (95% CI 21.8 to 23.4) and a severity The incidence of non-­concussion injuries decreased (ß: −1.25,
of 30 days (95% CI 29 to 31). The next highest burden injuries 95% CI −2.50 to −0.01, p=0.06, online supplemental figure
were muscle injuries followed by fractures (online supplemental S6). In contrast, the severity of all injuries (excluding concus-
figure S5). The most severe injury type was dislocation/sublux- sions) increased significantly (ß: 1.74, 95% CI 1.40 to 2.08,
ation, with a mean severity of 74 days (95% CI 65 to 83). p<0.001, online supplemental figure S7).
Examining the activities causing injury with the highest burden
Trends in match injuries from the 2002/2003 to 2018/2019 demonstrated no significant change in the incidence in tackle (ß:
season 0.73, 95% CI −0.05 to 1.51, p=0.07) or ruck (ß: 0.10, 95%
Table 1 reports the trends in match injury data from 2002/2003 CI −0.08 to 0.28, p=0.25). However, there was a significant
to 2018/2019. The incidence (figure 3A) of match injury was decrease in the incidence of running related injuries (ß: −0.31,
stable, while the mean severity (figure 3B), median severity 95% CI −0.57 to −0.05, p=0.02). Over time, the number of
(figure 3C) and burden (figure 3D) of match injuries rose tackle related injuries at the ankle rose significantly (ß: 1.17,
significantly. 95% CI 0.10 to 2.25, p=0.03), with no significant change at the
The 2–7 days injury severity category had the highest inci- knee (ß: 0.01, 95% CI −0.78 to −0.80, p=0.99) or shoulder (ß:
dence of injury (38/1000 hours, 95% CI 37 to 39; online supple- −0.08, 95% CI −0.85 to 0.70, p=0.83). The incidence of inju-
mental table S3), however, this significantly decreased over time ries to the hamstrings decreased significantly (ß: −0.14, 95% CI

West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529 3

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
Table 1 Long term trends in match injury in rugby (2002/2003 to 2018/2019 seasons)
Measure Beta P value Five season change Trend
Incidence −0.1 (−1.4–1.3) 0.93 −0.28/1000 hours →
Mean severity 1.2 (0.9–1.4) <0.01* 5.8 days ↑
Median severity 0.4 (0.3–0.6) <0.01* 2.1 days ↑
Burden 104.3 (58.6–150.0) <0.01* 521 days/1000 hours ↑
2–7 days incidence −1.3 (−2.1–-0.5) 0.01* −6.4/1000 hours ↓
7–28 days incidence 0.3 (−0.1–0.7) 0.17 1.5/1000 hours →
29–84 days incidence 0.5 (0.2–0.9) 0.01* 2.7/1000 hours ↑
84+ days incidence 0.5 (0.3–0.7) <0.01* 2.4/1000 hours ↑
Injury proportion −0.5 (−1.3–0.3) 0.24 −2.35% →
Concussion incidence 1.2 (0.8–1.6) <0.01* 6.0/1000 hours ↑
Concussion mean severity 0.5 (0.1–1.0) 0.04* 2.5 days ↑
Concussion median severity 0.1 (0.03–0.2) 0.02* 0.7 days ↑
Concussion burden 23.8 (10.9–33.9) 0.04* 119 days/1000 hours ↑
Incidence (excl. concussion) −1.25 (−2.50–0.01) 0.06 −6.3/1000 hours ↓
Severity (excl. concussion) 1.74 (1.40–2.08) <0.01* 8.7 days ↑
Beta values represent yearly change in respective values. Incidence (injuries per 1000 hours), severity (days absence per injury), burden (days absence per 1000 hours), proportion
(percentage of players with at least one injury). ↑ Rising, → Stable, ↓ Falling.
*P<0.05.

−0.24 to −0.04, p=0.01), however, the severity of hamstring and lineouts has declined.11 Other possible reasons for the
injuries increased (ß: 1.00, 95% CI 0.45 to 1.54, p=0.001), with changes in injury severity and burden include the increasing
no significant increase in burden (ß: 2.27, 95% CI −1.11 to mass and stature of players12 as well as improvements to injury
5.64, p=0.17). surveillance processes (improving the reporting (capture) of
injuries). We note that increasing severity (longer time to return
DISCUSSION from injury) will follow if clinicians employ more conservative
This is the longest running and largest injury surveillance study return to play strategies.13 Whether changes to the game or more

