Injuries in Netball-A Systematic Review

Background Netball is estimated to be played by more than 20 million people worldwide, but there is evidence of high injury incidence. A thorough understanding of the types and rates of netball injuries is essential for effective injury management and prevention strategies to be developed and implemented. This systematic review summarises the published findings with respect to injury types, participant characteristics and any identified risk factors for netball injuries. Methods A librarian-assisted computer search of seven scientific databases was conducted for studies reporting on netball injuries. Inclusion criteria were studies published in English, in peer-reviewed journals, which reported data on injuries and variables (e.g. age and competition level) that have been proposed as possibly associated with netball injury risk. Results Forty-six studies (43.5% prospective, 37% hospital/insurance records, 19.5% retrospective) from 45 articles were included after screening. The majority of studies (74%) were conducted in Australia or New Zealand. There was little consistency in the definition of ‘injury’. Elite or sub-elite level players were included in 69% of studies where the level of competition was reported. The duration of injury surveillance was generally related to the format of competition from which data were collected. Self-report questionnaires were used in 48% of studies and only 26% of studies used qualified health professionals to collect data courtside. Injuries to the ankle and knee were the most common (in 19 studies) although the incidence varied considerably across the studies (ankle 13–84% and knee 8–50% of injuries). Prevention of ankle and knee injuries should be a priority. Children sustained more upper limb injuries (e.g. fractures) compared with adults who sustained more lower limb injuries (e.g. ankle and knee sprains/strains). A large number of potential risk factors for injury in netball have been investigated in small numbers of studies. The main circumstances of injury are landings, collisions and falls. Conclusion Further studies should be directed towards recreational netball, reporting on injury incidence in players by age and utilising high-quality, standardised methods and criteria. Specific injury diagnosis and a better understanding of the circumstances and mechanisms of injury would provide more meaningful data for developing prevention strategies. Supplementary Information The online version contains supplementary material available at 10.1186/s40798-020-00290-7.


Introduction
The International Netball Federation (INF) has estimated that more than 20 million people play netball in more than 70 countries [1]. Netball is played predominantly in British Commonwealth countries and usually by females, although there is growing participation by males in both male and mixed competitions [2]. Players typically range in age from 5 to over 50 years and games are played on both indoor and outdoor courts, with a variety of court surfaces [3].
While netball is a sport with one of the highest participation rates, there is also evidence of high injury rates, with comparisons being made to other team sports including all football codes, basketball, field hockey and cricket [4][5][6][7][8][9]. Injuries are a concern in netball and injury prevention is a goal of the sport's governing bodies (INF; Netball Australia [10]). Injury surveillance is the first stage of the Translating Research into Injury Prevention Practice framework (TRIPP) [11]. Surprisingly there has been no published review synthesising the available data on netball injuries alone. A thorough understanding of the types and rates of netball injuries is essential in order to develop appropriate injury prevention and management strategies. Therefore, the purpose of this systematic review is to identify, collect and synthesise all the published data on netball injuries with respect to injury types, participant characteristics and any identified factors associated with the risk of injuries in netball.

Methods
The conduct and reporting of this review adheres to PRISMA guidelines [12].

Study Eligibility Criteria
Studies were included if they reported data on netball injuries sustained by netball players of any age, sex or competition level, were full-text peer-reviewed articles and published in English. Included study designs were observational and experimental. Case studies, reviews and conference abstracts were excluded. Included studies reported data on netball injuries in general and/or in relation to variables such as age, competition level or any other factors associated with netball injury or injury risk. Studies were included regardless of injury definition; i.e. any injury definition was acceptable provided the injury was related to netball matches or training.

Search Strategy
A biomedical librarian assisted with the development of the search strategy for a systematic search of the electronic databases Cinahl, Medline, Embase, Cochrane, Informit Health, Scopus and SPORTDiscus from 1985 to 20th January 2020 (completed on 20th January 2020). The search was limited to 1985 as the paper by Hopper [13] is regarded as the first peer-reviewed publication of a study of netball injuries. In an effort to be as comprehensive as possible, the search strategy used only the terms Netball* AND Injur*. A secondary search of the reference lists of included papers was conducted to identify additional articles for possible inclusion.

