A Narrative Review of Fatigue in Agriculture and Its Impact on Injury and Fatality in Australia

ABSTRACT

Objectives

To identify current knowledge about the role of fatigue in occupational injury in the agricultural sector and briefly assess potential intervention approaches.

Methods

Narrative review of peer reviewed literature (in English) from 2010 to 2022 focusing on fatigue in agricultural and other sectors. Data were extracted from Medline, Scopus, and Google Scholar.

Results

The initial search revealed 6,031 papers, of these 33 met the inclusion criteria. The literature unanimously agreed that fatigue contributes to occupational injury in agriculture and related industries. However, there was a scarcity of literature specific to Australia or agriculture. This limits the ability to draw conclusions about the true relationship between fatigue and injury.

Conclusion

While fatigue is likely a major contributor to occupational injury in Australian agriculture, the limited literature impedes the ability to draw conclusions and extrapolate interventions used by other industries that are practical and feasible, to agriculture. Future studies should establish the nature of the problem in Australian agriculture and consult with members of the sector to establish the best interventions to ameliorate the problem, then implement and robustly evaluate intervention studies.

KEYWORDS:

• Agriculture
• death
• injury
• sleep

Introduction

Agriculture is a major contributor to the Australian economy and underpins the social and economic structures of rural and regional Australia. Agricultural products were responsible for 12% of goods and services exported in 2020–21. This accounted for 1.9% of the total national Gross Domestic Product (GDP) and 2.5% of all employment.¹ Around 70% of all agricultural output is exported, with beef ranked (7th), meat excluding beef (14th) and wheat (16th) in relation to the total value of export incomes from all sources.²

Internationally, the agriculture sector is recognised as a high-risk industry.^3–5 This is also reflected in the Australian data with over 1,550 fatal non-intentional injury-related incidents on Australian farms between 2001 and 2020, with no change in the rates of death when assessed by hours worked.⁶ Furthermore, when compared to other industry sectors in Australia over the 2016–20 period, agriculture has the unenviable record of having the second highest occupational fatality rate (12.2/100,000 workers), behind road transport (12.4/100,000).⁷ By way of comparison, the construction and mining sectors recorded rates of 2.2 and 2.1/100,000, respectively. Additionally, the agricultural fatality rate (12.2), is almost 8.5 times higher than the all-industries average (1.45/100,00).⁷

A review of Australian workers compensation data over the 2014–2019 period examined non-fatal injury in the agricultural sector.⁸ It revealed that 70% of the claims involved body stressing (i.e. lifting, carrying, moving objects), being hit by a moving object (e.g. an animal or vehicle), or falls, trips, and slips. Furthermore, fatigue and its relationship to injury were identified as an issue by stakeholders in this study. However, the precise nature of this relationship could not be ascertained based on existing data, as fatigue is not captured in workers compensation data.⁸

Notwithstanding these data limitations, the relationship between fatigue and injury is well correlated across multiple industry sectors internationally.^9–11 Within agriculture, putative risk factors include the physically demanding nature of the work, use of a vast array of machinery and technology, the seasonal nature of some sectors (e.g. high work intensity sowing/harvest periods), plus overall chronic time demands (e.g. in dairying), and the all-encompassing nature of the job demanding a strong work ethic that sometimes portrays sleep as a weakness leading to further fatigue.^2,12,13 Added to this are factors outside the direct control of farmers such as the fluctuating economics of agricultural commodity prices, weather conditions, and natural phenomena that can unpredictably dictate many functions of a workday.¹² More recently, the COVID-19 epidemic has also exacerbated issues such as access to temporary staff (e.g. agricultural visa holders or back-packers), resulting in major labour shortages within the agricultural sector.¹⁴

Fatigue is a multi-factorial and multi-dimensional concept; within the literature the definition of fatigue varies based on context and required complexity.¹⁵ At its broadest definition fatigue is defined as “a biological drive for recuperative rest.”¹⁶ Most literature recognises both mental and physical causes and consequences to fatigue.^17,18 Fatigue can be defined as both a consequence of mental and/or physical exertion and a cause of mental and/or physical exertion. For the purpose of this paper, fatigue has been defined in accordance with the statements of Safe Work Australia “Fatigue is more than feeling tired and drowsy. It is mental or physical exhaustion that reduces your ability to work safely and effectively. It can also have serious short and long-term health effects on workers.”¹⁹

