A Summary of Fatal Injury Surveillance Methods in Australian Agriculture and Their Impact on Safety Policies and Practices

ABSTRACT

Agriculture is one of the most important and also hazardous industries in Australia. Having a sound knowledge and understanding of the circumstance of injury events is critical to developing evidence-based intervention programs. This paper aims to provide a brief historical snapshot of the development of data systems underpinning the assessment of fatal farm injury in Australia and how it has impacted on safety policy and practice. The first Australian studies used coronial information to explore agricultural fatalities, these studies reviewed paper-based records (in-situ) and collected the information for analysis and reporting. This task was laborious and costly. When the National Coronial Information System (NCIS) was established in 2000, this allowed access to coronial records online. Information provided about the deceased includes demographics, contextual details on the nature of the fatality and autopsy, toxicology, and police reports, as-well-as the coroner’s finding. Information from the NCIS, along with media reports, have been used to develop the farm fatality database. This information has been used to inform the safety goals and targets for farm commodity groups, identify key risks, provide long-term benchmark indicators and underpin the development of prevention materials and training resources. Without accurate, timely, concise and relevant data about injury occurring on farms, there is no evidence to drive policy and practice or to evaluate programs of work. As such, the continued utilization and extension of the NCIS data will prove crucial to further reducing the burden of preventable fatal injuries on Australian farms.

KEYWORDS:

• Agriculture
• injury
• death
• farm

Introduction

Agriculture is one of Australia’s most important industries, with the value of production reaching $90 billion in 2022–2023.¹ However, it constantly ranks among the most hazardous industries (2017–21), with the agriculture fatality rate (11.5 per 100,000 workers) being some eight times higher than the all-industries rate (1.4 per 100,000 workers).²

Having a sound knowledge and understanding of the circumstances of injury events, is critical to developing evidence-based intervention programs.^3,4 Such information can also provide a snapshot of current problems, identify patterns of risk specific to different agricultural commodities and track trends in a timely manner. From this, policy and practice initiatives to optimize safety outcomes can be identified to help influence safety.⁵

AgHealth Australia (initially the Agricultural Health Unit), was established in 1985 and commenced a variety of injury surveillance work to understand the nature and extent of the issue. This included examining locally based fatalities (coronial records);⁶ administrative health data from emergency departments, hospital separations and General Practice presentations;⁷ workers compensation data;⁸ as-well-as drawing on surveys⁹ and health screening.¹⁰

As a component of this approach, linkage with the International Classification of Diseases (ICD) was undertaken and a farm-specific data dictionary developed to guide and validate information.¹¹ AgHealth Australia remains the only dedicated agency within Australia that has maintained an agriculture specific long-term injury surveillance program. The National Farm Injury Data Centre, includes both work and non-work related fatalities. In the Australian context, the non-work fatalities are particularly important in the agricultural sector as the vast majority of farms are family owned and operated (~99%), resulting in family members and visitors also being exposed to injury risks.¹²

While there are a number of data types that can be used to assess injury and its impacts as outlined above, those that report on fatal incidents represent the most critical starting point for preventive actions, as they are obviously the most severe and also provide in-depth information into the circumstances surrounding the event. All deaths in Australia attributed to trauma, are reported to a coroner.¹³ Hence, this paper seeks to provide a brief historical snapshot of the development of data systems underpinning the assessment of fatal farm injury in Australia and how that has impacted on safety policy and practice.

The early years

The first Australian study examining agricultural fatalities, was part of a broader work-related assessment conducted by Erlich et al for the 1982–84 period that excluded suicides and medical misadventure.^14,15 At this time (1989), each of the eight Australian states and territories had separate coronial registration systems. This study involved the review of hard copy records (often onsite at the state coroner’s office), initially for their work-relatedness (n = 15,462) and secondly to determine if they were agriculturally based (n = 257 ~ 15%). ICD 9 Codes were assessed with the fatality incidence rate for the work-related cases being 19.4 per year per 100,000 workers. A subsequent study using the same methodology, was completed in 2001 and covered the 1989–92 period.¹⁶ It identified 373 farm-related deaths with the work-related incidence rate being 20.6 per year per 100,000 workers. This study was undertaken by AgHealth Australia under its National Farm Injury Data Centre.

