METHODS

The study area was restricted to six postcodes that fell wholly within the catchment area of the Latrobe Regional Hospital (3825, 3840–3842, 3844, 3869–3870). All residents aged over 4 years were eligible for inclusion.

Data on sporting and recreational injuries that received medical treatment were obtained from three sources: the Victorian Admitted Episodes Dataset (VAED; hospital admissions), the Victorian Injury Surveillance System (VISS; hospital ED presentations), and the ELVIS project (GP presentations). No sports medicine facility operated in the Latrobe Valley during the study period. All eligible patients recorded in these databases for the period 7 November 1994 to 6 November 1995 were selected. All pedal cycling related injuries were included, as it has been estimated that over 86% of all cycle use in Australia is for sport, exercise, or recreational purposes, rather than commuting.⁴

Hospital admissions

The VAED is a collection of data describing admissions to Victorian hospitals. Injury cases are coded with an external cause of injury code (E-code) under the International classification of diseases system (ICD-9-CM). Unfortunately, the classifications available to identify and describe cases of sport and recreation injury on the database are limited (table 1). The selection of sports injury cases is restricted to two main E-codes: E886.0 (fall on the same level from collision, pushing or shoving by or with other person in sports); E917.0 (striking against or struck accidentally by objects or persons in sports). The available codes for recreational injury are also restricted and only cover animal riding (predominantly recreational horse riding), pedal cycling, water skiing, swimming, and diving. All cases classified under E927.0 (overexertion and strenuous exercise) were included, although a small proportion may not be related to sport and active recreation. Private hospital admissions, readmissions within 30 days, medical injuries, and late/adverse effects of injury cases were excluded from the analysis.

The number of injury cases recorded on the VAED represents 100% of all actual injuries admitted to hospital. However, cases of sports injury are underestimated because of the limitations of the coding system mentioned above.

ED presentations

VISS, as it operated in the mid-1990s, collected data on injury presentations to the EDs of five participating hospitals, including the Latrobe Regional Hospital. A standard instrument for injury surveillance¹¹ which collected demographic data and information relating to the injury event such as the mechanism of injury and associated factors was used. On presentation to an ED, the injured patient (or parent/carer) and doctor completed sections of the form, on a voluntary basis.

An audit of ED injury surveillance in the Latrobe Regional Hospital¹² determined that the cases recorded on the VISS database represented 85% of all actual ED presentations for treatment of injuries and 100% of all actual cases subsequently admitted to hospital. However, there may be some discrepancy in the admission rate recorded on the VISS database compared with the VAED. This is because cases were selected on postcode of residence and therefore admissions recorded on the VAED may have occurred through hospital EDs that were not part of the VISS system.

The VISS used Injury Surveillance Intelligence System (ISIS) codes¹¹ to generate incidence distributions to compare the nature and extent of injury presentations. Sports injury cases were selected on the basis of ISIS context/activity codes 102 and 103 (bicycling) and 301–303, representing organised competition or practice, informal sport, and sport not specified respectively. Active recreation cases were identified by factor codes that covered injuries associated with horseback riding, roller skating, trampolining, swimming, skateboarding, and snow skiing. Up to three injuries may be recorded per case, and the system does not allow extraction by primary diagnosis.

GP presentations

The ELVIS project, established by the Central West Gippsland Division of General Practice, collected injury data through a research network of GPs.¹⁰ The process for VISS ED data collection was refined and modified for use in the general practice setting. Data on injury presentations to general practice were manually entered on a standardised form, different sections of which were completed by the patient and GP. Sixty four of the 66 GPs (97%) in the division participated in the ELVIS project through 18 general practices. An audit of the ELVIS system determined that the number of injury cases recorded on the ELVIS database represented 77% of all actual GP consultations for injury in the 12 month data collection period.¹⁰

ISIS classification was used, and cases were selected as described above for the VISS ED presentations. The ELVIS data were also manually assigned an ICD9 E-code for cause of injury, which allowed comparison with hospital admissions data on injury causation.

Data analysis

A descriptive analysis of the data for sports injuries from each of the three injury databases was undertaken. Data from these collections are presented as incidence and proportions, with associated 95% confidence intervals. It is not possible to combine data from the three databases because of the different methodologies and classification systems used to identify and code injury data, and there were no identifiers to enable data linkage.

