Article Text

PDF

RISUS study: Rugby Injury Surveillance in Ulster Schools
  1. H A P Archbold1,
  2. A T Rankin2,
  3. M Webb3,
  4. R Nicholas1,
  5. N W A Eames1,
  6. R K Wilson4,
  7. L A Henderson1,
  8. G J Heyes1,
  9. C M Bleakley5
  1. 1Royal Victoria Hospital, Belfast, UK
  2. 2Bupa Health Clinic, Belfast, UK
  3. 3Kingspan Stadium, Ulster Rugby, Belfast, UK
  4. 4Musgrave Park Hospital, Belfast, UK
  5. 5Ulster University, Sport and Exercise Science Research Institute, Newtownabbey, UK
  1. Correspondence to Dr CM Bleakley, Ulster University, Sport and Exercise Science Research Institute, Newtownabbey BT370QB, UK; c.bleakley{at}ulster.ac.uk

Abstract

Objective To examine injury patterns in adolescent rugby players and determine factors associated with injury risk.

Design Prospective injury surveillance study.

Setting N=28 Grammar Schools in Ulster, Ireland (2014–2015 playing season).

Participants 825 adolescent rugby players, across in 28 school first XV rugby squads; mean age 16.9 years.

Main outcome measures Injuries were classified by body part and diagnosis, and injury incidence using injuries per 1000 match hours of exposure. HRs for injury were calculated through Cox proportional hazard regression after correction for influential covariates.

Results A total of n=426 injuries were reported across the playing season. Over 50% of injuries occurred in the tackle situation or during collisions (270/426), with few reported during set plays. The 3 most common injury sites were head/face (n=102, 23.9%), clavicle/shoulder (n=65, 15.3%) and the knee (n=56, 13.1%). Sprain (n=133, 31.2%), concussion (n=81, 19%) and muscle injury (n=65, 15.3%) were the most common diagnoses. Injury incidence is calculated at 29.06 injuries per 1000 match hours. There were no catastrophic injuries. A large percentage of injuries (208/424) resulted in absence from play for more than 28 days. Concussion carried the most significant time out from play (n=33; 15.9%), followed by dislocations of the shoulder (n=22; 10.6%), knee sprains (n=19, 9.1%), ankle sprains (n=14, 6.7%), hand/finger/thumb (n=11; 5.3%). 36.8% of participants in the study (304/825) suffered at least one injury during the playing season. Multivariate models found higher risk of injury (adjusted HR (AHR); 95% CI) with: higher age (AHR 1.45; 1.14 to 1.83), heavier weight (AHR 1.32; 1.04 to 1.69), playing representative rugby (AHR 1.42; 1.06 to 1.90) and undertaking regular strength training (AHR 1.65; 1.11 to 2.46). Playing for a lower ranked team (AHR 0.67; 0.49 to 0.90) and wearing a mouthguard (AHR 0.70; 0.54 to 0.92) were associated with lower risk of injury.

Conclusions There was a high incidence of severe injuries, with concussion, ankle and knee ligament injuries and upper limb fractures/dislocations causing greatest time loss. Players were compliant with current graduated return-to-play regulations following concussion. Physical stature and levels of competition were important risk factors and there was limited evidence for protective equipment.

  • Rugby
  • Injuries
  • Surveillance

Statistics from Altmetric.com

Introduction

Following the Rugby World Cup (RWC) 2015, a new generation of players have been inspired to participate in a sport currently played by over 6.6 million people across 120 countries.1 The introduction of professionalism in 1995 saw a significant increase in injury incidence in Rugby Union.2 Other reports suggest that injury incidence in professional rugby is around 80 injuries per 1000 playing hours.3–5 In the UK and Ireland, Rugby Union is a key component of physical education and sports provision across many schools; however, there is concern that the risk of injury is too high. Previous reviews of youth rugby reported incidence rates of 12–22 injuries per 1000 match hours.6 Recently, a meta-analysis reported higher incidence figures of 26.7 per 1000 h of exposure, reflecting a 28% risk of injury over a single season.7

