Journal of the New Zealand Medical Association, 23-January-2009, Vol 122 No 1288
Obstructing the goal? Hospitalisation for netball injury in New Zealand 2000–2005
Pam Smartt, David Chalmers
Netball is a fast, demanding team game characterised by sudden direction changes, rapid acceleration, pivots, and elevating leaps.1,2 Injury is common and is reported to be one of the main reasons for young people dropping out of the sport.3,4 Reducing the risk of injury in netball is therefore an important goal for the sport, for New Zealand’s Accident Compensation Corporation (ACC) and, more generally, for public health.5
For the sport, injury affects performance and participation, for ACC it results in compensation claims, and for public health injury can negate the health benefits of physical exercise.6
Netball is the most popular team sport for women in New Zealand (NZ) and is played widely by women of all ages. Participation continues well into the childbearing years when active involvement by females in other sports is generally declining.7 Netball is popular with Māori who make up 26% of netball participants.7 The game is also played by a growing number of males,8 with 81% of male participants involved in the indoor game.7
In 1989–1990 there were an estimated 100,000 netball participants in NZ.9 According to Netball New Zealand, there are now an estimated 200,000 participants: 120,000 registered members and a further 80,000 playing in social competitions nationwide.8 This estimate is likely to be conservative as in 2003 SPARC7 estimated that 255,700 New Zealanders (club and non-club members/informal participation) aged 5 years and over played netball.
The last descriptive epidemiological study of netball injuries in NZ was published in 1993 (data from 1988).9 There have been several significant changes in the sport in that period. Netball became a recognised Olympic sport in 1995 (making its inclusion in future Olympics possible), and has grown in popularity in NZ both as a participant and a spectator sport. It is acknowledged that netball today is a faster, more physical game and that there is a closer contest for the ball. Attitudes toward sports injury have also changed and there has been a national netball injury preventive programme NetballSmart in place since 2005.10
The aim of this study was to examine the current profile of hospitalised netball injury and identify any the changes that have occurred since 1988. We have limited the study to injury treated in the public hospital system because our main focus is serious acute injury.
Study design and definitions—This study was undertaken as part of a larger study of hospitalised injury across 187 sporting and recreational activities.11 A novel search procedure was developed to identify injury cases associated with specific activities. The procedure comprised a search of both coded and narrative information in a linked hospital discharge and ACC entitlement claim dataset.
Sporting and recreational activity was defined as any pastime or game requiring physical effort undertaken for “amusement, diversion or fun” (Concise Oxford Dictionary). NZ’s National Minimum Dataset (NMDS) changed from ICD version 9 to ICD version 10-AM in 1999, with the first full year of implementation in 2000. There were subsequent updates in 2001 and 2004. With no single year likely to be free from the effect of coding changes, all of the available data coded to ICD version 10-AM was used to obtain proportions for comparison with the earlier study.
Datasets—NZ’s National Minimum Dataset (NMDS) for 2000–2005, coded to three editions of ICD-10-AM (ICD-10-AM-1-3), was the primary data source. The ability to identify sport/recreational injury cases directly from these data varied according to the edition of ICD-10-AM applying at the time (e.g. individual sports were not recognised in ICD-10-AM-112), and the hospital providing the data (e.g. narratives describing the circumstances of injury were not available for some hospitals).
To improve case ascertainment and sport-specific injury information, the NMDS data were linked to ACC sport/recreational injury entitlement claims data (2000-2005). The probabilistic linkage program AUTOMATCH13 was used to link the two datasets. Data were blocked into subsets of similar records and the records in each block matched on common variables which included National Health Index number, surname, first name, injury day and second initial. After each pass, successful matches were removed and the procedure repeated using a new set of match variables on the residual.
Approval for the use of these datasets in this way was provided by the Lower South Island Regional Ethics Committee (ref. OTA/99/02/008 & OTA/01/07/049) and the ACC Research Ethics Committee (ref.45).
Cases—Cases comprised discharges from public hospitals in the period 2000–2005 with an e-code in the discharge record. For the purposes of injury prevention, reporting is commonly restricted to diagnoses in ICD-10-AM Chapter XIX “Injury poisoning and certain other consequences of external causes” (diagnoses S00-T98).12, 14, 15 Because a number of sport and recreational injury diagnoses fall outside this chapter (e.g. musculoskeletal injuries, M00-M99), this restriction was not applied. However, injury diagnoses in the range A00-R99 (Chapters I-XVIII) are reported separately.
