Lesley Day, Voula Stathakis
Monash University Accident Research Centre
January 2004
RIRDC Publication No 03/139
RIRDC Project No UMO-22A
Introduction
Monitoring and evaluation of farm health and safety programs plays
an important role in reaching the common goal of improving health outcomes
for farmers, farm workers and their families. This role includes contributing
to development and refinement of programs, determining their effectiveness,
identifying unanticipated disadvantages and providing an evidence base
for prioritisation of resource allocation. Program based evaluation of
farm injury prevention activities has been occurring sporadically throughout
Australia. However, there is also a need for evaluation and monitoring
of regional programs at the state and national levels. Such work has been
initiated in New South Wales, where baseline data has been collected. In
Queensland, the impact of prevention activities has been monitored in a
series of farmer surveys that provide a summary of the status of farm safety
management practices in Queensland and demonstrate that injury reductions
have occurred in some agricultural sectors.
This project was undertaken to monitor changes to farm safety practices and behavior, and injury rates, in Victoria to support state-wide and regional programs. As similar monitoring also occurred in Queensland, there appeared to be an opportunity to compare results between two states where different approaches to farm injury prevention existed.
The objective of this study was to compare the impact of differing approaches to farm injury prevention in Victoria and Queensland. The main outcomes were intended to be:
Additional outcomes were anticipated to include:
Methods
Two complementary methodological approaches were employed:
In addition to the common variables, state specific questions, particularly
in relation to recall of prevention campaigns, were included.
Comparisons were made between the two time points in Victoria, between the four major commodity groups at the two time points, and between Victoria and Queensland for 1998. Comparisons were also made between those who had and had not been exposed to various farm safety activities and programs, to examine the association with improvements in safety related behavior, and practices, and injury occurrence.
Results
In addition to the obvious differences in geography, size and agricultural
activity, Queensland and Victoria also differed on their history of occupational
health and safety legislation. The key organisations involved in farm safety
are similar in both states, and the major farm safety strategies were similar.
However, a key difference exists in that legislation has been used in Queensland
to create an environment for change, whereas in Victoria, legislation actually
required change and has been coupled with enforcement campaigns. Implementation
in Victoria has tended to focus on single key issues such as tractor roll-overs,
while in Queensland implementation has tended to take a broader focus.
The Victorian surveys revealed that there were statistically significant
improvements during the study period in six variables measured:
Note that the decrease in ROPS purchases is due to earlier widespread
uptake.
There were no changes in the following variables:
Recall of the 2001 VWA farm safety advertising campaign, attendance
of at least one field day, seeking safety advice, and attendance at a safety
seminar or course were all independently associated with the presence of
at least one safety related action, over the 12 months prior to the survey.
Attendance of at least one field day, seeking safety advice, and attendance at a safety seminar or course were also independently associated with average or above average performance of farm safety behaviours. These results indicate an association only and no conclusions can be drawn about the temporal relationships or a possible causal nature of the association.
The purchase of at least one safety item and making at least one safety related change were both independently associated with an increased odds of having had at least one serious injury on the property in the previous 12 months. Again, no conclusions can be drawn about the temporal nature of the associations. These results could mean that purchasing a safety item, or making a safety related change on the property, occurred after the occurrence of an injurious event. Alternatively, the results could mean that these safety actions were already in place on properties on which an injury was subsequently reported.
Among the four major commodity groups (milk cattle, beef cattle, cereal grains, sheep for wool and meat), the milk cattle commodity group showed statistically significant improvements across the largest number of variables measured. This could partly be explained by the fact that this group, along with meat cattle, were the largest in the survey, resulting in increased power for the statistical testing.
However, the size of the improvement for some variables was much greater than those for the other groups. The milk cattle commodity group was also the leading group in 2001 on a large proportion of the parameters measured. Rates of serious injury per 100,000 hours worked decreased by 5% and 4% for milk and meat cattle respectively, and by 1% for cereal grains. None of these decreases were statistically significant.
With respect to the state comparison in 1998, Queensland farms were significantly higher on measures for: attendance rate per 100 staff for producer group talks, farm safety behaviour, and children who had been on a school farm safety tour in the previous 12 months. They also had higher proportions of farms conducting a formal safety check ever, or in the last 12 months, and purchasing a safety item in the last 12 months. However, this difference was not found when the four major commodity groups were compared. Victorian farms were significantly higher on measures for: attendance per 100 staff at TAFE courses, ROPS fitment, making safety related changes in the last 12 months, and properties with children under 5 years with secure fences.
There was no significant difference between states on measures for: seeking farm safety advice, proportion of properties with someone attending health and safety training, conducting a formal safety check ever, or in the last 12 months (for the four major commodity groups), or purchasing a safety item in the last 12 months (for the four major commodity groups).
The serious work-related injury rate per 100,000 hours worked for all farms was significantly higher in Victoria (8.5, 95% confidence interval 7.6-9.4) than in Queensland (3.1, 95% confidence interval 2.8- 3.3). This difference was also apparent across the four major commodity groups.