Most literature acknowledges a time awake and time active component to fatigue;^16,18,20 that is to say, fatigue is more likely to occur with more time since waking, or since the last rest break and with longer working hours. This concept is intimately linked to other expanded definitions of fatigue that include sleep quantity, quality, and homeostatic factors.²¹

Consequently, the aim of this narrative review is to establish what is known about the role of fatigue in occupational injury in the Australian agricultural sector. In addition, the review will assess contemporaneous data from other industry sectors, from which lessons to reduce fatigue-related injuries may be learned – such as forestry, maritime, oil, and mining. This information may then assist in providing a baseline to underpin future research and interventions in ameliorating the contributions of fatigue to injury.

Methods

This review was conducted with the intention of answering five research questions:

1. How does fatigue impact on injury in a work context?

2. What information is available on the impact of fatigue in relation to injury in western agricultural production systems (e.g. Australia, Canada, Europe, New Zealand, and USA)?

3. What steps have been taken to manage fatigue in the agriculture sector within Australia and internationally?

4. How have other industry sectors such as fishing, gas, mining, oil, forestry, and transport addressed fatigue management (what is considered best practice)?

5. Drawing on the information collated in questions 1–4, what steps are recommended for future study and/or implementation to manage fatigue in the context of the Australian agriculture sector.

The data used in this study are publicly available and used in accordance with the data custodians requirements, ethics approval was not required.

Three separate searches were completed to answer the five research questions as outlined above in a narrative style review. Medline, Scopus, and Google Scholar were employed to search for papers from 2010 to March 31, 2022. Papers were excluded if not written in English, were not peer reviewed, or if they were irrelevant to western agricultural production systems. The primary outcome focus of papers was required to centre on fatigue and/or specifically for the agricultural sector, and on antecedents that may be causative (e.g. sleep duration and quality).

The first search used the keywords “fatigue,” “injury,” “occupational injury” including all combinations of words. The second search employed the keywords “fatigue,” “agricult*,” “farm*,” “injury,” including combinations of all words. The third search included “fatigue,” “manage*,” “best practice,” and keywords related to relevant industries including “oil,” “gas,” “mining,” and “forestry.” Titles and abstracts were assessed for relevance and exclusion criteria by the principal author, full-text articles that were included were then read and further assessed for relevance and exclusion criteria by at least two authors. Additionally, bibliographies were interrogated for relevant literature complying with the inclusion and exclusion criteria.

Results

The results from the search are outlined in Figure 1. The initial three searches revealed 6,037 studies, which after consideration of inclusion and exclusion criteria were reduced to 33 papers. Of the papers included in the final review, 10 were directly related to agricultural production, with the remainder (n = 23), reporting on other industry sectors. Summary results are outlined in Table 1 (n = 10) and Table 2 (n = 23), reflecting those specific to agriculture and examining alternate sectors, respectively.

Figure 1. Study flow chart.

Table 1. Summary features of studies assessing fatigue in the agriculture sector from Jan 1, 2010 to March 31, 2022.