Identifying modifiable risk factors is central to injury surveillance and prevention initiatives; yet these are largely unknown in youth rugby. There is some concern surrounding the increased focus on strength and conditioning in young rugby players. The average weight of schoolboy rugby players in Ireland has increased by approximately 10 kgs since 1988.8 ,9 It may be that developing extreme physical profiles risks larger collision forces, greater mismatches in physical maturity and strength, and subsequently higher injury rates. A further moderating factor, unique in youth rugby, is that adolescents may be more likely to engage in risk-taking behaviours, such as overestimating their physical capabilities or adopting dangerous styles of play during sport.10 A large proportion of rugby injuries occur in the tackle situation which is central to modern styles of play. Other phases of play such as rucks, mauls, scrums and other collisions contribute to a players’ overall impact load. A related concern is the acute and long-term risk associated with concussive episodes. The reported rates of concussion in youth rugby vary significantly and may reflect under-reporting.6 ,7 Concussion can have devastating outcomes if not identified or managed accordingly. This is particularly pertinent in adolescent and young athletes, as developing brains undergo a more complex and protracted recovery postconcussion in comparison to an adult.11 ,12

Calls to urgently establish a prospective youth rugby injury surveillance project in the UK have recently been made.13 The Rugby Injury Surveillance in Ulster Schools (RISUS) was initiated in September 2014. Ulster is one of the four provinces of Ireland and the Ulster Schools Cup competition is the second oldest rugby competition in the world, having first been contested in 1876. The core aim of the RISUS project was to capture key epidemiology data to inform future injury prevention programmes in adolescent rugby. The primary objective was to record and classify injury patterns in schoolboy rugby players according to body part, diagnosis and inciting event. The secondary objective was to determine whether player demographics, biometrics, strength, injury history, level of play or use of protective equipment are associated with risk of injury.

Methods

Study design

This was a prospective injury surveillance study undertaken over a single season (2014/2015). Injury surveillance was conducted between 1 September 2014 and 31 March 2015 inclusive to cover the schools rugby season. All definitions and procedures used in the study aligned with the international consensus statement on injury surveillance studies for rugby.14

Recruitment

In April 2014, all 32 schools entered into the Ulster Schools Rugby Senior Cup 2014/2015 were contacted. This was followed up with an information day that was attended by Headmasters and/or Heads of Rugby from each school. A primary researcher (ATR) visited each school between August and September 2015. A total of 825 male Rugby Union players were recruited from across 28 different schools. To be eligible for inclusion in the study, players must have been part of their school's senior playing squad at the start of the 2014/2015 playing season. Ethical approval was obtained from the University of Ulster Ethics Committee (REC/14/0060). Individual consent was obtained from the players who participated and countersigned by their parents/guardians where appropriate.

Injury definitions

All injury definitions were consistent with the 2007 International Rugby Board (IRB) consensus statement.14 The primary injury definition used was for time-loss injuries, which are defined as ‘any injury that prevents a player from taking a full part in all training and match play activities typically planned for that day for a period of greater than 24 h from midnight at the end of the day the injury was sustained’. Injury severity was defined by the total number of days elapsed from the day of injury until a player returned to full fitness, with full fitness being defined as ‘the player being able to take a full part in training activities typically planned for that day and available for match selection’. Injury severity was classified according to the following subgroups: minor (1–7 days), moderate (8–28 days) and severe injury (>28 days).

Recording of injuries

Details of each individual injury occurring during the 2014/2015 season were recorded using an online reporting system. The following information was inputted for each injury: the date of injury, classification of the injury at two levels (body site, type of injury), information on the injury event, and the date of return from injury. Injury reporting was completed each week by a designated person at each school (data champion), directly onto an online data system. Prior to the start of the season, data champions attended a training evening to outline the study aims and provide a tutorial on data collection procedures. One author (ATR) collated and checked all injury data each week; if schools failed to register any data over the course of 1 week, they were sent reminders via text message or a follow-up phone call. A monthly audit of injuries was undertaken to screen for any unusual patterns of reporting. Data due to illness and/or non-sport-related medical conditions were not included in the study.

Covariates (predictor variables)

At the start of the season, data were recorded from each player under the following categories: demographic/biometric, injury history, use of protective equipment, strength profile, level of experience. Table 1 outlines all 15 covariates and their respective levels. These prognostic factors were initially selected based on current knowledge regarding risk of injury in rugby and consensus from the RISUS group.6 ,7

Table 1

Covariates

Height and body mass were measured on site by the primary researcher using a Height Measure (Marsden HM-250P Leicester Portable) and Mechanical Floor Scales (SECA 869). Participants were asked to remove both shoes and socks prior to measurement of height and body mass. Height and body mass data were rounded up to the nearest 0.01 m and 0.1 kg, respectively. The remainder of covariates in table 1 were collected via questionnaires that were administered and collected on the same visit. The primary researcher remained present at all times to answer any queries participants may have had relating to the questionnaire.

