https://orcid.org/0000-0001-5328-3161
ABSTRACT
Objective
Generation of reliable data underpins the effectiveness of Occupational Health and Safety (OHS) surveillance systems. Despite the importance of understanding OHS data systems, there are few papers that provide overviews of their structure and/or content. This paper introduces a basic framework for assessing OHS data systems that will be of use to researchers internationally. We applied this approach to assess the Irish OHS data system by undertaking a data mapping exercise.
Method
We developed a checklist based on recommendations of monitoring and measurement of OHS proposed by the National Academies of Sciences, Engineering, and Medicine (USA). An assessment of published reports that present systematic OHS surveillance data was undertaken to identify the institutions or organisations responsible for collecting and curating the data, their remit, and, associated with this, their respective case definitions. We then provide an overview of the variables collected and these are then mapped against the checklist.
Results
The assessment highlights that whilst the farm fatalities dataset provides complete coverage of all fatalities, regardless of age or employment status, the same is not true of the three non-fatal injuries datasets reviewed. There are important differences in the data collection methods and, associated with this, which populations are covered.
Practical Application
The assessment approach provides valuable insights into the strengths and weaknesses of a critical element of OHS surveillance systems, namely the production of datasets. This knowledge is important for researchers as understanding the data that informs their research is fundamental to good science. It is critical for policy-makers and other stakeholders to understand the strengths and weaknesses on which OHS policy, strategies, or education and training interventions are developed.
Introduction
Farming is a dangerous occupation, and farms are hazardous environments for those living and working on them.Citation1–4 In order to reduce injuries, Occupational Health and Safety (OHS) data systems, key elements of OHS surveillance systems, collect or collate data describing occupational injuries, fatalities, or diseases.Citation5 A variety of approaches have been developed to generate these data including surveillance of media reports,Citation6 analysis of prehospital care reports and hospital data,Citation7 social insurance systems,Citation8 and data generated via regulatory systems or survey,Citation9 among others. These data are used to monitor trends and identify vulnerable cohorts, risks, and hazards related to farm work or farms. This knowledge is central to developing a better understanding of injury incidents and developing tailored risk assessment/management tools, training programmes, education initiatives, and farm safety policies.Citation5 Whilst there are several papers describing the approaches used to collect data from media and hospital reports,Citation7 there is a significant gap in understanding data, and associated limitations, generated through other approaches, i.e. from social insurance or regulatory systems.Citation10 In order to support public health practitioners, particularly those involved in health and safety issues, we develop a “Practice History” that describes the data systems that systematically, i.e. on a regular and ongoing basis, collect, organise, and store records of fatal or non-fatal injuries associated with farming in Ireland. The Practice History is organised around three objectives: describe the institutions involved in the collection of systematic data; describe the data collected in terms of case definition; and describe and assess the variables collected.
Regulatory approaches to OHS surveillance generally rely on continuous monitoring of incidents by inspectors, data recorders, or through mandatory self-reporting to gather information regarding fatal or non-fatal farm work-related incidents.Citation5 Many countries including the USA, Ireland, the UK, Spain, and Australia, to name a few, have developed regulations to protect and improve OHS and associated with this they have developed OHS surveillance systems. Whilst these systems have the benefit of being considered authoritative, the data collection process and reporting procedures need to be carefully considered. This understanding is critical to policy-makers and stakeholders, and researchers, particularly those interested in understanding the relative level of risk of fatalities or injuries by comparing with other occupations or between countries.
The evaluation of OHS surveillance systems can be complex, resulting from the need to understand the roles, responsibilities, and relationships between multiple institutions or agencies, and assessment of the data system is relatively straightforward. Key measurement objectives for OSH surveillance systems have been set out in the National Academies of Sciences, Engineering, and Medicine (USA) (NASEM) report on “A smarter national surveillance system for occupational safety and health in the 21st century”.Citation5 The report’s authors concluded that the overall objectives of an OSH surveillance data system include the measurement of the burden of work-related injuries or illnesses and the monitoring of trends over time and space.Citation5 Associated with these objectives, there is a need to identify industries, occupations, and worksites that may present high risk for work-related injury, illness, or hazardous exposures. There is also the need to provide data that identifies populations, and their socio-demographic characteristics and work arrangements that are exposed to these risks; measure the severity or impact of injuries or illnesses including need for medical treatment, surgery, or number of lost workdays; and assess the short- and long-term economic and social costs.
Methods
To evaluate the Irish OHS data system, we developed a simple checklist based on the NASEM criterion to assess whether the data collected measure these features of fatal or non-fatal injuries. An assessment of published reports that present systematic OHS surveillance data is undertaken to identify the institutions or organisations responsible for collecting and curating the data, their remit, and, associated with this, the case definitions applied to determine which injuries and fatalities are counted. We then provide an overview of the variables collected and these are then mapped against the checklist. Before proceeding, we provide a brief introduction to the three organisations involved in the collection of systematic OHS data in the next section.