Although our original intention was to include injury arising from sports-specific paid-work, limited recording of occupation in the NMDS dataset precluded this and so cases coded as “paid work” were excluded. Other exclusions included day-patients (these are inconsistently recorded in the NMDS), readmissions for the same injury, cases of poisoning, self-harm, assault, and complications of medical and surgical care. ICD-10-AM classification groupings were used in the reporting of cases by body region, diagnosis and procedure. So for example, cases with “knee and lower leg” injury comprised all injuries with an ICD-10-AM diagnosis code in the range S80-S89. Where cases were reported according to medical procedure type, the first recorded procedure was used to group cases according to the invasiveness of the procedure. This classification separates cases undergoing ligament and tendon ”repair” from those undergoing more invasive “reconstruction” procedures.
Identification and assignment of cases—The linked NMDS/ACC dataset, comprising 328,802 first-time admissions for injury, was the starting point for the search.
Separate searches were undertaken which identified cases with:
A total of 56,144 new cases of sport and recreational injury were identified using these procedures. Cases were assigned to sport/recreational activities according to ICD-10-AM-3 activity codes U50-U72.14 Two netball codes are available in this classification: “Indoor Netball” (six players and excluding traditional netball played indoors) and “Netball, other and unspecified”.14
For the current study, cases classified to these two codes were extracted from the master dataset of 56,144 new sport and recreational injury in-patient cases. Conflicts between the datasets in the identification and assignment of cases were resolved by applying a hierarchy of preferences with (a) as the highest priority and (d) as the lowest.
Injury rates—Population estimates were obtained from Statistics New Zealand for the calculation of age-specific population rates. There were no reliable estimates of current participation in netball by age and gender. Participation was therefore estimated by applying the proportions of netball participants reported in the New Zealand National Surveys of Sport and Physical Activity (NZPAS), 1997–2001,16 to the average usually-resident population for 2001. These surveys were carried out by the National Research Bureau (NRB) for SPARC and gathered information about participation in sport and leisure activities through household surveys of NZ residents.
Thus the estimates used for calculating incidence rates were for participation in netball activity, not membership of netball clubs/associations; the former being likely to have come from the same population as the hospitalisation data which makes no distinction between members and non-members of clubs/associations.
Injury severity—For the purpose of this study a serious non-fatal injury case was defined as a case with an ICD-based Injury Severity Score (ICISS) of less than or equal to 0.941,17 i.e. cases whose injuries at admission give them a survival probability of 94.1% or worse (probability of death in hospital of at least 5.9%). The ICISS score for each case was computed as the product of the survival risk ratio (using survival probabilities based on 1999–2001 mortality data) for each of their injuries individually derived.18
56,144 (17%) of all injury cases were attributable to sport or recreation; 1126 (3%) of these cases were attributable to netball: 911 (81%) female and 215 (19%) male. The average age at injury was 29 years (range 5–82 years); males were slightly older (31 years, 7–59 years) than females (29 years, 5–82 years). Europeans comprised the largest ethnic group (61%), followed by Māori (26%). Pacific Island peoples accounted for 5% of cases.
Almost all cases (99%, N=1113) occurred in the 5–54 year age-range, with frequencies peaking in the 30–34 year age-group for both sexes (208 cases: 163 female, 45 male); see Figure 1. There was also a notable peak for females in the 10–14 year age-group (129 cases).
Figure 1. Incident hospitalisations due to netball by age-group and gender (New Zealand 2000–2005)
Assuming a participation rate of 7% and an average annual population of 3,974,483 for the period (2000–2005), the crude participant injury rate for regular players (i.e. adults and children who had played netball in the last 4 weeks) was 123 per 100,000 per year.
The highest injury rate was 325 per 100,000 participants per year for the 35–49 year age-group. The participant injury rate increased with age. The population injury rate was 5.0 per 100,000.
Body region and nature of injury—Across all diagnoses, the lower leg/knee (683 cases, 61%), forearm/elbow (151 cases, 13%), wrist/hand (91 cases, 8%), and the ankle/foot (73 cases, 6%), were the most commonly injured regions; see Table 1.
As far as could be determined from the coded data, injuries coded as lower leg/knee (S80-S89) included 73 cases (6% of all 1126 cases) with primary diagnoses relating to the knee.
Table 1. Incident hospitalisations due to netball injury by body region and nature of injury (New Zealand 2000–2005)
† This group includes injury to the distal areas of the radius and ulna; ‡ This group includes injury to the maleolus or ‘ankle bone” and injury to the Achilles tendon; * Including avulsion and traumatic rupture of joint cartilage, capsule, and ligament; a Includes connective tissue; b Includes abnormal clinical and laboratory findings NEC.