Data analysis

We performed statistical analyses with SPSS software (V.20.0; SPSS, Chicago, Illinois, USA). Descriptive statistics were used to analyse baseline participant characteristics. We presented injury counts and percentages based on severity (minimal, mild, moderate and severe injury), body site, type of injury and injury event. Injury severity was reported as median (days; 95% CI) number of days lost to injury. Match injury incidence was reported as the number of injuries/1000 player-hours of exposure. To do this, we calculated the total number of player-match exposures from each school (number of matches played×number of players in a school squad) and multiplied by 1.16 to provide the total number of match hours.

For the survival analyses, continuous covariates were categorised using 50th centiles. The primary outcome was the time to injury. In this model, the time (in days) from the start of the playing season, to a participants’ first event (injury) or the end of the follow-up period was the main end point. Initially, Kaplan–Meier survival curves were generated and log-rank tests were used to explore survival differences between levels of each covariate. Univariable and multivariable Cox proportional hazard models were then used to evaluate associations between covariates and hazard of injury. During the analysis, all candidate covariates were considered to be on an equal footing. Variables with p<0.1 in the univariable model were then analysed within a multivariate Cox proportional hazard model. Players with any missing data for the risk factors of interest were excluded from the multivariable analysis. HRs were presented with their 95% CI. We considered p<0.05 to be statistically significant. Non-proportionality was checked graphically by assessing Kaplan-Meier survival distribution for each level of the covariate; and stratum-specific log minus-log plots. We also extended Cox regression models to include time-dependent covariates and checked for significance (product of the time variable×covariate). We undertook additional survival models to analyse whether the protective equipment (shoulder pads, head guards, mouthguard) influenced risk of injury at respective body regions (shoulder, head/face).

Results

In total, 825 male rugby union players were recruited across 28 different schools. All participants entered the study at the start of the playing season. The average number of players recruited from each school was 27 (SD 6.3). All participants were male with an average age of 16.8 years (SD 0.8). Table 2 provides a summary of participants’ baseline characteristics.

Table 2

Baseline characteristics

Injury history

At the start of the season, the average number of previous injuries reported was 1.5 (SD 1.56), with n=91 participants reporting at least four previous injuries due to rugby. The four most common previous injuries incurred during rugby were: concussion (n=212/815, 25.7%), ankle ligament/sprain (98/815, 11.9%), knee ligament/sprain (55/815, 6.7%), shoulder sprain (40/815, 4.8%) and posterior thigh muscle injury (24/815, 2.9%).

Injury characteristics and severity (2014/2015 playing season)

A total of 426 injuries were reported prospectively across the 2014/2015 playing season. The three most common injury sites were head/face (n=102, 23.9%), clavicle/shoulder (n=65, 15.3%) and the knee (n=56, 13.1%). Sprain (n=133, 31.2%), concussion (n=81, 19%) and muscle injury (n=65, 15.3%) were the most common diagnoses made. A further breakdown of injury diagnosis by body region and by diagnosis is shown in tables 3 and 4, respectively.

Table 3

Injury by body region

Table 4

Injury by diagnosis

The most common specific diagnoses were: concussion (81/426; 19%), knee sprain (40/426; 9.4%), shoulder sprain/dislocation (38/426; 8.9%), ankle sprain (35/426; 8.2%), hamstring muscle strain (16/426; 3.8%). The average time lost from rugby was 23.8 days (median time loss 23 days (range 1–210 days), SD 19.5 days). Injuries were classified as being minor (n=43, 10.1%), moderate (n=173, 40.8%) or severe (n=208, 49.1%) based on time loss of <7, 7–28 and >28 days, respectively. Almost half of all injuries resulted in an absence of at least 28 days; these were most commonly due to fractures to the hand/finger/thumb (n=11/208, 5.3%), sprains to the ankle (n=14/208, 6.7%) or knee (n=19/208, 9.1%), sprains or dislocations of the shoulder (n=22/208, 10.6%), and concussion (n=33/208, 15.9%).

