Article Text

This article has a correction. Please see:

A unique insight into the incidence of rugby injuries using referee replacement reports
  1. J C M Sharp1,
  2. G D Murray2,
  3. D A D Macleod1
  1. 1Honorary Medical Advisors, Scottish Rugby Union
  2. 2Department of Medical Statistics, University of Edinburgh
  1. Correspondence to: Mr D A D Macleod, St John's Hospital, Livingston, West Lothian EH54 6PP, Scotland, UK


Objectives—To obtain further information on the incidence of injuries and playing positions affected in club rugby in Scotland.

Methods—Routine reports of injury (permanent) and blood (temporary) replacements occurring in competitive club rugby matches by referees to the Scottish Rugby Union during seasons 1990–1991 to 1996–1997 were analysed.

Results—A total of 3513 injuries (87 per 100 scheduled matches) and 1000 blood replacements (34 per 100 scheduled matches) were reported. Forwards accounted for 60% of the injury and 72% of the blood replacements. Flankers and the front row were the most commonly replaced forwards while wing and centre three quarters were the most vulnerable playing positions among backs. The incidence of injury replacements increased as the match progressed up until the last 10 minutes when the trend was reversed. Blood replacements showed a different pattern with 60% occurring during the first half of the match.

Conclusion—The most important finding of the study was reliability of referees in documenting the vulnerability of certain playing positions, and the timing when injuries took place, thus assisting coaches and team selectors when choosing replacement players for competitive club and representative rugby matches. This study re-emphasises the need for continuing epidemiological research.

  • rugby injuries
  • referee replacement reports
  • Scottish Rugby Union

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Take home message

Sports medicine must endeavour to make a positive contribution to player safety, irrespective of the sport involved. Recommendations to players, coaches, officials, and governing bodies designed to minimise illness or injury in sport should be based on reliable data. Inviting match officials to record basic details about players unable to complete a match because of injury has proved to be an accurate source of information.

During the 1970s, rugby union was criticised by the media alleging an increase in aggressive and deliberately dangerous play. Numerous clinical and epidemiological studies of rugby injury were subsequently reported from the British Isles15 and overseas,610 identifying an apparently greater incidence and severity of injury and the vulnerability of certain playing positions, in particular front row forwards.

The Scottish Rugby Union (SRU) led the way in the British Isles, following the introduction of competitive club rugby based on national leagues in Scotland in 1973, with a series of research projects which collected information about the nature, incidence, and causes of injury in an effort to enhance prevention and treatment. These studies included establishing the first national coordinated survey on injuries occurring during competitive club rugby11 in Scotland, reviewing the accident and emergency department records of all injured players in the Scottish Borders during the season 1990–1991 (H J Vander Post, personal communication), and encouraging clubs to report to the SRU all players admitted to hospital or dying as a result of playing or practising rugby. More recent epidemiological studies established jointly with Edinburgh University Department of Public Health Sciences reviewed injuries occurring in adult and school boy rugby in the Scottish Borders and Edinburgh schools.12

A further initiative was established in 1988–1989 when referees were asked to record all replacements made for injured players during club matches in the competitive leagues. This paper reports the findings from this study, which we believe to be unique, reviewing 3513 injury and 1000 blood replacement reports over seven seasons of Scottish Rugby (1990–1991 to 1996–1997).

Materials and methods

In the season 1988–1989, rugby referees in Scotland were required to notify the SRU of any player who had to leave the field as the result of injury or other “medical” reason—for example, hypothermia—for whom a replacement player was permitted. The quality of recording replacements was improved in the season 1990–1991 by the introduction of a redesigned prepaid postal match result card on which the referee could also record the playing position of the injured player and the time during the match that the replacement took place (fig 1). The number of players per team who could be replaced because of injury increased from three in 1990–1991 to four in 1995–1996. In addition, during 1993–1994, the laws of rugby also permitted temporary replacement for the treatment of players who were bleeding. Their position and time of injury were similarly recorded.

Figure 1

Scottish Rugby Union result card.