Discussion
Although the survey data could not be considered to be directly generalisable
to all Victorian farms, the majority of respondents in the surveys were
similar enough to the majority of all Victorian farms, to consider the
survey data as at least broadly indicative of Victorian farms. Non-response
bias did not appear to be present based on the limited information gathered
in a telephone follow-up of a sample of non-responders.
The period to which the Victorian farmer surveys relate covers from mid 1997 to mid 2001. This period was characterised by increasing farm safety activities and programs. The overall impression from the survey results for this period was one of small to modest changes in a relatively small number of farm safety behaviours and practices. The level of observed changes in farm safety behaviours and practices were associated with a 14% reduction in the rate of self-reported serious work related injury.
However, the link between these changes and the injury reduction is unclear. The surveys were crosssectional and therefore the temporal relationship between changes in behavior and practice, and injury reduction cannot be firmly established.
Among the four major commodity groups, milk cattle demonstrated statistically significant change across the greatest number of variables, and also had the highest levels of performance in 2001 on the greatest number of variables. It may seem curious then, that this group did not achieve a statistically significant injury reduction, and that the reduction (5%) was of a similar magnitude to that achieved by the meat cattle group (4%). The surveys may not be measuring all the variables associated with injury outcome. In addition, the baseline dairy specific injury rate (ie., 1998) was not significantly different to that for all farms, whereas the baseline injury rate for meat cattle was significantly higher than that for all farms (twice the all-farm rate, in fact). This being the case, it is possible that larger changes would be required by the dairy group, than by the meat cattle group, to have the same order of magnitude impact on injury rate.
The serious work-related injury rate per 100,000 hours worked for all farms, and for the four major commodity groups, was significantly higher (two to three times higher) in Victoria in 1998 when compared with Queensland. This appears to be somewhat unexpected, given that dramatic differences between the two states on farm safety practice and behavior as measured in the 1998 survey were not observed. It could be that the surveys did not measure those factors which influence injury occurrence, and therefore did not capture the factors responsible for lower rates in Queensland. Alternatively, systematic errors in estimating the rate numerators and denominators could also account for such a difference. The total average weekly hours worked per property in Queensland was higher than in Victoria (132.2 hours, 80.9 hours respectively). It is difficult to determine whether this is due to differences in reporting, or whether it reflects different agricultural practices and labour requirements.
However, differences of this magnitude were also observed in the average hours worked per week per property for each of the four major commodity groups, suggesting a systematic difference in reporting.
The cross-sectional nature of this comparison, and the possible influence of state-specific factors outside those considered in this study, limit the conclusions which can be drawn about the association between the differing approaches and outcomes in the two states.
Recommendations
That the impact of respondent perceptions of the organisation conducting
research on participation and data accuracy and quality be explored
That the feasibility of prospective cohort or longitudinal intervention
studies to determine whether improved safety behaviour and practice leads
to injury reduction be explored That ways to improve response
rates to mailed and telephone surveys on safety be developed and tested
That a systematic literature review of the effectiveness of farm safety
intervention programs be conducted and informative summaries of successful
programs be widely disseminated, to capitalise on the increased number
of intervention studies published in the peer-reviewed literature since
this project was commenced That the literature on uptake of
farm safety interventions be examined and summarised to inform farm safety
programs, and to identify research gaps That consideration
be given to targeting the following issues in intervention programs which
were identified as having low uptake in the 2001 survey: Power take off
guards Formal safety checks Use of hearing protection Machinery maintenance
Correct lifting technique Respiratory protection when handling chemicals
Managing Farm Safety training Seat belt wearing was also noted as being
relatively low. However, fitment of tractors with seat belts is low, and
there may be ergonomic problems with some seat belt configurations.
That the meat cattle and sheep industries receive some targeted interventions in view of the higher injury rates and lower safety performance That the presence of innovative and interactive safety displays and activities at farm field days be continued, given that approximately 40% of respondents attended field days in the last 12 months.
Even if those attending may be already actively interested in farm safety it would appear to be worthwhile to provide good quality advice and information to encourage continued and increasingly effective action.
That the need for, and feasibility of, an ongoing farm safety training program for rural suppliers and dealers be explored, given that approximately 30% of respondents sought farm safety advice in the last 12 months, and that rural suppliers and dealers were the two main sources of advice That the feasibility of establishing a formal network of farmers trained in providing farm safety advice to their peers be explored, on the basis that other farmers were another important source of advice.
That a set of standard indicators of farm safety management and performance in Australia be developed and validated That the feasibility of using a more reliable sampling frame for future monitoring and surveillance in Victoria be explored That measures of changes of risk exposure for the major commodity groups be developed That significant resources not be allocated to detailed state-based comparisons That broad level, ecological, associations noted by comparing different groups form the basis for hypotheses which may be tested using appropriate study design That this report be disseminated to the key stakeholder organisations in the Victorian agricultural industry, to members of the Victorian Farmsafe Alliance, and other interested parties That this report be disseminated to the key organisations in the agricultural industry in other states and territories.
That the results of this report be disseminated in an appropriate format to rural media That opportunities to present the results of these surveys to appropriate audiences be sought
 
|