The most common diagnoses were ACL rupture (22 cases), various disorders of the patella (17 cases), derangement and tears of the meniscus (12 cases), sprains and strains of collateral and other unspecified ligaments of the knee (9 cases), and chronic instability of the knee (5 cases). The remaining 8 cases had general or nonspecific conditions. Lower leg/knee injuries also included 500 cases of injury to the Achilles tendon (S86.0). A number of other strains and sprains and ruptures (22 cases) of the ankle area were coded as ankle/foot (S90-S99).
The distribution of cases by body region was similar for males and females except for the forearm/elbow (4% of all male vs. 15% of all female cases) and the ankle and foot (9% of all male vs. 5% of all female cases). For cases with a diagnosis in the range S00-T98, muscle and tendon injuries were the most common (46%, 518/1126 cases), followed by fractures (32%, 363/1126 cases); see Table 1.
Injury to the Achilles tendon (S86.0) was the most common individual diagnosis (500/1126 cases, 44.4%); followed by fractures of the proximal phalanx (34/1126, 3.0%), lower end of the radius (29/1126 cases, 2.6%) and lateral malleolus (29/1126 cases, 2.6%).
Five of the top 10 primary diagnoses were fractures of the distal radius. For cases with primary diagnosis codes in the range A00-R99 (67 cases), 35 cases (52%) had musculoskeletal injuries (M00-M99) including chronic instability and derangements of the ligaments of the knee, recurrent dislocations, nonunion/malunion of fractures; a further 12 cases (18%) had post trauma deep vein thrombosis of the lower extremities.
Age and the nature of injury—The distribution of injuries across age-groups by body region (detailed for injury diagnoses, S00-S99) is shown in Table 2. Injury to the forearm/elbow was the most common injury in the 0–14 year age-group (83 of 158 cases, 53%), while knee/lower leg injury was the most common injury in the 15–24 year age-group (98 of 201 cases, 49%), 25–54 age-group (523 of 753 cases, 69%) and the 55+ age-group (6 of 13 cases, 46%).
Wrist/hand injuries were relatively constant throughout the age-groups at 8–9%. The proportion of “non-injury” primary diagnoses varied across the age-groups, and (leaving aside 2 cases in the 55+ age- group), peaked at 9% in the 15–24 age-group.
The distribution of injury types across age-groups is shown in Table 3. Fractures dominated in the 0–14 year age-group (123/158 cases, 78%), in particular, fracture of the forearm (82 cases). Injury to the muscle and tendon dominated in the 25–54 year age-group (455 of 754 cases, 60%) and these were predominantly to the Achilles tendon (439 cases). Injury in the 15-24 year age-group was more evenly distributed between fractures (59 cases, 29%), muscle/tendon injury (57 cases, 28%) and strains/sprains (34 17%).
Overall, the most common diagnoses were Achilles tendon injury (S86.0, 470 cases) with a median age of 34 years, and forearm fracture (S52, 147 cases) with a median age of 13 years.
Table 2. Incident hospitalisations due to netball by age group and body region (New Zealand 2000–2005)
Table 3. Incident hospitalisations due to netball by age-group at admission and nature of injury (New Zealand 2000–2005)
† Primary diagnoses in the range A00-R99
Operations and procedures—168 cases (15%, and including 31 cases of injury to the Achilles tendon) had no coded procedures (standard procedures such as X-rays and plaster applications are not usually coded)
The remaining cases (958) underwent 166 different procedures.12,14,15 The type of first recorded procedure (usually the procedure performed for treatment of the principal diagnosis), by body region is shown in Table 4.
Table 4. First recorded procedures performed in new inpatient cases hospitalised for netball injury (New Zealand 2000–2005)
a The ICD-10-AM classification includes injury to the Achilles tendon in this group; b 97% (437 cases) of repairs were to the Achilles tendon; c 52% (160 cases) open reduction with or without fixation, 42% (130 cases closed reduction, 5% (14 cases) fixation only, unspecified 2 (1%).; d 27 reconstructions of the knee (22 with a primary diagnosis involving the ACL), 11 reconstruction of the Achilles tendon. Note: average number of procedures in cases with operations is 2 (1-8), (958 cases), 168 cases had no coded procedures; application of plaster is not a coded procedure.