With effect from November 1996, substitutes were permitted in rugby union for tactical reasons as well as replacements for injury, bleeding, or illness. As a result, the value of referee replacement reports in monitoring the occurrence of rugby injuries was diminished. Referees continued to report temporary blood replacements for a further season until the end of 1996–1997 but this has subsequently been discontinued.

The data presented are purely descriptive, without any formal statistical analysis. To make replacement rates comparable, they are summarised as replacements per 100 scheduled matches, or, when reporting on specific playing positions, as replacements per player per 100 scheduled matches. This allows for the fact that the reporting of injury and blood replacements does not span exactly the same seasons, that the number of matches per season varies, and that the different positions have one or two players per team (and hence two or four players per match).


Reports on 3513 injury replacements (87 per 100 scheduled matches) and 1000 temporary blood replacements (34 per 100 matches) were received by the SRU from referees of competitive club matches during the seven seasons 1990–1991 to 1996–1997. The number of reports of injury replacements increased from 410 in 1990–1991 (64 per 100 matches) to 554 in 1994–1995 (87 per 100 matches), when up to three replacements were permitted per match for each team. The total number of injury replacements increased to 989 in 1995–1996 (116 per 100 matches) when the number of replacements permitted had been increased to four per team and the playing season extended. Reports of temporary replacement of players because of bleeding varied from season to season, peaking at 321 in 1995–1996 (38 per 100 matches) but decreasing to 247 (29 per 100 matches) in the following season (table 1).

Table 1

Type of replacement by playing position (backs/forwards) and season 1990–1997

Of the 3513 injury replacements, 2112 (60%) involved forwards and 1390 (40%) backs; in 11 reports no playing position was identified. In every season, forwards were consistently injured more often than backs (interseasonal range 57–65%) even allowing for the 8:7 excess of forwards over backs. This pattern was even more apparent for the 1000 blood replacements, of whom 72% were forwards (interseasonal range 69–76%) (table 1).

Flankers (3.9 per individual player per 100 matches) followed by prop forwards (3.4) were the playing positions most often replaced as a result of injury. Among backs, wing (2.6) and centre three quarters (2.5) predominated. Full back (2.3), stand off (2.4), and scrum half (2.4) were the least commonly replaced players, although the replacement rates were very similar for all back positions. Flankers (1.8 per player per 100 matches) and the Number 8 (1.7), all in the back row of the scrum, were the players most often requiring attention to bleeding wounds, followed by each of the other forward positions. Scrum half (1.2) and centre three quarters (0.7) were the backs most commonly affected (table 2).

Table 2

“Injury” and “blood” replacement reports by individual playing positions, 1990–1997

Injury incidence increased as the match progressed up to the last 10 minutes, when the trend was reversed (fig 2). In the 3509 instances when the time of an injury was specified, 1068 (30%) occurred during the first half of the match and 2441 (70%) in the second half.

Blood replacements showed a different pattern, with 598 (60%) of bleeding injuries occurring during the first half of the match decreasing to 402 (40%) in the second half (fig 2).

While the incidence of injury and bleeding replacements did not vary substantially throughout the season, there was a notable decrease in the number of injury replacements, in particular during the later weeks of the season (fig 3, table 3).

Table 3

Injury replacement by division (I–VII), 1990–1991 to 1994–1995

Figure 3

Injury (1990–1995) and blood (1993–1995) replacements by match week number.

In the first five seasons (1990–1991 to 1994–1995) a total of 2524 injury replacements were reported by referees, with the greatest number recorded in division V (434) followed by division II (404), division IV (402), division III (392), and division I (340). The lowest incidence of reports was in divisions VI and VII (table 3) where the availability of replacement players was less likely. This distribution was less clear in the seasons after the changes in the number of replacements permitted per team (from three to four) and in the format of competitive club rugby in Scotland during season 1995–1996.

Comparison of blood replacement reports by divisions during seasons 1993–1994 and 1994–1995 and later seasons 1995–1996 and 1996–1997 similarly indicated a smaller number being used in lower divisions.


