Making Farm Machinery Safer Lessons from injured farmers

by Wayne Baker, Lesley Day, Karen Stephan, Don Voaklander,

Joan Ozanne-Smith, Jim Dosman, and Louise Hagel

February 2008

RIRDC Publication No 07/190 RIRDC Project No UMO-32A

ii

© 2008 Rural Industries Research and Development Corporation. All rights reserved. ISBN 1 74151 585 8 ISSN 1440-6845 Agricultural Machinery Design and Operation Safety Study: in-depth investigation of farm machinery injury Publication No. 07/190 Project No. UMO-32A The information contained in this publication is intended for general use to assist public knowledge and discussion and to help improve the development of sustainable regions. You must not rely on any information contained in this publication without taking specialist advice relevant to your particular circumstances.

While reasonable care has been taken in preparing this publication to ensure that information is true and correct, the Commonwealth of Australia gives no assurance as to the accuracy of any information in this publication.

The Commonwealth of Australia, the Rural Industries Research and Development Corporation (RIRDC), the authors or contributors expressly disclaim, to the maximum extent permitted by law, all responsibility and liability to any person, arising directly or indirectly from any act or omission, or for any consequences of any such act or omission, made in reliance on the contents of this publication, whether or not caused by any negligence on the part of the Commonwealth of Australia, RIRDC, the authors or contributors..

The Commonwealth of Australia does not necessarily endorse the views in this publication.

This publication is copyright. Apart from any use as permitted under the Copyright Act 1968, all other rights are reserved. However, wide dissemination is encouraged. Requests and inquiries concerning reproduction and rights should be addressed to the RIRDC Publications Manager on phone 02 6271 4165.

Researcher Contact Details Wayne Baker Accident Research Centre PO Box 70A Monash University VIC 3800 Australia Phone: 03 9905 1806 Fax: 03 9905 1809 Email: wayne.baker@muarc.monash.edu.au

Dr Lesley Day Accident Research Centre PO Box 70A Monash University VIC 3800 Australia Phone: 03 9905 1811 Fax: 03 9905 1809 Email: Lesley.day@muarc.monash.edu.au

In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form. RIRDC Contact Details Rural Industries Research and Development Corporation Level 2, 15 National Circuit BARTON ACT 2600 PO Box 4776 KINGSTON ACT 2604 Phone: 02 6271 4100 Fax: 02 6271 4199 Email: rirdc@rirdc.gov.au. Web: http://www.rirdc.gov.au Published in February 2008

iii

Foreword

Machinery is a major component of both fatal and non-fatal injury in agriculture. Reducing farm machinery related injury has been accorded priority by Farmsafe Australia and is specifically addressed by the National Farm Machinery Safety Strategy. This report complements the recent series of reports on farm injury by the Australian Centre for Agricultural Health and Safety (Farm Machinery Injury: Injury involving tractor run-over, Farm Machinery Injury: Injuries associated with posthole diggers, and Farm Machinery Injury: Power take-off shaft guards), outlining preventive strategies for specific items of farm machinery injury. This report undertakes a comparative analysis of the characteristics of farmers who have, and have not, recently sustained a farm work related machinery injury. This comparison was extended to include machinery characteristics. The inclusion of the machines in the study allowed for an in-depth epidemiological and engineering investigation of design features associated with injury. Recommendations have been made for the agricultural industry, and agricultural machine designers and manufacturers. The report recommends action in the following areas:

• Design solutions • Management of design issues • Education and training • Safe workplace systems initiatives • Standards • Further research.

This project was funded by the RIRDC-managed Joint Research Venture for Farm Health & Safety R&D program whose membership includes Rural Industries R&D Corporation, Grains R&D Corporation, Cotton R&D Corporation, Sugar R&D Corporation, Australian Wool Innovation and Meat & Livestock Australia. This report, an addition to RIRDC’s diverse range of over 1700 research publications, forms part of our Joint Research Venture for Farm Health and Safety R&D program, which aims to coordinate and support R&D to develop, implement, monitor and evaluate safe systems of work on farms across all rural industries. Most of our publications are available for viewing, downloading or purchasing online through our website: • downloads at www.rirdc.gov.au/fullreports/index.html • purchases at www.rirdc.gov.au/eshop Peter O’Brien Managing Director Rural Industries Research and Development Corporation

iv

Acknowledgments The Agricultural Machine Design and Safe Operation Study (AMDOSS) was undertaken as an additional component to a large case-control study of serious farm injury among men – the Farm Injury Risk among Men (FIRM) study, funded by the National Health and Medical Research Council. We are therefore most grateful to the members of the FIRM study team for their contribution to this work. As members of the FIRM investigator team, Prof Malcolm Sim, A/Prof Rory Wolfe and Prof John Langley made significant contributions to the scientific basis for the FIRM study. Voula Stathakis and Narelle Hayes co-ordinated the FIRM study, and provided valuable technical advice for AMDOSS. Jonathon Guy did a sterling job on data entry and cleaning. A large dedicated team of nurses and telephone interviewers recruited and interviewed the study participants. Dave Kenny assisted with the on-site machinery inspections. Staff in the emergency and trauma departments and surgical wards at the 19 participating hospitals made time in their busy schedules to assist with recruitment of injured farmers. Mr Rick Roberts, Principal Registrar at the Victorian Coroner’s Court facilitated recruitment of next of kin of fatally injured farmers. Ms Pam Simpson, Department of Epidemiology and Preventive Medicine, Monash University, undertook analyses for the FIRM study, upon which the analyses for AMDOSS was built. The Victorian Injury Surveillance Unit and the Victorian Farmers’ Federation gave their support and assistance to the study. Thanks also to Glenda Cairns for assistance with report formatting. Mr Wayne Jackson assisted with gaining media exposure for the project, and a number of industry and academic experts lent their expertise and provided advice on various aspects of the project, as outlined in Chapter 2. We would particularly like to acknowledge the support and assistance received from Vin Delahunty of the Tractor and Machinery Association, Peter Dunphy from the Farm and Industrial Machinery Association, John Curtis, Graeme Ford, Graham Prince and Eric Sharkey from the Victorian Farmers’ Federation, and the team at the University of Sydney’s Australian Centre for Agricultural Health and Safety. The design and conduct of AMDOSS benefited greatly from collaborative links with researchers involved with the Canadian case-control study of farm machinery injury. We are particularly grateful for the advice and expertise offered to this Australian study from our Canadian colleagues A/Professor Trever Crowe and Mark Ingram from the University of Saskatchewan, and Jim Wasserman from the Prairie Agricultural Machinery Institute in Humboldt, Saskatchewan. We would like to also acknowledge the assistance and advice from two other leading researchers from the USA in the field of agricultural machinery safety, Dr Eric Hallman, Director of the Agricultural Health and Safety Program at Cornell University, and Dr Mark Purschwitz from the National Farm Medicine Center in Wisconsin. Professor Tom Triggs of the Monash University Accident Research Centre (MUARC) provided insightful comment on the human factors aspects of this report. Finally, we are most grateful to the farmers who participated in this study, many of whom did so a short time after sustaining a major injury. The hospitality shown to our project engineer during the site visits was notable. We hope that the findings will be translated into prevention activities and contribute to ongoing improvements in the health and safety of Australian farmers.

v

Abbreviations ACAHS Australian Centre for Agricultural Health and Safety ADR Australian Design Rules AIS Abbreviated Injury Severity Score AS Australian Standards ASABE American Society of Agricultural and Biological Engineers (Formerly ASAE) ASAE American Society of Agricultural Engineers AMDOSS Agricultural Machinery Design and Operation Safety Study CI Confidence interval FEL Front end loader FIMDA Farm and Industrial Machinery Dealers’ Association FIRM Farm Injury Risk among Men FOPS Falling object protection system ISO International Organisation for Standardisation (from French) MUARC Monash University Accident Research Centre NFMRG National Farm Machinery Reference Group NZS New Zealand Standard OHS Occupational health and safety OR Odds ratio PPE Personal protective equipment PTO Power take off RIRDC Rural Industries Research and Development Corporation ROPS Roll-over protection system TMA Tractor and Machinery Association VEMD Victorian Emergency Minimum Dataset VFF Victorian Farmers’ Federation VSB Vehicle Standards Bulletins

vi

Contents Foreword ............................................................................................................................................... iii Acknowledgments................................................................................................................................. iv Abbreviations......................................................................................................................................... v Contents................................................................................................................................................. vi Executive Summary ............................................................................................................................ vii 1. Introduction ....................................................................................................................................... 1

Background....................................................................................................................................... 1 Farm machinery injury...................................................................................................................... 1 FIRM study....................................................................................................................................... 3 Objectives ......................................................................................................................................... 3

2. Methodology ...................................................................................................................................... 5 General overview.............................................................................................................................. 5 Definitions ........................................................................................................................................ 5 Case recruitment ............................................................................................................................... 6 Control recruitment........................................................................................................................... 7 Data collection and management – Stage 1 ...................................................................................... 9 Data collection and management – Stage 2 ...................................................................................... 9 Statistical analysis........................................................................................................................... 12 Industry and expert consultation..................................................................................................... 13

3. Stage 1 Results: Self-reported individual and machine characteristics ..................................... 16 Cases of farm machinery injury...................................................................................................... 16 Case control analysis ...................................................................................................................... 24

4. Stage 2 Results: In-depth investigation of machine characteristics............................................ 30 Recruitment for on-site machine investigations.............................................................................. 30 Univariate machine and machinery operation characteristics......................................................... 30 Multivariate machine and machinery operation characteristics...................................................... 31

5. Design changes for preventing, or reducing the severity of, farm machine related injury ...... 32 Role of design in machinery injury: epidemiological analysis ....................................................... 32 Role of design in machinery injury: engineering analysis.............................................................. 32

6. Discussion – Epidemiological analysis ........................................................................................... 40 Principal findings............................................................................................................................ 40 Strengths and weaknesses ............................................................................................................... 42

7. Discussion – Machine design analysis............................................................................................ 44 Systems approach to machinery design .......................................................................................... 45 The role of standards and legislation .............................................................................................. 47 Design issues determined from incident events.............................................................................. 48 Other agricultural machinery design issues .................................................................................... 57 Achievement of study objectives .................................................................................................... 59

8. Implications...................................................................................................................................... 61 9. Recommendations ........................................................................................................................... 63

Design solutions.............................................................................................................................. 63 Management of design issues ......................................................................................................... 63 Safe workplace system initiatives................................................................................................... 64 Standards......................................................................................................................................... 65 Education and training .................................................................................................................... 66 Further research .............................................................................................................................. 66

Appendix 1 ........................................................................................................................................... 68 References ............................................................................................................................................ 69

vii

Executive Summary What the report is about This report focuses on farm machinery injury. The work reported here identifies individual and machine characteristics that are associated with an increased risk of a serious farm work related injury. A comprehensive analysis of a series of farm machinery events is reported, and through the application of a human factors and systems approach, recommendations are made in relation to improving machinery design to reduce the potential for injury events to occur, and to reduce the severity of resulting injury when such events do occur. Who is the report targeted at? This report is expected to be of interest to the agricultural industry, farm machinery manufacturers and dealers, occupational health and safety authorities, and those involved in farm safety programs. Background Farm machinery related injury has been identified as a priority in Farmsafe Australia’s National Goals, Targets and Strategies, as this type of injury accounts for just over 20% of injury related deaths in agriculture. A unique opportunity to specifically examine risk factors for serious farm machinery injury arose several years ago within the context of a study of all types of unintentional farm injury among men – the Farm Injury Risk among Men (FIRM) study, funded by the National Health and Medical Research Council. The FIRM study recruited seriously injured farmers and farm workers from south-east Australia and collected information about themselves, their working life and the property on which they work. This information was compared with randomly selected farmers and farm workers who were not seriously injured to determine which personal, work and environmental factors were over-represented among those who were injured. This kind of farm injury study had not been conducted in Australia, as earlier studies did not have the benefit of a comparison group of uninjured farmers. Since the FIRM study would include a number of participants with machinery related injury, it provided an ideal platform for an additional in-depth study of the characteristics of the machinery involved, which would significantly enhance the evidence base from which preventive strategies could be developed. This opportunity arose at a time when the National Farm Machinery Safety Strategy had identified that research was needed in the areas of defining the injury problem and its causal factors, and improving machinery design. Aims/Objectives The aim of the Agricultural Machinery Design and Operation Safety Study (AMDOSS) was to identify the individual, farm and machinery characteristics associated with serious work-related farm machinery injury. A secondary aim was to explore the interaction between individual and machinery characteristics. The three hypotheses tested were that: • individual characteristics, such as previous personal farm injury, low level of education, less

than 10 years’ farming experience, and no previous safety training, tend to increase the risk of serious injury associated with farm machinery

• farm machinery associated with serious injuries does not meet current legal standards • design changes, beyond features required by current legal standards, can be devised to prevent,

or reduce the severity of, serious farm machinery injuries. The intended beneficiaries of the study are farmers, machinery manufacturers and dealers. Methods used The study had a case-control design. Farmers and farm workers who sustained a serious farm work related machinery injury (cases) were recruited via hospital emergency departments. Information was collected about themselves, their working life and the property on which they worked (Stage 1).

viii

Where the participant agreed, an on-site inspection of the machine was conducted, to collect information about the characteristics of the machine (Stage 2). A group of randomly selected farmers and farm workers who were not seriously injured (controls) were also recruited and comparable information was collected about them and a similar type of machinery to that involved in the case injury event. The study was restricted to men, who make up the substantial proportions of farmers who are fatally and seriously injured. It is important to consider some implications of the case-control study designs. These studies describe associations between risk factors and health outcomes. An association does not necessarily imply a cause and effect relationship. Other evidence may be required to establish whether there is a cause and effect relationship. Similarly, the finding of no association between a risk factor and a health outcome does not necessarily mean that an association does not exist – the correct interpretation is that there is no evidence for an association. Contemporary human factors systems analysis was used to examine the incident events and associated machinery to identify potential design changes to prevent, or reduce the severity of, serious farm machinery injuries. Several specific design outcomes as well as suggested strategies for hazard management are forthcoming from this means of incident and design investigation. Results/Key findings A total of 85 injured farmers and 205 age-matched uninjured control farmers participated in the study, among whom 37 and 71 respectively took part in Stage 2 – the on-site machinery inspection component. The first stage of the analysis examined the association of each specific hypothesised risk factor with farm machinery injury. The results showed that the following factors were statistically associated with farm machinery injury: • not having attended farm training courses • being less experienced in farming (i.e. having 1–4 years’ farming experience) • having had a hospital stay for a farm work related injury in the previous 3 years • having a medical condition requiring medication • being an employee or contractor • being engaged in seasonal farm work • using a machine that had not been purchased new. We also found that the odds of injury increased by 3–4% for each year increase in the age of the machine. Case farmers with farm machinery injury were statistically less likely to display the following features than were the control farmers who had not recently sustained a machinery injury: • have asthma • have had back pain in the previous 12 months • be using a machine in above average state of repair. One of the limitations of the above univariate analysis is that the influence of each factor on farm machinery injury risk were considered in isolation from each other. As some of these factors are correlated (e.g. having asthma and having a medical condition requiring medication), it would be helpful to be able to identify those factors which, independent of any other factors in the analysis, are associated with farm machinery injury risk. Therefore, logistic regression modelling was conducted to simultaneously examine the relationship between a selected number of these personal and machine characteristics. When the relationship between these variables was taken into account through the modelling, the following factors maintained statistical significance and were independently associated with machinery related injury: • being an employee or contractor • being engaged in seasonal farm work • using a machine that had not been purchased new.

ix

When these other factors were considered simultaneously, the odds of injury increased by 4% for each year increase in the age of the machine, compared with 3% in the first stage (univariate) analysis. In addition, results for three further factors, while not reaching conventional statistical significance, suggested that they may be of interest. There was some evidence that having had a hospital stay in the previous 3 years increased the odds of injury, and that having asthma, and having had back pain in the previous 12 months were associated with a decreased odds of machinery related injury. Several factors identified in the first stage analysis were no longer significant in the multivariable analysis. These include not having attended farm training courses, having 1–4 years’ farming experience, and having a medical condition requiring medication. This means that these factors were less important than the remaining factors above in their association with farm machinery injury. There was no evidence for the association of any of the design or safety related features with the odds of injury. The safety feature score, or performance of formal safety checks, were not associated with a decreased odds of injury. Previous modifications, inconvenient safety features, safety features hampering function or inconsistency with standards and regulations were not associated with increased odds of injury. Associations observed in this study do not necessarily imply a cause and effect relationship. Further, the finding of no association between a risk factor and farm machinery injury does not necessarily mean that an association does not exist – the correct interpretation is that there is no evidence for an association. We found that there is significant potential to address a number of specific risks that exist with current new and old agricultural machinery by advocating design interventions. In addition, systematic design and machinery management issues have been identified, with the intention of encouraging innovative and industry-relevant solutions to be developed from a number of suggested practical solutions. These designs, presented in the context of safe system design principles, serve to not only improve the safe working life of new and existing machines, but also lift the state of knowledge in the agricultural machinery industry. Implications This report aims to inform change at various levels of the agricultural and agricultural machinery industries, advocating a combined effort amongst stakeholders to enhance the state of knowledge by innovation, discussion, and implementation of solutions to key machinery design and machinery management issues. Machinery manufacturers, importers and suppliers will be encouraged that some of the injuries in this study resulted from design aspects that no longer exist on newer machines, and can learn from the many for which the design features are similar on new machines. There are specific actions recommended for occupational health and safety regulatory authorities, and there are implications for both the regulatory and prevention programs of these agencies. The work of Farmsafe Australia, and the National Farm Machinery Reference Group will be supported and enhanced by the findings and recommendations of this report. There are implications of this work for farmers and farm families regarding the ongoing utilisation of older farm machinery, and the associated risk management.

x

Recommendations The following recommendations from the study have been developed with the intention of providing direction for the benefit of the agricultural machinery industry, based on the evidence discussed in this and other studies.

Design solutions

1. Hazards that exist on many mobile grain augers can be controlled by the following commercially available retrofit components: • installing jockey wheels that have a lever to “walk” them into position removes a lifting

and moving manual handling hazard, and makes the job much easier • where required, guarding or a fine mesh cage should be installed over stationary engines

attached to the mobile auger. According to AS 4024.1801:2006, a 25 mm (one inch) mesh constructed 120mm from hazards on plant is sufficient to prevent access beyond the knuckle joint

• installing a hand winch that incorporates an automatic clutch to prevent the handle from flicking back

• guarding over the flighting according to the recommendations of the recently published “Grain Augers” Industry Safety Standard from NSW Workcover.

2. Both rollover and run-over hazards on tractors can be better controlled by maintaining the

operator in the protected zone with the use of seat belts, in addition to a roll-over protection system (ROPS) installation. While it is recognised that many farm workers may choose not to use a seat belt on all occasions, it is considered important that one is at least available for use. Development of a seat belt attachment system that is married to a retrofit seat for older tractors, for bolting on to existing seat mounts where applicable, is recommended.

3. It is recommended that agricultural machinery industry bodies consider means to address

apparent issues with the location of the tractor exhaust: • located on the right hand side of the bonnet, which creates a blind spot when travelling on

the left hand side of public roads • located in such a way that exhaust tends to enter the cabin.

Management of design issues

4. It is recommended that state workplace health and safety authorities mandate the installation of a falling object protection system (FOPS) on tractors that are fitted with a front end loader (FEL).

5. It is recommended that state workplace health and safety authorities advise of the various risks

of jump starting vehicles. Related advice should also be provided regarding the risk of explosion with non-sealed batteries that are located in a confined area.

6. It is recommended that Farmsafe Australia contact the manufacturers of the mobile grain silo

involved in the incident wherein a farm worker was struck by the wheel lifting lever, which occurred largely due to a poorly designed control lever mechanism.

7. It is recommended that all those involved in the agricultural industry give high priority to

addressing design and risk management issues associated with older machinery, given our finding that the risk of farm machinery injury increases by 3-4% with each year of increase in machine age, and that farm machinery tends to provide a long service life on Australian farms.

8. Farmers with older machinery, or machinery which has not been purchased new, should be

encouraged to undertake risk assessments and implement risk management strategies for this equipment. Farm machinery dealers could play a role in assisting this process. Various retrofit options are available to enable safe use of older machinery, although in some instances the

xi

most appropriate risk management may be to develop a plan for replacement of equipment. Other strategies could include using different equipment to undertake the task, upgrading guarding or other safety features, or regular preventative maintenance schedules.

9. It is recommended that resources be made available for a national body such as Farmsafe

Australia to develop a means to gather independent feedback from farmers about the ergonomics and other health and safety aspects of machinery design. This could contribute to an educational database of user experience to inform design improvements, and enhance management of current farm machinery risks. While research organisations are able to provide monitoring over time of the consequences of farm workplace incidents, it would be more effective from a design perspective to have access to direct feedback and ideas from farmers who are invariably intuitive and innovative. In addition to this, the same database could record the details of near misses and incident events.

Safe workplace system initiatives

10. It is recommended that agricultural machinery industry bodies consider the provision of practical and cost effective means to access higher sections of agricultural machinery that require occasional access. Currently, features on the machinery such as wheels, mud guards and implements are being used, where there is a risk of falls. Attention is drawn to the need for solutions to address a wide demographic, including older farmers. Suggestions to consider are: • An access frame with steps situated over the wheels to allow enough reach to clean

windows, and access the roof. • Provision through agricultural machinery dealerships of stable ladders of suitable height.