Consistent features across these studies, included the over-representation of older men and machinery-related incidents, with tractors being a prevalent agent of injury. As illustrated by both studies, there were significant time lapses between the occurrence of these events and their collation, approximately 5¹⁴ and 10 years,¹⁶ respectively. Additionally, all data retrieval was manual with no case indexing and required on-site assessment of hard copy records in each of the eight jurisdictions (states/territories).

A national collection system

In the early 1990s the Australian Coroner’s Society (ACS), led the way in establishing a national database to consolidate the coronial data from each of the states and territories into a unified dataset.¹³ It was recognized that such an undertaking would have major benefits in the timely collation and hence utilization of data to take preventive actions (setting priorities, justifying resource allocation, intervention development). Furthermore, the data would provide solid information for training initiatives and in defining research priorities. In response to a series of actions (see timeline below) and recommendations from the Royal Commission into Aboriginal Deaths in Custody,¹³ the National Coronial Information System (NCIS) was established in 2000.¹³ This was the first such system of its kind anywhere in the world.

Timeline for NCIS development

• 1990 - All eight jurisdictions have independent coronial offices.

• 1994 - the Australian Coroners’ Society commissioned the National Injury Surveillance Unit of the Australian Institute of Health and Welfare to undertake a feasibility study on a national database for coronial information. It recommended the establishment of a national database, and the recommendations were taken up by the Australian Coroners’ Society.

• 1997 - the Australian Coroners’ Society endorsed a business plan for the development and management of the National Coronial Information System.

• 2000 - NCIS established 2000.

• 2001 - first tranche of data becomes available.

• 2012 - New Zealand data included.

National coronial information system

The NCIS is a secure database of information on deaths reported to a coroner in Australia and New Zealand. Data includes demographics on the deceased, contextual details on the nature of the fatality and case reports including the coronial finding, autopsy and toxicology report, plus the police notification of death.¹³ Access to the NCIS by researchers requires approval through specified Human Research Ethics Committee(s). The use of the NCIS data is bound by strict ethical requirements that are reviewed regularly. A part of this requirement is to ensure all data is aggregated and no identifiable information is released publicly. Over the period, the data have been digitized and there has been ongoing improvements to the electronic access to these data, along with strengthening of the quality assurance processes.

AgHealth Australia continues to maintain a national farm fatality database drawing on information from incidents that have occurred on a farm since 2001 and is the sole agency assessing this information in an in-depth manner. The actions have necessitated the development of a relevant data coding system and linkage to the ICD – for Australian purposes, the ICD 10 Australian Modification is utilized.¹⁷ This coding system has been reviewed and updated on three occasions since its initial development in 1995.^18–21 To identify potential cases for inclusion, a commercial media tracking organisation (iSentia) is used to scan approximately 2500 daily, weekly and monthly publications Australia wide. Publications are assessed for various designated search terms (e.g. “farm*”, “agric*”, “growers”, “producers” and “horticulture”). Where a potential on-farm case is identified, the corresponding NCIS case file number is obtained for the case. As all cases of death due to an external cause are reported to a coroner and recorded in the NCIS, key word searches for the location of injury are also undertaken (in this context farm or agricultural property), to ensure cases are not inadvertently omitted. For each case, preliminary information is uploaded into the NCIS and these remain “open” until the coroner hands down a final determination and the case is then “closed”. In each of the NCIS cases, a cause of death is determined by a pathologist and recorded by a coroner, who also determines the work-relatedness of all cases, with specific cause of death details independently coded by the Australian Bureau of Statistics against the International Classification of Disease 10 (ICD-10AM).