Study Selection and Quality Assessment
After removal of duplicates, two reviewers independently screened titles and abstracts. If there was uncertainty regarding inclusion of identified articles, the full text was obtained and screened. Discrepancies were first independently re-assessed by each reviewer, then resolved following independent assessment by a third reviewer. Agreement between reviewers was calculated using Cohen's Kappa [14]. The quality of the studies was assessed using the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [15].

Data Extraction and Synthesis
Data from the included studies were extracted by one reviewer and verified by a second in a database template. Data included quality assessment information, country of origin, participant characteristics (age, competition level), data collection period and methods, injury definition, player exposure, injury rate or proportion, injury site, type and severity and key study findings. Results of studies were summarised qualitatively for comparison. Quantitative assessment was limited to tallying the number of studies reporting a particular issue.

Study Selection
The search identified 434 articles after duplicates were removed. After screening, 45 articles were included in the review (Fig. 1). The inter-reviewer agreement for the title/abstract and full-text screenings was considered to be good (κ = 0.59, 95% CI: 0.51-0.68) and very good (κ = 0.87, 95% CI: 0.74-1.01), respectively. The main reasons for exclusion were that no injury data were reported or publication was not peer-reviewed. Only two articles [13,16] were published between 1985 and 1989, with 11 in the 1990s, 14 in the 2000s, and 18 in the 2010s. One article [17] reported on two studies (one prospective, one retrospective); therefore, there were 46 studies. Langeveld et al. [18] and Coetzee et al. [19] used data from the same multiday tournament study. Hopper and Elliot [17] and Hopper [20] also used the same data source as one another but they reported on different research aims. There was overlap in the data used in Stevenson et al. [21] (one season) and Finch et al. [22] (two seasons). Also, there was an overlap in the data used in the articles of Hume [23] and Hume and Marshall [24].

Publication Quality
All studies included in this review had a clearly stated research question or objective (Fig. 2) and most of the studies (96%) clearly defined their study population. The participation rate of eligible players was adequate (i.e. > 50%) in 71% of studies. The time frame of exposure to netball was reported in 84% of studies. Sample size justification was reported in only 49% of studies. Exposure levels were clearly defined and determined using valid and reliable methods in 20%, but only reported in 18% of studies. Statistical adjustment (e.g. regression) for confounding variables (e.g. playing and training time) was adequate in only 9% of studies. The quality assessment results per individual study and criterion are detailed in Supplemental Table 1. Table 1 provides a summary of country of origin, sample size, participant age, competition level, and the type, duration and methods of data collection in the included studies.

Study Categories
The studies are presented in Table 1 in four categories: prospective studies over one or more playing seasons (n = 14, 30%), prospective studies at tournaments (n = 6, 13.0%), retrospective studies (n = 9, 20%) and studies accessing database sources, including hospital and insurance company records (n = 17, 37%). One article [17] reported both prospective and retrospective studies and the findings from these data are reported in their respective sections.

Number of Participants and Injuries
Among prospective studies during competition seasons, participant numbers ranged from 17 to 11,228, with 50% (n = 7) of studies having< 100 participants. Injury numbers ranged from 11 to 608, with 43% (n = 6) reporting < 25 injuries. In prospective studies over the course of a tournament, participant numbers ranged from 192 to 1280, and injury numbers from 52 to 205. In the retrospective studies, participant numbers ranged from 16 to 1084; 88% (n = 7) of studies had < 254 participants. Injury numbers ranged from 16 to 504 with most studies reporting < 100 injuries. Injury numbers in the hospital, clinic and insurance records studies ranged from 3 to 5, 263 with nine of the studies reporting > 1000 injuries.

Age of Participants
Age range or mean age were reported in most studies. Among the prospective studies, only Pringle et al. [34] focused on children < 16 years, whereas many studies [13,17,20,28,29,36] included participants > 12 years old, with the mean age typically in the 20-25-year range.
In the retrospective studies, Smith et al. [43] included only those 6-16 years and Whatman and Reid [45] investigated secondary school participants (16 ± 1 years) whereas the other studies covered wider age ranges [45].
Hassan et al. [50] in their hospital ED admission study focused on children 5-15 years whereas the other large database reports included both children and adults. However, these studies limited their reporting of results related to age [3,9,17,42,48,54,55] to only 'child versus adult' limiting the extraction of injury by specific age brackets. Although some studies did not report age specifically, it was inferred from the level of competition, which was often played in age categories in those studies.