Surgical repair was the first recorded procedure in 452 (40%) cases, mostly (437 cases) to the Achilles tendon. A further 324 (29%) cases had open or closed reduction and/or fixation of a fracture to the forearm (134 cases), lower leg/ankle (90 cases), wrist/hand (72 cases) or other areas (28 cases). “Other procedures” (77 cases) included wound drainage, removal of loose bodies in the joint, aspiration and other open procedures. Forty-two cases (4%) received more invasive reconstructive surgery to the knee (27 cases), Achilles tendon (13 cases), and hand and orbit (2 cases).
Of the remaining 140 cases, 32 had a CT or MRI scan, 17 had an arthroscopic washout and/or debridement, 11 had immobilisation or stabilisation procedures, and 3 cases had a discectomy.
Mechanism of injury—“Overexertion and strenuous or repetitive movements” was the most common mechanism of injury (517 cases, 46%), followed by a fall (314 cases, 28%); see Table 5. Narrative information suggested that falls were often the result of a heavy or bad landing after a jump and that “overexertion” was also associated with leaping, jumping and landing; there are no ICD-10-AM codes that adequately capture these “pre-fall” incidents or underlying mechanisms of injury.
Table 5. Incident hospitalisations due to netball injury by mechanism of injury (New Zealand 2000–2005)
Injury severity and length of stay (LOS)—Netball injury posed no “threat-to-life” (ICISS=1) in 669 cases (59%), while in 422 cases (37%) there was a threat-to-life of between 0.05% and 13.44% (ICISS scores 0.995-0.942); an ICISS score could not be calculated for 35 cases; see Figure 2.
The “threat-to-life” was less than 5.9% (ICISS score >0.941) in 416 cases. In the remaining 6 cases the injury sustained was deemed as a “serious threat to life” (ICISS score <=0.941 i.e. cases whose injuries at admission gave them a survival probability of 94.1% or worse or a probability of death in hospital of at least 5.9%), using the NZIPS definition of serious injury.17 These 6 cases all sustained fractures, 3 to the head/neck region and 3 to the femur; 5 were the result of a fall. The average age of players with a “serious injury” was 41 years, the gender split was even and the average length of stay (LOS) was 4.8 days. The average LOS for all 1126 cases was 2.3 days (range 1–21 days).
Figure 2. Incident hospitalisation due to netball injury by ICISS score (New Zealand 2000–2005)
While injury incidence rates are lower in netball than in many other sports, the popularity of the game and its ability to keep females actively engaged in sport make it an important target for injury prevention.
The last descriptive epidemiological study of netball injuries in NZ, for injury data in 1988, was published in 1993.9 In that study, hospitalised injury per 100,000 population per year was 4.3, the hospitalised injury per 100,000 participants per year was 143, the gender split was 89.5% female to 10.5 % male, the commonest sites of injury reported were the ankle (52.4%) and knee (18.9%), and the commonest ages at injury were 26, 25 and 17 years. The modal length of stay in hospital was 2 days and the proportion of serious injuries (AIS≥3) was 9%. Netball accounted for 3.1% of hospitalised sports injury and cost ACC $1.7 million. These results were based on day and inpatient cases hospitalised in 1988.
There are several notable points of difference between the earlier study9 and the current report. The proportion of male netball players hospitalised increased from 10% (1988) to 19% (2000-2005). This increase is consistent with an increase in the proportion of male participants; the SPARC physical activity survey of 1997–2001 reported that 21% of netball participants were male. A recent study of netball hospitalisation in Australia19 reported that 12% of hospitalised injury cases were male (participation rate 13.4%).
The importance of differentiating indoor from outdoor netball injury was acknowledged in the earlier study by Hume but injury coding at the time did not allow this distinction to be made. ICD-10-AM-3 (introduced July 2004) distinguishes “indoor” 6-player netball from “other netball” (including “traditional” netball played indoors). In 2005, the first full year of discharges coded to ICD-10-AM-3 which distinguished different types of netball, “indoor” netball accounted for 27% of cases, with an even gender split and an average age at injury of 34 years. In “other netball” females predominated (84%) and the average age of injury was 26 years.
Hume reported that 17, 25, and 26 year-olds were the most frequently hospitalised ages (5.6%), whereas in the current study 30, 31, and 32 year-olds were most frequently hospitalised ages (3.8-4.4%). One possible explanation for the apparent difference is the inclusion of day-cases in the earlier study—a recent Australian study,19 however, reported the peak age for netball hospitalisations as 25–34 years lending support to the impression that the average age of hospitalised injury cases may be increasing.