The Australian Bureau of Statistics 1995 census estimates that there were 69 663 people aged over 4 years usually resident in the six target postcodes approximately midway through the data collection period (June 1995). The incidence of medically treated injuries incurred during sport or active recreation in the Latrobe Valley was calculated per 10 000 population (aged over 4 years). The number of cases reported from each database was multiplied by the inverse of the expected capture rate from that source to estimate the true number of cases. For ED data, this was done separately for the cases in which the patient was admitted to hospital or not because of their differing case capture rates. It was not possible to take into account the age and sex distributions in this factoring up because the case capture rate information was only available as an overall figure. The ratio of the incidence of hospital admissions to the incidences of ED and GP presentations was calculated to generate a sports injury “pyramid”.

RESULTS

Hospital admissions

In the study period, there were 112 hospital admissions for injury from sport or active recreation in the Latrobe Valley (table 1). This corresponds to an annual incidence of 16 sports injuries requiring hospital admission per 10 000 population aged over 4 years (95% confidence interval (CI) 14.4 to 42.1).

In 70% of cases requiring hospital admission, the patient was male. Two thirds were in the younger age groups: 36.6% were in the 5–14 year old age group (65.9% were male); 30.0% were 15–24 year olds (73.5% male).

The major causes were “being struck or crushed during sport” (39.3%), cycling (19.6%), and “overexertion and strenuous movements” (18.8%) (table 1). Injuries occurred most often to the upper extremities (33.0%) and the head/face (29.5%) (table 2). The most common types of injury were fracture (44.6%), intracranial injury (17.9%), and dislocation (9.8%) (table 3).

The length of hospital stay provides an indication of the severity of the injury: 86% of stays were less than two days and only 3.6% were more than eight days.

ED presentations

There were 1179 ED presentations for sport injury in the Latrobe Valley in the study period. This corresponds to an annual incidence of 197 persons attending a hospital ED for treatment per 10 000 population aged over 4 years (95% CI 123.2 to 254.9).

The sports most often associated with ED presentations were Australian football (24.0%), cycling (15.7%), and basketball (13.8%) (table 4). In 73%, the patient was male. Cases were predominantly in three age groups: 10–14 years (29.5%), 15–19 years (19.5%), and 20–24 years (15.9%).

The VISS allows up to three separate injuries to be recorded per case. Most injuries occurred to the extremities and were fairly evenly divided between the upper (39.0%) and lower (31.2%) extremities (table 2). Sprain/strain (23.3%) was the most common type of injury (table 3). The most common specific injuries were ankle sprains/strains (9.0%), knee sprains/strains (4.0%), fractures of the radius/ulna (4.0%), face/scalp lacerations (4.0%), and finger fractures (3.0%).

A high proportion (80%) of people who attended an ED with a sports injury were subsequently admitted to hospital. Further examination of the data shows rugby injuries to result in the highest rate of hospital admission (16%), followed by trampolining (15%), soccer and horse riding (both 14%), cycling (11%), and football (5%). Just over half (52.8%) of ED presentations required major treatment—that is, they required follow up or referral usually to a GP (25.5%), a review in the ED (17.0%), or another type of referral (8.0%). Approximately one third of presentations (34.1%) required minor treatment, that is medical assessment only or treatment without follow up, and 5.5% required no treatment.

GP presentations

There were 1003 presentations to GPs for injuries associated with sport and active recreation. This corresponds to an annual incidence of 187 persons presenting to a GP for a sports injury per 10 000 resident population aged over 4 years (95% CI 160 to 214).

Australian football (22.0%), basketball (17.5%), and cycling (12.6%) were the sports most often associated with GP presentations (table 4). The major causes of injury were “overexertion and strenuous movement” (48.7%), “struck/crush in sport” (33.2%), and pedal cycling (11.6%) (table 1). Two thirds of patients were male. The age groups that accounted for most injuries were 10–14 year olds (34.2%) and 15–19 year olds (16.7%).

The most common types of injury presentation were sprains/strains (38.5%, mostly to the lower extremity), bruising (23.6%, commonly to the head/face), fracture (11.4%, mostly to the upper extremities), and inflammation (8.0%) (table 3). Overall, the lower and upper extremities were the most commonly injured body sites (39.3% and 38.4% respectively), followed by the head and face (11.8%), and trunk (8.4%) (table 2).

Almost as many GP presentations required minor treatment (41.8%, treated without further referral) as more major treatment (40.7%, treated with follow up or referral). An appreciable proportion required assessment only (17.5%). Two patients presenting to GPs were referred to an ED.

Injury pyramid

A pyramid of medically treated sports injuries was constructed for the Latrobe Valley. Over any 12 month period, for every 10 000 head of population in the Latrobe Valley, it can be expected that for each hospital admission for treatment of a sport injury, there will be 10.6 ED presentations and 11.7 GP presentations (1:11:12).