Results
Institutions collecting systematic data on fatal and non-fatal injuries
This Irish OHS surveillance system is grounded in legislation and regulation. Under the Safety, Health, and Welfare at Work Act (SHWWA) (2005), the HSA serves as the national statutory body responsible for documenting all work-related fatal and non-fatal incidents across all economic sectors, including “Agriculture, Forestry, and Fishing”. As part of this remit, the HSA is responsible for the collection, collation, and reporting of data on fatal injuries and mandatory self-reporting of non-fatal injuries.
There are two additional organisations active in the collection of systematic information relating to non-fatal injuries: the Central Statistics Office (CSO) – the national statistical office, responsible for collecting, analysing, and reporting statistics describing Ireland’s people, society, and economy, and Teagasc – the Irish Agriculture and Food Development Authority. The CSO, in order to comply with the requirements of European Council Regulation (EC) No. 577/98, collects and reports data describing work-related injuries across all sectors, including agriculture. Teagasc undertakes an annual National Farm Survey (NFS).
Farm fatalities
Case definition, data collection process, and variables
The HSA is the only organisation involved in the collection of data relating to farm fatalities. The HSA implements the Health, Safety and Welfare at Work Act 2005, which defines a workplace incident as a fatal or non-fatal injury arising out of or in the course of employment. The HSA further defines a farm fatal incident as an unintentional injury resulting in death that occurs as a consequence of farming activities. The combination of these two definitions results in the HSA applying a broad definition of their remit, which covers any death that occurs on a farm or in the course of farm work being undertaken. Consequently, all fatal incidents related to farm work activities, including those involving farmers/farm operators, farm workers, family members, and bystanders, are reported in the national statistics. This comprehensive approach ensures coverage of all vulnerable cohorts and distinguishes it from regulatory systems that only report incidents affecting employees, working for payment or profit who are over 15 years of age.
The HSA conducts an investigation into notification of a fatal farm injury. This can be reported by either farmers/farm operators, other family members, the police, or the local coroner. An inspector is tasked with visiting the location of the fatality and undertaking an investigation to establish the cause(s) of the incident. As part of this process, the inspector records demographic and incident characteristics. These variables include personal characteristics (age in years, gender, type of farm, e.g. dairy, beef, tillage, etc., location of farm and location of incident on the farm), injury characteristics (cause of incident, cause of injury, time of injury (hour, day, month, and season)), and a brief qualitative description of the incident ().
Table 1. Farm fatality dataset (HSA).
Farm injuries
Case definition, data collection process, and variables
According to the Safety, Health and Welfare at Work (General Application) Regulations 2016, which is implemented by the HSA, all employers and self-employed individuals, including farmers, are legally required to self-report farm-related injuries that result in four or more days of lost work by their employees. Additionally, injuries sustained by non-workers, such as family members, that require medical treatment in a healthcare facility must be reported to the HSA. Notwithstanding this, the HSA has concluded that there is a significant level of underreporting of injuries that occur on farms.Citation11 This issue is not unique to farming. An assessment by the HSA concluded that underreporting of work-related incidents is more prevalent in sectors dominated by smaller enterprises and self-employed people who are less likely to report incidents than larger enterprises.Citation11 In 2021, 120 injuries that occurred on farms were reported to the HSA.Citation12 As an EU Member State, Ireland must comply with EU occupational health and safety regulations. This include the application of the European Statistics on Accidents at Work methodology, which sets out which variables have to be collected by the HSA. These include demographic and cause and mode of injury variables (). Whilst the dataset is not publicly available, the HSA provides a detailed analysis and breakdown of the figures in regular reports (see referenceCitation11).
Table 2. Self-reported (mandatory) farm injury dataset – HSA.
The CSO collects data on work-related injuries from all sectors, including agriculture, as part of the Labour Force Survey (LFS) work-related injury and illness module, which is carried out in the first quarter of each year since 2003. Whilst the survey focuses on work-related injuries and illnesses that result in four or more days of work loss, it also captures data on injuries or occupational illnesses that result in any loss of work. The module is limited to those working for payment or profit that are, at least, 15 years old. Therefore, the large workforce of unpaid family members are not covered in this survey. In 2021, the number of persons participating in the injuries module was 391.Citation11 Whilst the CSO weights these responses to estimate the total number of injuries by economic sector, the small sample size associated with each sector means that the results need to be treated with considerable caution. The data reported by the CSO differ from those reported by the HSA. Whilst, overall, they are not as detailed, they do provide information about minor injuries that do not result in the loss of any working time and occupational illnesses. Broad categories of variables include demographic and cause and types of injury or illness variables (). The CSO do not publish these data directly, rather they provide them to the HSA who publish them in conjunction with their analysis of self-reported injuries (see referenceCitation12).