Injury incidence

The total time lost to injury in this sample was 18 091 days, with an average of 646.1 player days lost per school, per season. We were unable to obtain reliable training exposure; however, injury incidence was estimated for matches. In total, 319 injuries occurred during match situations, which equates to 29.06 injuries per 1000 match hours. Fifteen concussions occurred during training. The remainder (n=66) reported during matches, with an incidence of 6.01 per 1000 match hours. Table 5 shows the estimated seasonal incidence figures per school, for commonly occurring injuries.

Table 5

Seasonal injury incidence (2014/2015 playing season)

Inciting event

Twenty-five percent of injuries (n=107) occurred during training, with the remainder occurring during a match situation. Sixty per cent of injuries occurred in the second half of a given exposure, with 32.9% of all injuries occurring in the third quarter. In 105 cases, the phase of play associated with injury was not reported. However, in the remainder of cases, the most frequent inciting event for injury was the tackle situation, when players were either tackling (117/426) or being tackled (87/426). Sixty-six injuries occurred during collisions, with few injuries occurring during set plays such as scrums (n=7) or lineouts (n=5).

Medical verification

Sixty-one per cent of all injury diagnoses were made by a qualified health professional, such as a physiotherapist (28.6%), emergency medicine practitioner (23.2%), family doctor (8.1%) or school nurse (0.7%).The remainder were diagnosed by a coach (21.4%), parent (1.6%) or self-diagnosis (8.7%). In 32 injuries, it was not clear who made the diagnosis.

Survival analysis

Fifty-seven cases were excluded due to incomplete or missing values for covariates. This left a total of 768 participants in the survival analysis; 290 (35.3%) were injured during the playing season. In total, 478 were right censored; of which 444 remained in the study until the end of the season, with 34 censored at various time points during the season due to dropout. No time-varying effects were observed and the proportional hazard assumption was satisfied for all covariates. Table 6 demonstrates the Kaplan-Meier survival analysis of candidate variables.

Table 6

Kaplan–Meier log rank

Cox proportional hazards model

The univariate and adjusted HR, 95% CI and p value for each variable in the Cox’s proportional hazard model are outlined in table 7. Seven candidate variables were entered into the multivariate model, based on p<0.1 in the univariate analysis. The multivariate model showed that the following factors were associated with a significantly higher risk of injury: higher age (adjusted HR (AHR) 1.448; 95% CI 1.144 to 1.832); higher weight (AHR 1.324; 95% CI 1.039 to 1.686); undertaking regular weight training (AHR 1.654; 95% CI 1.110 to 2.466); and playing representative rugby (in addition to school rugby) (AHR 1.424; 95% CI 1.065 to 1.903). Also, players who were playing for a middle (AHR 0.670; 95% CI 0.497 to 0.904) or lower tier schools (AHR 0.766; 95% CI 0.578 to 1.019) were less likely to get injured compared with players at top tier schools; and players who regularly wore a mouthguard were significantly less likely to get injured (AHR 0.704; 95% CI 0.538 to 0.916) compared with those that did not. There were trends that a history of concussion increased the risk of injury to any body part; however, this was not significant when adjusting for the other important covariates in the multivariate model (AHR 1.266; 95% CI 0.983 to 1.623).

Table 7

Results of Cox proportional hazard analysis

Region-specific protection

Players wearing shoulder pads were at similar risk of shoulder injury compared with those who did not wear any shoulder protection (log rank=0.099; df 1; p=0.753). The risk of head/face injury (log rank=0.327; df=1; p=0.567) and concussion (log rank=0.022; df=1; p=0.882) were similar regardless of head guarduse. However, players who regularly wore a mouthguard were significantly less likely to injure their head/face region (log rank=6.790; df=1; p=0.009) compared with those that did not.

An overall summary of injury data from the 2014/2015 School's Cup season is listed in the online supplementary RISUS infographic.

Discussion

Injury surveillance in schoolboy rugby is essential to inform an evidence-based framework to minimise risk through preventative strategies. This is the first paper to provide a comprehensive description of, and the associated risk factors for, the injuries sustained over a season in a senior schoolboy rugby playing population.