Sex of Participants
Seven studies indicated that they included males, who represented 5.7-10.5% of hospitalisations [21][22][23], and 2-19% of injuries [52] [48,52,55]. No other information was provided on the sex of participants, although it was presumed to be only female in most studies, based on the level of competition or competition categories.

Data Collection Methods
In the prospective studies, physiotherapists diagnosed and recorded injuries in 10 (50%) studies [13, 17, 20, 25-29, 33, 37], although in three of these studies, only elite players' data were provided by physiotherapists, with non-elite players' data provided by selfreport [25,33,37]. Five (25%) of the prospective studies [21,22,30,32] used only self-report data, and training injuries were often dependent on player notification to a therapist [29]. In the other prospective studies, physiotherapists or students/trainers/first Hospital inpatients related to netball injury; data also linked to injury entitlement claims Participation in netball estimated from national survey data.

Comparison of Injury Rates in Matches, Tournaments, Training and Warm-up
Hopper et al. [29] reported no injuries during training, Attenborough et al. [25] reported nine ankle sprains during matches compared with two during training, and Chong et al. [47] reported three ACL injuries during matches and one during training. Langeveld et al. [18] found that 95% of injuries occurred in matches, 3% in warm-up and 2% in practice and Joseph et al. [3] reported that 92% of injuries occurred in matches. Hume [23] also commented that injuries during warm-up (3%) and practice (1.5%) were low compared with matches. Langeveld et al. [18] noted that injury rates were higher in tournaments compared with regular season competition.

Injury Sites
Forty-three studies (96%) provided some data on injury site in terms of body area. There were various methods of categorising body areas among the studies, some more specific (e.g. ankle sprains, Achilles injury), others more generalised (e.g. lower limb injury) ( Table 2). Fifteen studies (33% of all 46 studies) were limited to specific body sites or regions based on the aim of their study and eight of those (17% of all studies) were limited to a single structure or injury type, for example, ACL   Intervention reduced training injuries but not match injuries or re-injuries. Number needed to treat to reduce lower limb training injuries was 3.

Severity of Injury
Sixteen studies (34.8%) used descriptive categories, for example, mild, minor, moderate, severe, new, recurrent and overuse. Many authors stated that most netball injuries are minor, but the rates and definitions of minor, moderate and severe injuries varied among studies. Ferreira et al. [27] reported rates of 35% minor, 56.5% moderate and 8.7% severe, and Finch et al. [22] reported 60-70% of injures as moderate severity, indicating that treatment by a health professional was required. Stevenson et al. [21] indicated that 20% of netball injuries were moderate to severe but few required hospitalisation. Hopper and Elliot [17] used a grading system modified from Oakes (1981) [56] and reported 71% of tournament injuries were Grade 1, 15.4% Grade 2, and 13.5% were Grade 3 which included fractures. Langeveld et al. [18] found 70% of injuries were minor (not missing a game), 11.4% resulted in missing 1-2 games, and 2% resulted in missing 3-5 games. Hopper [13] found that knee injuries resulted in greater disability than other injuries, which was supported by Hopper et al. [29] who also found that there were more severe knee injuries than ankle injuries. Hopper et al. [28] reported that 15% of ankle and 27% of knee injuries required referral to hospital and that 10% of ankle injuries were fractures. Whatman et al. [45] reported that 10% and 24% of players had substantial knee and ankle problems, respectively. Otago et al. [54] concluded that injury costs were highest for knee injuries. McKay et al. [31] found there were more severe injuries in netball than basketball, and Finch et al. [39] reported that children were more likely to require treatment by a health professional than adults. The hospital/clinic/insurance records studies reported primarily on more severe injuries, although Kirkwood et al. [8] noted that there was only one admission to hospital for a severe injury related to netball in their study of 157 injuries. Finch et al. [6] reported fractures were 22% of injuries in children and 13.3% of injuries in adults. Flood et al. [48] found 29.5% of netball hospital presentations were fractures, 27.8% strains/sprains and 10.1% dislocations, which was similar to hospital and emergency department (ED) presentations reported by Hume [23]. Smartt et al. [55] found 32% of netball hospital admissions were due to fractures. Joseph et al. [3] and Kirkwood et al. [8] found 15% and 19.5% of injuries due to fractures, respectively, and King et al. [52] reported 16.7% of injuries were due to fractures or dislocations.