Wide base ladders would need to be used, as the tractor roof is often not suitable to prevent the ladder slipping sideways.

• Provision through agricultural machinery dealerships of suitable long-reach tools tailored for specific aspects on machinery that require attention, such as cleaning windows.

11. It is recommended that agricultural machinery industry bodies consider means to address risks

with lifting tillage equipment to carry out periodic maintenance such as replacement of consumables. Cost effective and practical design solutions are required in order to provide a secondary load path to allow for the failure of the mechanism, such as hydraulics, that are used to lift the equipment. One way of achieving this could be to attach components to machinery (including retrofit) that can provide the secondary load path. For some machines these components could also serve to provide the lifting force.

12. It is recommended that agricultural machinery industry bodies consider means to address risks

when accessing areas underneath components held up by hydraulics. While manual hydraulic lock-out devices are commonplace, it was widely reported by farmers in this study that they were never used. Means of establishing lock-out systems that allow for this and other human factors with the use of hydraulics is required. Other industries that use hydraulics would also stand to benefit from such systems.

13. It was found that several farmers with serious injuries could have died had it not been for help

arriving in time, either by coincidence, or by others on the farm noting their absence. Farm workplaces should establish means to ensure that communication is available to make arrangements for periodic contact to be made such that an absence is noted within a reasonable time frame.

Standards

14. It is recommended that agricultural machinery industry bodies continue to consider means to ensure the availability and widely understood interpretation of various machinery standards. While current standards should not be seen as perfect, and are subject to continual

xii

improvement, they do provide advice toward a reasonable state of knowledge and some minimum design criteria.

In particular, the industry should continue to make themselves aware of the expected levels of safety of equipment under state OHS law, which has been tested in a number of pertinent cases in recent years. These cases make it clear that the operator protective devices used, including guarding and sensor switch systems, are expected to be generally more thorough than the requirements of Australian Standard AS 2153.1:1996 – Tractors and machinery or agricultural and forestry – Technical means for ensuring safety. Attention is drawn to the recently revised Australian Standard AS 4024:2006 – Safety of Machinery.

15. For a number of specific machines particular to agricultural applications there is ambiguity

with the interpretation of various standards and other safety requirements. The best means to control risks needs to be determined and disseminated in a way that recognises compliance and promotes national uniformity. It is recommended that a discussion paper be developed to consider the potential for the uniform provision of standards for agricultural machinery, which may draw from, but be independent of, existing guidance and standards regimes. A key function of this discussion paper would be to canvas options to determine the best way to achieve national uniform compliance. One model to consider is that of the Australian Design Rules (ADR). An ADR series for agricultural machinery could be developed over time in close consultation with the industry to promote uniform continual improvement to machinery safety, enforce minimum standards with clear guidelines, and encourage the development and uptake of innovative solutions to problematic hazard control issues.

16. Australian agricultural machinery designers and professionals who purchase and import

agricultural machinery could consider better consolidating their skills and knowledge base by developing a professional body or negotiating equivalent relationships with existing organisations. Existing organisations that provide a similar role in enhancing professional standing are: • Engineers Australia, and the Society for Engineering in Agriculture • American Society of Agricultural and Biological Engineers • Tractor and Machinery Association • Farm and Industrial Machinery Dealers’ Association • The Kondinin Group

17. Funding should be made available to develop and support Farmsafe Australia’s National Farm

Machinery Safety Reference Group to provide for ongoing national initiatives to advocate and promote: • nationally coordinated programs, including working closely with state workplace health

and safety authorities and state farmer representative organisations • reduction of duplication of agricultural machinery injury prevention programs • development of shared resources for the benefit of agricultural machinery safety.

Education and training

18. The design and implementation of error tolerant systems in agricultural practice and agricultural machinery design could be viewed as an important element of risk management within the industry. Machinery design education, agricultural workplace training, as well as stand alone OHS education and training curricula, should incorporate the principles of error tolerant systems. Farmers as innovators in the agricultural machinery industry are a very important target group for such educational initiatives.

xiii

Further research 19. There appears to be considerable potential to improve machinery safety with sensors and

interlocks. An engineering issue exists with the introduction of multiple node sensor systems, such as interlocks, on complex machinery such as headers and balers. Each new sensor introduces another failure mode for the machine. This is particularly an issue when interlocks should be designed to fail safely – i.e. if the interlock fails, it should preclude the function of interest from occurring. Methods to overcome this issue have known problems. • One way to overcome reliability issues is to introduce redundant wiring or other systems

that assists with fault diagnosis, resulting in minimal down time in the event of a sensor failing safely. This comes at a cost that must be passed on to machinery owners.

• Increasing reliability by making sensor systems more robust and higher quality also comes at additional expense, and does not guarantee that problems will not occur throughout the life of the machine.

The best way to address this problem of safety systems introducing multiple failure modes could be to use both of the above strategies, and other approaches. This design issue could benefit from further research to find ways of introducing effective operator protection systems that are cost effective and reliable.

20. Further research is required to understand the mechanism by which those with a previous farm

work related injury severe enough to require hospital admission may be at increased risk for a subsequent injury. In the interim, patients being discharged from hospital following a farm work related injury should be advised of their possible increased risk of a subsequent injury. Patients likely to have an ongoing physical limitation may benefit from occupational therapy assessment and support during the initial return to work phase. In instances where a physical limitation persists over a longer period, regular assessments may be required.

21. Further research may also be required to understand the mechanism by which employees,

contractors and seasonal workers are more likely to sustain a machine related injury. This finding could be related to their increased exposure to machine use, or more risky machinery related tasks, than owners/managers. It could also be related to a lower prevalence of having attended agricultural training courses among this category of farm worker, since this study found that injured farm workers were less likely to have undertaken agricultural training courses.

22. Many mechanisms are possible for the continued development of safe design innovations for

agricultural machinery. One of these is Australian Standards. Notwithstanding their important role, standards can also limit innovation, and are subject to sometimes lengthy review processes. Consequently, improvements in standards (and similar instruments such as industry codes of practice), are not seen as the only means to bring about change. Some design issues require serious research and development commitment to ensure that solutions have been well considered, and properly evaluated, in a way that is more thorough than an expert Standard or Code committee can provide. The industry would benefit from research that seeks to optimise the process for designing, evaluating, and translating safe machine innovations to end users, in the agricultural machinery context. Such an investigation would take into account national, international, and hypothetical models for the process.

23. This study has demonstrated the value of an in-depth analysis of incident events, using an

approach that uses the most contemporary incident prevention theory as a benchmark for hazard management. Further research of incident events is required to build on this body of knowledge.

1

1. Introduction Background

In 2005, Australia had 129,900 agricultural businesses, occupying an estimated 58% of the total land area (Australian Bureau of Statistics 2007). The agricultural industry is an important source of employment for Australians living in rural and remote areas. In February 2003, 329,000 people (including unpaid family members) worked in agriculture and services to agriculture (Australian Bureau of Statistics 2003b). In the 2004/05 financial year, the estimated gross value of agricultural production in current prices was $35.6 billion (Australian Bureau of Statistics 2007). Although Australian agriculture no longer constitutes a large proportion of gross domestic product (averaging around 3% in recent years), its presence and activity is significant. The national commitment to occupational health and safety reflected in the National Occupational Health and Safety Strategy 2002–2012 is particularly relevant to the agricultural industry (Australian Safety and Compensation Council 2002). This sector has the second largest number of work-related deaths after transport and storage and the fifth highest occupational death rate after forestry and logging, fishing and hunting, mining and transport (National Occupational Health and Safety Commission 1998). The injury burden not only has a direct human impact to the health of individuals, but also can threaten the viability of the farm business. The Australia-wide cost of injury and illness in the agricultural sector is estimated to be between $0.52 and $1.29 billion annually, with almost 1.7 million working days lost over a twelve-month period as a result (Fragar & Franklin 1999), in addition to the social burden. Consequently, Federal and State governments, and the agricultural industry through Farmsafe Australia, have accorded priority to health and safety on Australian farms. Farm machinery injury

Mechanisation of agriculture delivered large advances in productivity, but also exposed the agricultural work force to a new range of hazards. Machinery is a major component of both fatal and non-fatal injury in agriculture. Tractors are the leading single machinery type, accounting for 14.8% of farm-related fatalities in Australia (Franklin et al. 2000). Each year in Australia, there are on average 36 deaths, and an estimated 270–500 hospital admissions, and 705–6700 emergency department presentations related to farm machinery injury (Farmsafe Australia 1998). While the reduction of fatalities is an important prevention target, non-fatal farm machinery injury tends to be severe, resulting in increased lost work time and hospitalisation periods compared with other farm injury (Day & McGrath 1999; Griffith 1994). The literature on farm machinery injury in Australia, and other mechanised countries, is dominated by studies characterising the frequency, type and distribution of such injury. Mobile farm machinery and plant accounts for 20.8% of farm fatalities in Australia (Franklin et al. 2000). Tractors contributed to 71% of these deaths, followed by grain augers (5%) and post-hole diggers (3%). Agricultural machinery accounts for between 12% and 16% of admissions to hospital for a farm related injury (Fragar & Franklin 2000; Day & McGrath 1999; Day et al. 1997). Machinery type is not systematically identified, but it appears that tractors are also a leading cause of non-fatal machinery injury (Day & McGrath 1999; Day et al. 1997). Powered machinery and equipment is prominent in workers’ compensation claims, accounting for 24% of all claims in the agriculture and horticulture industries in Australia (Fragar & Thomas 2005). A body of work has emerged in Australia which characterises fatal and serious injury events associated with those items of machinery most commonly involved. Coronial files for tractor fatalities from 1989 to 1992 were examined in a national work-related fatality study (National Occupational Health and Safety Commission 1998). This study provided an excellent profile of tractor related fatalities in Australia including a summary of safety and design issues highlighted in coronial findings.

2

Factors associated specifically with tractor run-over fatalities have been identified using a mixed method approach including reference to Australian and overseas literature, and analysis of existing data sources (such as the national work-related fatality study). Recommendations for prevention were then derived from this evidence base, complemented with a review of existing standards, using the hierarchy of control framework (Miller & Fragar 2006). Similar approaches have been used to produce recommendations for prevention of injury from grain augers (Athanasiov et al. 2006), and posthole diggers (Miller et al. 2006). This body of work, upon which the current preventive strategies rely, has been derived from descriptive case series studies, in which data were sometimes incomplete. Such studies are useful for describing the scope of the issue, and identifying patterns, trends and issues for further research (Layde 1990). In some instances, sufficient evidence is available to move directly to a preventive strategy. A complementary approach, which can provide a more sophisticated understanding of aetiological or causal factors, is that provided by analytic epidemiological studies. These types of studies examine the presence or absence of possible causative factors (risk factors) among groups of people with and without the health outcome of interest, facilitating the identification of factors which are over-represented among those with the health outcome under study, and providing some insight into cause and effect (Layde 1990). It is important to consider some implications of these types of study designs. These studies describe associations between risk factors and health outcomes. An association does not necessarily imply a cause and effect relationship. Other evidence may be required to establish whether there is a cause and effect relationship. Similarly, the finding of no association between a risk factor and a health outcome does not necessarily mean that an association does not exist – the correct interpretation is that there is no evidence for an association. To date, all analytic studies for farm machinery injury have been undertaken overseas. Four studies in the USA have identified, in multivariable analyses, independent risk factors specifically for farm machinery injury (i.e. those factors which remain statistically significant after adjustment for other related factors). A number of the independent farm machinery risk factors are associated with exposure to farm work and machinery use (Layde et al. 1995; Gerberich et al. 1998; Sprince et al. 2002; Carlson et al. 2005). These include: • increasing hours of work (except for one study which focussed on tractors) • operating/working with an auger • most time spent on field crops. Demographic or personal characteristics independently associated with an increased risk of machinery injury included: • male gender • being married or having ever been married • fewer years of farming experience • prior agricultural injury • wearing a hearing aid • problem drinking (Gerberich et al. 1998; Sprince et al. 2002; Carlson et al. 2005). Farm risk factors included the presence of non-resident workers on the farm, while factors associated with more modern and affluent farms – having registered cows on the farm, and feeding cows in the barn in summer – were protective for injury (Layde et al. 1995). None of these studies collected data on machinery characteristics or involved machinery inspections to examine design features, and collected limited information regarding machinery exposure. However, a recently completed Canadian case-control study to identify individual, farm and machinery characteristics associated with farm machinery injury (Ingram et al. 2001) did collect data on machinery characteristics and included a machinery inspection component. Cases of farm

3

machinery injury admitted to hospitals in Western Canada, and non-injured comparison farmers from the same area were interviewed to collect information about themselves and their farm. Machinery inspections were conducted on randomly selected cases and the corresponding control participants. Data collection has concluded, although complete results have not yet been published. While some farm injury risk factors may operate across cultures and commodity groups, it is highly likely that some risk factors will vary according to agricultural practices, commodities produced, climate, growing season and culture. Overseas research findings, therefore, may not be readily generalised to the Australian setting, necessitating that analytic studies of farm machinery injury be undertaken in Australia. Further, farming has strong cultural roots, therefore Australian based studies may be more salient and meaningful to Australian farmers. FIRM study

A unique opportunity to specifically examine risk factors for serious farm machinery injury arose several years ago within the context of a study of all types of unintentional farm injury among men – the Farm Injury Risk among Men (FIRM) study, funded by the National Health and Medical Research Council. The FIRM study recruited seriously injured farmers and farm workers from south-east Australia (Victoria, southern New South Wales and eastern South Australia), and collected information about themselves, their working life and the property on which they work. This information was compared with randomly selected farmers and farm workers who were not seriously injured to determine which personal, work and environmental factors were over-represented among those who were injured. This kind of farm injury study has not been conducted previously in Australia. As the FIRM study would include a number of participants with machinery related injury, it provided an ideal platform for an additional in-depth study of the characteristics of the machinery involved. The personal, work and environmental information for these participants could then be integrated with information on the machine characteristics, such as specific safety design features (or lack thereof). The potential therefore existed to capitalise on the FIRM study and include an additional module to examine characteristics of farm machinery involved with serious injuries, which would significantly enhance the evidence base from which preventive strategies could be developed. This opportunity arose at a time when the National Farm Machinery Safety Strategy (Farmsafe Australia 1998) had identified that research was needed in the areas of defining the injury problem and its causal factors, and improving machinery design. Objectives

The aim of the Agricultural Machinery Design and Operation Safety Study (AMDOSS) was to identify the individual, farm and machinery characteristics associated with serious work-related farm machinery injury. A secondary aim was to explore the interaction between individual and machinery characteristics. The hypotheses were that: • individual characteristics, such as previous personal farm injury, low level of education, less than

10 years’ experience farming, and no previous safety training, tend to increase the risk of serious injury associated with farm machinery

• farm machinery associated with serious injuries does not meet current legal standards • design changes, beyond features required by current legal standards, can be devised to prevent, or

reduce the severity of, serious farm machinery injuries.

4

Specific outputs were expected to include: • the identification of injury reducing design features for new equipment • recommendations for modification of relevant standards • recommendations for structurally sound modifications which could be made to existing machinery

to reduce serious injuries • implications for farm safety educational and training resources. For example, if certain individual

characteristics are found to increase the risk of farm machinery injury, specific educational and training resources may be required for sub-groups within the farm workforce.

5

2. Methodology General overview

The study had a case-control design. Farmers and farm workers who sustained a serious farm work related machinery injury (cases) were recruited via hospital emergency departments as a subset of those recruited for the FIRM study (serious farm work injury of any type). Information was collected about themselves, their working life and the property on which they worked (Stage 1). Where the participant agreed, an on-site inspection of the machine was conducted, to collect information about the characteristics of the machine (Stage 2). A group of randomly selected farmers and farm workers who were not seriously injured (controls) were also recruited as part of the FIRM study. Comparable information was collected about them and a similar type of machinery to that involved in the case injury event. The study was restricted to males who comprise 95% of fatal (Franklin et al. 2000), and 71% of farm-work related injuries admitted to Victorian hospitals (unpublished data from the Victorian Injury Surveillance System). Definitions

The study region comprised the catchment areas (defined by postcode) of the major regional hospitals in the state of Victoria, Australia. This area included Victoria, southern New South Wales, and eastern South Australia. The study base was adult males (16 years and older) working on study region farms. Agriculture was defined as crop and livestock activities classified according to the Australian and New Zealand Standard Industrial Classification (Australian Bureau of Statistics 1992), and excluded forestry and fishing. The commodity range produced by Victorian farms reflects most commodity types produced in Australia (Australian Bureau of Statistics 2003a). Cases were members of the study base who sustained a serious farm-work related injury involving machinery, after January 2003. Farm-work included paid or unpaid work related to the agricultural livelihood of the person, their employer or relative. Serious injuries were initially defined as those with an Abbreviated Injury Severity Score (AIS) of 2 or higher (Association for the Advancement of Automotive Medicine 2001). The AIS is a numerical measure of damage due to injury, and ranges from 1 for minor injury, to 6 for an un-survivable injury. It is based on anatomical location, and the actual type of injury, rather than physiological measures or functional outcome. The rationale for the AIS 2 threshold was that these injuries would be expected to result in at least restriction of normal activities. However, within 6 months of commencement, it became clear that some AIS 1 injuries were also in this category, and the case definition was extended to include AIS 1 injuries admitted to hospital, or that subsequently became infected beyond the immediate area of injury. Machinery was defined as agricultural machinery (e.g. tractor, tiling, planting or harvesting implement or attachment), other mechanically powered machinery, equipment or implements used for farm business. Controls were randomly selected members of the study base who had undertaken paid or unpaid farm work in the previous month, involving exposure to the same type of machinery as the corresponding case, and who were age matched to the case within a ± 5 year age range. In instances where an age match within this 10 year age range was not recruited within one month of case notification, the age range was expanded to ± 10 years, but the minimum age of 16 years was retained.

6

Case recruitment

Fatally injured cases were recruited prospectively through the Principal Registrar at the Victorian Coroner’s Court, where all fatal work related incidents are reported and investigated. A letter of invitation to participate was dispatched 6–8 weeks after the date of death to the case next of kin. Next of kin wishing to participate returned the consent form to Monash University and were interviewed as proxy respondents. Non-fatally injured cases were recruited prospectively through the emergency departments of the 14 participating regional hospitals, and 5 metropolitan hospitals including those with major trauma centres and those with specialist services such as microsurgery and ophthalmology, to which these 14 hospitals would refer the most serious cases (Table 2.1).

Table 2.1 Participating hospitals, Agricultural Machinery Design and Operation Safety Study, 2003–2006

Regional hospitals

Metropolitan hospitals

Barwon Health (Geelong) Bendigo Health Care Group Alfred Hospital South West Healthcare (Warrnambool)

Echuca Regional Health Royal Melbourne Hospital

Western District Health Service (Hamilton)

Mildura Base Hospital Dandenong Hospital

Ballarat Health Services Swan Hill District Hospital Royal Victorian Eye and Ear Hospital

Wimmera Health Care (Horsham) Central Gippsland Health Service (Sale)

St Vincent’s Hospital

Goulburn Valley Health (Shepparton)

Latrobe Regional Hospital (Traralgon)

Wangaratta District Base Hospital West Gippsland Healthcare Group These 19 study hospitals treat 74% of farm-work related injuries occurring in the study region and admitted to hospitals in Victoria (unpublished data from the Victorian Admitted Episodes Database). All except four of these hospitals had an electronic injury surveillance system at the time of the study and contributed to the Victorian Emergency Minimum Dataset (VEMD) under an agreement with the Victorian Department of Human Services. The VEMD records details of injuries treated at hospital emergency departments according to the National Data Standards for Injury Surveillance. Consequently, details on the location where each presenting injury has occurred are sought as a matter of routine in each of the hospitals contributing to the VEMD. Male patients over 16 years of age who sustained an injury on a farm were then approached by hospital staff seeking approval to give patient contact details to Monash University for the purposes of contacting them regarding the study, as required by the Health Records Act 2001 (Victoria). Once consent to contact was given, the remainder of the screening and recruitment process was conducted by study staff. Farmers with a machinery related injury were recruited as part of this process. A hospital records audit conducted after the first 5 months of recruitment in the FIRM study (and therefore AMDOSS) revealed that up to approximately 50% of potentially eligible patients were not being approached for participation during their treatment in the regional hospital emergency departments (Day et al. 2007). Therefore, an additional component to the recruitment process was developed. Hospital staff generated regular (e.g. monthly) lists of potentially eligible cases using the injury surveillance system to identify males, 16 years and over presenting with an unintentional injury sustained on a farm during work-related activities. Patients who had not been approached were sent a letter from the hospital providing the opportunity to opt-out of telephone contact by study staff. The hospital then provided contact details for those who did not opt-out, and study staff completed the recruitment process. One study hospital required an opt-in system, for which the letter provided the opportunity to actively opt-in to the study by contacting the study staff directly.