Netball has been characterised as a game prone to knee and ankle injury, with knee injury highlighted because of the high cost and associated disability.5,9,20,21 Only 6% of hospitalised injury in the current study involved the knee, compared to 19% reported by Hume. It is not possible to determine if knee injury (and in particular ACL rupture) has become less common or if differences between studies are due to changes in treatment and management since 1988. There is anecdotal evidence to suggest that same-day arthroscopic reconstruction of the knee carried out in private hospitals under ACC cover may account for the difference.
There also appear to be large differences between the two studies in the proportion of ankle injuries reported. The most likely explanation is that Achilles tendon injury was classified as “ankle injury” in the earlier study and as “Injury to the knee and lower leg” in the current study. Because the mechanisms of injury are likely to be very different for Achilles injury and ankle sprain/strain, reporting specific diagnoses is arguably more important than general “body areas” or injury “type” for targeted injury prevention.
Forearm fractures were much less frequent in the earlier study (5 cases, 3.5%) than the present study (149 cases, 13%); with the young median age (12 years) of cases in the current study being of concern. There was also a notable overall injury peak (129 cases) for females in the 10-14 year age-group in the current study that was not observed in the earlier study. There are a number of possible reasons for the increases. It has been argued elsewhere that children are starting competitive sport too young, increasing their risk of injury;22 70% of children presenting to Australian emergency departments with a netball injury sustained the injury during formal competition.
Changes in the tempo and physicality of the game since 1988, and the increase in indoor netball and the type of game played on indoor courts, may also have increased the risk of injury. However, it is also possible that the success of the national netball team, the silver ferns, in the 2000s has increased the popularity of the sport with young females and the observed number of injuries in this group.
Injury rates varied between the studies but not appreciably. Census data were used in both studies to calculate population injury rates. Participant injury rates were calculated using estimates from two physical activity surveys—i.e. the Life in New Zealand (LINZ) study and the New Zealand Sport and Physical Activity Survey (NZSPAS). A slightly higher population injury rate (5.0 vs. 4.3 per 100 000) was reported in the current study but a slightly lower participant injury rate (123 vs. 143 per 100 000).
Netball injury accounted for only 3% of all sports injury in both studies. There were few cases in either study where injury posed a high threat to life and the most frequent LOS was 2 days in both studies. The proportion of “serious” injuries was 9% in the earlier study compared to <1% in the current study, however, different severity measures and thresholds were used in each study. A case-by-case examination of injuries in the current study suggested that approximately 90% of players with lower leg injuries could be “out of action” for 3-6 months. In the sports injury setting, a severity measure based on a “threat of disability” rather than “threat to life” is required.
There were a number of study limitations. In restricting eligible cases to in-patients our population is not representative of all netball-related injury cases; however, it is likely to represent cases with the greatest consequences for the health system and the individual. Routinely collected census and sport participation data were used to estimate injury incidence rates in both studies. However, although the sport participation surveys (LINZ, NZSPAS) were conducted in a similar manner and were funded by the same organisation (Hillary Commission/SPARC) they were conducted at least 10 years apart and were subject to the limitations associated with such surveys;23, 24 the reported rates therefore should be interpreted with caution.
Netball is a relatively safe sport;2 however, the physical demands of the game and its popularity have grown. The differences highlighted in this study, between hospitalised netball injury in the late 1980s and the present, suggest that (a) the average age of hospitalised netball injury cases may be increasing, (b) forearm fractures in young netball players are a cause for concern, (c) Achilles tendon injuries and/or the surgical repair of these injuries appears to have increased while knee ligament injuries requiring hospital admission/surgical repair appear to have decreased, and (d) the indoor version of the game and male players may be important targets for injury prevention.
Competing interests: None known.
Author information: Pam Smartt, Research Fellow; David Chalmers, Deputy Director; Injury Prevention Research Unit, Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago, Dunedin
Acknowledgements: The New Zealand Accident Compensation Corporation (ACC) funded this research. Data for the study was supplied by the New Zealand Health Information Service, ACC, and the New Zealand Sport and Recreation Association. Helpful comments on an earlier draft of this paper were received from ACC, and Chris Lewis (Information Analyst, New Zealand Health Information Service) provided very helpful comments on the final manuscript.
Correspondence: Pam Smartt, Injury Prevention Research Unit (IPRU), Department of Preventive and Social Medicine, PO Box 913, Dunedin, New Zealand. Fax: +64 (0)3 4798337; email: firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals