Discussion

Horse-riding and handling horses is a common activity, with over 250,000 Australians actively engaged in sport and recreational horse riding in 1997 (Australian Bureau of Statistics 1999). This report describes injury related to horses, considering cases resulting in death or admission to a hospital. These sources do not describe the whole picture, as they do not include non-hospitalised morbidity. Furthermore, some types of injury such as bites, and from handling horses, are not identifiable as horse-related.

The epidemiology of horse-related injuries has been well described in numerous studies, primarily overseas (Edixhoven, Sinha et al. 1980), (Berhang and Winslett 1983), and (Bixby-Hammett and Brooks 1990). A majority of these were prospective or retrospective studies of cases presenting to regional emergency departments or admitted to hospitals rather than national population studies (Buckley, Chalmers et al. 1993), (Hobbs, Yealy et al. 1994), (Chitnavis, Gibbons et al. 1996), and (Campbell-Hewson and Robinson 1999). These and other horse-related injury studies have been summarised in detailed reviews of local and international literature (Masters 1991) and (Watt and Finch 1996).

Pounder (1984), extrapolating from South Australian horse riding deaths and population in South Australia at the time of his survey, estimated the national death rate to be one death per million population, per year. In 1998 population terms, this would be equivalent to 19 deaths from horse riding activities, similar to the figure (20 deaths per year) calculated from national mortality unit record data.

Buckley et al. (1993) using New Zealand Accident Compensation Corporation (ACC) figures from a 1983-84 survey of participation in sporting activities, estimated that 5 deaths per 100,000 population per year occurred in horse riders falling from horses. This rate is probably an underestimation of the death rate in New Zealand as falling from horses represents about 85 per cent of horse-related accidents (Barber 1973). In the United States, using medical examiner and national population figures, horse-related deaths were estimated to be 219 deaths per year (Bixby-Hammett 1987).

The major cause of death reported in Australia and overseas is head injury, accounting for: 77.8 per cent (n=14) in Australia (Pounder 1984); 61.1 per cent (n=33) in New Zealand (Buckley et al. 1993); and 60.5 per cent (n=124) in the United States (Bixby-Hammett and Brooks 1990). According to these studies, young females and older males are particularly at risk of sustaining fatal head injuries, with most of these deaths occurring in young, female, amateur riders who fall off their horses, followed by male professional riders aged 30 to 50 years. Head injury was prominent among the hospitalised ages reported in this report.

A limitation of present mortality and hospital morbidity data collections is that they provide very little detail concerning the circumstances and mechanisms of occurrence of injury. For example, these sources provide no information about the type of riding, the experience of the rider, nor whether a helmet was in use. As noted above, some types of horse-related injuries are not identified as such. Because injuries from kicking, biting, or being crushed (typical injuries from handling horses) cannot be attributed to horses and, therefore were excluded, horse-related injuries reported for the period 1996-97 were based on national estimates of horse rider injuries in Australia

The 10^th revision of the International Classification of Diseases (ICD-10), recently introduced for coding deaths and hospital morbidity cases in Australia, should result in some improvement in potential to identify and characterise horse-related cases. The ICD-10 external cause code V80 (`animal rider or occupant of animal drawn vehicle injured in transport accident') is superior to the current ICD-9 E-code categories for animal related injury used in this report. For example, V80.0 describes the mechanism of injury (fall or being thrown from), important in understanding the injury event and for planning interventions to reduce injury. In addition, the ICD-10 supplementing classification of activity at the time of injury has potential to enable better identification of cases associated with sport and recreation (though the activity classification is not presently applied to deaths).

Most studies of horse-related deaths in Australia and overseas have used surveys, State Coronial records or National Medical Examiner records (in the USA) to obtain more detailed information on the medical cause and circumstances that led to the death. The value of coroner records and the benefits of improving their utility for purposes including injury prevention research has been recognised in Australia. In 1997, the National Coronial Information System (NCIS) was established by a consortium called the Monash University Centre for Coronial Information (MUCCI). When fully operational, the NCIS will hold information on all deaths referred to a Coroner in Australia.

The draft NCIS Data Dictionary, published in February 2000, specifies data standards for the NCIS Core Data Items. A review of the Data Dictionary indicated that horse-related deaths in Australia will be readily identifiable and national statistics on horse-related deaths, including details on the injury event and cause of death, will be eventually available for research purposes.

The following Core Data Items would be particularly useful for identifying horse-related deaths:

• Description of circumstances of death,
• Activity,
• Sport or active recreation,
• Mechanism of injury,
• Mode of transport,
• Counterpart,
• Object or substance producing injury, and
• Pathological cause of death.

In addition, the ability to conduct text string searches on coroners' findings and circumstances of death will provide further information on the mechanism and factors that contributed to the horse-related injury and medical cause of death--invaluable in the development and evaluation of injury prevention programs.

The lack of national statistics on horse-related injury is acknowledged in the literature. Many studies use localised population centres especially associated with emergency departments, and investigate horse-related injuries using retrospective or prospective casenote reviews. Although these studies describe horse-related injuries epidemiologically, they are limited by information available in the casenotes, particularly the nature and circumstances of the injury event. Detailed information on the mechanism and factors that contributed to the injury event is fundamental in the development of prevention programs.