7

Control recruitment

When a case was enrolled, matching controls (at least two per case – see Data collection and management for further explanation) were recruited by telephone using contact numbers from the study areas selected randomly from the electronic white pages. Respondents who were potential controls (i.e. working on a farm in previous month) were screened for age matching, and then for a match on exposure to the same type of machinery as the case. Respondents were eligible if they had worked on a farm in the previous month, their age was within the required range of the case age, and they had been exposed in the previous month to the same type of machinery as the case. Machinery was classified as a match if it performed the same general function and had a similar hazard profile for the user, as the machine involved in the case event. Machinery classification (Table 2.2) was based on the Farm Injury Optimal Dataset (Fragar et al. 2000). Proxy interviews with the next of kin were conducted for controls matched to a fatally injured case. An unusual feature of control recruitment was a “dynamic pool”, introduced as an efficiency measure (Day et al. 2005). Eligible controls that did not age- or machinery-match a case were placed on a list for possible matching with cases occurring in the future. These controls remained on the list for a maximum of one month after their initial contact, to ensure that controls were exposed during approximately the same season as the cases. When possible, selection of controls was made from the list in order of decreasing time in the pool. Following this, if controls were still required for a case then recruitment reverted to the random telephone number approach. Recruitment of cases and controls was supported early in the project by seeking media coverage of AMDOSS. In addition, greater public awareness of the study amongst farm machinery end users would help to publicise and disseminate results of the study. A media strategy was developed with the assistance of the Monash University professional media office. A public figure, Wayne Jackson, the well known former CEO of the AFL, who had recently been injured in a farm machinery incident was recruited to help promote the project. Extensive media coverage was achieved, the most important of which included: • Weekly Times article (Victoria and southern NSW’s leading rural newspaper) • Victorian Farmer article (Membership magazine for Victorian Farmers’ Federation) • ABC Radio interview, distributed to over 20 local rural stations • Commercial radio interview, for ACE radio network, with distribution to 7 AM and 6 FM regional

Victorian stations.

8

Table 2.2 Machinery categories, Agricultural Machinery Design and Operation Safety Study, 2003–2006.

Main Group Machinery Examples Case Machines

Powered hand tools Air pruner Air secateurs Shearing handpiece

Air pruner Air secateurs Shearing handpiece

Harvesting equipment

Harvester/header: Grain harvester Potato roller/harvester Silage harvester Hay baler (pulled by tractor) Small square hay baler Round baler Large square hay baler Baler

Harvester Hay baler and tractor Hay baler Potato roller Header Baler

Tillage equipment

Air seeder Direct drill seeder Bed forming implements Harrows Plough Scarifier Other tillage equipment

Air seeder Bed forming implement Grader/seeder Harrow tyne Plough Scarifier

Augers, elevators Elevators Augers

Hay Elevator Auger

Cherry picker Cherry picker Cherry picker

Field bin Potato field bins Grain field bins Storage hoppers

Mobile field bin

Tractor-based lifting devices

Front end bin/bucket Front end loader

Front end bin/bucket Front end loader Moore scoopmobile

Stationary engine driven plant

Pump Water pump Tub grinder (chaff cutter)

Pump Water pump Tub grinder (chaff cutter) Sprayer (pump) with PTO

PTO driven plant

Hay feeder Slasher Sprayer (pump) with PTO added Tractor (with slasher) Binder

Hay feeder Slasher Tractor (with slasher) Binder

Fence post implements Post driver Post hole digger

Post driver Post hole digger

Tractor

Non-articulated tractors, with or without cabins FEL 3 point linkage PTO FOPS ROPS Hydraulics Articulated tractor

Non-articulated tractors, with and without cabins FEL 3 point linkage PTO FOPS ROPS Hydraulics Articulated tractor

Wool press Wool press Wool press Irrigation machinery (non-pump) Linear irrigator No cases

Earthmoving equipment Bulldozer, grader, laser grader No cases

9

Data collection and management – Stage 1

Demographic, farm and exposure data for Stage 1 were collected by interview from all cases and controls using a structured questionnaire. This questionnaire collected more data than that required for AMDOSS, as the participants were also involved in the broader study of all serious farm injury (the FIRM study). For cases, the interview occurred as soon as possible after the injury had occurred. If possible, the farm owner/manager was interviewed for the farm characteristics section if the case was unable to answer these questions. The questionnaire consisted of three sections:

1. Farm characteristics 2. Personal characteristics 3. Injury incident.

The farm characteristics included details on major commodities grown, farm size, income, management structure, farm work-force, injury history on the property, presence of hazards presence of protective equipment, safety policies and practices, and recent major changes to size, commodities produced, staff, equipment or production methods. The personal characteristics section included details on the individual’s employment in the agricultural sector, relationship to farm owner, age, gender, education and specific agricultural training, farm injury history, use of glasses/hearing aids, health conditions, medications, and sleepiness. The injury incident and exposure section included details of the injury event and the machinery involved, and data on exposure to specific risk factors at the time of the injury event. For controls, this third section included details of the circumstances of the most recent day when the item of machinery was used, and data on exposure to specific risk factors at the time of using the machine. Data collection and management – Stage 2

Permission was also sought from the non-fatal cases, and their matching controls, for participation in an in-depth investigation component (Stage 2) for which the study engineer visited the farm and undertook an additional data collection component which included inspection of the machine involved in the injury event. If the case did not agree to this component, then participation of the control in this component was not sought. If the case did agree, matching controls were recruited and interviewed until two controls had agreed to participate in the in-depth investigation. In some instances therefore, more than two controls were recruited for a case. If the participating case or control was an employee, their consent was sought to contact their employer to seek permission to inspect the machine (with the exception of those participants recruited from one of the regional hospitals, where approval was not given for this step). Data for Stage 2, the in-depth investigation, was collected using a structured worksheet which consisted of seven sections:

1. Injury event 2. Human factors 3. Safety and training 4. Machine use 5. Machine repair 6. Machine maintenance 7. Design.

The development of the worksheet benefited from the input of researchers involved with a similar Canadian case-control study of farm machinery injury (Ingram et al. 2001, 2003). Table 2.3 provides a summary of the purpose and type of data collected, for each of these sections. To avoid ambiguity, questions were taken to apply to the specific injury event or the equivalent mechanism on the control machine. A detailed protocol ensured consistent application and interpretation of the questions. Photographs of each case and control machine were taken. Where

10

possible (where other machinery or material was not obstructing the camera location), a photo was taken on all 4 sides, and from the four corners, in the horizontal plane. Photos were also taken from other angles where possible, as well as targeted photos of aspects of interest on the machine, particularly including hazards to operators. Components that were involved in the injury event were of particular interest, and equivalent or comparable components were photographed on the control machines.

Table 2.3 Components of the in-depth machinery inspection worksheet, Agricultural Machinery Design and Operation Safety Study, 2003–2006

Section No.

Section Title Purpose and data collected

1 Injury event

the injury event (or most recent day worked with the machinery for the controls) section served to assist the participants in recalling the circumstances of the event or most recent day

2 Human factors

physical, visual, auditory, or cognitive distractions which may have been present at the time

3 Safety and training

formal safety or operational checks, awareness of regulatory requirements, and training with the machine

4 Machine use

frequency of machine use, modifications, and risk management strategies, irrespective of whether this was relevant to the injury event

5 Machine repair

previous repairs and their potential impact on machine design and function, when relevant to the injury event

6 Machine maintenance

maintenance philosophy (condition-based or prevention-based) and maintenance routines, frequency of preventive maintenance, and circumstances surrounding any condition based maintenance

7 Design

current safety features, their impact on function, suitability for Australian conditions, and design features which may have prevented, or reduced the severity of, the injury

In order to test the study hypothesis that farm machinery associated with serious injuries does not meet current legal standards, each machine was separately assessed for consistency with current standards and regulations, according to the criteria detailed in Table 2.4.

11

Table 2.4 Criteria used to assess consistency with specifications in standards and regulations, Agricultural Machinery Design and Operation Safety Study, 2003–2006

Specification Interpretation

Agricultural Machinery Australian Standards: • AS /NZS 2153 – Tractors and machinery for

agriculture and forestry – Technical means for ensuring safety

• AS1064 Agricultural and light industrial equipment – operator controls – symbols

• AS1211 Guards for Agricultural PTO Drives • AS1636 Tractors – Roll over protective structures –

Criteria and tests

• emphasis is on moving parts • guard need not prevent all human access in some

cases (e.g. Shield guard) • strength – suitable to withstand 1200N • if guard is sometimes used for a step, must be

suitably designed for step requirements (non-slip, level, etc.)

• interlock need not be installed in some cases • suitable decals and control symbols in place • no specific emphasis on control of risks for

maintenance and other secondary use scenarios Machinery Safety Standard AS 4024 – Safety of Machinery

• rigorous guarding requirements including ensuring that persons (including non-operators of the machine) are excluded from the danger area entirely

• interlocking requirements, including the need for the interlock to either remove the hazard (e.g. stopping the machine) or disallow the guard to open until hazard is removed (e.g. the machine runs down to a stop)

• also note – design principles, ergonomics, and display requirements.

• no specific emphasis on control of risks for maintenance and other secondary use scenarios

Occupational Health and Safety Law • Victorian OHS Act 2004 • Victorian OHS Plant regulations 1995 • Codes of Practice • Guidance Notes

• risk must be controlled adequately using hierarchy of controls and subject to a risk assessment. There needs to be evidence of control of the risk using such a methodology, particularly those that are to be expected on the machine. In cases where PPE and training (administrative controls) could control the risk, this is taken into account, but not to the extent that behavioural based training can overcome an inherently hazardous machine

• guarding principles • design life • non-primary function operational safety such as

maintenance, and clearing of blockages A separate assessment was made for an interpretation of requirements for the first party (the primary machine operator) and any third party. Assessment of consistency with these standards and regulations for the role of the first party was made with respect to the operator in the usual operating position to perform the primary function for which the machine was designed. Assessment of consistency with these standards and regulations for the role of any third party was made with respect to the machine in all states of operation (i.e. whether or not the engine is running, or whether components are coupled to the engine and running). Consideration was given to the possible physical location of people (and in some cases, livestock) near the machine. This took into account: • assistants to the primary operator • contractors, workers, family members, visitors, or members of the public who may come nearby

for any reason e.g. making enquiries of the primary operator, where the machine is being used near a road (such as slashing)

12

• the primary operator in the event that they leave the usual operating position for the machine’s primary function to inspect the machine or approach it for any reason. This consideration takes into account the hazards whether or not the engine is running, or whether components are coupled to the engine and running, unless something precludes the existence of a hazard if the operator is not in the primary operating position.

In many cases, the assessment of whether these regulations and standards have been met requires a degree of judgement. This subjective assessment has been made across all cases and controls by the project engineer after all cases and controls were recruited and the on-site inspections completed.. Standards of interpretation were used according to the level of rigour demonstrated in court, using case summaries from instances in which occupational health and safety law has been tested. These have been accessed via the Victorian Workcover Authority website during the study period. All assessments of standards and regulatory requirements were systematically reviewed for consistency after all site visit data had been collected. Assessment of consistency with the standards and regulations was made with reference to the injury event. Therefore, aspects of the machine that were not related to the incident did not influence the judgement as to whether the machine met current standards and regulations. Equivalent mechanisms were inspected and assessed on the control machines, therefore the same principal applied to the control machines. Statistical analysis

As the questionnaire for Stage 1 collected more data than that required for AMDOSS, the statistical analysis utilised only those variables relevant to the specific hypotheses for AMDOSS. Statistical analyses were performed using Stata Version 8.2 (Statacorp LP, College Station, Texas, USA). Descriptive analyses were conducted to describe the cases in terms of their personal characteristics, the farm on which they worked, circumstances of injury and machinery involved. Univariate logistic regression was conducted to determine the relationship between each of the individual personal or machine characteristics and risk of machinery injury, using the data collected via questionnaire in Stage 1. Further univariate logistic regression analyses were performed to investigate the association between machinery and operation characteristics and injury risk, using data collected during the on-site inspections in Stage 2, for which the sample size was less than that for Stage 1. Multivariate logistic regression analyses were conducted to determine which personal and machine characteristics were independently related to risk of farm machinery injury, after adjustment for other related factors. Age, season of injury, machine type, and all variables with a p-value of less than 0.2 in the univariate analyses were entered into the model. An interaction term was also included to determine if age of respondent and age of machinery had an interactive effect on risk of machinery injury. Backward stepwise regression was conducted manually to determine which variables remained as independent predictors of outcome, with a p-value for removal from the model of 0.10. Once the final set of predictors was obtained from the stepwise procedure, each of the other variables (both those that were originally excluded because their p-value was greater than 0.2, and those that were removed during the stepwise procedure) was entered into the model individually to determine whether they confounded the association between any of the predictors and outcome. If the estimates for any of the variables changed by more than 10% upon addition of a particular variable to the model, likelihood ratio tests were performed to determine whether the addition of that potential confounder improved the fit of the model.

13

Industry and expert consultation

Comprehensive industry and expert consultation took place throughout the project, with the aim of ensuring that the industry was engaged, informed, and had the opportunity to provide feedback, and to seek expert input to improve study outcomes. Agricultural Machinery Design and Operation Safety Study Working Group A stakeholder-focussed Working Group was formed for AMDOSS, providing a formal means of obtaining feedback from stakeholders within the industry. Three Working Group meetings, including the use of teleconference facilities for interstate and overseas members, provided a valuable forum for discussion about aspects of the project, and the implications that were arising from preliminary results. The membership of the Working Group is shown in Table 2.5 – changes of the membership that occurred during the project are noted.

Table 2.5 Membership of the working group, Agricultural Machinery Design and Operation Safety Study, 2003–2006

Name Position and address Role in/expertise for working group

Mark Ingram Project Officer Institute for Agricultural Rural and Environmental Health, University of Saskatchewan, Canada

M Ag Eng awarded for similar research conducted in Canada

James Houlahan

Deputy Director Australian Centre for Agricultural Health and Safety

OHS risk management practitioner, consulting on OHS for agricultural sector, considerable knowledge of farm machinery and its operation, expertise in OHS training in agriculture, farmer, translating research into practice

David Phillips, later, John Curtis

Project Officer Victorian Farmsafe Alliance Victorian Farmers’ Federation

Occupational health and safety training, dairy farmer, considerable knowledge of farm machinery and its operation

Trevor Crowe Assistant Professor Department of Agricultural and Bioresource Engineering University of Saskatchewan, Canada

Interest and expertise in agricultural machinery, instrumentation and automation

David Logan Senior Research Fellow Monash University Accident Research Centre (MUARC)

Mechanical engineering expertise, considerable experience in motor vehicle occupant protection research, motor vehicle design rule testing, vibration testing

Michael Regan

Senior Research Fellow MUARC

Applied experimental psychologist with specialist expertise in human factors and ergonomics

Eric Sharkey Grains Group Victorian Farmers’ Federation

Grain farmer, considerable experience in farm machinery operation

Wayne Baker Research Fellow, Project Engineer MUARC

Mechanical engineering, agricultural experience, project engineer, OHS and safety systems specialist

Lesley Day Senior Research Fellow, Principal Investigator MUARC

Injury epidemiology and prevention, expertise in farm injury research, project manager and chair of working group

National Farm Machinery Safety Reference Group Both the Principal Investigator, Dr Lesley Day, and Project Engineer Wayne Baker are members of the National Farm Machinery Safety Reference Group (NFMSRG). This group (Table 2.6) regularly convenes in Sydney (usually one or two times per year) under the leadership of the Australian Centre for Agricultural Health and Safety. An initial interactive presentation was given to the group and subsequently regular progress summaries were provided.

14

Table 2.6 Membership of the National Farm Machinery Reference Group

Name Organisation

Vin Delahunty Tractor and Machinery Association

Peter Goodwin Goodwin Kenny

Wayne Baker MUARC

Donald Sutherland Chairman – Farmsafe Australia

Assc/Prof Ann Williamson UNSW – NSW Injury Risk Management Research Centre

Justin Crosby NSW Farmers’ Association

Prof. Lyn Fragar Australian Centre for Agricultural Health and Safety

John Temperley Australian Centre for Agricultural Health and Safety

Denita Wawn National Farmers’ Federation

Antonia Hawkins Australian Centre for Agricultural Health and Safety

Sam Beechey Australian Workers Union

Kirrily Pollock Australian Centre for Agricultural Health and Safety

Assc/Prof Lesley Day MUARC

Bruce Marshall Workcover NSW

Damien Bromly Workcover NSW

Bruce Gibson Victorian Workcover Authority

Keith Ferguson Workcover Queensland Tractor and Machinery Association The Tractor and Machinery Association (TMA) is Australia’s most significant industry group for manufacturing, importing and supplying agricultural machinery. The association has been directly engaged throughout the project. Executive director of the TMA, Vin Delahunty, visited MUARC to discuss the project, and project engineer Wayne Baker has attended the TMA annual general meetings in 2004 and 2006. At the 2006 annual general meeting, a presentation was given to provide an update on the ongoing project, and to garner interest in the forthcoming results. This was well received by the industry, resulting in positive feedback and commentary. Farm and Industrial Machinery Dealers Association The Farm and Industrial Machinery Dealers Association (FIMDA) is Australia’s most significant industry group specifically for dealers of agricultural machinery. This group represents a key stakeholder in the industry, as dealers have an important role to play in the supply and maintenance of safe agricultural machinery, as well as in the provision of information and training to farmers. FIMDA has been directly engaged throughout the project. Information and updates was provided to the FIMDA members with the assistance of FIMDA division manager Peter Dunphy and Chairman Ian Goding, and Warren Mills, the managing director of Occupational Health and Safety training and service provision company, CR Management Systems. Farmsafe Victoria and Victorian Farmers’ Federation Principal Investigator, Dr Lesley Day, is a member of the Farmsafe Victoria Committee, and Project Engineer, Wayne Baker, is an observer on the committee. Progress reports have been provided to the Committee and input on project directions and recommendations have been made by the Committee.

15

MUARC Expert Panel An internal Expert Panel was formed to review a set of case and controls and discuss in detail the design, ergonomic, and regulatory implications of the designs, as a mechanism for drawing on the depth of experience that MUARC has with incident investigation. The process for the case and control review was based on the procedure used by the MUARC motor vehicle crash investigation team. These two areas of activity, investigation of on-road motor vehicle crashes, and in-depth investigations of machinery incidents, provided an opportunity to cross-fertilise ideas between the two project teams. A trial review was conducted in which one case and the two matched controls were presented to the panel, the key machine features summarised, photographs presented, and the injury event was described. The panel then made determinations of the error mode (using a standard framework including human error), and the contribution of the design to hazardous operational procedures. Following this, there was discussion of design changes or practices that may have prevented the injury event, or reduced the risk of the event occurring. The panel provided feedback as to the case and controls under investigation, and were able to advise on the methodology used to consider human factors, and regulatory implications of design changes. The format for case and matching control review was not repeated thereafter, as it was agreed it would be more efficient for individual members to be approached for advice on specific issues. Membership of the review panel consisted of:

- Dr Michael Regan, Senior Research Fellow, MUARC, human factors psychologist - Dr David Logan, Senior Research Fellow, MUARC, mechanical engineer experienced in crash

investigation - Mr Dave Kenny, Technical Officer, MUARC, experienced in crash investigation - Mr Wayne Baker, Research Assistant, MUARC, mechanical engineer and machinery project

officer, experienced in operation of agricultural machinery - Dr Lesley Day, Senior Research Fellow, MUARC, epidemiologist and machinery project

manager Australian Centre for Agricultural Health and Safety The University of Sydney’s Australian Centre for Agricultural Health and Safety (ACAHS), based in Moree, NSW, is Australia’s foremost centre for research into rural health and safety. Staff at the centre also provide executive support for a number of relevant committees including Farmsafe Australia and the National Farm Machinery Reference Group. The ACAHS have been kept informed of the progress of AMDOSS through these existing networks. In addition, further collaborative opportunities were provided when project engineer Wayne Baker assisted the ACAHS on a number of projects related to agricultural machinery and workplace safety. Visiting experts from the USA The Monash University Accident Research Foundation, as part of an ongoing program to encourage networking from leading overseas experts, sponsored a visit from two leading researchers from the USA in the field of agricultural machinery safety, Dr Eric Hallman, Director of the Agricultural Health and Safety Program at Cornell University, and Dr Mark Purschwitz from the National Farm Medicine Center in Wisconsin. In addition to presenting a joint seminar, Drs Hallman and Purschwitz joined the project engineer on a field visit for some on-site inspections. This networking opportunity provided valuable feedback and commentary about the project.

16

3. Stage 1 Results: Self-reported individual and machine characteristics Cases of farm machinery injury

A total of 258 male farmers aged 16 years and over and injured during farm work were interviewed for the FIRM study, among whom 85 had sustained their injury in association with farm machinery, forming the case group for AMDOSS. Slightly more than one quarter were recruited from hospitals in the Central region (Table 3.1).

Table 3.1 Regions of Victoria, recruitment of machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Region of Victoria Cases Central 24 Western 17 Mallee 16 Northern Country 9 Wimmera 9 Northeast 8 West and South Gippsland 2 Total 85

More than half the cases were owners or managers (62.4%). The average of age of the farmers with machinery related injury was 48.4 years. The age distribution ranged from 19 to 78 years, and followed a normal distribution, despite an apparently higher than expected peak in the 64 to 69 year age group (Figure 3.1).