Table 3. Labour force survey: occupational (farm) injury dataset – CSO.
The Teagasc farm injuries dataset is collected as part of a larger survey, the National Farm Survey, undertaken with a nationally representative sample, by size and system, of farms with a “Standard Output” greater than €8,000. The total sample size in 2021 was 795 representing 85,000 farms nationally.Citation13 There were 135,037 farms in Ireland in 2020 and, of these 48,356 (36%) are estimated to have <€8,000 Standard Output.Citation14 Trained data recorders conduct face-to-face interviews with farm operators, generally the “farm holder” or manager. The survey of farm workplace injury levelsCitation15 has been conducted every 5 years since 1996 and annually since 2022. This data collection instrument has previously been described.Citation15
In the NFS survey, a farm injury is defined as any accident-causing injury, irrespective of the severity of the injury. Injuries affecting any member of the farm household or any worker on the farm are recorded. Variables relating to injuries include demographic, socio-economic, type of worker, injury type and location, level of medical attention required, and number of days absent from work (). Teagasc makes these data available to researchers on a case-by-case basis.
Table 4. Farm injury dataset1 – Teagasc national farm survey.
With the exception of the estimation of the long-term social and economic costs, the HSA’s fatalities dataset fulfils the criteria proposed by the NASEM for OHS surveillance systems. These data provide a comprehensive overview of fatal injuries and are used to produce regular reports and presentations monitoring trends and identifying and ranking hazardsCitation11 (). These analyses also consider emerging hazards identified through an analysis of both categorical data and qualitative descriptions of the events resulting in a fatality.
Table 5. Evaluation of OHS data system in Ireland.
The evaluation of the non-fatal injuries datasets highlights variance in the number and type of variables collected by the HSA, CSO, and Teagasc. Each dataset is updated annually, and summarises socio-demographic characteristics (Age, Gender), Work arrangements (Self-employed, Employed), Worksite (Farm), Cause/Trigger (type of injury), and Impact (>4 workdays lost). The HSA and Teagasc datasets contain additional categories of data relating to type of farm enterprise where the injury occurred and employment status of victims. The CSO and Teagasc provide additional details on the number of workdays lost whilst the CSO is the only dataset that captures occupational illnesses. Teagasc is the only organisation that collects variables describing the type of medical treatment required and provides six categories of worksite. The HSA is the only organisation that collects a variable identifying if the victim is not Irish. None of the organisations collect or report data estimating the long-term social or economic costs associated with fatal or non-fatal injuries.
Discussion and conclusions
Despite the importance of understanding OHS data systems, few papers provide overviews of their structure and/or content.Citation16 This paper has sought to present a simple framework for assessing OHS data systems that will be of use to researchers internationally. We applied this approach to assess the Irish OHS data system by undertaking a data mapping exercise. This involved the identification of the institutions or organisations responsible for collecting and curating the data, their remit, and, associated with this, the case definitions applied to determine which injuries and fatalities are counted. We then provided an overview of the variables collected as a means of highlighting key differences in the data systems associated with estimating non-fatal farm injuries. The approach provides valuable insights into the strengths and weaknesses of a critical element of OHS surveillance systems, namely the production of datasets. This knowledge is critical for researchers as understanding the data that informs their research is fundamental to good science. As important, it is critical for policy-makers and other stakeholders to understand the strengths and weaknesses on which OHS policy, strategies, or education and training interventions are developed.
The assessment presented above highlights that whilst the farm fatalities dataset provides complete coverage of all fatalities, regardless of age or employment status, the same is not true of the non-fatal injuries datasets. Each of these is fundamentally different. With respect to the HSA dataset, it covers farm workers and non-workers. The data are detailed, and they provide rich insights into the triggers of injuries and their impacts. Counterbalancing this, the underreporting of injuries to the HSA means that the total number of cases reported annually is small. The CSO also collects data from employed and self-employed workers, and the number of surveys collected from farmers is estimated to be small, i.e. they must be a small fraction of the total sample of the 391 responses covering all economic sectors in 2021. Implementing the survey through the LFS means that the population covered is limited to those whose main occupation is farming or agriculture work. Unlike the HSA dataset, however, the CSO considers minor injuries and, importantly, occupational illnesses. The Teagasc injuries survey is explicitly targeted at farmers, implemented through an established and accepted data collection system, and covers most types of farm systems. It is also part of a wider survey meaning that there are a large number of social, economic, environmental, and farm performance variables that can be used in analyses to identify risk factors.Citation15 The NFS does not, however, cover smaller farms, i.e. those with less than €8,000 Standard Output. A further point that should be considered in the interpretation results generated from this type of data system is that it is not representative of farmers, rather it is a nationally representative sample of farms by farm size and farm system.