DISCUSSION

This is the first Australian study to report the incidence of sports injuries for a defined population. Available data indicate that there were at least 2300 medically treated sports injuries in the Latrobe Valley population of about 70 000 aged over 4 years during the period 7 November 1994 to 6 November 1995: 112 hospital admissions, 1179 ED presentations, and 1003 GP presentations. This corresponds to 272 cases per 10 000 persons.

This study confirms the findings of others that a substantial number of sports injuries are not treated at hospitals or EDs; instead they present to community based sources such as GPs or sports medicine clinics.^4,6,13–16 One British study found that about 75% of all reported sports injuries were not treated at an ED or admitted to hospital.¹⁷ In a study of Australian football injuries in children and adolescents, EDs treated only 28% of all reported injuries.¹⁵

An injury pyramid was constructed to describe the profile of injuries within the Latrobe Valley. It is estimated that over a 12 month period, per 10 000 population, 16 people will be admitted to hospital for treatment for a sports injury, 169 will present at an ED, and 187 people will receive treatment from a GP (adjusted proportions 1:11:12). These figures suggest that most sports injuries presenting to clinical services are mild to moderately severe, which is also indicated by the short stay (fewer than two days) of most admitted cases. As would be expected, fractures and intracranial injuries comprise a higher proportion of hospital admissions than ED and GP presentations. Injuries presenting to the ED and general practice are usually sprains and strains, although fractures are not uncommon. The substantial number of fracture cases treated by GPs is noteworthy as it is often assumed that injuries of this severity generally present to hospital EDs.

Our data suggest a higher rate of hospital admission for sports injuries than other studies, although they are consistent with previous Australian data.

Finch et al⁴ described sport and active recreation ED presentations across Australia over a four year period. Some 8% of the injured adults were subsequently admitted to hospital for further treatment, compared with just over 14% of children aged <15 years. In contrast, among sports injury cases presenting to an ED, only 2% of cases in Glasgow,¹⁸ 5% of child cases in Canada,¹⁹ and 2% of 10–19 year olds presenting in Columbia, USA²⁰ required admission to hospital.

Men and boys were much more likely to present for medical treatment of a sports injury than women or girls, and younger age groups were also overrepresented in injury data, but these differences may just reflect the greater participation of these groups in sport and active recreation.

The available data, limited to ED and GP presentations, show that Australian football is the sport most often associated with medically treated injuries in Latrobe Valley, accounting for 22–24% of presentations. Cycling, basketball, netball, cricket, and soccer were also associated with appreciable (and similar) proportions of injuries presenting to both EDs and GPs. Care should also be exercised in drawing inferences from these findings because comparisons are based solely on injury incidence. A fairer but more complex method of determining the relative risk of injury in the various sporting and recreational activities would factor in participation data (to estimate the risk per participant) and, preferably, exposure or time at risk data (estimating risk per time unit).²¹ These adjustments are important because, in general, sports with the highest numbers of participants have the highest incidence of injuries, especially in communities where a small number of sports dominate. Nonetheless, popular sports that produce sizeable proportions of the injury problem in a given community are legitimate targets of prevention efforts, even if less popular sports are shown to carry greater risks of injury when data are adjusted for exposure.

Although the three data sources used for this study provide rich information on the size and nature of the sports injury problem, they have limitations.² Firstly, the VAED underestimates hospital admissions for sports injuries because, under ICD9-CM, sports specific E-codes are limited and restrict the counting of hospital admissions. Secondly, the data only cover the injuries for which medical treatment was sought from hospitals and GPs. Many sports injuries, particularly overuse and other soft tissue injuries, are attended to by other practitioners such as sports medicine specialists, physiotherapists, chiropractors, masseurs, and sports first aiders, or are self treated. The available statistics therefore probably represent a fraction of those injured in sport and active recreation. Indeed, comparison of the health sector data with results from a population survey conducted in the same region indicates that the available injury databases describe less than 30% of all self reported sports injury cases.²² Despite these limitations, health sector data are essential for quantifying and describing injuries that are severe enough to require medical treatment, and they assist the identification of high risk sport and active recreation activities.²³

In summary, this study shows that routine health sector data collections in defined populations can provide useful information on the size, distribution, and characteristics of the sports injury problem at the community level. However, all current health sector systems for collection of injury data require attention to improve case capture and identification and data quality. Recent coding changes are improving the potential to identify sports injury cases among all cases admitted to hospital. The new ED surveillance system operating across 25 public hospitals has extended Victoria’s capacity to provide injury data for defined populations, but identification of the sport or recreation activity involved is inconsistent at present. There has been some expansion of routine collection of injury data at the level of general practice in Victoria, modelled on ELVIS. These clinical sources provide some of the information needed to guide programme planning and implementation in the area of community sports safety.