copyright.
in professional rugby, capturing nearly 125 000 hours of match conservative management strategies are the driving force behind
exposure and 10 851 time loss match injuries over 16 seasons the rise in severity is unknown—we speculate that both are likely
between 2002/2003 and 2018/2019. This equates to a mean at play. Injury surveillance alone cannot determine the causal
of 57 injuries per club per season and 1 injury per team per nature of this change, and different study designs are required
match. The incidence of injury was stable, while mean severity, to answer this question. Further research might incorporate
median severity and injury burden rose significantly. Concussion mixed methods (both qualitative and quantitative) to establish
incidence, severity and burden also all rose significantly. The the causes of increasing severity of injury.
lower limb was the most commonly injured body region; the
knee represented the anatomical region with the highest injury Concussion
burden. The tackle was the game event most commonly asso- While concussion was not the focus of this study, it has emerged
ciated with injury contributing 43% of all injuries (25% player as the most common injury in rugby since 2011. The study mean
being tackled, 18% tackling player). concussion incidence does not portray the changes in concussion
reporting over time, nor the potential importance of changes
Trends in match injury in concussion recognition tools and more conservative return
A key benefit of longitudinal injury surveillance is the ability to to play strategies. In our study, concussion increased in inci-
track trends in injury over time. Over the 16 season period, the dence, but also in severity and thus burden (online supplemental
overall incidence of injury did not change, however, the inci- figure S2) and this extends findings reported in other elite rugby
dence of injuries excluding concussion decreased by a mean settings.6
of 1.2/1000 hours each year (online supplemental fgure S7). In an effort to improve concussion recognition and manage-
The incidence of injury (87/1000 hours) was higher than that ment a number of tools have been introduced, including an
in professional rugby in Australia (66/1000 hours3), but lower off-­field in-­game assessment as part of the Head Injury Assess-
than South Africa (100/1000 hours6). The severity of injuries ment (HIA) protocol14 in 2012/2013, real-­time pitch-­side video
increased over time, with an injury on average lasting 1.2 days review for medical staff of head injury events in 2017 and the
longer per year (including concussions) and 1.7 days longer introduction of independent match day doctors to identify
per year (excluding concussion). Mean severity of injury in this significant head injury events and supervise the application of
study was lower (25 days) compared with Australian professional the HIA protocol in 2018. We opine that these interventions
players (40 days3). increased clinicians’ recognition of concussion in professional
Explaining the changes demonstrated in this study is difficult, rugby, however it is still to be determined whether the increase
in part because the game has evolved, with more contact events in reported concussions is due to this awareness alone, or due to
occurring in tackles and rucks.11 Both of these match events are an increase in the rates of concussion itself.15 It is incontrovert-
considered among the highest burden events (figure 1), given ible that the current rates of concussion in professional rugby
their unpredictable nature and high speed contact. Over the are consistently high (accounting for an average of 21% of all
same period, exposure to lower burden events such as scrums injuries in the five seasons from 2014/2015 to 2018/2019) and

4 West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
120 A

Incidence (number per 1000 hrs)

100

80

60

40

20

0
B
40

35
Mean Days absence per injury

30

25

20

15

10

0
16 C
14
Median severity (days absence)

12

10

copyright.
4

4000
D
3500
Days absence per 1000 hrs

3000

2500

2000

1500

1000

500

80 E
70
Proportion of injured players (%)

60

50

40

30

20

10

0
20 3

20 4

20 5

20 6

20 7

20 8

20 9

20 0

20 1

20 2

20 3

20 4

20 5

20 6

20 7

20 8

9
-0

-0

-0

-0

-0

-0

-0

-1

-1

-1

-1

-1

-1

-1

-1

-1

-1
02

03

04

05

06

07

08

09

10

11

12

13

14

15

16

17

18
20

Mean Upper Limit Lower Limit

Figure 3 Trends in match injury incidence (A), mean severity (B), median severity (C), burden (D), proportion (E). No data were collected during the
2004–2005 season. Dotted grey lines represent 2 SD from the mean for the whole period (depicted by grey dashed line).