Figure 3.1 Age distribution of machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

05

1015

Freq

uenc

y

19 24 29 34 39 44 49 54 59 64 69 74 79

17

Of the 80 cases who reported property size (Figure 3.2), over half (58%) worked on properties that were less than 500 hectares, while just over a third of the properties were between 500 and 2500 hectares (38%). Properties over 2500 hectares were rarely reported (5%). Five cases did not report the property size; four of these did not know the size of the property, while the other case preferred not to answer.

Figure 3.2 Reported property size (hectares), machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

02468

10121416182022242628

0-99 100-499 500-999 1000-2499 Over 2500

Property size (hectares)

Num

ber o

f cas

es

Approximately a quarter of the cases worked on farms on which the main commodity produced was grains (27%) (Table 3.2). The next most common commodity was horticulture and fruit growing (22%), of which grape growing (13% of total) was the most frequent type. Sheep farming accounted for 18% of the machinery injuries, while almost 10% of the cases were injured by machinery on beef cattle farms.

18

Table 3.2 Main commodity group produced on the farm, machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Main Commodity Subgroup

Number of cases (%)

Grains only (wheat, barley, oats, etc.) 23 (27.1) Horticulture & fruit growing

Plant nurseriesPotatoes

VegetablesGrapes

Fruit

19 (22.4) 1 (1.2)5 (5.9)1 (1.2)

11 (12.9)1 (1.2)

Sheep farming only WoolMeat

Wool & meat

15 (17.7) 3 (3.5)5 (5.9)7 (8.2)

Beef cattle farming only 8 (9.4) Grain & animal farming

Grain & sheepGrain & beef cattle

Grain/sheep/beef cattle

7 (8.2) 4 (4.7)1 (1.2)2 (2.4)

Dairy cattle farming only 7 (8.2) Poultry farming

MeatEggs

3 (3.5) 2 (2.4)1 (1.2)

Other livestock Horse farmingGoat farming

2 (2.4) 1 (1.2)1 (1.2)

Sheep & beef cattle farming 1 (1.2) Total 85 Cropping related activities (41%) were the most commonly reported activity at the time of the injury, with harvesting, picking and cutting accounting for over a quarter of those injuries (12% of total) (Table 3.3). Activities related to maintenance of buildings, structures, equipment and land accounted for over a third of the injuries (38%), with machinery maintenance being the most common specific activity conducted at time of injury (21% of total).

19

Table 3.3 Activity at time of injury, machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Activity at time of injury Specific activity

Frequency

Cropping related activities Cultivating, ground preparation, mulching

Pesticide application, ground rigPlanting, seeding

IrrigatingPruning

Harvesting, picking, cuttingElevating, augering

Boxing, carting, gradingHaymaking

Mechanical loading, unloadingInspecting

Other crop related activity (NEC)

35 (41.2) 1 (1.2)1 (1.2)2 (2.4)2 (2.4)4 (4.7)

10 (11.8)4 (4.7)1 (1.2)2 (2.4)4 (4.7)1 (1.2)3 (3.5)

Farm maintenance of buildings, structures, equipment and land Maintenance of other structures

Machinery maintenanceEarth moving, bulldozing

FencingInspecting

Weed controlOther maintenance (NEC)

Machinery modification

32 (37.7) 1 (1.2)

18 (21.2)1 (1.2)6 (7.1)3 (3.5)1 (1.2)1 (1.2)1 (1.2)

Animal related activities Feeding, watering animals

ShearingMulesing

Baling, pressing

10 (11.8) 2 (2.4)3 (3.5)1 (1.2)4 (4.7)

Transport related activities Other vehicle (NEC)

5 (5.9) 5 (5.9)

Forestry Trimming

Loading

2 (2.4) 1 (1.2)1 (1.2)

Other codes Other unspecified activity (NEC)

1 (1.2) 1 (1.2)

Total 85

20

By far the most common cause of injury was machinery in operation (52; 61%) (Figure 3.3). Note that some injuries caused by machinery in operation occurred during maintenance activities. Fewer than 10% of injuries were caused by machinery during maintenance or modification (not while machinery was in operation), or falls, or being struck by an object or person (8; 9% respectively).

Figure 3.3 Cause of injury, machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

0

10

20

30

40

50

60

Mac

hine

ry in

oper

atio

n

Falls

Stru

ck b

yob

ject

or

pers

on

Mac

hine

rydu

ring

mai

nten

ance

or

mod

ifica

tion

Tran

spor

tre

late

d

Cut

ting/

pier

cing

obje

ct

Oth

er (i

ncl.

heat

sou

rce,

smok

eel

ectri

city

)

Num

ber o

f cas

es

Tractors accounted for the largest proportion of injuries (25; 29%), followed by harvesting equipment (13; 15%) (Table 3.4). Powered hand tools, tillage equipment and stationary engine driven plants each accounted for approximately 8% (7) of injuries, while augers/elevators accounted for 7% (6) of injuries.

21

Table 3.4 Machine type involved in the injuries, machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Machine type Subgroup

Frequency

Tractor 25 (29.4) Harvesting equipment

HarvesterHay baler and tractor

Hay baler Potato roller

HeaderBaler

Windrower

13 (15.3) 6 (7.19)

1 (1.2)1 (1.2)1 (1.2)2 (2.4)1 (1.2)1 (1.2)

Powered hand tools Air pruner

Air secateursShearing handpiece

7 (8.2) 1 (1.2)3 (3.5)3 (3.5)

Tillage equipment Air seeder

Bed forming implementGrader/seeder

Harrow tynePlough

Scarifier

7 (8.2) 2 (2.4)1 (1.2)1 (1.2)1 (1.2)1 (1.2)1 (1.2)

Stationary engine driven plant Pump

Water pumpTub grinder (chaff cutter)Sprayer (pump) with PTO

Boxing machine

7 (8.2) 3 (3.5)1 (1.2)1 (1.2)1 (1.2)1 (1.2)

Augers, elevators Hay elevator

Auger

6 (7.1) 1 (1.2)5 (5.9)

Fence post implements Post driver

Post hole digger

5 (5.9) 4 (4.7)1 (1.2)

PTO driven plant Hay feeder

SlasherTractor with slasher

Binder

4 (4.7) 1 (1.2)1 (1.2)1 (1.2)1 (1.2)

Tractor-based lifting devices Front end bin/bucket

Front end loaderMoore scoopmobile

4 (4.7) 1 (1.2)

2 (2.42)1 (1.2)

Cherry picker 2 (2.4) Wool press 3 (3.5) Field bin 1 (1.2) Irrigation machinery (non-pump)

Linear irrigator 1 (1.2)

1 (1.2) Total 85

22

An inspection of the data on nature of injury and body part involved (Table 3.5) revealed that a large proportion of injuries involved crushing or traumatic amputation of the hand (32%). There were also a large number of multiple injuries involving more than one body location (13%). Fractures were common (18%), and occurred over several body parts. Two of the cases were fatally injured (2%). Most of the cases (67; 79%) were admitted to hospital following their injury while the remaining 16 (19%) were not admitted. A further indication of injury severity is provided by the Abbreviated Injury Severity Score which ranges from 1 for minor injury to 6 for an un-survivable injury. The majority of cases sustained injuries at the lower end of this scale (4.7% AIS 1, 71.8% AIS 2). However, 20% sustained an injury of AIS 3, and 3.5% of AIS 5.

23

Table 3.5 Nature of injury and body part involved, machinery cases, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Open

wound (excl. eye)

Fracture (excl tooth)

Dislocation Sprain or strain

Injury to nerve

Crushing injury

Traumatic amputation

Burn or corrosion

Foreign body in external eye

Foreign body in soft tissue

Electrical injury

Multiple injuries of >one nature

Total

Head (excl. face)

1 1 2

Face (excl. eye)

1 1 2

Thorax 1 3 2 6 Lower back 2 1 3 Pelvis 1 1 2 Upper arm 1 1 Elbow 1 1 Forearm 2 1 3 Wrist 1 1 Hand (incl. Fingers)

5 12 15 3 35

Thigh 1 1 Lower leg 2 1 1 4 Ankle 1 1 Foot (incl. Toes)

1 1

Multiple injuries (involving more than one body location)

1 3 1 2 1 1 11 20

Eye 2 2 Total 10 15 1 1 1 16 16 2 2 1 1 19 85

24

Case control analysis

When examining the results below, it must be remembered that case-control studies describe associations between risk factors and health outcomes. An association does not necessarily imply a cause and effect relationship. Other evidence may be required to establish whether there is a cause and effect relationship. Similarly, the finding of no association between a risk factor and a health outcome does not necessarily mean that an association does not exist – the correct interpretation is that there is no evidence for an association. Two hundred and five age and machinery-matched controls were recruited for the study. Of the controls, 30.7% were matched to within two years of the case age, 68.3% were matched to within five years, and 100% were matched to within ten years. The mean age of controls was very similar to that of cases (Table 3.6). With respect to farm characteristics, cases were more likely to be employees, rather than owners/managers of farms, compared to controls. The distribution of property size was similar in both groups, as were the main commodities of the farms on which they worked. The one exception to this was that controls were marginally more likely to work on dairy cattle farms only than cases.

Table 3.6 Age and farm characteristics of machinery cases and controls, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Characteristic Cases N [%]

Controls N [%]

Age mean [sd] 48.4 [13.2] 48.2 [12.0] Position on farm Owner/manager Employee

53 [62.4] 32 [37.7]

169 [82.4] 36 [17.6]

Farm size (hectares) 0-99 100-499 500-999 1000-2499 > 2500

21 [24.7] 25 [29.4] 12 [14.1] 18 [21.2]

4 [4.7]

49 [23.9] 82 [40.0] 23 [11.2] 40 [19.5] 10 [4.9]

Main commodity Poultry farming Horticulture and fruit growing Grain only Grain and animal Sheep and beef Sheep only Beef cattle only Dairy cattle only Other livestock Services to agriculture Other

3 [3.5]

19 [22.4] 23 [27.1]

7 [8.2] 1 [1.2]

15 [17.7] 8 [9.4] 7 [8.2] 2 [2.4]

0 0

0

27 [13.2] 35 [17.1] 12 [5.9] 9 [4.4]

35 [17.1] 29 [14.2] 39 [19.0]

4 [2.0] 3 [1.5]

12 [5.85] Univariate analysis The association between individual personal and machine characteristics and machinery injury was assessed using univariate logistic regression models. The measure of effect obtained from logistic regression is the odds ratio (OR, i.e. the odds of exposure in the cases divided by the odds of exposure in the controls). The OR measures the relative odds of cases being exposed to a risk factor compared with controls. An odds ratio of one indicates no difference in the odds of cases and controls being exposed to the risk factor. An odds ratio of greater than one indicates the odds of cases being exposed is greater than for controls. An odds ratio of less than one indicates the odds of cases being exposed is smaller than for controls. The precision or accuracy of the odds ratio estimate is indicated by 95% confidence intervals, where narrower confidence intervals indicate more precise estimates. The p-

25

value indicates whether the observed association is likely to be due to chance. For example, a p-value of less than 0.05 indicates that the probability of obtaining an odds ratio with a magnitude at least as large as the one obtained when there truly was no association between risk factor and outcome is less than 5%. This is the usual threshold for assessing statistical significance. The association of various demographic and experience factors with farm machinery injury is shown in Table 3.7. Cases were more likely not to have attended farm training courses, and more likely than controls to have one to four years of farming experience, suggesting lack of specific agricultural training and relative inexperience are risk factors for farm machinery injury.

Table 3.7 Demographic and experience factors, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Cases (N=85) n [%]

Controls (N=205)

n [%]

Odds Ratio [95% CI]

p-value

Education University TAFE High school (completed) Primary school (completed)

9 [10.6] 26 [30.6] 12 [41.1] 38 [44.7]

21 [10.2] 41 [20.0] 42 [20.5]

100 [48.8]

1 1.48 [0.59, 3.72] 0.67 [0.24, 1.83] 0.89 [0.37, 2.11]

- 0.41 0.43 0.79

Attended farm training courses Yes No

54 [63.5] 31 [36.5]

158 [77.1] 47 [22.9]

1

1.93 [1.11, 3.34]

-

0.02 Farming experience Farm experience 20 years+ 10–20 5–9 years 1–4 years < 1 year

62 [72.9] 15 [17.7]

2 [2.4] 6 [7.1]

0

154 [75.1] 38 [18.5] 10 [4.9] 3 [1.5]

0

1

0.98 [0.50, 1.91] 0.50 [0.11, 2.33]

4.97 [1.20, 20.49] -

-

0.95 0.38 0.03

- Grew up on a farm Yes No

61 [71.8] 24 [28.2]

165 [80.5] 40 [19.5]

1

1.62 [0.90, 2.91]

-

0.11 Experience with index machine <20 hours 20–100 hours 100–200 hours >200 hours

4 [4.7]

9 [10.6] 12 [14.1] 53 [62.4]

12 [5.9]

28 [13.7] 20 [9.8]

140 [68.3]

1

0.96 [0.25, 3.75] 1.80 [0.47, 6.87] 1.14 [0.35, 3.68]

-

0.96 0.39 0.83

Previous safety training Previous safety training Yes No

34 [40]

36 [42.4]

99 [48.3] 89 [43.4]

1

1.18 [0.68, 2.04]

-

0.56 The association of a comprehensive range of health related factors and farm machinery injury is shown in Table 3.8. Few of these health related factors appeared to be associated with machinery injury. The odds of case having had an injury in the previous three years that resulted in a hospital stay were almost four times that for controls. Cases were also more likely to suffer from a medical condition requiring medication than control participants. Cases were less likely to suffer from asthma or to have had back pain in the last 12 months.

26

Table 3.8 Health related factors, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Cases (N=85) n [%]

Controls (N=205)

n [%]

Odds Ratio [95% CI]

p-value

Previous personal farm injury Previous injury in last 3years No Yes

69 [81.2]

15 [17.7]

163 [79.5] 42 [20.49]

1

0.84 [0.44, 1.62]

-

0.61 Hospital stay injury in last 3years No Yes

79 [92.4]

6 [7.1]

200 [97.6]

4 [2.0]

1

3.80 [1.04, 13.82]

-

0.04 Previous multiple personal injuries No Yes

81 [95.3]

4 [4.7]

195 [95.1]

9 [4.4]

1

1.07 [0.32,3.57]

-

0.91 Previous multiple hospital stay injuries No Yes

84 [98.8] 1 [1.18]

204 [99.5] 0

Cannot be calculated

Medication use Regular use of stomach remedies No Yes

84 [98.8] 1 [1.2]

197 [96.1] 8 [3.9]

1 0.29 [0.04, 2.38]

-

0.25 Regular use of heart/circulatory medications No Yes

69 [81.2] 16 [18.8]

179 [87.3] 26 [12.7]

1 1.60 [0.81, 3.16]

-

0.18 Stopped analgesics in last 12 months No Yes

73 [85.9] 4 [4.7]

197 [96.1] 7 [3.4]

1 1.54 [0.44, 5.42]

-

0.50 Stopped arthritis medications in last 12 months No Yes

72 [84.7] 2 [2.4]

194 [94.6] 4 [2.0]

1 1.35 [0.24, 7.51]

1 0.73

Existing health problems Ulcer/stomach upsets No Yes

80 [94.1]

5 [5.9]

185 [90.2]

20 [9.8]

1

0.58 [0.21, 1.59]

-

0.29 High blood pressure No Yes

73 [85.9] 12 [14.1]

169 [82.4] 36 [17.6]

1

0.77 [0.38, 1.57]

-

0.47 Heart attack No Yes

82 [96.5]

3 [3.5]

198 [96.6]

6 [2.9]

1

1.21 [0.29, 4.94]

-

0.79 Arthritis No Yes

68 [80.0] 17 [20.0]

166 [81.0] 39 [19.0]

1

1.06 [0.56, 2.01]

-

0.19 Asthma No Yes

83 [97.7]

2 [2.4]

181 [88.3] 24 [11.7]

1

0.18 [0.04, 0.79]

-

0.02 Urinary incontinence/disturbance urinary system No Yes

83 [97.7] 2 [2.4]

194 [94.6] 11 [5.4]

1 0.42 [0.09, 1.96]

-

0.27 History chronic medical condition

27

Risk Factor Cases (N=85) n [%]

Controls (N=205)

n [%]

Odds Ratio [95% CI]

p-value

No Yes

52 [61.2] 33 [38.8]

105 [51.2] 100 [48.8]

1

0.67 [0.40, 1.12]

-

0.12 Medical condition requiring medication No Yes

51 [60.0] 33 [38.8]

150 [73.2] 55 [26.8]

1 1.76 [1.03, 3.02]

-

0.04 Back pain in last 12 months No Yes

47 [55.3] 38 [44.7]

72 [35.1]

133 [64.9]

1

0.44 [0.26, 0.73]

-

0.002 Alcohol-problem user <7 on audit tool 8+ on audit tool

69 [83.1] 14 [16.9]

157 [77.0] 47 [23.0]

1

0.68 [0.35, 1.31]

-

0.25 The association of a number of work related factors and farm machinery injury is shown in Table 3.9. Cases were significantly more likely to be farm employees than controls. The odds of cases being involved in seasonal farming, rather than all year round, was over seven times that for the control participants. Increasing hours of farm work did not appear to be associated with machinery injury.

Table 3.9 Work related factors, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Cases (N=85) n [%]

Controls (N=205)

n [%]

Odds Ratio [95% CI]

p-value

Farm work hours per week – mean [sd]

46.3 [19.75]

50.6 [19.95]

0.99 [0.98, 1.00]

0.10

Farm work hours per day – mean [sd]

8.55 [2.88]

8.58 [2.66]

1.00 [0.91, 1.09]

0.93

Position on farm Owner/manager Employee/contractor

53 [62.4]

32 [37.7]

169 [82.4] 36 [17.6]

1

2.83 [1.61, 5.00]

-

<0.0001 Nature of involvement in farming All year round Seasonal

60 [70.6] 12 [14.1]

176 [85.9]

5 [2.4]

1

7.04 [2.38, 20.81]

-

<0.0001 Pesticide/herbicide use previous day No Yes

78 [91.8] 6 [7.1]

176 [85.9] 14 [6.8]

1 0.97 [0.36, 2.61]

-

0.95 The association of a number of machinery characteristics and farm machinery injury is shown in Table 3.10. The odds of injury increased by 3% for every year increase in the age of the machine. Cases were more likely to have been working on a machine that was not purchased new than control participants.

28

Table 3.10 Machinery characteristics, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Cases (N=85) n [%]

Controls (N=205)

n [%]

Odds Ratio [95% CI]

p-value

Machine age Median [IQR]

15 [27]

10 [17]

1.03 [1.01, 1.05]

0.008

Presence of safety features Yes No

53 [62.4] 23 [27.1]

158 [77.1] 42 [20.5]

1

1.63 [0.90, 2.96]

-

0.11 Length of time in use on farm Median [IQR]

5 [14]

7 [11.5]

1.00 [0.98, 1.03]

0.99

Machine purchased new Yes No

30 [35.3] 43 [50.6]

124 [60.5] 73 [35.6]

1

2.43 [1.41, 4.21]

-

0.001 Previous modifications No Yes

58 [68.2] 19 [22.4]

157 [76.6] 44 [21.5]

1

1.17 [0.63, 2.17]

-

0.62 State of repair Excellent Above average Average Below average

43 [50.6] 22 [25.9] 13 [15.3]

2 [2.4]

87 [42.4] 77 [37.6] 33 [16.1]

3 [1.5]

1

0.58 [0.32, 1.05] 0.80 [0.38, 1.67] 1.35 [0.22, 8.38]

-

0.07 0.55 0.75

In summary, when the various factors studied were considered individually (ie., any relationship between the factors aside), there were a number of factors associated with injury occurrence. Factors associated an increased odds of machinery injury were: • not having attended farm training courses • having 1–4 years’ experience (compared with more than 20 years’ experience) • having a hospital stay for a farm work related injury in the previous 3 years • having a medical condition requiring medication • being an employee or contractor • being engaged in seasonal farm work • machine age • using a machine that had not been purchased new. Factors associated with a decreased odds of machinery injury were: • having asthma • having had back pain in the previous 12 months • using a machine in above average state of repair (compared with an excellent state of repair) Multivariate analysis: personal and machine characteristics One of the limitations of the above univariate analysis is that the influence of each factor on farm machinery injury risk is considered in isolation from each other. As some of these factors are correlated (eg., having asthma and having a medical condition requiring medication), it would be helpful to be able to identify those factors which, independent of any other factors in the analysis, are associated with farm machinery injury risk. Therefore, logistic regression modelling was conducted to simultaneously examine the relationship between a selected number of these personal and machine characteristics.

29

Several factors identified in the univariate analysis were no longer significant in the multivariable analysis. These include not having attended farm training courses, having 1–4 years’ farming experience, and having a medical condition requiring medication. The final model identified eight factors of interest in relation to machinery injury (Table 3.11). Five variables reached statistical significance as independently associated with machinery injury. The odds of injury increased by 4% for each year of increase in machine age (compared with an increase of 3% each year in the univariate analysis). Other factors that increased the odds of machinery injury for farmers were: using a machine that was not purchased new, being involved in seasonal farming, and being an employee, rather than an owner/manager. Curiously, having machinery that was rated as being in above average or average state of repair, compared to excellent state of repair, appeared to be associated with a decreased odds of injury. The remaining variables (hospital stay in the last 3 years, asthma, and back pain in the last 12 months) did not reach the conventional level for statistical significance as independent factors. However, the borderline p-values suggested that they may be of interest and could be considered in further studies. In addition, the inclusion of these factors improved the fit of the model to the available data.