It is evident that whilst the primary shortcoming associated with each of the non-fatal injuries datasets is the relatively low number of reports or responses to surveys. To complement these datasets and overcome this issue and extend coverage of the data surveillance system to the entire farm population, it is recommended that the potential of additional non-fatal injuries datasets are assessed, particularly those maintained by primary healthcare providers, to establish if they can provide a more comprehensive understanding of the prevalence, causes, and impact of injuries. Finally, we call on researchers concerned with farmer health and safety to undertake similar assessments of the data systems in their countries or regions. Evaluations of OHS data systems are of critical importance in promoting learning and the identification of good practice in the design of data systems to support effective OHS surveillance. The evaluation system developed for this study was adapted from the NASEM. Whilst this is suitable for Ireland, it may not be appropriate for all jurisdictions, and researchers should consider adapting this model to their context. We strongly recommend that those applying this model or undertaking similar work pay careful consideration to which cases are included, and who is excluded from OHS data systems.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Funding
References
- Angioloni S, Jack C. Farm fatalities in Northern Ireland agriculture: what fifty years of data tell us. Econ Hum Biol. 2022;46:101122. doi:10.1016/j.ehb.2022.101122.
- Lower T, Peachey KL, Rolfe M. Farm-related injury deaths in Australia (2001-20). Aust J Rural Health. 2023;31(1):52–60. doi:10.1111/ajr.12906.
- Meredith D, Mohammadrezaei M, McNamara J, O’Hora D. Towards a better understanding of farm fatalities: identification and estimation of farming fatality rates. J Agromedicine. 2023;28(2):239–253. doi:10.1080/1059924X.2022.2113196.
- Merisalu E, Leppälä J, Jakob M, Rautiainen R. Variation in Eurostat and national statistics of accidents in agriculture. Agron Res. 2019;17(5):1967–1983. doi:10.15159/AR.19.190.
- National Academies of Sciences E, Medicine. A Smarter National Surveillance System for Occupational Safety and Health in the 21st Century: Washington, DC: National Academies Press. 2018. doi:10.17226/24835.
- Weichelt BP, Gorucu S, Burke RR, Salzwedel MA, Murphy DJ, Lee BC. Agricultural youth injuries: an updated review of cases from U.S. news media reports, 2016-2021. Front Public Health. 2022;10:1045858. doi:10.3389/fpubh.2022.1045858.
- Scott EE, Krupa NL, Sorensen J, Jenkins PL. Electronic merger of large health care data sets: cautionary notes from a study of agricultural morbidity in New York state. J Agromedicine. 2013;18(4):334–339. doi:10.1080/1059924X.2013.826608.
- Rautiainen RH, Ledolter J, Donham KJ, Ohsfeldt RL, Zwerling C. Risk factors for serious injury in Finnish agriculture. Am J Ind Med. 2009;52(5):419–428. doi:10.1002/ajim.20688.
- Weichelt B, Scott E, Burke R, et al. What about the rest of them? Fatal injuries related to production agriculture not captured by the Bureau of Labor Statistics (BLS) census of fatal occupational injuries (CFOI). J Agromedicine. 2022;27(1):35–40. doi:10.1080/1059924X.2021.1956663.
- Patel K, Watanabe-Galloway S, Gofin R, Haynatzki G, Rautiainen R. Non-fatal agricultural injury surveillance in the United States: a review of national-level survey-based systems. Am J Ind Med. 2017;60(7):599–620. doi:10.1002/ajim.22720.
- Health and Safety Authority (HSA). A Review of Work-Related Fatalities in Agriculture in Ireland 2011-2020. Dublin, Ireland: Health and Safety Authority (HSA). 2021.
- Health and Safety Authority (HSA). Annual Review of Workplace Injuries, Illnesses and Fatalities 2020–2021. Dublin, Ireland: Health and Safety Authority (HSA). 2022.
- Dillon E, Donnellan T, Moran B, Lennon J. Teagasc National Farm Survey 2022. Ireland: Athenry; 2023.
- Central Statisitics Office (CSO). Census of Agriculture 2020 Detailed Results. Dublin, Ireland: Central Statisitics Office (CSO). 2021.
- McNamara J, Kinsella A, Osborne A, Blake C, Meredith D, Kinsella J. Identifying farmer workplace injury risk factors in Ireland using farm accounts data. J Agromedicine. 2021;26(4):411–419. doi:10.1080/1059924X.2020.1837704.
- Yang L, Branscum A, Kincl L. Understanding occupational safety and health surveillance: expert consensus on components, attributes and example measures for an evaluation framework. BMC Public Health. 2022 Mar 14;22(1):498. doi:10.1186/s12889-022-12895-6.