Timing of Occurrence of Injuries
Ten (21.7%) studies reported on some aspect of timing of netball injuries. Five studies (10.9%) included data on the time and/or the quarter in the game in which injuries occurred [3,13,18,23,28]. Hopper et al. [28] reported that most injuries occurred in the 1st quarter of matches, while Hopper [13] reported that more new injuries occurred in the 1 st quarter whereas more recurrent injuries occurred in the 2nd quarter. Langeveld et al. [18] reported that most injuries occurred in the 2nd and 3rd quarters with a linear increase from the 1st to 3rd quarters. Stuelcken et al. [44] investigated ACL injuries in 16 players and found that these injuries occurred most frequently in the 1st and 4th quarters (n = 6 and 4, respectively). Hume [23] found no difference between injury rates in the first (43.3%) and second (44.8%) halves of matches. Joseph et al. [3] found slightly more injuries in the 2nd quarter but the main trend was for there to be less injuries in the 4th quarter compared with the other three quarters. In reporting injury incidence in four sports, including netball, Stevenson et al. [21] found there was an overall elevated level of injury incidence in the first month of the season with an injury rate of 28.9 injuries per 1000PH compared with the season mean of 12.1 injuries per 1000PH.

Environmental and Surface Conditions
Four studies (8.7%) reported on playing surface in relation to injury frequency. Studies reported a lower injury incidence on synthetic surfaces compared with concrete [19], bitumen [13] and grass [13]. Pillay et al. [41] and Joseph et al. [3] stated that the most common surface on which injuries occurred was outdoor concrete, and Coetzee et al. [19] reported this was the surface associated with most (88.9%) serious injuries. Purdam [16] also suggested that the training surface may contribute to injuries.

Player Age and Injury Risk
Fifteen studies (32.6%) included data relating injury incidence to age. Most studies reported a higher injury incidence with increasing age [3,9,17,21,37,39,42,43,48,54,55], one found more injuries in younger players [36] and in others it depended on the specific injuries being investigated [6,13,48,54,55]. Hopper [13] found more hand and knee injuries in junior players. Finch et al. [6] found more upper limb injuries in children (54.5%) compared with adults (27.5%) but more lower limb injuries in adults (63.9%) compared with children (37.3%). Finch et al. [39] noted that children had a lower overall injury rate compared with adults. Flood et al. [48] found forearm fractures were highest in the 5-14-year age group, ACL ruptures highest in the 25-34-year age group, and Achilles ruptures highest in the 35-44-year age group, and Otago et al. [54] stated that Achilles injuries occur primarily in those over 25 years.

Competition Level and Injury Risk
A higher injury incidence was reported in elite or higher grade players compared with lower grade players [13,28,29,41]. Hopper [13] reported that higher grade players had more injuries than lower grade players at both junior and senior levels. Pillay et al. [41] found that statelevel players had a higher incidence of ankle and knee injuries than both club level and national level players.

Player Position and Injury Risk
Seven studies reported on the specific position or a group of positions (attacking or defending) being played at the time of injury [17,18,23,28,41,42,44]. The player positions more likely to be injured varied with the study. Stuelcken et al. [44] reported that ACL injuries varied with court position with more injuries in the attacking third (44%) and centre third (37%) compared with the defensive third (19%). Hopper and Elliot [17] also reported that more injuries (54%) occurred when playing an attacking position. In contrast, an earlier study [13] reported no association between player position and injury.