Table 3.11. Multivariate analysis using personal and machine characteristics, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Odds Ratio [95% CI] p-value Cases N=45 Controls N=130 Machine not purchased new 6.11 [2.38, 15.69] <0.0001 Nature of involvement in farming 5.29 [1.01, 27.79] 0.049 Position on farm 3.11 [1.13, 8.53] 0.028 Machine age 1.04 [1.00, 1.07] 0.067 State of repair Excellent Above average Average Below average

1

0.31 [0.11, 0.85] 0.16 [0.04, 0.66]

0.60 [0.01, 33.13]

-

0.023 0.011 0.804

Hospital stay injury in last 3years 5.67 [0.84, 38.30] 0.075 Asthma 0.10 [0.01, 1.06] 0.056 Back pain in last 12 months 0.45 [0.19, 1.04] 0.063

Note: Each odds ratio has been adjusted for all other risk factors in the table. Epidemiological research is observational and other factors may explain the relationship between an exposure and an outcome, for example confounding variables. Confounding variables are related to both the exposure or risk factor of interest, as well as being independently associated with the outcome, and therefore can obscure the relationship between the exposure or risk factor of interest, and the outcome being studied. There were no confounding factors identified in this study.

30

4. Stage 2 Results: In-depth investigation of machine characteristics Recruitment for on-site machine investigations

Of the 85 farm workers who sustained a serious farm work related machinery injury, 37 (43.5%) had inspections performed on the machinery involved in their injury, compared to 71 (34.6%) of the matched controls. The cases whose machines were inspected (mean age 50.3 years) were slightly older than cases whose machines were not inspected (mean age 47.0 years), however this difference was not statistically significant. Controls who were inspected were also older (mean age 51.1 years) than those that were not inspected (mean age 46.6 years), and this difference was statistically significant. Comparisons revealed that there was no difference in injury outcome between the inspected and the uninspected cases. Both inspected cases and controls were significantly more likely to be owners, rather than employees, than uninspected participants. In terms of factors specific to the machines being investigated, the age of the machine did not differ between inspected and uninspected cases or controls. See Appendix 1 for data showing characteristics of cases and controls who did and did not have machine inspections. Univariate machine and machinery operation characteristics

The association of a number of machinery and machine operation characteristics reported during the on-site inspections and farm machinery injury is shown in Table 4.1. None of these were significantly associated with the risk of injury.

31

Table 4.1 Machine and machinery operation characteristics, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Risk Factor Cases (N=37) n [%]

Controls (N=71) n [%]

Odds Ratio [95% CI]

p-value

Previous modifications No Yes

27 [73.0] 9 [24.3]

48 [69.6] 21 [30.4]

1

0.76 [0.31, 1.90]

-

0.56 Inconvenient safety features No Yes

27 [73.0] 6 [16.2]

54 [78.3] 13 [18.8]

1

0.92 [0.32, 2.70]

-

0.88 Safety features hamper function No Yes

29[78.4] 4 [10.8]

60 [87.0] 8 [11.6]

1 1.03 [0.29, 3.72]

-

0.96 Condition based maintenance at index time No Yes

Data only available for

cases

Days since last maintenance Mean [sd]

161.5 [672.6]

116.6 [208.3] 1.00 [1.00, 1.00] 0.63

Type of maintenance tended towards this season Preventative Condition-based

19 [59.4] 9 [28.1]

51 [77.3] 14 [21.2]

1 1.73 [0.64, 4.64]

-

0.28 Formal safety check ever Yes No

3 [8.1]

34 [91.9]

6 [8.6]

64 [91.4]

1

1.06 [0.25, 4.52]

-

0.94 Formal safety check in last 12 months Yes No

0 3 [100]

2 [40.0] 3 [60.0]

Cannot be calculated

Safety feature score Percentage score [sd]

69.7% [25.0%]

73.6% [25.5%]

0.54 [0.11, 2.80] 0.47

Machine involved in previous injury event No Yes

36 [97.3] 1 [2.7]

69 [97.2] 2 [2.8]

1 0.96 [0.08, 10.93]

-

0.97 Machine meets AS4024 Yes No

15 [40.5] 20 [54.1]

36 [50.7] 32 [45.1]

1

1.5 [0.66, 3.41]

-

0.33 Machine meets Ag Standards Yes No

24 [64.9] 13 [35.1]

50 [70.4] 20 [28.2]

1

1.35 [0.58, 3.17]

-

0.49 Machine meets OHS obligations Yes No

20 [54.1] 17 [46.0]

48 [67.6] 22 [31.0]

1 1.85 [0.82, 4.21]

-

0.14 Multivariate machine and machinery operation characteristics

Backward stepwise logistic regression was conducted to determine which of the variables investigated during the on-site inspection were independently associated with risk of farm machinery injury. The results revealed that none of the variables were independently associated with risk of farm machinery injury.

32

5. Design changes for preventing, or reducing the severity of, farm machine related injury Role of design in machinery injury: epidemiological analysis

The epidemiological analysis presented in Chapter 4 examined the association of a number of design related features with the odds of farm machinery injury (Table 4.1). There was no strong evidence for an association of any of the design or safety related features with the odds of injury. The safety feature score, or performance of formal safety checks were not associated with a decreased odds of injury. Previous modifications, inconvenient safety features, safety features hampering function or inconsistency with standards and regulations were not associated with an increased odds of injury. There was weak evidence to suggest that not meeting the requirements of occupational health and safety obligations was associated with an increased odds of machinery injury (Table 4.1). An important consideration for interpretation of these results is that absence of any evidence of an association does not necessarily prove there is no actual association of these features with the odds of machinery related injury. These findings were based on a relatively small number of cases (n=37) and controls (n=71) for which data were available on these features. However, in the larger sample which included those cases and controls that did not agree to the on-site inspection component, we did find that the odds of injury increased by 4% for every year increase in the age of the machine, even after several important personal demographic factors that may be associated with machine age (such as gender, age, and farming experience) were taken into account either by the study design (restricted to males and age-matched) or shown in the multivariable analysis to have no influence on this finding. Older machinery does not have the benefit of advances in design, is less likely to meet current safety standards and regulation, and may not be maintained as well as newer machinery. Post-hoc analysis showed that machines that did not meet the agricultural or occupational health and safety standards we assessed were significantly older than those machines that did meet these standards. The engineering analysis of the inspected machines, in the next section, complements the epidemiological analysis by providing further insight into the role that design may play in the aetiology and prevention of machinery related injury. Role of design in machinery injury: engineering analysis

The following summary of case injury events has been organised into groups according to common injury mechanism, with description of themes within each group. Chapter 6 provides more general discussion of these results with suggestions for prevention. The injury mechanism groups are as follows: • injured by moving parts on operating machinery (13) • struck by machine/equipment (8) • shear/pinch points (6) • fall from stationary machinery (4) • equipment falling on farmer (4) • fall from mobile machinery (2). The intention of this analysis is not to apportion blame, liability or fault, rather to examine the cases for common themes and consider the wider picture of machinery management and farm safety culture.

33

The majority of the pictures were taken up to two months after the incident, so the machinery may be in a different location, and in some cases in a different configuration from the incident event. Care has been taken to protect participant anonymity by obscuring registration plates or other potentially identifying aspects in the photographs. Important factors that influenced the injury outcome, such as how the farmer obtained first aid and hospitalisation after the event, are not discussed in detail here, but will be the subject of more thorough discussion elsewhere. Of particular interest are the eight cases where, had it not been for a chance discovery of the injured farmer by a relation or neighbour, the farmer most likely would have died. This is obviously an aspect worthy of serious consideration when considering injury management strategies, and points to the need for effective communication methods to be used by the farm business and farm families. Injured by moving parts on operating machinery There were 13 cases among which the injury occurred due to moving parts on operating machinery, comprising 35% of the 37 inspected cases. In six of these, the injured participant was undertaking maintenance activity, such as oiling a chain, wiping oil from a pulley, or checking a slipping belt (Figure 5.1). These mechanisms were running at the time. Three participants were injured while clearing a blockage (Figure 5.2). In three of the above cases, the farmer considered the machine to be stationary, but this was either not the case, or the machine inadvertently moved. Two participants were pulled into rotating shafts when clothing became caught (Figure 5.3). One of these was a guarded PTO shaft. The participant was saved from life-threatening injuries due to a fault in the tractor which stalled. Two injuries were to “third party” participants in that the injured farmer was not the primary machine operator. One case involved the use of an interlock on a wool press, which effectively limited the injury severity. Of this sub-group of 13 cases, one third (4) of the machines involved were assessed as substantially meeting the requirements of AS4024 – Safety of Machinery, where this standard applied (this standard is not applicable to PTO guarding). Half were assessed as substantially meeting agricultural standards such as AS/NZS 2153, and the same half met the basic requirements under OHS legislation for risk management. In 7 cases, the design feature associated with the injury would be either equivalent on all or some similar types of new machinery. In four cases, the injury causing feature has been designed out of similar types of new machinery. In two cases, the machine had been modified.

34

Figure 5.1 Examples of machinery with which participants were injured while undertaking maintenance activities, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Figure 5.2 Examples of machinery with which participants were injured while clearing blockages

35

Figure 5.3 Examples of machinery into which participants became caught by clothing, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Struck by machine/equipment There were eight cases among whom the injury occurred when machinery, or components of machinery, struck farmers who were nearby, comprising 21% of the 37 inspected cases. In most of these eight cases, the release of energy was not expected, so farmers did not consider themselves to be at risk of injury at the time. This meant that generally, no personal protective equipment was being worn. However, in most cases, basic personal protective equipment would not have influenced the injury outcome. Three cases involved farmers being injured when poorly configured machinery suddenly released energy from: a PTO shaft, hot water from a burst pipe attachment on a pump, and battery explosion while connecting jumper leads (Figure 5.4). A further three cases involved farmers injured while operating machinery, and an energy release was not anticipated. These involved a spring-loaded handle on a grain silo (Figure 5.5), a hand starter on an old tractor, and a dislodged piece of steel from a shear on an air seeder. In the remaining two of the cases, the farmer was seriously injured when attempting to free components (scrub clearing tines, light harrows) from being stuck, and the components sprung back. In five of these cases the design feature associated with the injury would be equivalent on all or some similar types of new machinery. In one case, the injury causing feature has been designed out of similar types of new machinery. In one case involving a pump, poor configuration of the pump by the farmer led to a component failure, which caused the injury.

36

Figure 5.4 Examples of poorly configured machinery suddenly releasing energy, causing injury, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Figure 5.5 Example of machinery (spring loaded handle on grain silo) where unanticipated energy release during operation resulted in injury, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003-2006

37

Shear/pinch points There were six cases among whom the injury occurred where the mechanism was a shear point or pinch point, resulting in amputations, lacerations, and crushing injuries (Figure 5.6). Four of these injuries involved pinch and shear points when assembling or maintaining machinery. The remaining two injuries involved mechanised hand tools: a pneumatic pruner and a sheep shearing handpiece. In all of these cases, the design feature associated with the injury is equivalent on most similar types of new machinery.

Figure 5.6 Examples of pinch and shear points which resulted in injury when assembling or maintaining machinery, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Fall from stationary equipment There were four cases among whom the injury occurred when farmers fell from stationary equipment, all of which involved cleaning or maintaining equipment (Figure 5.7). In two of these cases, the farmer was attempting to clean or maintain the tractor windows and roof by standing on the wheel guard, where there is no provision for access to these areas of the tractor. One case involved falling from a header while conducting an oil level check, and another was a fall from an old tractor with a custom rear lifting mechanism. In the former three cases, the design feature associated with the injury would be equivalent on similar types of new machinery.

38

Figure 5.7 Examples of machinery from which farmers fell during cleaning or maintenance activities and were injured, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Equipment falling on farmer There were four cases where injuries occurred when equipment fell onto farmers. The nature of the injuries was generally very serious and potentially life threatening. In most of these cases, personal protective equipment would not have prevented the injury or reduced the injury severity. In all cases, different machinery design using current technology could have prevented the incident from occurring. Two cases involved machinery (a modified auger frame, and a set of harrows) collapsing onto the farmer, in the latter case due to structural failure (Figure 5.8). In addition, there were two cases where a tractor with a front end loader (FEL) but no falling object protective structure (FOPS), and using forklift style tines to lift round bales, resulted in the round bale falling onto the operator, resulting in serious injury. In these two cases, the design feature of having a FEL without the use of FOPS is equivalent on similar new machinery. In one of the two other cases, the design feature has been designed out of similar types of new machinery; the remaining case involved a substantially modified auger.

39

Figure 5.8 Examples of machinery which fell on farmers resulting in serious injury, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Fall from moving machinery Two farmers were seriously injured when they fell or jumped from moving machinery. In one case, the tractor was losing traction and slipping on sloping ground and the farmer jumped clear. In the second case, a farmer was thrown clear of a tractor and crushed under the rear wheel, when the right brake was inadvertently activated causing the tractor to hit a stump. In both these cases, similar new machinery has equivalent design features.

40

6. Discussion – Epidemiological analysis Principal findings

In this study of serious farm machinery injury among men, we found a number of factors associated with injury occurrence. Factors associated an increased odds of machinery injury were: • not having attended farm training courses • having 1–4 years’ experience (compared with more than 20 years’ experience) • having a hospital stay for a farm work related injury in the previous 3 years • having a medical condition requiring medication • being an employee or contractor • being engaged in seasonal farm work • machine age • using a machine that had not been purchased new. Factors associated with a decreased odds of machinery injury were: • having asthma • having had back pain in the previous 12 months • using a machine in above average state of repair (compared with an excellent state of repair). When any relationship between these variables was taken into account through multivariable logistic regression the following factors maintained statistical significance and were independently associated with an increased odds of machinery related injury: • being an employee or contractor • being engaged in seasonal farm work • machine age • using a machine that had not been purchased new. Having machinery that was rated as being in above average or average state of repair, compared to excellent state of repair, appeared to be associated with a decreased odds of injury. In addition, results for three further factors, while not reaching conventional statistical significance, suggested that they may be of interest. There was some evidence that having had a hospital stay in the previous 3 years increased the odds of injury, and that having asthma, and having had back pain in the previous 12 months were associated with a decreased odds of machinery related injury. Associations observed in this study do not necessarily imply a cause and effect relationship. Further, the finding of no association between a risk factor and farm machinery injury does not necessarily mean that an association does not exist – the correct interpretation is that there is no evidence for an association. Our findings confirmed previous observations that having a prior agricultural injury increased the odds of injury (Carlson et al. 2005). There are several potential mechanisms by which a previous injury could increase the risk of a subsequent injury. A previous injury could be a marker for risk taking behaviour, or it could result in some residual difficulty in performing farm work, either of which may then increase the risk of a subsequent injury. The remaining independent risk factors have not been reported in previous case-control studies of farm machinery related injury. Being an employee or contractor, or being engaged in seasonal farm work were significantly associated with farm machinery injury. These findings are not explained by employees or contractors working more hours than other farm workers, as hours worked per week was not a significant variable in our study. Similarly, these results cannot be explained by employees or seasonal workers being less experienced in farming, or using older machinery, as machinery age was adjusted for in the multivariable analysis, and

41

experience did not have any influence on the results for employees or seasonal work. Experience with the index machine, which is another potential explanation, did not appear to influence injury outcome. However, employees or seasonal workers may spend a greater proportion of their working hours engaged in tasks requiring machine use, or the types of machinery related work undertaken by employees or seasonal workers may be of a higher risk than that undertaken by owners/managers and those engaged in farming all year round. It is also possible that employees, contractors and seasonal workers may be less likely to have undertaken farm training courses. This factor was significant in the univariate analysis, but did not remain independently significant in the multivariable analysis. This study cannot provide any insight into these, or other potential explanations. Previous similar studies of farm machinery injury have not examined machine age as a continuous variable. We found that the odds of injury increased by 4% for every year increase in the age of the machine, even after several important personal demographic factors that may be associated with machine age (such as gender, age, and farming experience) were taken into account either by the study design (restricted to males and age-matched) or shown in the multivariable analysis to have no influence on this finding. Older machinery does not have the benefit of advances in design, is less likely to meet current safety standards and regulation, and may not be maintained as well as newer machinery. Post-hoc analysis showed that machines that did not meet the agricultural or occupational health and safety standards we assessed were significantly older than those machines that did meet these standards. Similarly, we found the odds of injury increased if the machine had not been purchased new, even when machine age, length of time in use on the farm, and personal demographic factors had been taken into account. The mechanism by which this machine characteristic might increase injury risk is not clear. The finding that the odds of injury are decreased for machinery in a state of above average repair, when compared with machinery in an excellent state of repair is counterintuitive. It would be expected that the odds of injury would decrease as the state of repair increases from below average to excellent. However, the data for state of repair were self-reported and this finding could be compromised by reporting bias. A high proportion of both cases and controls reported that the machine was in excellent state of repair (51% and 42% respectively), while much lower proportions reported an average state of repair (15%, 16% respectively). The finding that asthma and back pain in the previous 12 months were associated with a decreased odds of machinery related injury are interesting. Farmers with these conditions may work fewer hours, however hours worked was not significant in our study and therefore does not explain these findings. It is possible, however, that farmers with asthma and back pain may select the type of machinery related work which is least irritating for their condition which may at the same time reduce their risk of machinery related injury (e.g. farmers with asthma may not engage in harvest related activities unless using cabined machinery). Four previous overseas analytic studies of farm machinery injury have found an increased risk of injury associated with increasing hours of farm work (Layde et al. 1995; Lee et al. 1996; Gerberich et al. 1998; Sprince et al. 2002). Only one of these (Layde et al. 1995) analysed hours worked as a continuous variable as in our study, but unlike our study, included controls who had not worked any hours on a farm during the referent period. One previous study of tractor related injury did not find any evidence for an independent association between increasing farm work hours and tractor related injury (Carlson et al. 2005). We found no evidence for an association between increasing farm work hours and machinery related injury. It is possible that average hours worked is associated with age, and as our study was matched on age, this could explain the lack of association between hours worked and odds of injury. However, there was no evidence in our study for a relationship between age and hours worked. Controls were also matched to cases on machine type (controls had to have used a similar machine in the previous month), and post-hoc analysis of our data revealed that hours worked

42

on the farm is associated with machine type. Therefore, the matching on machine type is likely to have also matched sufficiently on hours worked to obscure any observation of association between farm work hours and machinery injury, if such a relationship truly exists. Other independent risk factors for farm machinery injury reported in these overseas studies described above, but not observed in our study include fewer years of farming experience, and problem drinking (Sprince et al. 2002). We did find that having 1–4 years’ experience was associated with an increased odds of machinery related injury, but this did not remain significant in our multivariable analysis. The previous study that reported problem drinking as a risk factor used a different method to measure problem drinking (Sprince et al. 2002). Strengths and weaknesses

This was the first Australian study to examine factors associated with farm machinery injury that has included both injured and uninjured farmers, providing the benefits of a control or comparison group. Previous Australian studies have all been based on the case series design. In addition, this study was one of the first world-wide to examine the contribution of individual and machine characteristics to the odds of sustaining a farm machinery related injury. The study design, which included only males and matched cases and controls on age, removed two of the dominant risk factors for injury in general. This allowed for the examination of other potentially important factors that may otherwise be obscured by the overwhelming influence of gender and age. Most studies that rely on human participants are at risk of selection and non-response bias. Our study was embedded in the FIRM study which relied on emergency department staff to approach potential participants. As a hospital records audit conducted early in the FIRM study revealed that up to approximately 50% of potentially eligible patients were not being approached for participation in the FIRM study during their treatment in the regional hospital emergency departments (Day et al. 2007), it is possible that some selection bias has occurred with respect to case recruitment. There is minimal information available on those who were not approached (including whether the injury was associated with machinery), and therefore assessment of the extent, or impact, of that bias is not possible. Comparing the profile of non-fatally injured farmers who were admitted for their injury and participated in the FIRM study, with the profile of males with non-fatal farm work related injury in the state hospital admissions data provides some insight into the extent of selection and response bias. There was no difference on age, month, or type of injury between the participating non-fatally injured farmers and those in the state hospital admissions data. However, among those injured farmers who were approached for participation in the FIRM study, 33% did not give consent to be contacted by study staff (unpublished observations). In slightly more than 50% of those who declined to be contacted, we were able to collect data on farming experience. Non-fatal cases participating in the FIRM study were significantly more experienced (i.e. had more than 20 years’ experience) than the potentially eligible cases who declined to be contacted (90% compared with 65%, p<0.001, unpublished observations). The effect of this response bias would be that the study might under-estimate the odds for inexperience. Inexperience was a significant factor in the univariate analysis (OR 4.97, 95% CI 1.20–20.49), but did not remain significant when other factors were taken into account in the multivariable analysis. As inexperience is also likely to be associated with being an employee or being involved in farming on a seasonal basis, the study may also under-estimate the odds of injury for employees, or those involved in farming on a seasonal basis. As these two factors remained significant in the multivariable analysis, the odds ratio may be an under-estimate. The potential for detection and selection bias in control recruitment for the FIRM study appears relatively low. Unpublished data for the FIRM study indicate that overall 12% of valid numbers were not contactable, and a small proportion (5%) of households terminated the telephone call before their status could be determined. Therefore, a relatively small number of eligible controls would not have

43

been invited to participate in the FIRM study. Among those eligible controls who were invited to participate, 10% declined. Another point in our study where it would be possible for response bias to occur was in Stage 2, in which cases and controls were asked for permission to undertake a machinery inspection. However there did not appear to be any evidence for response bias between the cases and controls who agreed to inspection, as any differences were present to approximately the same extent for both the cases and controls. Comparisons revealed that there was no difference in injury outcome between the inspected and the uninspected cases. In terms of machine specific factors, machine age did not differ between inspected and uninspected cases or controls. An inherent difficulty for case-control studies, particularly of acute injury, is that the cases have had a significant life event which may influence their recall of associated details, some of which will relate to study exposures. In addition, the interviewers and engineer were aware of the case or control status of the participants when undertaking the interviews and on-site inspections, which may have some unconscious influence on the manner in which the questions were asked. Our study had a relatively small number of cases (85 who participated in stage 1, and 37 who participated in stage 2) which has implications for statistical power. The control to case ratio of 2:1 compensated for the low case numbers to some extent. However, study power to examine some of the variables would have been low. Therefore, the absence of effect for some of the factors studied should not be interpreted to mean that there is no association of these factors with machinery injury, rather that this study did not provide any evidence for an association.