Study	Location	Study design	Industry	Fatigue measurement or definition	Methods	Major findings	Strengths & limitations
Norasi et al. 2022^Citation22	USA	Cross sectional	Agriculture	Range of tests for lab (9) and field (4) assessments of heavy vehicle operation i.e. tractor over a 2 hr period. Included subjective, performance and physiological measures.	In the laboratory study, participants (n = 6) maintained a seated posture for a period of 2 hours and experienced a controlled seat motion profile consistent with that of farm machinery working in a moderately rough field. Field study participants (n = 6) experienced an 8–10 h ground preparation (heavy tillage with chisel plow) workday.	It is plausible to have a battery of tests to identify and quantify fatigue during operation of machinery.	Strengths Specific to agriculture Validated quantitative and subjective data measures Limitations Small sample size Cross-sectional study design, therefore no causal conclusions.
Ramos et al. 2021^Citation23	USA	Cross sectional	Agriculture	Iowa Fatigue Scale and Need for Recovery Scale.	A total of 228 interviews were conducted with participants.	Lower self-reported health, experiencing physical pain, not handling animals and decreased decision latitude were directly related to increased physical fatigue. Shorter tenure working on cattle feedyards, lower educational level, experiencing physical pain, and increased job demands were directly related to heightened mental fatigue. Being female, experiencing physical pain, an elevated average of hours worked per day, increased job demands, and less decision latitude were directly related to an increased need for recovery and indirectly related to both physical and mental fatigue.	Strengths Specific to agriculture Limitations Response bias Cross-sectional study design, therefore no causal conclusions.
Tone & Irwin 2021^Citation24	Ireland	Qualitative	Agriculture	Identified situational awareness (SA) lapses	Interviews with 15 subjects.	General SA failures and lapses at Level 1 (perception) and Level 2 (comprehension), were present in all incidents reported. The data also indicated the potential negative impact of stress and fatigue on SA, both generally and in the context of critical incidents.	Strengths Specific to agriculture Limitations Small sample size Response bias Cross-sectional study design, therefore no causal conclusions.
Ramos et al. 2020^Citation25	USA	Cross sectional	Agriculture	Iowa Fatigue Scale	Face-to-face interviews utilizing a 208 item questionnaire with 241 participants.	Being female, pain, hours of sleep, and job demands were significant predictors of fatigue. Fatigue and job-related injury were the only significant predictors of pain.	Strengths Specific to agriculture Limitations Response bias Cross-sectional study design, therefore no causal conclusions.
King et. al. 2016^Citation26	Canada	Cross sectional	Agriculture	Self-reported hours of sleep	Using data from the phase 2 baseline survey of the Saskatchewan farm injury cohort, multivariable logistic regression was used to examine associations between self-reported BMI and injury.	Among women there was an increased odds of injury associated with being overweight, the same was not true for men. It is speculated that being overweight and obese can result in higher rates of sleep apnoea and increase fatigue.	Strengths Specific to agriculture Large sample size Limitations Response bias Cross sectional design, therefore no causal conclusions
Dosman et al. 2015^Citation27	Canada	Cross sectional	Agriculture	N/A – self-reported farm injury	A self-reported mail survey of 1,196 farms. Survey questions were used as proxy measures of the farm owners practices relevant to steps of the hierarchy of risk controls (HOC). This included hazard identification, risk assessment, personal protection, administrative controls, engineering controls and elimination of hazards.	When 4 of the 6 steps of the HOC steps were adhered to there was a significant protective effect (OR: 0.32).	Strengths Specific to agriculture Large sample size Limitations Response bias No causal conclusions Proxy measures for HOC implementation
Clay et. al. 2015^Citation28	NZ	Qualitative	Agriculture	Subjective, self-reports of fatigue.	Qualitative content analysis of 216 face to face interviews with farmers regarding risk taking, unrealistic optimism and fatalism as risk factors for quad bike loss of control events.	Risk taking elements were more prevalent in conversation than unrealistic optimism or fatalism. Additionally, fatigue, stress, multitasking, inexperience and bike faults, were identified as likely contributors.	Strengths Specific to agriculture and quad bike incidents Large sample size Limitations Response bias Qualitative study design, therefore no causal conclusions
King et. al. 2014^Citation29	Canada	Cross- sectional	Agriculture	Epworth Sleepiness Scale score and self-reported sleep apnoea.	A questionnaire was completed by 2,849 individuals living on farms regarding self-reported sleep, demographics, farm injuries, and general health. Logistic regression was used to quantify relationships between daytime sleepiness, injury, and health.	After adjusting for confounders, individuals with excessive daytime sleepiness were more likely to experience farming related injury and poor physical health.	Strengths Specific to agriculture Large sample size Use of a validated measure Limitations Response bias Large potential for confounders No causal conclusions
Dosman et. al. 2013^Citation30	Canada	Prospective cohort	Agriculture	Physician diagnosed sleep disorder	5,502 representatives from 2,390 farms completed baseline questionnaires regarding demographic information, sleep apnoea diagnosis, snoring and agricultural injury. Survival analyses were used to relate sleep patterns to subsequent injury.	Loud snoring was only associated with a consistent increase in risk for farm injury outcomes. Diagnosis of sleep apnoea was not associated with injury, presumably because diagnosed individuals received treatment.	Strengths Specific to agriculture Large sample size Limitations Response bias Large potential for confounders No causal conclusions.
Lilley et. al. 2012^Citation20	Canada	Cross-sectional	Agriculture	Self-reported hours of sleep	Self-reported survey data (n = 2,390 farms) assessing work related injury in the last year, hours of sleep, working hours, alcohol consumption and worry. Incidence of injury was grouped according to season of occurrence. Logistic regression models identified associations between sleep quantity and farm injury.	Peak production season was not associated with increased odds of injury. <5 hours of sleep per night in non-peak season has 2.4× the odds of injury compared to those sleeping >7 hours. Daily economic worry, >60 hours of work per week and daily alcohol consumption, was associated with <5 hours of sleep.	Strengths Specific to agriculture Large sample size Limitations Response bias Cross-sectional study design, therefore no causal conclusions.