Other Factors Associated with Increased Injury Risk
A number of factors were reported as increasing risk of injury. Attenborough et al. [25] reported 4.04 times greater odds of sustaining an ankle sprain during netball if a player's preseason performance on the posteriormedial direction of the Star Excursion Balance Test (SEBT) was ≤ 77.5% of their leg length. Pickering Rodriguez et al. [33] found a relationship between higher soleus and Achilles stiffness (assessed with a validated hop test on a force platform) and higher incidence of lower body injuries in elite netballers. Smith et al. [43] reported significantly increased injury incidence in junior netball players with higher levels of hypermobility. Also, having a lower limb asymmetry in turn performance > 10% was identified as increasing risk of a lower limb injury [30]. Hopper et al. [29] found that more athletic players, as evidenced by higher vertical jump and aerobic fitness, and less endomorphy, were at higher injury risk. Ferreira et al. [27] found that excess body fat, poor balance and agility, and limited explosive power may have contributed to injuries. Chronic ankle instability contributed to increased risk of recurrent ankle sprains in recreational players [38]. A lack of warm-up [32,36] and a previous injury [19,32] were also identified as increasing injury risk. Coetzee et al. [19] reported that a lack of core stability and poor utilisation of neuromuscular and proprioceptive training increased injury risk. Hume et al. [36] reported that players failed to adopt recommended injury prevention strategies, such as using wobbleboards, using ankle supports or wearing high-cut netball shoes. Further, 54.7% reported finishing the game before seeking treatment [36]. All of these behaviours may have contributed to higher risk of injury or more severe injury. Smyth et al. [37] reported that 27% of players arrived at their tournament with an existing injury or illness.

Factors Not Associated with Increased Injury Risk or Protective
Hopper and colleagues investigated a number of podiatric variables [17,40], as well as the somatotype [20] and the landing techniques of players [17], and found no relationships with injury risk. Flexibility training was also found to be of limited benefit in preventing injury [19]. Whatman et al. [45] found no relationships between movement competency, landing technique, ankle range of motion or jump performance on overuse knee and ankle injuries. McManus et al. [32] found that not having an injury in the previous 12 months and training more than four hours per week were protective against injury.

Comparison of Netball Injuries with Other Sports
A number of studies compared netball injury presentations to those of other sports. McKay et al. [31] reported that netball had 3.3 times more severe injuries compared with basketball. Finch et al. [22] found netball had a lower injury rate than Australian Rules Football (AFL), field hockey and basketball. Finch et al. [39] reported less injuries with netball compared with cricket, horse riding and soccer but more than AFL, basketball and tennis. Fernando et al. [9] compared the top 20 sport and recreational activities in terms of injury-related ED presentations; netball ranked 13th in terms of injury rate and 19th in terms of cost per presentation. Kirkwood et al. [8] found netball ranked 2nd to trampoline for sports-related ED attendances in 0-19 year-old females. Cassell et al. [46] found netball ranked 4th in sportsrelated ED and general practitioner presentations after AFL, cycling and basketball in Australia. Hume et al. [24] also found netball ranked 4 th in presentations after rugby union, rugby league and soccer in New Zealand. Hassan et al. [50] found netball was responsible for the largest percentage of fractures presenting to the ED in girls after rollerblading, and Hon et al. [51] found netball accounted for 33% of sports related fractures in females. Gwynne-Jones et al. [49] found netball accounted for 31% of all sports related Achilles injuries and 54% of all Achilles injury ruptures at hospital in females aged 15-40 years. Love et al. [53] found netball was one of the top 10 sports responsible for sports-related dental insurance claims. More recently, King et al. [52] compared the moderate to serious sports-related insurance claims by females across five sports (netball, rugby union, rugby league, football, cricket) and found by far the most claims were for netball but the relative of risk of injury was lowest in netball. Mean cost per claim was 4th in netball, only higher than football for moderate to serious claims but 2nd to rugby union in serious claims.

Interventions to Reduce Injuries in Netball
Only two studies [26,35] reported on the outcomes of an injury reduction program. Elphinston et al. [26] used an integrated multidisciplinary sports science and sports medicine approach incorporating screening and evaluation of players followed by individualised technical training and conditioning to achieve a reduction from 1.3 to 0.2 injuries per player. Zulkarnian et al. [35] found that the Down to Earth safe-landing technique program reduced training injuries but not match injuries.