44

7. Discussion – Machine design analysis The in-depth investigation and on-site inspections of 37 of the injured farmers, along with 71 comparable controls, highlighted a range of factors associated with machine design, and the potential for design to reduce the frequency and severity of farm machinery injury. Much of the following discussion will centre on specific potential interventions that may be considered by the industry as a means to reduce the rate of serious injury. In addition, study of the control and other machinery has provided a large amount of information that has more general machine design and machine management implications. This will also be presented with the intention of providing ideas to inform a growing knowledge base for agricultural machinery design. Current theory and methodology was used to examine the machinery and incident events. This is introduced to the reader as an explanation of how various suggested interventions were generated, and to enhance the state of knowledge within the industry, advocating a contemporary and thorough means to assess risk. However, there are two general observations which will be considered first. Awareness of design interventions All case and control farmers were specifically asked whether they knew, or could think of, any design aspects (including design changes, modifications, replacements, and retrofits) that might make the machine safer to use. Farmers were also asked if they knew of changes of procedure that would be safer. Generally it was found that farmers were able to demonstrate exceptional knowledge of a range of hazards and design interventions, some of which were innovative solutions to problems. In many cases, these ideas had not been taken up, and for good reasons that commonly included cost, time, or unavailability of parts in the region, particularly for retrofit components. In some cases farmers knew of a simple procedure or design improvement that had not been implemented despite any clear barriers to doing so. These cases show that there is room for improvement, possibly by targeting risk perception and safety culture issues. Comparison with new machines Some of the design issues that are apparent with the case machinery have been “designed out” of modern machinery. Of the machines that had not been modified or incorrectly configured by the farmer, one fifth of the injuries would not have occurred on an equivalent new machine. An additional one quarter may not have occurred, depending on which new machine design was being used (for those cases where the aspect of interest has been designed out on some equivalent new machines but not on others). These observations from the on-site visits are consistent with the epidemiological findings from the larger sample of machinery cases and controls that the risk of injury increases by 4% with each year of increased machinery age. Like other manufacturing industries, such as light and heavy vehicle manufacturing, many safety advances with agricultural machinery have been made. Comparing machinery produced over previous decades demonstrates that agricultural machinery is subject to continual improvements. In addition to recognition within the agricultural machinery industry of the importance of safety by design, there are a number of regulatory measures also encouraging these improvements. These include:

• the introduction and enforcement of occupational health and safety laws from around the 1980s worldwide

• increasing globalisation of the agricultural machinery marketplace • increasing commitment to, and improvement of international standards within the agricultural

machinery industry. The most important of these are American Society for Agricultural and Biological Engineers (ASABE) Standards, International Standards Organisation (ISO) (widely referred to as European standards), and Australian Standards (In many cases for agricultural machinery, these are identical to ISO standards).

45

Of the machines that were not modified or were incorrectly configured by the farmer, over half of the specific design issues relating to the injury event are equivalent on a new machine. While not all of these cases call for a design change on new machine, for many of the injury events it is clear that design changes, in some cases beyond the requirements of current standards and occupational health and safety law, can be implemented for safer machine use. Some suggestions to this end are a key objective of this report, and are discussed below. Systems approach to machinery design

Methods of incident prevention are usually determined by the perception one has of the means by which incidents occur, and what array of measures are possible to prevent them from happening. An understanding of the approach that has been taken by investigators on AMDOSS will help to inform and educate readers of this contemporary perspective. There has been a historic, but now outdated, view that eliminating human error provides a path to incident reduction. In the 1930s, a landmark study by Heinrich determined that around 88% of incidents were caused by human error, 10% were caused by the operator environment, and 2% were acts of god (Heinrich 1931, 1959). Given that human error was deemed to be an important factor in so many incidents, Heinrich advocated controlling the human element – behavioural training, discipline and awareness – as the best way to reduce the number of incidents. This idea is similar to the commonly held belief that “stupidity” was to blame for the injury, or a lack of “common sense”. Whilst in hindsight it is tempting to make such judgements, the fact remains that, irrespective of known risk factors such as lack of education and training, risk awareness, intelligence, time of day, levels of distraction, or fatigue – humans can, and will, make errors. It is now considered that while training and awareness is a critical aspect of incident prevention, this approach on its own generally has a limited ability to achieve significant incident rate reduction unless combined with other preventive strategies (National Committee for Injury Prevention and Control 1989). Incident prevention theory since the Second World War has increasingly emphasised the need to consider the human as part of a system, and to change both the human and system to eliminate, or at least minimise the probability of an incident (Reason 1997). Human factors is the scientific discipline concerned with understanding interactions between humans and other elements of a system. The central principle is that the human operator is part of the system, not apart from it. The system then includes the human operator, the work environment, and equipment or machinery used by the operator. The primary goal of human factors is to optimise system performance and safety through the application of theory, principles, data and methods to the design products, machines, environments, and procedures. This is achieved by a thorough understanding of the abilities, limitations and characteristics of the people working within the system (Human Factors and Ergonomics Society 2005). A key difference between the person-based approach and the systems-based approach is that the person-based approach assumes that the system itself is inherently safe, and that the issue relies solely with the human operator. This leads to the faulty assumption that increasing the competence and skill of the operators alone will reduce accidents. Yet, accidents and incidents may be a result of the system not supporting the operator to safely achieve their objectives. A systems based approach to safety considers human error to result from latent system conditions (Reason 1990). These might include equipment that is inadequate for the task to be performed, poor design, poor maintenance, operators with insufficient training or supervision, or faulty procedures (Dekker, 2000). Contemporary theory suggests that incidents are caused by a combination of latent conditions and human error. Operator skill is only one modifiable aspect, and injury prevention activities focused only on training operators in isolation are inadequate. Even the most highly trained operators make errors. A crucial objective is to minimise the occurrence of human error, and the consequences of error if it does occur. In order to minimise human error, the latent conditions that lead to error must be considered. One of these is the machine, which must be designed to minimise human error (Gardner-Bonneau 2000).

46

Consideration must be given to the abilities, limitations and characteristics of all intended users, taking into account such factors as age, gender, education and experience, and physical attributes, like visual acuity, hearing, height, reach distance, and health conditions. Preventive efforts are thus aimed at a broad range of factors, such as the machinery involved, the environment in which the machinery is used, as well as the human operator. A system that is not designed to preclude the possibility of an incident in the event of human error ensures that incidents will occur, given sufficient opportunity over time. The systems approach to safety has been embraced by many high-risk industries, such as aviation, nuclear power, and more recently, road safety and medicine (e.g. Salmon, Regan & Johnston 2005, 2006). An effective systems-based approach to the design of machinery for incident prevention benefits from making the following assumptions:

• The machine should be designed for safe use for all intended purposes, for its entire life. The same level of rigor should be applied to optimising the machine’s safe use for all scenarios, including manufacturing, assembly, transport, commissioning, testing, training, operation, maintenance, cleaning, inspection, repair, reconditioning, decommissioning and disposal. Anyone involved with the machine is considered as an operator, that is, not only the owner, or well-trained personnel, but also others, such as maintenance workers.

• The machine will be operated by a large number of people for a large amount of time. For the

purposes of operational safety, the designer must consider the machine being used for an arbitrarily large number of man-hours in various conditions.

• The machinery operator is not perfect. Human error can, and will, occur irrespective of the

skill and training of the operator. Human-machine interaction should be considered in light of this potential for human error.

• The machinery operator and others within proximity of the machine are vulnerable to the

levels of energy that are typically associated with machinery. The design must be such that the operator is inherently protected from those energy levels that can cause serious injury or death.

• Generally it cannot be assumed that operators will wear personal protective equipment (PPE).

• Generally it cannot be assumed that the operator is familiar with the hazards or operation of

the machine.

• Generally it cannot be assumed that the operator has, or can read the manual or decals. The machine must be designed such that where there is a reliance on the lower level safety management techniques (the latter three dot points above), the result of human error is not severe – i.e. an incident cannot occur, or if the possibility of an incident cannot be eliminated, the injury severity is limited by design to a manageable and non-fatal level. The corollary of these design assumptions is a design philosophy that encourages machinery design to be tolerant of human error. In addition, other considerations must also inform the design process under a systems-based approach – particularly procedural, environmental, and organisational factors.

47

The role of standards and legislation

Where applicable, the role of agricultural machinery standards and legislation is examined throughout the following discussion of injury events, and is intended to inform discussion about the role of standards within the agricultural machinery industry. Further comments are made in Recommendations. Machinery design standards should not generally be regarded as a benchmark that dictates perfect solutions to design challenges. Rather, standards usually serve as a minimum performance guideline, and in some cases, simply offer advice on methods for determining solutions. With the exception of specific design features, the most applicable agricultural machinery Australian Standard is AS /NZS 2153:1997 Tractors and machinery for agriculture and forestry – Technical means for ensuring safety. Other machinery standards and legislation include:

• AS 4024.1:1996 Safeguarding of Machinery, or its successor, AS4024:2006 Safety of Machinery, as applicable according to time of manufacture.

• State Plant Regulations and related OHS law, including the expectations of the courts in enforcing the respective state OHS Act.

• State Plant Codes of Practice, Guidance and other state workplace health and safety organisations’ publications

• Various international standards including ASAE standards (American Society of Agricultural Engineers)

• Road transport regulations and associated requirements including Vehicle Standards Bulletins and Australian Design Rules, as applicable.

The chief difference between AS/NZS 2153:1997 and other resources is that provision is made throughout the standard to allow for exceptions to some requirements to allow for the unique agricultural environment. Examples of these in AS/NZS 2153:1997 Chapter 6 are:

• “Unless it is clearly inappropriate …” • “Guards shall normally …” • “In some circumstances it may be necessary …”

There is nothing to preclude the applicability of all of the above standards and legislation to agricultural machinery, in addition to AS/NZS 2153:1997. In general, there are no specific legal requirements for the design of machinery to meet Australian Standards. Observation of the case and control machinery throughout this study has shown that a machine that meets Australian Standards is not necessarily inherently safe to use, and likewise, a machine that does not meet Australian standards is not always inherently dangerous. The role of, and ability of standards to reduce the risk of injury with agricultural machinery, is a key deliverable of this study, and commentary to that end is made throughout the following discussion and recommendations.

48

Design issues determined from incident events

Operator protection during machinery operation Several injuries resulted due to the inadequacy or absence of guards or protective switches. Guards and other means of operator protection, such as ROPS and FOPS, and the use of switches such as interlocks, can all be discussed under the same theme of operator protection. The design principles that are put forward in this section have been tailored for agricultural machinery, and takes into account observations from the 108 case and control machines studied, as well as many other machines observed throughout the duration of the study. These principles are further articulated in a number of readily available sources, including AS 4024–2006 Safety of Machinery, as well as state OHS Plant Regulations. Any discussion of operator protection devices must keep in mind the importance of the need in the industry for well designed operator protection equipment that is suitable for retrofit (such as guards, and sensors including interlocks). There are many machines in use that are functionally capable of many more years’ service to Australian agriculture, but present a risk to operators. Well considered, cost effective modifications to these machines can give them a much longer safe operational life without the need for premature retirement and replacement of equipment. In most cases operator protection that is retrofit may exceed the quality of original measures. Guards From a safe system design perspective, physical guarding must be 100% effective for preventing operator contact with the hazard. This can be achieved by designing for hazards, including hazards created by disintegrating moving parts or product flow:

• fully enclosed guard around hazards • guard with apertures to allow for inspection or product flow, situated at a distance away from

the hazard, proportional to the size of aperture, according to ergonomic data available in AS4024–2006 Safety of Machinery

• displacement guard such as a walkway barrier that prevents human access to machinery within a proximity that is determined according to ergonomic data available in AS4024–2006 Safety of Machinery

• enclosure of machinery, such as in a cage. This prevents all access near the moving parts, and may be more cost effective if it is suitable to negate the need for installing several smaller guards by installing an interlock.

Switches Protective switches come in many forms, such as electrical, mechanical, or hydraulic, and can be manually operated, or integrated into a machine for automatic triggering. Protective switches that are integrated into the machine design, known as an interlock, are usually used to turn off an engine or disengage a power transfer mechanism such as a drive belt, gear, PTO shaft, drive chain, or hydraulics. Some examples of commonplace effective protective switches are as follows:

• electrical safety switches • microwave oven doors • washing machine lids • electrical fuse • kill switch on workshop machines and motorbikes • proximity device such as reversing sensors that use ultrasonic waves • light and movement detectors • fire alarm/water systems • brake systems on trucks – air brakes hold the brakes “off”, so if the air pressure fails, brakes

are applied • dead man seat • movement sensing devices.

49

Designing an effective protection system that incorporates an interlock can be very challenging, and becomes more challenging as the machine complexity increases. Each new interlock is another component that can break down, so reliability, functionality, and the ergonomic design of interlocks is critical. Design principles Operator protection should be designed from a safe system approach, the principles of which are outlined above. Following are some safe system design principles that also encompass the installation, maintenance and life considerations of a machine.

• The design life should be such that it outlasts the machine. This takes into account: o fatigue o corrosion o embrittlement due to exposure to UV radiation o embrittlement due to exposure to fuels, oils, and solvents o additional loads that may be incurred during the life of the protection, such as a guard

being used as a step or lifting point.

• Design features should not be able to be easily defeated. This takes into account: o the protection is not easy to remove or discard o the protection is not located or designed in such as way as to hinder the normal

operation of the machine, therefore bringing about the need to remove it.

• Designed with the machine’s functionality in mind, the protection must: o allow for product flow o allow for inspection of the machine o allow for maintenance.

Protective guards and switches must be designed to fail safely – if anything goes wrong, then the machine should end up in a safe state. They must either not require maintenance, or the maintenance must be within the reasonable capacity and expectations of the user. Protective switches such as interlocks must be sufficiently effective and reliable and not lead to lost functionality of the machine, so that the user is not inclined to circumvent it during the life of the machine. The best interlocks become integrated, normal aspects of a machine. A key functional feature of all protective switches is that they must ensure that the hazard is controlled before a person can be harmed. For example, this may involve ensuring that it takes longer to remove a guard than it does for a machine to wind down. In addition, many interlocks are designed to protect the machine itself.

Any manually operated protective switches must take into account the operator, and be in the logical location on a machine. For example, on a wool press, where it is possible to operate the machine from either side, the kill switch should be able to be operated on the same side of the machine relative to the operator, or on both sides of the machine (this was a factor in an injury event that was examined as part of AMDOSS).

A sensor that relies on an operator’s response, such as a smoke alarm, or one that relies on the potential victim’s response, such as a reverse beeper, must be designed with due consideration to human factors. For example, a smoke alarm may be less effective if it has a tendency to cause false alarms, and reversing beepers may not be sufficiently effective to prevent running over children or animals. There are several common examples where a safe system design philosophy cannot easily be married to existing farm workplace infrastructure and farming methods. These areas are particularly challenging for designers.

50

Table 6.1 consists of a selection of common machines that do not preclude, by design, access by the operator to the hazard, some of which were inspected both as case and control machines as part of the AMDOSS study. The table takes into account both new and existing machines of all makes and models, and it is recognised that some machines are safer than others. The agricultural machinery industry and research community have responded to high injury and fatality rates with these machines, and introduced various measures to reduce the injury burden. Some of the most recent of these initiatives are listed accordingly.

Table 7.1 Agricultural machines that present safe system design challenges, and recent initiatives

Agricultural machine Design notes and recent initiatives

Mobile grain auger NSW Workcover has recently released an “industry safety standard” document that was designed to address the safety issues with this machine, particularly the guarding requirements over the intake flighting, where compromise mesh dimensions are put forward The University of Sydney’s ACAHS has completed a report for RIRDC, Publication number 06/034 (Athanasiov et al. 2006)

Post driver NSW Workcover has recently released an “industry safety standard” document that was designed to address the safety issues with this machine, particularly guarding requirements The University of Sydney’s ACAHS has completed a report for RIRDC, Publication number 06/035

PTO power transfer components

Draft versions of proposed new Australian Standards for PTO guarding are available, which offers some improvements to the 1983 Standard

Post hole diggers The University of Sydney’s ACAHS has completed a report for RIRDC, Publication number 06/036 (Miller et al. 2006)

Wool press Notes: One injury event involved an interlock engaging to reduce the injury severity, but not preclude the incident from being able to occur

Harvesting machines

Hay baling machines Agricultural implements Notes:

Agricultural implements with moving parts powered by hydraulics and PTO – protection of third party persons with the use of interlocks or proximity devices

Front End Loaders (FEL) TMA has recently released an industry code of practice for front end loaders for use on agricultural tractors.

51

Headers, hay balers, and various harvesters The safety of headers is of particular interest if one compares the integrated operator protection against machines manufactured for factory workplace environments, which are generally designed to comply with the guarding requirements of AS 4024:2006 Safety of Machinery, and equivalent international standards. In particular, it is common for various reasons that farmers access the “running gear” areas of the header, while the machine is running, in close proximity to moving belts and chains. This occurred to one farmer involved in AMDOSS, who had a finger partially amputated. In order to comply with the principles of AS 4024–2006 (as well as its predecessor, AS 4024–1996), changes to the design would need to include:

• Interlocks would need to be installed on various access points and guards. • Guards would need to provide a more effective separation from moving parts, precluding

access altogether, taking into account ergonomic dimensions. • In some cases, proximity devices would need to be considered. Examples of a proximity

device is a dead man seat, which is installed on some headers, and reversing alarms on cars, which alert the driver to the presence of objects behind the vehicle.

• It is possible that more sophisticated measuring and sensing equipment could negate the need for manual inspection in the future. This could include temperature measurement devices for bearings, and RPM and movement sensors for running gear.

• The increasing use of grease banks and automatic oilers saves time and reduces the operator’s exposure to risk.

Complex machines with many hazards, such as harvesters, stand to benefit most from safe system design; however, it is because of this complexity that the design task is challenging. There are a number of design challenges when introducing a safe system approach for headers. Guards can encourage, or host, accumulation of waste product. It is relatively common for header bearings to fail and then become hot, which can start a fire. This poses a risk to the farm and the community if the fire is not controlled during the hot summer harvest. Attempting to prevent this happening involves operators inspecting around unprotected parts, by physically feeling adjacent to the bearings, checking for refuse build-up, and monitoring running gear. This exposes farmers to a risk – when a risk is continuously taken by a large number of people, a portion of these operators will be injured due to human error. This also impacts individuals, families, and the community. The technology exists to integrate electronic temperature sensors on headers for bearings, sensors on running gear, and interlocks on guards, but the cost of these systems would need to be absorbed by the industry, and the reliability would be increasingly critical with each sensor potentially adding another failure mode to the machine. Proximity devices such as a dead man seat for the operator could be more widely used to disengage running gear in the event that a “third party” person or the operator is near a hazard. Dead man seats are already installed on some new headers. However, this type of sensor can introduce problems of their own. For example, control farmers reported that a dead man seat meant that the operator cannot stand or stretch while he or she works, and cannot walk around the machine while it is operating in order to inspect bearings and other moving parts. As a result, the dead man seat is usually circumvented by re-wiring or placing ballast on the seat as required. Post drivers A hydraulic, tractor mounted post driver resulted in an injury that led to the partial amputation of three fingers. This case highlights known issues with post drivers. The control location is often within reach of the post, and there is nothing preventing hands from entering the hazardous region. Hands can be caught on top of the post, or in between the post and the machine, during ramming, leading to serious hand injury that can permanently affect the functional capacity and quality of life of the farm worker. The control location should be of sufficient distance from the hazard, or separated by means of an effective guard, to ensure that the operator cannot reach

52

the post while operating the controls. A form of proximity device for the machine would reinforce a separated control location. For example, controls that require that both hands to be separated from the hazard area (such as by the use of dead man switch and/or proximity devices) in order to engage the ram. Such a control method may reduce the design requirement of the guard if the designer is confident that operator’s hands are already precluded from reaching forward.