West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529 5

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
governing bodies should continue to investigate strategies to Finally, before data were analysed, all injuries were checked for
lower the risk. duplicates and inconsistencies. We gained final approval of the
included injuries from the medical lead in each club. These prac-
tices ensure data quality and are in-­line with those included in
Activity causing injury: tackles and more
the 2020 IOC consensus statement on injury surveillance.8
The most common activity causing injury was the tackle (43% of
all injuries, 25% to the player being tackled, 18% to the tackling
player: figure 1). This mean figure is slightly lower than those Summary and conclusion
documented in South Africa, with 50% of all injuries linked to Over the course of 16 seasons, the incidence of injury remained
the tackle in that setting,6 however, the yearly proportion in stable, while injury severity and burden rose steadily. Our study
this study ranged from 36% to 52%. The tackle accounts for an provides insight for future investigation into specific areas of
injury burden of 987 days absence per 1000 hours of match play, interest, including activities such as the scrum, ruck and tackle
which is nearly four times higher than the next highest burden as well as specific injury types such as concussion. These data
activity (running: 270 days absence/1000 hours, figure 1). provide an important resource for practitioners working within
Nevertheless, efforts to reduce the injury risk in the tackle are the sport, as to the most common and highest burden injuries.
hampered by the frequency and dynamic and unpredictable We call for studies examining the specific types and activities
nature of this game event. Despite this, several law variations for causing injury associated with commonly injured body locations
reducing the risk of injury in the tackle are being considered,16 such as the knee, shoulder and ankle. We also call for continued
with one law variation to reduce the legal height of the tackle support from all stakeholders for this important prospective
having been evaluated in the second tier of professional rugby in study which informs injury prevention strategies and thus, has
England.17 With a lower legal tackle height, incidence of all inju- the potential to influence player welfare in professional rugby.
ries did not change and neither did overall concussion incidence,
but the incidence of concussion while tackling increased, which What are the findings?
demonstrates the challenges associated with improving safety in
this high risk component of the game. While the focus of many ►► This is the largest injury surveillance study globally in
injury prevention strategies have been on concussion, it must not professional rugby union, with nearly 11 000 injuries recorded
be overlooked that the three highest burden injury locations in over a 16-­season period.
rugby are the ankle, knee and shoulder (figure 2). Given that a ►► Injury incidence and the proportion of players injured each
high proportion of injuries to these body locations are associ- year remained stable between the 2002/2003 and 2018/2019
ated with the tackle (ankle: 46%, knee: 45% and shoulder 65%), seasons.
investigations and injury prevention strategies targeted at these ►► Between the 2002/2003 and 2018/2019 seasons, there was a

copyright.
sites should be considered alongside those specific to concussion. significant rise in the mean and median severity of injuries.
Running and the ruck make up the second and third highest ►► The incidence, severity and burden of concussion rose
burden activities causing injury in rugby, with the incidence of dramatically between 2010/2011 and 2016/2017.
running injuries decreasing significantly over the study period
(figure 1). No significant changes have occurred in the inci-
dence of ruck related injuries; however, given recent evidence How might it impact on clinical practice in the future?
suggesting inconsistent law application in the professional
►► Identifies the need for more focused studies examining
game,18 this may be an area that could benefit from an injury
prevention focus.
specific body locations including the knee, ankle and
shoulder.
►► Provides medical teams with expected injury rates
Limitations and severities for rugby related injuries, to inform
We acknowledge several limitations. Over the course of the 16 interdepartmental return to play strategies within clubs.
seasons in this study, the project has been led by a number of ►► Supports the ongoing need for injury prevention strategies to
researchers using a number of methods in line with the consensus mitigate risk in the tackle.
statement for injury surveillance in rugby union.7 We aimed to
ensure the continuity of methods among researchers over this
Twitter Stephen W West @westy160991, Simon Kemp @drsimonkemp and Keith A
time period but it is likely that small systematic changes occurred Stokes @drkeithstokes
over time.
Acknowledgements The authors would like to thank the medical and sport
Larger changes in the system include the change from a paper-­ science staff at each of the Premiership clubs who partook in the study.
based recording system to a web based platform in 2012, as
Contributors Each of the authors was involved in the original conception of the
well as the change from a three digit Orchard code (OSICS-8) paper, data collection process, analysis of results and interpretation of findings. SWW
to a four digit Orchard code (OSICS-10).19 One of the diffi- and LS drafted the original manuscript and all other authors provided significant
culties associated with studies of this type is validating the data feedback and comments in refining the final manuscript.
input into the system (while avoiding double entry of data).20 Funding This project was funded by the Rugby Football Union and Premier Rugby.
To address this, a number of quality control processes have Competing interests SK and KAS are employed by the Rugby Football Union. MC
been added to the data collection. During each match, a match is employed by Premier Rugby.
report card is completed by an official, which notes the reasons Patient and public involvement Patients and/or the public were involved in the
for substitutions (tactical, injury, blood substitution, HIA, etc). design, or conduct, or reporting, or dissemination plans of this research; however, the
These report cards are cross-­referenced against match injuries project steering group includes a range of stakeholders and practitioners.
entered into the database to ensure that all injuries sustained are Patient consent for publication Not required.
captured. Furthermore, concussions reported in the database Ethics approval The study was subject to ethical approval from the host academic
were crosschecked with CSx (concussion management mobile institutions (University of Leicester (2002–2007), University of Nottingham
application) data to ensure all concussions are logged correctly. (2007–2012) and University of Bath (2011–2019).