Discussion
This systematic review presents the first summary of the published literature on netball injuries. Forty-six studies published in 45 papers met the inclusion criteria. Twenty studies were prospective, nine were retrospective, and 17 were analyses of hospital/clinic records or insurance claims. Not surprisingly, all the studies were from British Commonwealth countries with most conducted in Australia, New Zealand or South Africa. Injury rates over a netball season of 11.3-14/ 1000PH were reported in recreational players whereas in elite players 19.35/1000PH was reported [33]. Injuries were predominantly to the lower limb, particularly the ankle, with ligament sprains the most common injury type. One study reported younger players had more injuries to the upper rather than lower limbs [6]. Severe injuries (those resulting in hospitalisation) were typically fractures, especially in children, and ACL or Achilles ruptures in adult players. There were few head injuries reported in netball. Player characteristics, such as age and level of competition, may influence injury rates. There are few data on male netball players, although the game is still played predominantly by females. Larger prospective studies would be beneficial to further improve our understanding of possible risk factors for injury and the types of injury in netball.
A number of points are clear from the studies included in this review. Firstly, 87% of the studies were conducted in Australia, New Zealand or South Africa, with very little representation from the other 73 countries who have netball organisations. Australia and New Zealand have, in recent times, been the two dominant nations in international netball competitions. This success may reflect the high proportion of these nations' young and recreational populations participating in netball. High participation incurs an associated injury impact and cost, and therefore, possible available funding from governments to research injury incidence, type and risk factors in netball.
Second, most studies included players across a wide age range with little reporting of the results by age. Children (< 16 years) were included in 92% of the database studies, 63% of the retrospective studies, but only 39% of the prospective studies. It is generally acknowledged that hospital and insurance data do not provide a complete picture of the volume or type of injuries that occur in a sport but do identify those that are the most serious [6]. Finch et al. [6] provide the most comprehensive hospital emergency department data comparing child and adult injury rates in netball. They reported children experiencing a much higher percentage of upper limb injuries (54% vs 27.5%) and fewer lower limb injuries (37.3% vs 63.9%), more fractures (22% vs 13.3%) and fewer sprains/strains (43.5% vs 57.5%) compared with adults. Only Smartt et al. [55] reported their hospital admission results across a number of age ranges. Among those < 18 years, they identified a higher injury rate in the 10-14-year group; however, most injuries occurred in those 25-39 years with the highest number of cases in the 30-34-year age group. Less serious injuries, which are the most numerous, are usually managed in other ways and at non-hospital locations. Pringle et al. [34] focused on children (6-15 years) in their prospective study and concluded that the incidence of injury in children playing netball is low. There is clearly a need to improve our understanding of less severe injury types and rates by age in netball.
Third, competition level was reported to influence injury rates, which has also been described in other sports [57,58]. Studies of elite players are often overrepresented relative to less competitive players in sports injury studies [59]. Elite or sub-elite players were the focus of 63% of the retrospective studies, 50% of the prospective studies but only one (8%) of the database studies. Four studies [13,28,29,41] reported that injury rates were higher in players in elite compared with lower competition levels. One limitation when comparing injury rates in elite vs recreational players is that injury data collection methods may differ with competition level, even within the same study [25,33]. No studies have investigated whether injury rates or types vary with competition levels when elite/sub-elite players are excluded.
A number of potential risk factors for injury in netball, in addition to age and competition level, have been investigated, mostly in only one or two studies. Those reporting increased risk of injury included poor balance [31,44] and agility [44], joint stiffness [40] or hypermobility [45], lower limb asymmetry [33], chronic ankle instability [35], lack of core stability [20], lack of warm-up [39,43], previous injury [20,39] and playing when injured [32,43], and not adopting recommended injury prevention strategies [20,43]. Whether player position influences risk of injury is not yet clear [16,18,19,24,29,37,54]. Further studies are required to determine the impact of these factors on injury risk and whether interventions targeting these factors can reduce injuries.