Given this approach, it is hypothesized that the data capture rate for on-farm cases is close to 100%. In this manner, the data have been validated and provide a level of consistency now dating over 20 years since the first tranche of NCIS data were accessed by AgHealth.¹⁶

Practice

The fatality data have underpinned work in several strategic ways:

• Informing goals and targets for critical industry farm groups e.g. Farmsafe Australia, National Farmers Federation and the Research and Development Corporations (RDC’s) representing each of the major commodity sectors.^22–24

• Identification of key risks and developing targeted prevention materials and approaches that are specific to a diversity of commodity sectors e.g. grains, dairy, sugar cane, horticulture, cotton, beef and sheep/wool production.^8,25–27 These extend to both workers, farm families and visitors.

• Identifying and tracking new and/or emerging safety risks e.g. ATVs, GPS technology and round bales in cotton.^28,29

• Providing long-term benchmarking indicators (20+ years) for selected commodity sectors, interventions (e.g. ROPS on tractors) and more broadly across all agricultural enterprises.^30,31

• Providing factual detail to incorporate into training resources.³²

• In conjunction with other datasets and information (hospital data, workers compensation records, General Practice clinic data), provide overviews of safety risks.⁸

In relation to outputs, the data have been central to ensuring the evidence-base within the peer-reviewed literature. To date, the information has resulted in 25 peer reviewed papers (334 citations), 14 reports (26 citations), numerous government submissions, five PhD completions, four inquest submissions and the development of an extensive array of free and publicly available training materials covering a diversity of issues and commodity sectors. Many of these training resources continue to be available and widely accessed online e.g. in 2022 the farmers resources section (https://aghealth.sydney.edu.au/resources/resources-for-farmers/) was visited 5,640 times and the research reports (https://aghealth.sydney.edu.au/projects-and-reports/research-reports/) 3,629 times. The primary school information package (RIPPER)³³ was downloaded on 2,215 occasions, with those for Managing the Pressures of Farming (Mental Health)³⁴ and Safety of Quad Bikes and Side by Side Vehicles on Australian Farms³⁵ being downloaded on 1,658 and 1,463, respectively.

Policy

The data from NCIS plays a pivotal role in informing and guiding future polices for the people working on the farm and those around the farm environment (children, visitors). By harnessing relevant data, policy makers can gain valuable insights to formulate targeted strategies and interventions aimed at reducing risks and improving safety outcomes. The impact of these data on policy and practice to improve safety outcomes, can be clearly evidenced by the following examples involving tractors, grain augers and quads (All Terrain Vehicles – ATVs) outlined in Table 1.

Table 1. Examples of the use of fatal injury data to support safety improvements.

Agent	Issue	Response	Outcome
Tractors	Events involving tractor rollovers were the leading agent of fatal injury.	Regulatory requirements mandated that all tractors must be fitted with a Rollover Protection Structure (ROPS). Concurrently, state jurisdictions offered financial incentives for the fitment of ROPS to tractors.	The number of tractor fatalities involving a rollover have reduced by more than 70% over a 30-year period.^Citation31 Of the 88 tractor deaths in the 1989–92 period, 48 involved a rollover. In 2001–04, this reduced to 74 incidents with 25 being rollovers. Most recently (2018–21), there were 44 deaths with nine involving rollovers.
Grain augers	The outer mesh guards on augers were being removed and not replaced as they impeded grain flow and caused blockages, increasing the risk of entanglement with the auger flight.	A new and more practical Grain Auger Industry Standard (2008)^Citation36 was developed along with guidance to fitting the guard.^Citation37	Fatal incidents have dropped from six (1989–92), to four (2001–04) and one (2018–21).
Quads (ATVs)	The data clearly indicated a greater risk of rollover in work-related fatalities for farmers compared to recreational use.	The Australian Competition and Consumer Commission (ACCC) developed and introduced (October 2020), a new Quad Bike Safety Standard to require improved vehicle design (lateral stability/Operator Protection Devices).^Citation38	Work-related deaths have reduced from 33 in the 2001–04 period, to 31 in 2010–13 and 23 in 2020–23.*

*Data on quads (ATVs) are only available from 2001, plus the data for 2023 are current through to 30 June 2023.