Every epidemiological surveillance system has individual advantages and disadvantages, with under-reporting a consistent feature. In this unique study using routine referee reporting, compliance was almost 100% because of their requirement to complete result cards and forward them to the SRU after every championship match. Any shortfall was attributed to reports being telephoned in lieu of written completion of a card. In addition, the data obtained from routine reports was inexpensive and straightforward to administer. The disadvantage was its failure to obtain information with regard to the phase of play and the nature and cause of the injury sustained as well as being limited to first XV teams playing in the senior leagues of club rugby in Scotland.

In the season 1995–1996, the format of competitive rugby in Scotland was changed from the previous structure of seven divisions each including 14 clubs (a total of 98 clubs) to four premiership divisions each of eight clubs and seven national leagues each of 10 clubs, thus increasing the total number of clubs involved in competitive rugby to 102. As a result the comparison of the frequency of replacements at different levels of competitive rugby was thereafter more difficult to undertake.

This study identified that 50% of injury replacements occurred in the six matches played in September and October, the first two months of the season. The remaining matches in the league competitions were spread out during the rest of the season. Garraway and Macleod12 showed that injuries were most common during the first two months of the season, quoting a period prevalence of 15.2 per 1000 playing hours, amounting to 60% of all injuries which occurred in the season, compared with 12.3% in March and April. Similar early season injury patterns have been reported including a South African analysis of injuries in school boy rugby.8

The number of temporary and permanent injury replacements reported at different levels of rugby nevertheless suggest that, although there is a greater incidence of injuries in the upper echelons of Scottish rugby, this is not a major feature. Although a full complement of replacement players is less likely to be a feature at lower levels of competitive rugby, a degree of under-reporting by referees in the lower leagues must also be borne in mind.

The observed decrease in the number of injury replacements during the last 10 minutes of matches can reasonably be attributed to no further replacements being available and/or injured players playing on with only a few minutes of a match remaining. Similarly a reduced availability of replacements for bleeding injuries during the later stages of a match would not be entirely unexpected.

In addition to confirming the feasibility of this approach to analysing rugby injuries, perhaps the second most important finding from the study was to show once again the incidence of injury among certain playing positions, in particular the flankers and front row forwards, thus assisting team coaches and selectors when choosing the replacements most likely to be required for injured players. It is not surprising that bleeding injuries more or less followed a similar pattern. In effect, the use of up to 87 permanent replacements for injuries per 100 scheduled matches and 38 temporary replacements for bleeding injuries was remarkably low, before the law changes in the game of rugby entitling the use of replacements for tactical reasons.

There remains a continuing need for further epidemiological studies into the incidence, distribution, and nature of injuries in rugby union,12, 13 along with the establishment of properly funded case registers reporting key injuries.14 The referee in rugby union has an increasingly wide range of responsibilities because of the complex nature of the laws of the game. It is becoming increasingly difficult for the referee to maintain an accurate record of replacements of players for injury or tactical reasons as a result of recent law changes. It is suggested therefore that a touch judge, or fourth official in those matches where one was available, could record the timing and position of injured or bleeding players, adding this information to the referee's report card at the end of a match. This invaluable source of epidemiological information would therefore not be lost as a research tool and perhaps could be developed further at representative and international level.


We thank the SRU and their referees for their support and assistance throughout the period of the study. Our thanks are also due to Mrs Derena Ritchie for typing the manuscript.

Take home message

Sports medicine must endeavour to make a positive contribution to player safety, irrespective of the sport involved. Recommendations to players, coaches, officials, and governing bodies designed to minimise illness or injury in sport should be based on reliable data. Inviting match officials to record basic details about players unable to complete a match because of injury has proved to be an accurate source of information.


Supplementary materials

  • A unique insight into the incidence of rugby injuries using referee replacement reports
    J C M Sharp, G D Murray, and D A D Macleod. Br J Sports Med 2001;35:34-37

    Missing figure
    Figure 2 was inadvertently omitted from the printed version of this article. The figure is shown below (the error is regretted).

    GIF Image

    Figure 2    Injury (1990-96) and blood (1993-97) replacements by stage of match.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and British Association of Sport and Exercise Medicine