Injury incidence

A recent meta-analysis, based on pooled data from five prospective studies (>4000 participants) reported an injury incidence of 26.7 per 1000 player-hours of exposure in youth rugby.7 Match day incidence in the current study was similar, with 29.1 injuries per 1000 match hours. Although incidence figures are much higher in the professional game, the body parts affected show similarities with the youth game.2–5 In accordance with previous reports, we found a high percentage of injuries occurring in the head, shoulder, knee and ankle regions.2–5 A key finding was the high proportion of severe injuries, and that approximately 90% of injuries resulted in at least 7 days of time loss. Indeed, the total time loss to injury was 18 091 days, with an average of 646 player days lost per school, per season. This is equivalent to 92 weeks of time loss, or each school losing approximately four players for an entire season. Approximately, half of all injuries resulted in more than 28 days’ time loss; with the majority due to fractures of hand/finger/thumb, sprains to the ankle or knee, sprains or dislocations of the shoulder or concussion. Fractures and dislocations were the most severe injuries in terms of requirement for surgical intervention and extensive time lost from participation, which is consistent with previous reports.15 A key concern is that time loss injuries may also necessitate extended school absence and have greatest potential for long-term morbidity; however, this was not measured in the present study.

Concussion

At the start of the playing season, the lifetime prevalence of concussion was 26% which is comparable to other recent reports in schoolboy rugby players.16 In our prospective study, around one in every five injuries reported during the season were due to concussion. The reported incidence of 6.01 concussions per 1000 match hours, or approximately 3 concussions occurring per team per season. These figures are higher than most previous reports in youth rugby.15 Young athletes are more vulnerable to concussion and may be affected by more complicated recovery times and higher risk of adverse outcomes.11 ,12 In addition to the potential of acute catastrophic traumatic brain injury, there have been increasing concerns of potential long-term adverse effects of sports-related concussion particularly when multiple concussive blows are sustained over a lifetime.17–19 A key consideration is that the apparent increase in concussion incidence could relate to better awareness and recognition of concussion/suspected concussion. Recently, there have been a number of educational initiatives aimed at educating the public around the recognition and management of concussion. Prior to the start of the 2014/2015 season, Ulster Rugby in conjunction with the Irish Rugby Football Union (IRFU) delivered a number of roadshow workshops and disseminated educational material aimed specifically at young rugby players in the province. These were informed by recent refinements of concussion evaluation and return-to-play guidelines, and promoted a low index of suspicion for concussion diagnosis and take home messages such as: ‘Stop, Inform, Rest, Return’ and ‘if in doubt, sit them out’.20 Although we did not record data on the nature of return-to-play advice postconcussion, 96% of players returned to play after a minimum of 24 days, which is consistent with the 23 days minimal graduated return-to-play guidelines of the IRFU. This is a positive finding as previous studies have reported that adolescent rugby players poorly comply with return-to-play regulations postconcussion.21

Important inciting events and risk factors

The tackle situation and collisions were the most dangerous facets of play. This aligns with previous systematic reviews where 39.6–64% of injuries occurred because of tackles.6 ,7 In the professional game, high-speed and high-impact tackles carry a greater propensity for injury, particularly collision-type tackles with head-to-head or neck contact.22 ,23 Although we did not examine the specific factors associated with tackles that led to injury it would appear, with the trend of head/upper body injuries, that it is critical that referees protect players by consistently penalising collisions and tackles above the line of the shoulder. These events are more likely to result in injury and are specifically identified in the Laws of Rugby as foul play.24