Table 2. Summary features of studies assessing fatigue in non-agricultural sectors from Jan 1, 2010 to March 31, 2022.

Study	Location	Study design	Industry	Fatigue measurement or definition	Methods	Major findings	Strengths & limitations
Alroomi & Mohamed 2022^Citation17	Kuwait	Cross-sectional	Offshore oil and gas	Self-reported 14 item questionnaire. Items were derived from the fatigue assessment scale, fatigue questionnaire, and the multidimensional fatigue inventory.	Data from 387 responses were used for regression analyses testing the hypothesised relationship between physical isolation, fatigue and safety behaviour.	Loneliness impacts mental fatigue which contributes to reduced compliance and participation in safety behaviour. When mental fatigue was added to the regression analysis, loneliness and mental fatigue accounted for 10.4% of safety participation.	Limitations Response bias Cross sectional design, therefore no causal conclusions Large potential for confounders
Hinze et al. 2021^Citation18	New Zealand	Retrospective cohort study	Forestry	Fatigue was implied by time since last meal break.	Data from 6,584 forestry incident reports over 8 years assessing time of day (indicating fatigue contribution) and incident distribution.	70% of incidents were caused by worker failure (human error) and 78% of incidents peaked at 10:00 and 14:00 hours, indicating fatigue contributions.	Strengths Large sample size Limitations Response bias Cohort study design, therefore no causal conclusions Inconsistencies in data quality e.g. 89% of data was missing location of injury. Assumptions of fatigue.
Ghasemi et al. 2021^Citation31	Iran	Cross-sectional	Fire fighting	The multidimensional fatigue inventory	Online questionnaires were completed by 310 firefighters assessing demographic information, fatigue, safety behaviour and incidents. The data was used to construct 2 regression models looking at the relationship between fatigue and safety climate & behaviour.	Fatigue was found to have a significant negative effect on safety behaviour i.e. the more fatigued a person is, the lower the perceived safety climate.	Limitations Response bias Small sample size Cross sectional study design, therefore no causal conclusions
Drews et al. 2020^Citation32	USA	Qualitative	Mining	Subjective reports	Eight focus groups of 44 mining operators were employed to discuss their experience of operator fatigue, issues surrounding fatigue in the workplace and current best practices to reduce fatigue.	Participants reported that fatigue causes altered perception, inability to pay attention, memory impairments, poor higher executive function & decision making, causes grumpy or irritable moods, slows reaction time, and contributes to imprecise vehicle control. With recognition of distal factors including diet, overtime hours, commute time, sleep, stress, physical fitness, and illness.	Limitations Response bias Qualitative study design; therefore, no causal conclusions
Friedman et al. 2019^Citation9	USA	Retrospective cohort	Mining	Working hours (>9 hours)	US Mine Safety and Health Administration reports from 1983 to 2015 were used to identify injuries associated with long working hours, defined as an incident occurring 9 or more hours after the start of the shift.	The average annual incidence rate per 100 FTE for injuries occurring >9 hours into a shift was 6.54 compared to 5.49 for injuries in the first 8 hours of a shift.	Strengths Large sample size Limitations Response bias Cohort study design; therefore, no causal conclusions Inconsistencies in data
Bowen et al. 2019^Citation33	New Zealand	Cross sectional	Forestry	Cognitive response times (Deary-Liewald reaction time test)	This study captured physiological data (n = 15) from a fitbit to assess level of activity, reaction times using the deary liewald reaction time test as a measure of fatigue, and ambient temperature. As well as a questionnaire regarding sleep quality to identify correlations between levels of physical activity and fatigue and differences based on roles with high physical load and high cognitive load.	Results revealed there was no defined pattern across each working day and no clear indication that higher physical activity was related to higher levels of fatigue among forestry workers. Authors were unable to find significant fatigue based impairments	Limitations Response bias Small sample size Cohort study design, therefore no causal conclusions Unable to collect data at peak incident times due to restraints of the job.
Shortz et. al. 2019^Citation34	USA	Tool developed	Oil & gas	Multiple	This study developed a Fatigue Risk Assessment and Management in high-risk environments (FRAME) survey using systematic gathering of existing fatigue scales, refining the survey using the Delphi Consensus technique, and further refinement through employee focus groups.	The FRAME survey has 26-items that assesses 4 dimensions of fatigue including sleep, shift work, physical, and mental fatigue.	Limitations Specific high-risk environments Validity & reliability not assessed
Bauerle et al. 2018^Citation35	USA	Scoping review	Mining	Various	A scoping review examining what is known about fatigue, injury and interventions to reduce worker fatigue.	Hours of service (HoS) limitations (including a limit of 14 per 24 hours, 60 hours per week, and 48 hours of weekly hours per year). And enforced breaks (including 1 continuous 24 hour period of rest per week and 30 minute breaks per 5.5. hours) were observed as strategies for fatigue prevention in mining. However, additional factors such as leadership, safety culture, reporting systems, work/life balance, commuting time, opportunities for recovery should be considered as contributors and opportunities for intervention.	Limitations Limitations associated with individual studies.