Even if the control location is adequately excluded from the hazard area, a third party is not excluded from the hazard area. The third party (including children) may be present for a number of reasons, but the most common reason might be to assist the operator, for example to help line up the post relative to the machine. The former point requires consideration from manufacturers and suppliers, with a clear need to exclude the operator from the hazard zone. NSW Workcover has recently released an “industry safety standard” document that advises guarding requirements for this reason. It is operationally desirable to be near the post ram as it is faster and more accurate to line up the post by hand. Equipment that is commonly used to attach the post to the rammer is not as fast or effective, such as a chain and lever (this is particularly important for the viticulture industry, ramming hundreds of posts per day). The improvement of this methodology is also a key factor in reducing the risk of operating these machines. If a better method was available to do so, the hazard controls would be less likely to be circumvented. Monitoring and unblocking of product flow Several injuries occurred when farmers were attending to a blockage of material flow with a machine. These necessary operations with the machines result in exposure to hazards. Other common causes of direct human exposure to moving parts include sampling grain or other flowing material, checking of parts such as bearings on machinery, and inspection of components. On some operations where such access is commonplace, good practice could be to use a tool in order to prevent the need for body parts to come close enough to be in danger. If the problem can be reasonably anticipated, tailored tools for the task could be fitted to the machine in the toolbox or near to where it would be needed, possibly attached with a lanyard.

53

Operator protection from falling objects FOPS and FEL Two cases were studied where small tractors with front end loaders (FELs) were inherently hazardous by design, but were legally compliant. Round bales fell onto the operator, causing very severe injuries. Common factors involved with the two were:

• The FEL attachment was lifting fork tines (similar to those used on an industrial loader), was not self-levelling, and the frame at the rear of the forks was lower than the centre of mass of the round bale being lifted.

• The tractor did not have any structure to protect the operator, such as a falling object

protective structure (FOPS). In both cases, compulsory ROPS were fitted, but the ROPS frame was ineffective in protecting the operator from a round bale falling off the FEL.

Two potential strategies could be used to approach this design issue.

1. Mandate FOPS (whether designed to current standards or to an original design that is developed with the consensus of stakeholders and safety experts and engineers) for all tractors fitted with a FEL. This would be regardless of whether the FEL is self levelling, and regardless of the FEL attachment that is fitted at the time of sale, based on the argument that at some stage during the life of the machine, there will be attachments and load scenarios, including adjacent hazards (for example, a hay stack falling down while attempting to retrieve bales) that can threaten the safety of the user.

2. Mandate FOPS for all tractors fitted with a FEL with lifting fork tines.

The former strategy would appear more appropriate, given that during the life of a tractor:

• many different attachments may be used by a number of different users • FOPS also protect the operator from some other hazards and other falling objects • farmers may innovate to create original fork tines or other attachments • the operation of a FEL is subject to human error • the operation of a FEL is subject to mechanical failure (such as hydraulic failure) • The stability of a load on a FEL is also subject to the dynamics of travel, such as accelerating,

decelerating, undulating terrain, slope, slippery terrain, and other sub-optimal operating conditions.

Several older control tractors with FELs had original cabins and shelter structures fitted that did not meet standards for either ROPS or FOPS. The control farmers were either unsure or unaware that these structures, which were originally designed purely for protection from weather, were ineffective as a risk control for these purposes.

54

Safe work environment during maintenance Access to higher sections of tractors and headers Four separate injury events were associated with falls from stationary machinery. Access to areas of these machines that seldom require attention was poor, this feature was also noted on a number of the larger machines reviewed in this study. One way to approach this issue from a safe system point of view is to offer better guards, railings and platforms to facilitate effective access throughout the life of the machine. However, it may also be feasible to ensure that safe access is possible when required without incurring this potentially prohibitive expense by ensuring the practical alternatives are available to end users. The two options presented below are intended to stimulate discussion. Of the many potential solutions, the prevailing principle for all machinery remains - if access is required by design, then the design should ensure safe access.

• Two farmers slipped from the mud guard on tractors, and one from similar guarding on a header. According to Australian Standard AS/NZS 2153.1 Tractors and machinery for agriculture and forestry – technical means for ensuring safety. Part 1: General (which is identical to ISO 4254–1:1989), “Where a guard is in such a position that it may occasionally be used as a step, it shall be capable of withholding 1200N”. A corollary from this requirement must then be interpreted by applying the requirements of state plant regulations under state OHS law. If it is to be used as a step, then the capability of 1200N must be measurable in the vertical direction, and would increase if there was a reasonable expectation that more than one person could use the step at the same time. Consideration would also need to be given to the need for a non-slip level surface, and hand rails if required. The dimensions of the step/guard would also need to be a suitable, for ergonomic requirements, and the steps that lead to it would need to meet the step requirements also detailed in AS/NZS 2153.1/ISO4254–1:1989.

• Safe access may be feasible by designing suitable ladder access. However, many current

designs of tractors may not favour safe ladder access, as the upper roof surface on tractors is smooth and concave, which is an unsuitable mount for most commercially available ladders. Consideration could be given to ensuring the availability of ladders that have a wider, more stable base, or offering a retrofit component for tractor roofs that prevents “off the shelf” ladders from slipping at the top, causing the ladder to tip over.

Tillage implements Several injuries occurred with tillage machinery. Tillage machinery includes various seeders, scarifiers, ploughs, harrows, and similar earth working machinery. All of these machines have a similar hazard profile, with common features for the purposes of risk assessment:

• tillage machines are either towed implements, or attached to the hydraulic three point linkage • towed tillage implements often have hydraulically lowered wheels for transport • implements need to be serviced occasionally to replace consumable parts • tillage implements without hydraulic components often require other hydraulic means to lift

them in order to access the machine, such as a jack, FEL, or three point linkage on a tractor. During maintenance access, a common hazard exists in that an hydraulic system is used to lift the mass of the machine while an operator is nearby. An hydraulic system should never be relied upon to hold up a mass above a person, due to an inherent risk of it failing. It is desirable to find means to consistently ensure that tillage machinery can be safely lifted to allow access in close proximity and underneath, using stable “locked out” lifting devices. One way of achieving this might be to design fixed or removable lifting devices, tailored for the safe lifting of tillage machinery. Features of such devices would need to include:

55

• inherent stability, and tolerance of high horizontal forces without tipping or failure • the ability to be both automatically and manually locked out • the status of the locking out mechanism/s would need to be apparent with simple visual

inspection • failure of components must be anticipated such that the jacking device fails safely.

Air seeder tines Hardened air seeder shears are consumables, and need periodic replacement. Instruction or recommendation of appropriate tools and PPE for their replacement may have prevented the injury that occurred. Tests could be conducted to determine a material that could be recommended to absorb some energy from the impact of a hammer such that the possibility of shatter is low, for example a brass spacer. Battery One farmer was seriously injured, when a battery that he was attempting to jump-start exploded in his face. The problem was apparently caused by a build up of hydrogen gas in the vicinity of the battery, which was located in an unventilated area under the seat of the tractor. This allowed an accumulation of gas to build up, which was ignited when the leads were connected. This incident emphasises the need for the following:

• Batteries should not be located in the operator’s cabin. • If the battery is not of the sealed type, then it must be located in a well ventilated area to avoid

explosion due to build up of gas. • If the battery is not in a ventilated area, then it should be a sealed battery. • If a battery is specified to be a sealed battery for safety reasons, it is important to ensure that

replacement batteries meet the same specification. Means to advise mechanics and operators would include, but should not be limited to, decals and the operator’s manual.

• Irrespective of battery location, battery explosion remains a risk when jump starting. Manufacturers of equipment that include batteries should make end users aware of these risks through the usual means such as decals and owner’s manuals. In addition, this risk should be addressed in tractor use training material.

Farmer-modified or manufactured machinery Several farmers (6) were injured on machines that had been customised for their purpose. In all of these cases, the modification was a key factor in the injury event. Innovation from farmers is integral to Australian agriculture. Customisations are often the only means of efficiently solving problems, and the ability to generate original designs and configurations is an important part of Australian farmers’ skills base. The injury causing mechanism was different in all of these modifications. However, a common theme was that the modifications created a workflow that was not tolerant of human error, i.e., an innocent mistake on the part of the farmer led to serious injury. A greater appreciation of the importance of designing with the assumption that error can and will occur, regardless of the skill of the operator, may support this activity. One injury case provides a good illustration. An engine and pump were assembled for irrigation purposes in a shed. A standard PTO shaft connected the engine and pump, but a guard was not placed on or over the PTO shaft. During a subsequent maintenance activity a farmer became entangled in the PTO when his clothes were caught by the PTO, resulting in multiple leg fractures.

56

Other design issues determined from incident events Dual brake lever arm Serious injuries resulted when a farmer in our study inadvertently pressed a tractor brake pedal while traveling along a dirt road. From a design point of view, the reasons are not particularly important – in any case, it is an example of human error. The dual rear brakes were missing the lever arm that would have engaged both rear brakes simultaneously. Therefore, when only the right rear brake was engaged, the tractor careered off course, and the driver was thrown from the tractor into the path of the rear wheel. This case can be considered an example the uneven application of the rear brakes, which can occur for a number of other reasons. The rear brakes lever may be left disengaged since both rear brakes can usually be effectively operated by the foot straddling both brakes. Even if the brake lever is engaged, uneven application can still occur due to differential drum brake wear during paddock work.

Generally, this design and machine control issue is not a problem when a tractor is travelling in a paddock at slow speed. However, when travelling at higher speeds, there exists a real risk of serious incident. Potential design solutions for newer machines include:

• Farmers could choose to purchase tractors that have disc brakes, which self adjust and apply a force proportional to the constant hydraulic pressure.

• The maximum speed of tractors could be limited to minimise the potential consequences of a loss of control – this could be done by introducing workplace rules, or by retrofitting a governor to limit the use of higher gears.

• Seat belts combined with ROPS would keep operators located in a protected area. • More regular maintenance of brakes could be done to ensure even or substantially even wear

between pads. These measures all affect older, smaller tractors more than others, so the cost implications may be prohibitive, resulting in a reliance on the administrative control of limiting the use of higher gears in older or affected tractors. This issue of braking capacity, as well as stability and control at speed, is likely to increasingly be the subject of scrutiny in the future, as more farmers contract out major functions of the farm business. An increase in the extent, and the speed, of machinery travelling on public roads means that the overall risk of incident will increase with both an increased hazard and increased exposure. Mobile silo levers One case involved serious facial and head injuries when a spring assisted mobile silo lever sprung back after the farmer lost his grip. The design of the particular lever involved in the injury event was such that:

• the force required for normal operation was in excess of ergonomic suitability • the posture required when applying the force was not ergonomically desirable • the spring assisted the function of the mechanism for 90° of travel and hindered for the

remaining 90° of travel.

This particular design was inherently flawed, and posed a hazard to any operator, both for the risk of acute injury, as well as a manual handling related injury. Other mobile silo levers reviewed as part of the study did not present the same extent of specific hazard, but could benefit from more consideration of the manual handling perspective.

57

Other agricultural machinery design issues

Mobile grain augers Older mobile grain augers can be useful for many years on farms, but they generally do not contain some of the features of newer augers that increase the safety and ease of use. Many of these older machines can have their life safely extended with the use of retrofit items that can address these design issues. Key issues with old (and some new) augers which became apparent in our study are:

• The cable winch used to change the angle and height of the auger flighting on older augers often does not incorporate an auto-clutch, which serves to hold the load when the operator lets go of the winding handle. A control farmer in the study reported twice having knocked himself out when he lost grip on the handle, and it swung around and hit him. Replacing these winches with ones that incorporate an auto-clutch would improve their safe use.

• The movement of a mobile auger after it is disconnected from a tractor or utility becomes a

manual handling issue, in that it must be moved short distances over uneven ground. This can be aided by the retro-fit of a “walking jockey wheel”, removing this hazard. These components are available off-the-shelf, and are commonly retrofitted to caravans and boats.

• The “weight on the pull” (mass as measured at the trailer coupling component) of an empty

auger was reported by two control farmers to be troublesome, in that it was too light, and the auger was prone to lifting further over bumps. This issue is simple to rectify by adding mass, but a solution would need to take into account the manual handling issues raised above.

• It is common for the pulleys, belts, and exhaust on engines that are attached to augers to be

unguarded. It was reported from participants that guards were either not replaced after having been taken off for maintenance, or had failed to last as long as the rest of the machine, and had fallen off due to fatigue and corrosion. Retro-fitment of suitable fully enclosed permanent guards would be suitable to control this hazard.

Auger guarding near the flighting is does not currently preclude access to the flighting area for hands and feet. NSW Workcover has recently released an “industry safety standard” that provides guarding requirements for the intake flighting, where compromise mesh dimensions are put forward. The University of Sydney’s ACAHS has completed a report on auger safety (Athanasiov et al. 2006). Tractor exhaust Several control farmers noted that the exhaust pipe on their tractors, located on the right hand side of the bonnet, obscured the view of oncoming traffic when driving on the road. It would seem that many European and US tractors were made primarily for right hand side driving on roads. Rectifying this issue on existing machines by rerouting the exhaust to the opposite side would be a fairly expensive retrofit; the design of such a retrofit would need to be considered by a competent mechanic on a case by case basis.

Two control farmers reported that exhaust fumes found their way into the cabin on tractors due to a poorly designed exhaust pipe location. One of these farmers had modified the exhaust to solve the problem. Noise of tractor exhaust was an issue for some farmers, who preferred that the manufacturer install a more effective muffler system in order to prevent the constant need for hearing protection. Clearly there is the potential to develop generic retrofit exhaust systems, to improve performance in this area. This would need to be designed by a competent person, and could be done using existing components and technology, such as utilising the “baffle” style muffler design currently common for stages of light vehicle exhaust systems.

58

Post hole diggers PTO driven, 3-point linkage mounted post hole diggers were reported by control farms to be troublesome to operate in conditions where down force in addition to the digger’s self weight is needed. Farmers reported adding mass by sometimes climbing onto the digger, or by hanging off a crowbar. This problem was particularly common in the dry soil conditions present in the current drought. It is possible to add “dead weight” to a post hole digger by contriving various means of holding rocks, water, or steel ballast that is designed for tailoring tractor axle loads.

There is nothing precluding access of third party persons to the moving parts of PTO driven post hole diggers, with both the PTO shaft and the drilling auger presenting serious hazards, and no means to stop the machine either automatically or manually, should a person be caught in these parts. A recently development of a PTO interlock that is triggered by a magnetically attached lanyard is now commercially available. This is a good example of innovation for retrofit to tractors, in order to allow safer use of older machines with known hazards. Safe tractor access Several control tractors were of the design that allows access by means of steps in front of the rear wheels, and without any guard in front of the rear wheels. Gear controls were generally located in the middle of the tractor, which would need to be stepped past in order to climb on and off the tractor. There is a risk of a rolling tractor, as kicking the gear lever on the way past could engage the tractor. One known intervention for this risk is the safe tractor access platform (Day & Rechnitzer 2004; Miller & Fragar 2006). These are designed for retro-fit to tractors with the chief intention of making access to the tractor more ergonomic and safe for the operator, while precluding the possibility for run-over to occur if the farmer falls in front of the wheel of a moving tractor. It is possible that the safe access platform design could be retrofit to tractors to allow both safe and ergonomic tractor access if this is not facilitated by any other means. Anecdotal feedback from industry representatives has indicated that in some cases, the safe tractor access platform can encourage farmers to jump on and off the moving tractor in order to perform various tasks such as fodder distribution. There was no evidence to this effect in the evaluation of the safe tractor access platform (Day & Rechnitzer 2004). Nonetheless, there is the potential to further improve the efficacy of the safe tractor access platform by developing retrofit designs such incorporate simple interlocks at access points that prevent tractor and/or PTO operation by providing feedback to sensors that switch the tractor controls as required. Greater role for effective administrative controls Many older tractors observed throughout the study served important operational functions on farms. In some cases, smaller properties could not justify purchasing a newer equivalent machine. Regardless of property size, older tractors may also be dedicated to niche purposes such as towing/transporting implements, feeding, and post hole digging. In addition, older tractors are used as a backup to newer machinery to ensure minimum down-time in the event that a newer tractor is not operational. As many older tractors are mechanically robust and have similar operational specifications to new tractors, it is expected that these will continue to be a feature of the machinery fleet. Several hazards exist on older tractors that have been “designed out” of newer tractors. In many cases, the cost (both time and money) of retrofitting components to eliminate these hazards is prohibitive, in some cases of similar cost to the tractor itself. For this reason, operators would benefit from undertaking risk assessments and develop strategies to manage the risks accordingly. In some instances, this will require administrative controls.

59

An example of this approach is an older tractor wherein the tractor access does not prevent run-over, leading to a risk of fatality if the farmer accidentally kicks the tractor into gear or the tractor rolls when getting on or off the machine. The best approach would be to install a safe tractor access platform (see above), but if this is not possible, then the tractor’s engine should be off every time the steps are used. This makes no difference mechanically to most tractor engines, takes little time, and eliminates the hazard when it is done. Note that even with a safe tractor access platform, the tractor should be stationary when getting on and off the machine. Tractor seat belts Tractor rollover is a well known safety issue, and the widespread uptake of ROPS, encouraged by various incentive programs, has lowered the risk of fatality for farmers (Day & Rechnitzer 1999). This study investigated one case of tractor run-over, which involved a farmer being thrown in front of the rear wheel as the tractor careered out of control. A measure to keep the occupant in a protected zone in the event of tractor instability is to combine the use of ROPS with the use of seat belts. For most older tractors, seat belts would need to be retrofitted. A retrofit could be developed that consists of replacing the original seat with a new incorporated seat, and a seat belt lap sash. It would need to be installed such that the seat belt moves in sympathy with the seat, particularly where original seats have had the benefit of a seat suspension system. Tractor stability will be a growing issue in the future, with tractor speeds increasing to meet the demands of larger properties, and the greater prevalence of contracting. The use of seat belts in combination with ROPS (or AutoROPS – ROPS that remain low and spring up if the tractor starts to roll over) could serve to provide a safer workplace for those seasonal and contract workers who choose to wear the seat belt during operations where the risk of tractor instability is considered to be sufficiently high. Achievement of study objectives

The objectives were to identify the machine factors associated with farm machinery injury, and to explore the interaction of machine factors with human factors. These objectives were met. Personal and machine factors associated with farm machinery injury were identified in the epidemiological analysis (sections of the report entitled “Stage 1 Results: Self-reported individual and machine characteristics”, “Stage 2 Results: In-depth investigation of machine characteristics” and “Design changes for preventing, or reducing the severity of, farm machine related injury”). The interaction of some of these factors was explored in the multivariable analysis (sections of the report entitled “Stage 1 Results: Self-reported individual and machine characteristics”, “Stage 2 Results: In-depth investigation of machine characteristics”). The ability to draw strong conclusions for some aspects of these analyses was restricted to some extent by limited statistical power. The results and subsequent discussion sections (sections of the report entitled “Design changes for preventing, or reducing the severity of, farm machine related injury”, “Discussion: Epidemiological analysis” and “Discussion: Machine design analysis”) detail a thorough investigation of human factors and machine factors. This analysis includes a comprehensive discourse of contemporary incident investigation, which serves two purposes. It details the methodology that has been used to arrive at the various systematic and specific suggested interventions, as well as advocates a way of thinking about agricultural machinery incidents. This means systematic analysis with an emphasis on the role of machinery design for human factors, is something of a new paradigm for the agricultural machinery industry, and emulates similar shifts in thinking evident in other industries such as road transport, aviation, and medicine.

60

The intended deliverables were: • the identification of injury reducing design features for new equipment • recommendations for modification of relevant standards • implications for farm safety educational and training resources • contribution towards meeting the objectives and strategies of the National Farm Machinery Safety

Strategy, which specifically includes recommendations for improved machinery design and modification of relevant standards and/or practice, and identification of target groups for education and training programs

• recommendations for structurally sound modifications which could be made to existing machinery to reduce serious injuries.

All intended deliverables have been produced. Thorough discussion of these deliverables occurs both in the Discussion and Recommendations sections of the report.