Why these data are important and what they illustrate?

The NCIS fatality data fills a void as other systems in Australia are either incomplete (such as workers compensation) and/or lack detail (Australian Bureau of Statistics – Causes of Death). In terms of preventing and/or ameliorating fatal incidents, these data are crucial to identify the leading agents of injury and to provide a more targeted approach for evidence-based interventions drawing on the hierarchy of risk controls. By way of example, a previous AgHealth study identified a significant disconnect between what Australian farmers perceive as the risks on their farm and what hazards and risks cause the highest rates of fatalities in Australian agriculture. Whilst tractors and quads (All Terrain Vehicles – ATVs), have long been the leading two causes of fatal incidents on Australian farms, these were rated by farmers (n = 335) as 20^th and 27^th most risky, respectively.³⁹

The Hierarchy of Controls are placed in order of their effectiveness to eliminate or reduce risk of injury and are the cornerstone of the Work Health and Safety legislation in Australia.⁴⁰ Data drawn from the NCIS indicate that of the six leading agents causing fatal injury (tractors, quads (ATVs), water/dams, farm utilities (pickups), motorcycles and horses), at least 36% of the cases had the potential to be prevented with the use of evidence-based controls in the 2012–2016 period.⁴¹ Further, the longer term trends for 2001–20 indicate a reduction in death rates involving all on-farm fatal incidents (both work and non-work) per 10,000 farms (p = .015) and work-related rates per 100,000 workers (p = .015), with both illustrating considerable fluctuation on a year by year basis. Despite these reductions, there was no change in the work-related rates when assessed against hours worked (p = .276).³⁰

What would make it better?

While this paper outlines the importance of the NCIS and the impact on safety policy and practice, more could be done to improve the information available. In response to a notifiable fatal injury at a place of work, workplace inspectors are required to attend and investigate circumstances of the incident. At this time, information is recorded and gathered (incident details, demographics, employment and job characteristics), along with conducting interviews. From this information a Workplace Health and Safety Investigation Report is compiled. Within the NCIS there is a section to ascertain if a workplace health and safety investigation has been undertaken, however reports are not contained within the electronic database. If such reports could be accessed electronically once cases are closed and any potential legally proceedings completed, this would likely assist with more definitive statements pertaining to events, which may have been missed elsewhere.

Limitations

While this paper focuses on fatal incidents, it would be remiss if some discussion around serious injuries and the wider data landscape for agricultural health and safety was not highlighted. As mentioned earlier, there are a number of sources where farm injury data are available. Currently access to this information on an ongoing basis is either not possible or resource intensive. Developing a system, which makes this information more readily available for research and prevention purposes, would help augment the fatality data.

Conclusion

Without providing accurate, timely, concise and relevant data about injury occurring on farms or due to agricultural work, there is no evidence to drive policy and practice or to evaluate programs of work. This brief report outlines the significant impacts that the availability of the NCIS data has had in Australia. Whilst other data sources such as Workers Compensation and hospital presentations are also important, as they extend the coverage of data captured to inform preventive options, fatality data remains the major focal point for action. As such, the continued utilization and extension of the NCIS data will prove crucial to further reducing the burden of preventable fatal injuries on Australian farms.

Acknowledgments

We acknowledge the indispensable work of those involved in establishing and maintaining all elements of the coronial system across Australia over the past three decades. A particular note of thanks to The Victorian Department of Justice and Community Safety as the source organisation of the data and the National Coronial Information System as the database source of the data.

The Victorian Department of Justice and Community Safety as the source organisation of the national coronial data within Australia and New Zealand and the National Coronial Information System.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.