Jones & James 2018^Citation36	Australia	Case control pilot study	Mining	The Need for Recovery after Work Scale	This study assesses the feasibility, benefits and challenges of implementing a task rotation schedule in a coal mine over 6 months with 2 work crews. There was rotation of 5 main tasks, 3 times per shift (9 hours). Questionnaires were completed at 3 and 6 months and a focus group was conducted at the completion of the study.	Significant improvements across the psychological & environmental domains, fatigue, & musculoskeletal pain scores. The participants reported feeling “mentally fresher”, “didn’t do the same task twice in a row”, and “had more variety making the shift go faster”. Task rotation was also feasible, regarded positively and practical.	Limitations Response bias Small sample size Limited amount of time
Yung et al. 2017^Citation37	Canada	Cohort	Multiple industries	Range of quantified and reliable fatigue measures consisting of biomechanical, physiological, cognitive and sensorimotor measurements.	Participants (n = 16) performed four conditions on four separate days: (1) physically demanding work-Whole Body Vibration (WBV) -physically demanding work, (2) mentally demanding work -WBV- mentally demanding work, (3) concurrent physical and mental work-WBV-concurrent physical and mental work, and (4) control quiet sitting-WBV-control quiet sitting.	Significant time effects during the control condition were identified in the physical (11), mental (7), and concurrent conditions (12). WBV combined with different demands do not have simply additive effects on fatigue, but are often synergistic or antagonistic. Arranging and scheduling work duties appears to be contingent on the inter-dependent demands of the tasks.	Strengths Use of a range of quantitative validated measures Limitations Small sample size
Rosecrance et al. 2017^Citation38	USA	Qualitative	Forestry	Subjective, self-reported discussion of fatigue.	A focus group of 16 forestry workers for 90 minutes was conducted and moderated by investigators. Injury data from workers compensation claims was shared with the group and used as the basis for focus group questions regarding contributors to injury in forestry. Trends and perceptions of participants were analysed.	Working on steep slopes, below freezing temperatures and physical fatigue from strenuous work for 8–10 hours, were identified by workers as the main contributors to occupational injury.	Limitations Response bias Small sample size Bias associated with workers compensation data Qualitative study design, no causal conclusions
Khoshakhlagh et al. 2017^Citation39	Iran	Cross-sectional	Mining	Piper Chronic Fatigue Scale	Data concerning demographics, fatigue and occupational injuries over 3 years collected from 453 male mining workers. The data was interrogated for associations between fatigue and occupational injury.	The mean work-related fatigue score for subjects with occupational injury was 3.75 ± 0.98 and 2.86 ± 0.7 for those without history of occupational trauma. Employees reported the main cause of work-related fatigue was high amounts of strenuous work & inappropriate workplace conditions.	Limitations Response bias Cross sectional data, therefore no causal conclusions
Zhang et al. 2015^Citation40	USA	Cross-sectional	Construction	Self-reported reports of tiredness	Regression models were built using data from construction workers who completed the 2010–2011 national health interview survey to predict physical and cognitive difficulties associated with self-reported fatigue while controlling for confounders.	Chi-squared test results showed significant differences in reported physical and cognitive functions between workers who reported tiredness or exhaustion and those who did not.	Strengths Large sample size Limitations Response bias Cross sectional study design, therefore no causal conclusions Convenience sampling
Lombardi et al. 2014^Citation41	China	Cross-sectional	Multiple industries	Self-reports of quality of sleep & tiredness	703 hospitalised workers were interviewed within 4 days of their injury to assess their rest breaks, age, gender, work hours, work start time & duration, injury day & time, duration & quality of last sleep, and alertness/sleepiness. Cox regression was used to compare time into work shift and nature of rest breaks.	After adjusting for significant covariates, participants with rest breaks of 1–30, 31–60, and >60 minutes were able to work longer into their shift without injury when compared to those who had no rest breaks.	Strengths Large sample size Limitations Response bias Cross sectional study design, therefore no causal conclusions
Eiter et al. 2014^Citation42	USA	Case Study	Mining	Self-reported fatigue	A case study of a small mining organisation is presented after the implementation of a fatigue risk management system following a fatality in the mine. Mine safety personnel were interviewed to discuss the changes made.	The fatigue risk management system included changes to work environment (e.g. road widths & maintained berms), prioritisation of safe work procedures, formation of a safety committee, reduced work hours and mandated rest breaks and vacation leave, encouragement of stretching & exercise. Participants highlighted that the implementation was not without monetary and productivity cost, however reported the cost was small compared to loss of life.	Limitations Case study design, therefore no causal conclusions