61

8. Implications Agriculture in Australia is significant as an export income earner, employer, and supporter of rural communities and farming families. It is therefore of concern that the farm workplace is currently over-represented in workplace incident statistics, meaning that this section of the community is exposed to higher workplace safety risk than its counterparts in other industries (NOHSC 1998). Approximately one quarter of all farm workplace incidents are agricultural machinery related (Fragar & Thomas 2005). This report aims to inform change at various levels of the industry, advocating a combined effort amongst stakeholders to enhance the state of knowledge by innovation, discussion, and implementation of solutions to key machinery design and machinery management issues. A number of these issues have been identified in this report, largely based on the following methods: • statistical analysis against a number of design and operator safety hypothesis, using a case-control

methodology • study of injury event case studies, and comparing these machines and operators to control

machinery and operators • consultation within the agricultural machinery industry • evaluating the machinery that has been studied in the context of available literature in the field of

incident prevention, particularly as it applies to agricultural machinery. Several systematic issues were identified by viewing injury events with the benefit of the most contemporary incident analysis techniques. In particular, agricultural machinery design stands to benefit from the myriad solutions that flow from a systems based approach to injury prevention, with due consideration of human factors. An increasing emphasis in agricultural machinery on ergonomic design has flow on benefits for several reasons. Generally speaking, ergonomic design aims for maximum efficiency, and inherently generates maximum output from operators of machinery. Machinery that is more ergonomic can therefore be safely operated for longer hours, more efficiently, adding to farm productivity. Many solutions and design changes to specific machinery components have been put forward to this end, and are intended to drive discussion toward development of interventions to prevent injury. New machinery designs were compared to injury event machines to compare the components that were involved in the injury event. Machinery manufacturers, importers and suppliers will be encouraged that some of the injuries resulted from design aspects that no longer exist on newer machines, and can learn from the many for which the design features are similar on new machines. The role of various standards and regulatory mechanisms in Australia and internationally has been discussed, with some suggestions as to how standards could be used to drive innovation and best practice into the future. An unanticipated finding from this investigation was the important role of incident response, communication systems, and monitoring on the farm workplace. Various suggestions regarding enhanced communication and teamwork at the farm workplace are put forward. A number of the machinery design issues that were encountered during this investigation are already the subject of a proactive effort within the industry. Examples are post drivers, silos, grain augers, front end loaders, and PTO guard design. These various interventions are commensurate with the National Farm Machinery Injury Prevention Strategy, which shows that specific hazardous machinery that is systematically responsible for ongoing injury on farms is being addressed at a national level. An ongoing coordinated effort for this purpose is encouraged.

62

The most significant beneficiary from this study is the family farm. Family farms will benefit from having an industry that has a greater state of knowledge with regard to safe design and operation of machinery. As farms become safer by upgrading to safer machinery, and retrofitting existing equipment to be inherently safe for a wider cross section of workers, increased access to the labour market can offer a number of cost and logistical benefits to the farm business. Most importantly, any reduction in the overall risk to the farm workplace has direct benefits by reducing the injury burden – serious injury or fatality on a family farm can often jeopardise the viability of the farm, which in turn has an immense impact on the local community.

63

9. Recommendations The following recommendations from the study have been developed with the intention of providing direction for the benefit of the agricultural machinery industry, based on the evidence discussed in this and other studies. The following recommendations are either: • strong recommendations, where there is a clear opportunity to lower risk • identification of systematic design issues, for which a number of potential solutions exist.

Suggestions are been made, from which specific machinery modifications and tailored solutions can be developed.

This study provides a number of initiatives to inform a continuing process of improvement in the field of agricultural machinery safety. The health and safety of farm workers will benefit from the availability of a range of design solutions to ensure a work environment that is as safe as it can be, and certainly as safe and without risks to health as other occupations.

Design solutions

1. Hazards that exist on many mobile grain augers can be controlled by the following commercially available retrofit components: • installing jockey wheels that have a lever to “walk” them into position removes a lifting

and moving manual handling hazard, and makes the job much easier • where required, guarding or a fine mesh cage should be installed over stationary engines

attached to the mobile auger. According to AS 4024.1801:2006, a 25 mm (one inch) mesh constructed 120mm from hazards on plant is sufficient to prevent access beyond the knuckle joint

• installing a hand winch that incorporates an automatic clutch to prevent the handle from flicking back

• guarding over the flighting according to the recommendations of the recently published “Grain Augers” Industry Safety Standard from NSW Workcover.

2. Both rollover and run-over hazards on tractors can be better controlled by maintaining the

operator in the protected zone with the use of seat belts, in addition to ROPS installation. While it is recognised that many farm workers may choose not to use a seat belt on all occasions, it is considered important that the product is at least available for use. Development of a seat belt attachment system that is married to a retrofit seat for older tractors, for bolting on to existing seat mounts where applicable, is recommended.

3. It is recommended that agricultural machinery industry bodies consider means to address

apparent issues with the location of the tractor exhaust: • located on the right hand side of the bonnet, which creates a blind spot when travelling on

the left hand side of public roads. • located in such a way that exhaust tends to enter the cabin.

Management of design issues

4. It is recommended that state workplace health and safety authorities mandate the installation of FOPS on tractors that are fitted with FEL.

5. It is recommended that state workplace health and safety authorities advise of the various risks

of jump starting vehicles. Related advice should also be provided regarding the risk of explosion with non-sealed batteries that are located in a confined area.

64

6. It is recommended that Farmsafe Australia contact the manufacturers of the mobile grain silo

involved in the incident wherein a farm worker was struck by the wheel lifting lever, which occurred largely due to a poorly designed control lever mechanism.

7. It is recommended that all those involved in the agricultural industry accord high priority to

addressing design and risk management issues associated with older machinery, given our finding that the risk of farm machinery injury increases by 3–4% with each year of increase in machine age, and that farm machinery tends to provide a long service life on Australian farms.

8. Farmers with older machinery, or machinery which has not been purchased new, should be

encouraged to undertake risk assessments and implement risk management strategies for this equipment. Farm machinery dealers could play a role in assisting this process. Various retrofit options are available to enable safe use of older machinery, although in some instances the most appropriate risk management may be to develop a plan for replacement of equipment. Other strategies could include using different equipment to undertake the task, upgrading guarding or other safety features, or regular preventative maintenance schedules.

9. It is recommended that resources be made available for a national body such as Farmsafe

Australia to develop a means to gather independent feedback from farmers about the ergonomics and other health and safety aspects of machinery design. This could contribute to an educational database of user experience to inform design improvements, and enhance management of current farm machinery risks. While research organisations are able to provide monitoring over time of the consequences of farm workplace incidents, it would be more effective from a design perspective to have access to direct feedback and ideas from farmers who are invariably intuitive and innovative. In addition to this, the same database could record the details of near misses and incident events.

Safe workplace system initiatives

10. It is recommended that agricultural machinery industry bodies consider the provision of practical and cost effective means to access higher sections of agricultural machinery that require occasional access. Currently, features on the machinery such as wheels, mud guards and implements are being used, where there is a risk of falls. Attention is drawn to the need for solutions to address a wide demographic, including older farmers. Suggestions to consider are: • An access frame with steps situated over the wheels to allow enough reach to clean

windows, and access the roof. • Provision through agricultural machinery dealerships of stable ladders of suitable height.

Wide base ladders would need to be used, as the tractor roof is often not suitable to prevent the ladder slipping sideways.

• Provision through agricultural machinery dealerships of suitable long-reach tools tailored for specific aspects on machinery that require attention, such as cleaning windows.

11. It is recommended that agricultural machinery industry bodies consider means to address risks

with lifting tillage equipment to carry out periodic maintenance such as replacement of consumables. Cost effective and practical design solutions are required in order to provide a secondary load path to allow for the failure of the mechanism, such as hydraulics, that are used to lift the equipment. One way of achieving this could be to attach components to machinery (including retrofit) that can provide the secondary load path. For some machines these components could also serve to provide the lifting force.

12. It is recommended that agricultural machinery industry bodies consider means to address risks

when accessing areas underneath components held up by hydraulics. While manual hydraulic

65

lock-out devices are commonplace, it was widely reported by farmers in this study that they were never used. Means of establishing lock-out systems that allow for this and other human factors with the use of hydraulics is required. Other industries that use hydraulics would also stand to benefit from such systems.

13. It was found that several farmers with serious injuries could have died had it not been for help

arriving in time, either by coincidence, or by others on the farm noting their absence. Farm workplaces should establish means to ensure that communication is available to make arrangements for periodic contact to be made such that an absence is noted within a reasonable time frame.

Standards

14. It is recommended that agricultural machinery industry bodies continue to consider means to ensure the availability and widely understood interpretation of various machinery standards. While current standards should not be seen as perfect, and are subject to continual improvement, they do provide advice toward a reasonable state of knowledge and some minimum design criteria.

In particular, the industry should continue to make themselves aware of the expected levels of safety of equipment under state OHS law, which has been tested in a number of pertinent cases in recent years. These cases make it clear that the operator protective devices used, including guarding and sensor switch systems, are expected to be generally more thorough than the requirements of Australian Standard AS 2153.1:1996 – Tractors and machinery or agricultural and forestry – Technical means for ensuring safety. Attention is drawn to the recently revised Australian Standard AS 4024:2006 – Safety of Machinery.

15. For a number of specific machines particular to agricultural applications there is ambiguity

with the interpretation of various standards and other safety requirements. The best means to control risks needs to be determined and disseminated in a way that recognises compliance and promotes national uniformity. It is recommended that a discussion paper be developed to consider the potential for the uniform provision of standards for agricultural machinery, which may draw from, but be independent of, existing guidance and standards regimes. A key function of this discussion paper would be to canvas options to determine the best way to achieve national uniform compliance. One model to consider is that of the Australian Design Rules (ADR). An ADR series for agricultural machinery could be developed over time in close consultation with the industry to promote uniform continual improvement to machinery safety, enforce minimum standards with clear guidelines, and encourage the development and uptake of innovative solutions to problematic hazard control issues.

16. Australian agricultural machinery designers and professionals who purchase and import

agricultural machinery could consider better consolidating their skills and knowledge base by developing a professional body or negotiating equivalent relationships with existing organisations. Existing organisations that provide a similar role in enhancing professional standing are: • Engineers Australia, and the Society for Engineering in Agriculture • American Society of Agricultural and Biological Engineers • Tractor and Machinery Association • Farm and Industrial Machinery Dealers’ Association • The Kondinin Group

66

17. Funding should be made available to develop and support Farmsafe Australia’s National Farm Machinery Safety Reference Group to provide for ongoing national initiatives to advocate and promote:

• nationally coordinated programs, including working closely with state workplace health

and safety authorities and state farmer representative organisations • reduction of duplication of agricultural machinery injury prevention programs • development of shared resources for the benefit of agricultural machinery safety.

Education and training

18. The design and implementation of error tolerant systems in agricultural practice and agricultural machinery design could be viewed as an important element of risk management within the industry. Machinery design education, agricultural workplace training, as well as stand alone OHS education and training curricula, should incorporate the principles of error tolerant systems. Farmers as innovators in the agricultural machinery industry are a very important target group for such educational initiatives.

Further research

19. There appears to be considerable potential to improve machinery safety with sensors and interlocks. An engineering issue exists with the introduction of multiple node sensor systems, such as interlocks, on complex machinery such as headers and balers. Each new sensor introduces another failure mode for the machine. This is particularly an issue when interlocks should be designed to fail safely – i.e. if the interlock fails, it should preclude the function of interest from occurring. Methods to overcome this issue have known problems. • One way to overcome reliability issues is to introduce redundant wiring or other systems

that assists with fault diagnosis, resulting in minimal down time in the event of a sensor failing safely. This comes at a cost that must be passed on to machinery owners.

• Increasing reliability by making sensor systems more robust and higher quality also comes at additional expense, and does not guarantee that problems will not occur throughout the life of the machine.

The best way to address this problem of safety systems introducing multiple failure modes could be to use both of the above strategies, and other approaches. This design issue could benefit from further research to find ways of introducing effective operator protection systems that are cost effective and reliable.

20. Further research is required to understand the mechanism by which those with a previous farm

work related injury severe enough to require hospital admission may be at increased risk for a subsequent injury. In the interim, patients being discharged from hospital following a farm work related injury should be advised of their possible increased risk of a subsequent injury. Patients likely to have an ongoing physical limitation may benefit from occupational therapy assessment and support during the initial return to work phase. In instances where a physical limitation persists over a longer period, regular assessments may be required.

21. Further research may also be required to understand the mechanism by which employees, contractors and seasonal workers are more likely to sustain a machine related injury. This finding could be related to their increased exposure to machine use, or more risky machinery related tasks, than owners/managers. It could also be related to a lower prevalence of having attended agricultural training courses among this category of farm worker, since this study found that injured farm workers were less likely to have undertaken agricultural training courses.

67

22. Many mechanisms are possible for the continued development of safe design innovations for agricultural machinery. One of these is Australian Standards. While Australian Standards will continue to have an important role in influencing agricultural machinery design, it is recognised that standards can also limit innovation, and are subject to a review process that takes a long time. With this in mind, and acknowledging that the same limitations can apply to other forms of pseudo – standards (such as industry codes of practice), improvements in standards are not seen as the only means to bring about change. Some design issues require serious research and development commitment to ensure that solutions have been well considered, and properly evaluated, in a way that is more thorough than an expert Standard or Code committee can provide. The industry would benefit from research that seeks to optimise the process for designing, evaluating, and translating safe machine innovations to end users, in the agricultural machinery context. Such an investigation would take into account national, international, and hypothetical models for the process.

23. This study has demonstrated the value of an in-depth analysis of incident events, using an

approach that uses the most contemporary incident prevention theory as a benchmark for hazard management. Further research of incident events is required to build on this body of knowledge, likewise looking at actual causal mechanisms, and seeking further evidence of the hazard management themes that are articulated in this report.

68

Appendix 1 Table A.1 Characteristics of cases and controls who did and did not have the index machine inspected, Agricultural Machinery Design and Operation Safety Study, Victoria, 2003–2006

Characteristic Inspected (cases n=37

controls n=71) n [%]

Un-inspected (cases n=48

controls n=134) n [%]

Odds Ratio [95% CI]

p-value

Age Cases Controls

50.3 [12.6] 51.1 [10.8]

47.0 [13.6] 46.6 [12.3]

1.02 [0.99, 1.05] 1.03 [1.01, 1.06]

0.26 0.01

Injury outcome (cases only) Fatality Admitted Not admitted

0

32 [86.5] 5 [13.5]

2 [4.2]

35 [72.9] 11 [22.9]

N/a 1

0.50 [0.15, 1.59]

-

0.24

Position on farm – cases Employee/contractor Owner/manager Position on farm – controls Employee/contractor Owner/manager

5 [13.5]

32 [86.5]

4 [5.6] 67 [94.3]

27 [56.3] 21 [43.8]

32 [23.9] 102 [76.1]

1

8.23 [2.73, 24.76]

1 5.25 [1.78, 15.54]

-

<0.0001 -

0.003 Machine age Cases Control

19.4 [18.4] 12.6 [11.1]

17.9 [16.2] 12.8 [12.5]

1.01 [0.97, 1.04] 1.00 [0.97, 1.03]

0.74 0.93

69

References Association for the Advancement of Automotive Medicine, The Abbreviated Injury Scale, 2001 Revision, The Association, Des Plaines, IL. Australian Bureau of Statistics 1992, Australian and New Zealand Standard Industrial Classification, Catalogue No. 1292.0, Australian Bureau of Statistics, Canberra. Australian Bureau of Statistics 2003a, Agricultural Commodities, Australia, 1999–2000, Catalogue No. 7121.0, Australian Bureau of Statistics, Canberra..

Australian Bureau of Statistics 2003b, Labour Force, Catalogue 6203.0, Australian Bureau of Statistics, Canberra.

Australian Bureau of Statistics 2007, Year Book Australia 2007, Catalogue 1301.0, Australian Bureau of Statistics, Canberra.

Australian Safety and Compensation Council 2002, National OHS Strategy 2002–2012, Commonwealth of Australia, Canberra.

Athanasiov, A, Fragar, L & Gupta, M 2006, Farm Machinery Injury: improving the safety of grain augers in Australia, RIRDC publication no. 06/034, Rural Industries Research and Development Corporation, Canberra.

Carlson, K, Gerberich, S, Church, T, Ryan, A, Alexander, B, Mongin, S, et al. 2005, ‘Tractor-related injuries: a population based study of a five state region in the Midwest’, Am J Ind Med, 47: 254–264.

Day, L, Ashby, K & Stathakis, V 1997, ‘Non-fatal farm injury in Victoria’, Hazard, 33, Monash University Accident Research Centre, Melbourne.

Day, L, Langley, J, Stathakis, V, Wolfe, R, Sim, M, Voaklander, D, et al. 2007, ‘Challenges recruiting farm injury study participants through hospital emergency departments’, Injury Prevention, 13: 88–92.

Day, L, Langley, J, Voaklander, D, Sim, M, Wolfe, R, Dosman, et al. 2005, ‘Minimising bias in a case control study of farm injury’, Journal of Agricultural Safety and Health, 11(2): 175–184.

Day, L & McGrath, A 1999, Unintentional machinery injury on farms in Victoria, Report no. 148, Monash University Accident Research Centre, Melbourne.

Day, L & Rechnitzer, G 1999, Evaluation of the tractor rollover protective structure rebate scheme, Report no. 155, Monash University Accident Research Centre, Melbourne.

Day, L & Rechnitzer, G 2004, ‘Safe Tractor Access Platforms: from guidance material to implementation’, Journal of Agricultural Safety and Health, 10(3): 197–209.

Dekker, S 2002, ‘Reconstructing human contributions to accidents: the new view on human error and performance’, Journal of Safety Research, 33: 371–385.

Farmsafe Australia Inc. 1998, National Farm Machinery Safety Strategy, June 1998, Australian Agricultural Health Unit, Moree.

Fragar, L & Franklin, R, The health and safety of Australia’s farming 2000, community, a report prepared for the National Farm Injury Data the Farm Safety Joint Research Venture, May 2000.

Fragar, L, Franklin, R & Coleman, R 2000, The Farm Injury Optimal Dataset RIRDC ver.1.2, publication no. 00/10, Rural Industries Research and Development Corporation, Canberra.

Fragar, L & Thomas, P 2005, Machine Injuries on Australian Farms, RIRDC publication no. 05/050, Rural Industries Research and Development Corporation, Canberra.

70

Franklin, R, Mitchell, R, Driscoll, T & Fragar, L 2000, Farm-Related Fatalities in Australia, 1989–1992, Australian Centre for Agricultural Health and Safety, National Occupational Health and Safety Commission and Rural Industries Research and Development Corporation, Moree.

Gardner-Bonneau, D 2000, ‘Machine design’, In Kadzin, A.E. (Ed) Encyclopedia of psychology, Vol 5 (90–94), American Psychological Association, Washington DC.

Gerberich, S, Gibson, R, French, L, Lee, T, Carr, W, Kochevar, L, et al. 1998, ‘Machinery-related injuries: regional rural injury study – 1 (RRIS–1)’, Accident Analysis & Prevention, vol .30, no. 6, pp793–804.

Griffith, G 1994, The economics of farm safety in Australian agriculture, Final report DAN–62A, Rural Industries Research and Development Corporation, Canberra.

Heinrich, H W 1931, Industrial Accident Prevention, McGraw – Hill, New York.

Heinrich, H W 1959, Industrial Accident Prevention A Scientific Approach, (4th ed.) McGraw – Hill, New York.

Human Factors and Ergonomics Society, 2005, Definitions of human factors and ergonomics, Downloaded from the world wide web, May 4th, 2007, http://www.hfes.org/Web/EducationalResources/HFEdefinitionsmain.html).

Ingram, M, et al. 2001, A protocol for data collection in a case-control study of farm machinery injuries, Paper 01–405, American Society Agricultural Engineering Annual International Meeting, Sacramento, California, USA.

Ingram, M, Crowe, T, Wassermann, J, Hagel, L & Dosman, J 2003, ‘Case reports of on-site investigations of auger related injuries’, Journal of Agricultural Safety and Health, 9(2): 133–142.

Layde, P, Nordstrom, D, Stueland, D, Brand, L & Olson, K 1995, ‘Machine-related occupational injuries in farm residents’, Annals of Epidemiology, 5(6): 419–426.

Layde, P, 1990, ‘Beyond surveillance: Methodologic considerations in analytic studies of agricultural injuries’, American Journal of Industrial Medicine, 18: 193–200.

Lee, T, Gergerich, S, Gibson, R, Car, W, Shutske, J & Renier, C 1996, ‘A population-based study of tractor related injuries: regional rural injury study –1 (RISS–1)’, Journal of Occupational and Environmental Medicine, 38(8): 782–793.

Miller, J, Fragar, L & Franklin, R 2006, Farm Machinery Injury: injuries associated with posthole diggers, RIRDC publication no. 06/036, Rural Industries Research and Development Corporation, Canberra.

Miller, J & Fragar, L 2006, Farm Machinery Injury: Injury involving tractor run-over, RIRDC publication no. 06/033, Rural Industries Research and Development Corporation, Canberra.

National Committee for Injury Prevention and Control 1989, Injury Prevention: Meeting the Challenge, Oxford University Press, New York.

National Occupational Health and Safety Commission 1998, Work-related traumatic fatalities in Australia, 1989–1992, Summary Report 1998.

Reason, J 1990, Human error, Cambridge University Press, New York. Reason, J 1997, Managing the risks of organizational accidents, Ashgate, Aldershot.

71

Salmon, P, Regan, M & Johnston, I 2005, Human error and road transport: Phase 1 A framework for an error tolerant road transport system, Report no. 256, Monash University Accident Research Centre, Melbourne.

Salmon, P, Regan, M & Johnston, I 2006, Human error and road transport: Phase 2 A framework for an error tolerant road transport system, Report no. 257, Monash University Accident Research Centre, Melbourne.

Sprince, N, Park, H, Zwerling, C, Lynch, C, Whitten, P, Thu, K, Gillette, P, et al. 2002, ‘Risk factors for machinery-related injury among Iowa farmers: a case-control study nested in the agricultural health study’, International Journal of Occupational and Environmental Health, Oct–Dec 2002; 8, 4 Health Module p332.