The method of injury data collection used in the studies varied, with many studies using unqualified personnel, for example students or team managers, and this reporting was particularly influenced by the level of competition of the players. In the season-competition prospective studies, those conducted on recreational players only obtained their data by self-report [21,22,32]. In contrast, in studies of elite players, almost all injuries were assessed and recorded by physiotherapists [17,20,[26][27][28][29]. In studies investigating injuries across a broad range of competition levels, some were recorded by physiotherapists [13,28,29] while others were selfreported or recorded by physiotherapy or sports science students or first aid staff [18,19,25,31,36]. In the tournament prospective studies, medical staff contributed to injury recording in two studies [17,20] whereas a range of personnel provided the treatment or completed the injury records in the other studies [18,19,36]. Thus, there is uncertainty in the accuracy of injury records obtained from some of these varied methods. The method of data collection most likely to capture all injuries that occur on courts in competition (at all levels) is to have data collected on the courts at the time of play by appropriately qualified sports medicine professionals, with follow-up and specialist referral when necessary. To increase the accuracy and depth of injury data collected, future studies should ensure that all personnel collecting data are suitably qualified and trained in standardised injury reporting, and/or that each injury is reviewed by a qualified health professional.
Injury rates were reported using a number of different metrics and some studies focused on particular types of injury only; consequently, consolidating the findings on injury rates is limited. Some of that inconsistency may be related to the fact that 13 of the articles included in this review were published before 2000 [6,13,16,17,20,23,24,28,29,31,34,40,53], and two before 1990 [13,16]. Reporting injuries by exposure (/1000PH) was limited in the studies included in this review but data from four prospective studies [21,22,32,34] in recreational players provided consistent values of 11.3 to 14 injuries/ 1000PH. Injury rates in tournaments or in more elite players appear to be higher. Pickering Rodriguez et al. [33] reported injury rates of 19.35/1000PH for elite players compared with only 7.13/1000PH for sub-elite players. Higher injury risk at higher levels of competition appears to be consistent with several other sports (e.g. tennis, soccer, and rugby union 15 s) [60][61][62].
Future studies should aim to use standardised methods, such as those outlined by Fuller et al. [63] in their consensus statement for the collection of injury data in rugby union, to further build on the existing evidence provided by the studies in this review. This consensus statement provides operational definitions for injury, severity, injury recurrence, match and training exposure, as well as recommended methods for collecting and classifying injury data. The development of a consensus statement by the International Netball Federation would assist in this process.
Although the proportion of injuries to each body region varied among studies, most netball injuries in those over 15 years were to the lower limb, particularly the ankle and knee. Ankle injuries were far more frequent than knee injuries with the exception of Maulder et al. [30]. There were reports of injuries to other parts of the lower limb [17,18,30,33,36,40,41] and back [17,20,22,26]; the upper limb, particularly the hand [13,18,28,32,34,43,50,55], wrist [18,32,34,42,55] and forearm [55]; and relatively smaller numbers of injuries to the head, face or neck [22,23,26,27,31,36,48,53,54] although those to the head or face, including dental injuries, were evident in insurance claims [23,53]. Interestingly, while only four studies included in this review [8,22,37,52] provided any data on concussion injuries in netball, a report on sports-related concussion identified that the rate of concussion in netball has been increasing [64]. Netball Australia has developed a concussion policy [65] based on the Concussion in Sport Australia Position Statement [66] and the Consensus Statement of the International Conference on Concussion in Sport [67]. This concussion policy may be justified as 'collision' is a frequently cited circumstance of injury in netball, and females may be more vulnerable to and take longer to recover from a concussion [66].
It was not possible to determine the extent to which there are multiple injuries associated with a single injury incident. For example, it is possible that slips, trips, falls or collisions could result in injuries to both upper and lower limbs and possibly the head. It is not clear in the prospective and retrospective studies whether all injuries are recorded or only the most severe or memorable injuries. Similarly, in hospital records and insurance claims, it is likely only the most severe injury is recorded.