Mooren et al. 2014^Citation43	Australia	Literature review	Heavy vehicle transport	Not specified	Literature review was conducted to assess evidence of effective safety management characteristics that reduce crashes & injury in heavy vehicle transport.	The characteristics associated with reduced crash & injury risk included safety training, management commitment, scheduling or journey planning, worker participation in OHS, incentives, and safety or return to work policies	Limitations Limitations associated with individual studies
Sneddon et al. 2013^Citation44	UK	Cross-sectional	Offshore oil drilling	A 13-item scale developed by the Australian Maritime Safety Authority (AMSA)	A self-reported scale measuring work situational awareness (WSA) and its relationship with stress and fatigue was completed by 185 offshore oil drillers to test relationships between the variables using Pearson’s product moment & regression analyses.	High stress levels and fatigue are linked to lower levels of WSA. Which in turn, indicates increased participation in unsafe work behaviours and higher accident risk.	Limitations Response bias Small sample size Cross sectional design, therefore no causal conclusions
McGuinness et al. 2013^Citation45	Norway	Retrospective cohort study	Fishing	Fatigue implied by hours worked	Statistical analysis of reported occupational injuries from the Norwegian fishing fleet from 2000 to 2011, assessing important characteristics and traits of injuries.	The main cause of incidents accidents were ship motion & conditions on the ship (slippery floors etc). However, it was determined that fatigue from excessive hours & unregulated hours contributed. This was determined by injury time peaks at 10:00–13:00 and 14:00–17:00.	Limitations Response bias Cohort study design, therefore no causal conclusions Assumptions of fatigue
Arlinghaus et al. 2012^Citation21	USA	Cross-sectional	Multiple industries	Self-reported weekly hours worked and sleep duration.	Data from 89,366 employed workers taken from the national health interview survey was used to develop a structural equation model which examined fatigue related direct and indirect potential risk factors for occupational injury. Multiple confounders were simultaneously examined.	Longer working hours and short sleep duration independently increased risk of occupational injury. Risk of injury escalated by 14% with every 10-hour increase in weekly working hours. For every one hour decrease in sleep, risk of injury increased by 10%.	Strengths Large sample size Limitations Response bias Cross sectional design, therefore no causal conclusions Multiple industries unrelated to agriculture
Gander et al. 2011^Citation46	Multiple	Literature & Policy review	Trucking, rail, aviation, maritime	Not specified	A literature review was conducted to assess work related fatigue and strategies that minimise this safety hazard in the transport sector.	Revealed three levels of fatigue risk management 1. Regulatory responsibility 2. Industry responsibility and 3. Individual responsibility. Themes across industries included hours of service limitations, fatigue risk management systems (FRMS) including procedures for reporting, investigating, recording fatigue related incidents, education, and safety culture.	Strengths Comprehensive report of one industry
Williamson et al. 2011^Citation16	Australia USA France UK NZ	Literature review	Multiple industries	Fatigue is defined as “a biological drive for recuperative rest”	Literature review.	Impaired sleep homeostasis including sleep deprivation & time since waking, causes changes in cortical functioning leading to fatigue and to reduced performance with safety impairment. Task related factors, particularly sustained or continuous attention & monotonous tasks, unmask underlying sleepiness and cause fatigue.	Limitations Limitations associated with individual studies.
Chan et al. 2011^Citation47	China	Qualitative & cross sectional	Oil & Gas	Self-reported	320 workers from 4 oil & gas construction projects participated in a survey questionnaire & interviews were conducted with workers who had experienced actual fatigue related incidents.	All those involved ranked fatigue as the most critical risk to incidents.	Limitations Response bias Cross sectional/qualitative study design, therefore no causal conclusions
Kongsvik et al. 2011^Citation48	Norway	Cohort study	Maritime	Self-reports of fatigue & sleep/rest	2 maritime crews (n = 43) were allocated to different watch keeping regimes over 4 weeks to assess sleep, fatigue, safety perceptions, incidents & near misses. One included 6 hours on duty and 6 hours rest (6–6 regime). The other included 8 hours duty, 8 hours rest, 4 hours duty, 4 hours rest (8-8-4-4 regime). Cognitive & physiological testing, surveys & sleep diaries, and accelerometers/actigraphy was used to assess cognitive & physical performance, objective & subjective sleep quality, and demographic information.	The 8-8-4-4 regime had significantly better sleep efficiency and more continuous sleep, with less sleep episodes during shift. However, there were no reported differences between fatigue or safety parameters.	Limitations Response bias Small sample size Cohort study design, therefore no causal conclusions