6 West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529

 Original research

Br J Sports Med: first published as 10.1136/bjsports-2020-102529 on 12 October 2020. Downloaded from http://bjsm.bmj.com/ on October 13, 2020 at UCL Library Services. Protected by
Provenance and peer review Not commissioned; externally peer reviewed. 9 Brooks JHM, Fuller CW, Kemp SPT, et al. Epidemiology of injuries in English
professional rugby Union: Part 2 training injuries. Br J Sports Med 2005;39:767–75.
Data availability statement All publicly available data are included in the article
10 Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful
or uploaded as online supplemental information.
approach to multiple testing. Journal of the Royal Statistical Society: Series B
ORCID iDs 1995;57:289–300.
Stephen W West http://o​ rcid.​org/​0000-​0001-​5800-​7000 11 World Rugby. Rugby world cup 2015 statistical report, 2016. Available: https://www.​
Simon Kemp http://o​ rcid.​org/​0000-​0002-​3250-​2713 playerwelfare.w ​ orldrugby.​org/?​documentid=​156;
John H M Brooks http://​orcid.​org/​0000-​0003-​2935-​7858 12 Fuller CW, Taylor AE, Brooks JHM, et al. Changes in the stature, body mass
and age of English professional rugby players: a 10-­year review. J Sports Sci
2013;31:795–802.
REFERENCES 13 Kemp S, West S, Brooks J, et al. England professional rugby injury surveillance project
1 Roberts SP, Trewartha G, Higgitt RJ, et al. The physical demands of elite English rugby 2016-17, 2018. Available: https://www.​englandrugby.​com/​mm/​Document/​General/​
Union. J Sports Sci 2008;26:825–33. General/​01/​32/​91/​95/​Inju​rySur​ vei​llan​ceRe​port2016-1​ 7_​English.​pdf;
2 Williams S, Trewartha G, Kemp S, et al. A Meta-­Analysis of Injuries in Senior Men’s 14 World Rugby. Head injury assessment (Hia) protocol, 2015. Available: https://www.​
Professional Rugby Union. Sports Med 2013;43:1043–55.
sareferees.c​ om/​uploads/fi​ les/​Eugene%​202017/​180326%​20HIA%​20Protocol%​
3 Whitehouse T, Orr R, Fitzgerald E, et al. The epidemiology of injuries in Australian
20and%​20Procedures.​pdf
professional rugby Union 2014 super rugby competition. Orthop J Sports Med
15 Cross M, Trewartha G, Kemp S, et al. Concussion in rugby Union: improved
2016;4:232596711663407.
reporting, a more conservative approach or an increased risk? Br J Sports Med
4 Brooks JHM, Fuller CW, Kemp SPT, et al. Epidemiology of injuries in English
2017;51:309.2–309.
professional rugby Union: Part 1 match injuries. Br J Sports Med 2005;39:757–66.
5 Schwellnus MP, Thomson A, Derman W, et al. More than 50% of players sustained a 16 World Rugby. World rugby to consider injury-­prevention law trials, 2019. Available:
time-­loss injury (>1 day of lost training or playing time) during the 2012 Super Rugby https://www.​world.​rugby/​news/​432632
Union Tournament: a prospective cohort study of 17,340 player-­hours. Br J Sports 17 Stokes KA, Locke D, Roberts S, et al. Does reducing the height of the tackle through
Med 2014;48:1306–15. law change in elite men’s rugby union (The Championship, England) reduce the
6 Schwellnus MP, Jordaan E, Janse van Rensburg C, et al. Match injury incidence during incidence of concussion? A controlled study in 126 games. Br J Sports Med 2019.bjsp
the super rugby tournament is high: a prospective cohort study over five seasons orts-2019-101557.
involving 93 641 player-­hours. Br J Sports Med 2019;53:620–7. 18 Kraak W, Bam J, Kruger S, et al. Sanctioning of illegal and dangerous Ruck Cleanouts
7 Fuller CW, Molloy MG, Bagate C, et al. Consensus statement on injury definitions and during the 2018 super rugby competition. Front Psychol 2019;10:803.
data collection procedures for studies of injuries in rugby Union. Clinical Journal of 19 Rae K, Orchard J. The orchard sports injury classification system (OSICS) version 10.
Sport Medicine 2007;17:177–81. Clin J Sport Med 2007;17:201–4.
8 Bahr R, Clarsen B, Derman W, et al. International Olympic Committee consensus 20 Ekegren CL, Gabbe BJ, Finch CF. Sports injury surveillance systems: a review of
statement: methods for recording and reporting of epidemiological data on injury methods and data quality. Sports Med 2016;46:49–65.
and illness in sport 2020 (including STROBE extension for sport injury and illness 21 Fuller CW. Injury risk (burden), risk matrices and risk contours in team sports: a review
surveillance (STROBE-­SIIS)). Br J Sports Med 2020;54:372–89. of principles, practices and problems. Sports Med 2018;48:1597–606.

copyright.

West SW, et al. Br J Sports Med 2020;0:1–7. doi:10.1136/bjsports-2020-102529 7