Body size (height and body mass) has been highlighted as a discriminating factor between successful and less successful teams in the Rugby Union World Cups, 1987 and 2007.25 Mismatch in size is a complex issue in adolescent sports, and is often thought to be an important risk factor for injury.26 Perhaps surprisingly, we found that players with smaller body masses were less likely to get injured. It may be that smaller players actively avoid physical confrontation with larger or heavier opponents, or in certain circumstances they can tailor the type of tackle or contact that they initiate in order to minimise the collision. The trend for an increased risk of injury in higher level, older, heavier schoolboy players who regularly undertake weight training is worrying. It is likely that this increased risk is multifaceted and due to a number of external, personal and behavioural factors. It is likely that these higher level, larger players are used more during games, are higher risk takers, and play at a higher intensity and that their increased injury risk relates to exposure or a higher volume of collisions. With a peak of injuries seen in early season, it is our recommendation that this subgroup of players should have their match load and training modified over this period to lessen injury risk. More young players are aspiring to play professionally and there may be increasing pressure for schoolboy players to increase their mass through the use of various dietary and training strategies. Currently, some schools in Ireland employ professional coaches and/or strength and conditioning personnel. A rapid or disproportionate development of physique and power by a combination of nutritional supplementation and training during rapid growth and development is perhaps a significant contributory factor towards the risk of injury in contact sports, but merits further investigation. Players selected for a provincial squad represent some of the most talented players within the province of Ulster and Ireland; yet they were significantly more likely to sustain an injury. Others have warned that talented players are at greater risk as they are often involved in more than one training squad and are competing at a higher level of competition.27 We did not record physiological training and match load; however, sophisticated inertial sensing systems are available and could be incorporated into future surveillance methods to obtain a more accurate marker of training and match load.

We found mixed patterns on the effect of protective equipment in rugby. Approximately 33% of players reported regular use of shoulder pads or head guards, but there was no evidence that these had any influence on injury risk. This is the first study to report that wearing shoulder pads does not implicate risk of shoulder injury, and a previous randomised controlled study concurs that head guards did not affect concussion risk in youth rugby.28

Previous studies have reported that wearing a mouthguard reduces the risk of dental injuries.29 Our study found 79% players reported regularly wearing a mouthguard, and this subgroup were significantly less likely to get injured. This was an interesting finding, given that these patterns relate to any injury, not just those involving the head or face. It may be that choosing not to wear a mouthguard reflects a wider behavioural characteristic of greater risk taking. Adolescents are a key population in this regard and are much more likely to engage in risky or reckless behaviours during sport, with many harbouring a sense of invincibility.10 Others note the importance of sociocultural factors within paediatric sport and recreational injury risk.30 These patterns perhaps suggest that future studies in youth injury should look beyond biological risk factors.

Study strengths and limitations

RISUS is the first injury surveillance system across the UK and Ireland in schoolboy rugby and highlights some the challenges of collecting epidemiological data in schools rugby. Findings are based on a large cohort of players competing in the same cup competition, over an entire playing season. At the start of the study, our sampling frame was all 32 schools and 28 were successfully recruited (87%), of which 1 dropped out mid-way through the playing season. Injuries were reported prospectively, and generally inputted on a weekly basis. The injury incidence of 29.1 injuries per 1000 match play hours is likely to be lower than the true figure due to varied compliance across schools and potentially under-reporting of injury. Candidate risk factors were selected through consensus from our research group. Candidate variables were reported at the start of the season and it is likely that some of the biometric data could change over the course of the playing season. We were unable to record individual player exposure or team training exposure. Medical verification occurred in 61% of injuries. The remainder were verified through the player, the coach or a parent; although this may affect the accuracy of the diagnosis, this is not likely to affect the validity of other key injury data, for example, body part or time to return to play.

What are the findings?

  • Rugby injuries have increased in adults since the introduction of professionalism in rugby union in 1995.

  • A large number of young people play rugby across the UK and Ireland, but there is a concern that the risk of injury is too high.

  • Concussion may be under-reported in youth rugby.

How might it impact on clinical practice in the future?

  • There was a high incidence of severe injuries in adolescent rugby players.

  • Concussion, ankle and knee ligament injuries and upper limb fractures/dislocations caused greatest time loss.

  • One in five injuries was a concussion, with players generally compliant with return-to-play regulations.

  • Larger physical stature and higher levels of competition are important risk factors for injury.

  • Head guards and shoulder pads were not shown to protect against injury in this study.

  • Future risk models should consider sociocultural factors associated with risk taking.

  • Please refer to the Rugby Injury Surveillance in Ulster Schools (RISUS) infographic for a summation of results.

References

View Abstract

Review history and Supplementary material

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors HAPA, ATR, MW, RN, NWAE, RKW, LAH, GJH and CMB were all substantially involved in the conception, consent process, data collection, data upload, data interpretation and drafting of this paper.

  • Funding Financial support was received by the following institutions: Irish Rugby Football Union, MITRE Trust and Ulster Rugby.

  • Competing interests None declared.

  • Ethics approval University of Ulster Ethics board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

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.

Linked Articles