In relation to the first and second study questions, the literature supports the proposition that fatigue contributes to injury occurrence in a work context. Ten of the identified studies were conducted solely on agriculture, none of which were Australian.^20,22–30

Fatigue increased the risk of injury in several studies.^{9,18,20,21,28,29,38–40,44,45} A cross-sectional survey of 89,366 US workers found that long working hours and short sleep duration both independently increased the risk of occupational injury; with each 10-h increase in weekly working hours and each 1-h decrease in sleep time, there was an increase in injury risk by 14% and 10%, respectively.²¹ An analysis of 8 years, including 6,584 reports of forestry incident data in New Zealand concluded that 70% of incidents were caused by worker failures and 78% of worker failures showed indications of fatigue. There was a bimodal distribution of the time of incidents, with peaks at 1000 and 1400 h. The authors speculated that this was indicative of fatigue due to patterns of hours worked and break times. More specifically; 1000 and 1400 h tended to correlate with >4 h since the last rest break.¹⁸ The relationship between fatigue and injury is also reported in qualitative studies with farmers, forestry plus oil and gas workers.^28,38,47

A literature review assessing the link between fatigue and safety suggested that fatigue is a consequence of sleep homeostatic dysregulation and leads to reduced performance and safety impairment.¹⁶ This was reinforced by an offshore oil drilling study that found stress, sleep disruption, and fatigue were all associated with lower work safety awareness and increases in unsafe behaviour.⁴⁴ More recently, it was identified that isolation and loneliness contribute to fatigue generation and both physical and mental fatigues are negatively correlated with safety compliance and participation in oil and gas workers.¹⁷

With specific regard to agricultural studies (research question 2), there were two that quantified an increased risk of injury from fatigue, both emanating from the Saskatchewan Farm Study.²⁰. The cross-sectional survey involved 2,390 Saskatchewan farms and found that <5 h of sleep increased the injury odds by 2.42 (95% CI 1.04–5.59) compared to >7 h of sleep.²⁰ This study also established that those participants reporting economic worry, working more than 60 h per week and daily alcohol intake, were more likely to report less than 5 h of sleep and furthermore that alcohol significantly increased the rate of injury.²⁰ In assessing differences between peak and non-peak farming season and injury rates, no difference between seasons was found.²⁰ However, once confounding variables were accounted for, a reduced sleep duration in non-peak season was associated with increased injury rates, even though shorter sleep duration was more common in peak production seasons than non-peak production seasons. The authors speculated that this unexpected result may be related to the nature of the task; with non-peak season tasks tending to be more monotonous, less varied, and more hazardous.²⁰ This is in keeping with data from other sectors.¹⁶ King (2014) also found that farmers with excessive daytime sleepiness were more likely to be injured (OR 1.34 [95% CI 0.92 to 1.96]) and report poorer physical health (OR 2.19 [95% CI 1.45 to 3.30]).²⁹ Additionally, loud snoring was associated with higher rates of injury, however physician diagnosed obstructive sleep apnoea was not.³⁰ It was proposed that this may reflect that once apnoea is diagnosed and treatment is received, injury risk returns to baseline. Additionally, despite relatively small sample sizes, recent studies suggest women may be at higher risk from fatigue.^23,25