There has been limited standardisation of reporting of COI. In the Coetzee et al. [19] study, 60.8% of injuries were attributed to 'physical contact' and Hon and Kock [51] also used this terminology, as did Pringle et al. [34] and Hopper et al. [29]. 'Collision', a similar term, was used by Singh et al. [42] and McKay et al. [31]. Landing 'incorrectly' and 'awkwardly' were used by a number of authors but without a clear definition of a 'correct' landing. A more specific system of describing COI may lead to further associations with specific injuries and, as a result, enable equipment, coaching and training strategies to minimise these injuries. This should include intrinsic and extrinsic factors as well as differentiating between acute and overuse injuries [16]. It should also be noted that many studies used questionnaires to gather this information from players, which introduces recall bias and player opinion about what their COI might have been. Players may need training in the terms to use to describe these circumstances to improve consistency of description. Stuelcken et al. [44] was the only study to review video footage of games, providing the opportunity to assess possible mechanisms of ACL rupture, in addition to COI. Injury risk is likely a multi-factorial phenomenon; however, if the circumstances associated with injuries are reported in a detailed, standardised manner, then further studies can examine these more closely. These findings could provide the basis for strategies to be developed for modifiable factors, such as neuromuscular control in ACL injuries in a range of sports [68] and implemented in training for the purpose of reducing the incidence of those injuries.
Currently, most recreational netball competitions are played outdoors on concrete/asphalt 'all weather' courts, and so players are exposed to a variety of weather conditions that may influence the type and incidence of injuries [3]. Elite competitions are also often played on these courts, as well as indoors on sprung timber courts, which may produce different injury profiles. Indoor competitions mostly play on synthetic surfaces over concrete bases, again possibly affecting the pattern of injury and frequency. Playing surface has been shown to influence injury rates in netball [3] and other sports [69][70][71]. Although concrete playing surfaces may be a risk factor for injury in netball [3,16], data on playing surfaces were limited in the studies in this review and so the impact of playing surface type on injury incidence needs further investigation.
Training data should be reported in future studies to clarify which practices associated with training are associated with injury risk. This should include the type, duration, intensity and frequency of training activities, as well as any injuries that may occur during training. These studies may then be able to identify training practices associated with high injury risk, or the effect of changed training loads on the incidence of injury during games, enabling coaches and trainers to modify training practices and reduce injury risk.
The main strengths of this review are the comprehensiveness of the search, the good agreement between reviewers for inclusion of studies, the stratification of the studies into categories and the rigour of data extraction. The main limitations are those that relate to the quality and extent of reporting in the included studies, particularly differences in injury definitions, limited or lack of clarity regarding the use of health professionals to collect data or verify injury diagnosis in recreational players, reliance on self-report data in many studies, and limited reporting of injuries by age or sex.
Recommendations for research and practice are 1) for data on recreational players to incorporate a diagnosis by a qualified health professional; 2) for netball injury data to be collected by people trained in injury reporting and ideally stored in a centralised national database; 3) to report data by age group, especially among those < 18 years of age, such as (a) < 10 years, (b) ≥ 10 ≤ 15 years and (c) > 15 ≤ 17 years; 4) to provide data on male players separately if included; 5) to provide more detail on the types of head injuries sustained by netball players, including the extent and severity of concussion injuries; 6) to provide greater documentation of injuries relative to training and/or match load; and 7) to provide stronger evidence of risk factors, particularly in relation to circumstances and mechanisms for netball injuries that can be targeted by prevention strategies and provide teachable suggestions for coaches. This would improve the data available to netball clubs, organisations and sports medicine personnel working with netball communities, and enable them to provide appropriate resources and equipment to manage injuries and reduce injury risk in this large, varied population of netball players.

Conclusions
This review found that most netball injuries are to the lower limb, particularly the ankle, and that injuries to the upper limb have a higher rate in those < 16 years old. Severe injuries are fractures in younger players and ACL and Achilles injuries in adults. Injury rates appear higher in elite players than recreational players. Factors often associated with injury include circumstances such as landings, slips/trips/falls, collisions/player contact and being hit by the ball. Further studies should be directed towards recreational netball, reporting on injury incidence in the under 15-year population and utilising high-quality, standardised methods and criteria. Specific injury diagnosis and a better understanding of the link with circumstances of injury would provide more meaningful data for developing prevention strategies. Further classification of data into acute/ recurrent/chronic and timing of injuries may provide links to elements in training and games which increase or decrease injury risk. Studies identifying successful injury prevention strategies are also needed.