In relation to the third and fourth study questions, there were only three intervention studies that have directly addressed the issue of fatigue and been systematically evaluated.^36,42,48,49 While none of these was in the agricultural sector, potential ameliorating factors include the implementation of a broad ranging Fatigue Risk Management System (FRMS) consisting of several aspects (e.g. reduced work hours, mandated rest breaks and vacation leave, task rotation, encouragement of stretching & exercise).⁴² A review of multiple industries indicated that HoS limitations were effective for fatigue prevention and provided sufficient opportunity for rest.⁴⁶ This is reinforced by a cohort study comparing different watchkeeping routines on maritime workers' sleep, fatigue, and safety. The cohort working the 8-h work and 8-h rest routine had significantly better sleep efficiency compared to the 6-h work and 6-h rest cohort, however no difference in fatigue or safety was found.⁴⁸ This suggests that both limitations on hours worked and sufficient time for rest are beneficial to the prevention of fatigue and quality of rest. Additionally, a study interviewing 703 hospitalised workers compared the time of injury to work shifts and rest breaks. Workers with rest breaks of >30 min were able to work significantly longer without injury, compared to those without a rest break.⁴¹

Discussion

The literature unanimously agrees that fatigue contributes to work injury across multiple sectors. While there is a relative paucity of agriculturally focused studies (n = 10),^20,22–30 the data supports this contention.

A range of factors were defined as being potentially associated with this increase across sectors, including long working hours and reductions in sleep time,²¹ periods between break times,⁴¹ stress,⁴⁴ sleep homeostasis,¹⁶ and social isolation,¹⁷ among others. More specifically for the agricultural sector, the additional issues highlighted included daytime sleepiness,²⁹ loud snoring,³⁰ less than 5-h sleep, economic worry, working more than 60 h per week and daily alcohol intake.²⁰ There was also a suggestion that women are at increased risk.^23,25

As evidenced in this review, the overwhelming majority of studies (both agricultural and in other sectors), have been descriptive cross-sectional assessments looking at quantifying and explaining the potential causative mechanisms and impacts of fatigue on injury. Further, there are only three studies within the papers assessed that have conducted an intervention study with an accompanying evaluation, none of which involved agriculture.^36,42,48 This suggests significant potential to scale-up more robust intervention studies to address this issue within agriculture.

When other industries have attempted to address fatigue, an approach drawing on a holistic Fatigue Risk Management System (FRMS) has typically been adopted, reflecting the multifactorial nature of fatigue. Several of the primary industries assessed in this review have internationally coherent strategies to address fatigue based on a FRMS that are then underpinned by either aligned national and/or state-based strategies. From an international perspective, this includes agencies such as the International Maritime Organization,⁵⁰ the International Association of Oil and Gas Producers⁵¹ and the International Aluminium Institute.⁵² The forestry sector has a dated Code of Practice (1998),⁵³ but does have current national approaches such as those in Australia,⁵⁴ Canada,⁵⁵ and New Zealand.³³

Conceptually, strategies to address fatigue and injury can be divided into three modifiable factors: factors related to occupation, to the individual and the environment.⁵⁶ Occupation factors include implementation of hours of service (HoS) restrictions, enforced rest periods and vacation time, timing of shift work, and task rotation.

Factors related to individuals include lifestyle modifications such as diet and exercise,^32,56 limited alcohol consumption,^20,56 a sleep space, and habits conducive to sleep.⁵⁶ Management of medical conditions that may cause fatigue and sleepiness, such as sleep apnoea and obesity, have also been highlighted as beneficial for fatigue prevention.^26,28,30

Environmental factors amenable to intervention include provision of conditions that are conducive to a safe workplace and limit the requirements for mental energy to navigate the physical workplace.⁴² Additionally, the implementation and encouragement of a safety culture and FRMS are considered important. A safety culture describes a workplace that approaches safety lapses as opportunities to improve safety in a workplace.⁴⁶ Meanwhile, a fatigue risk management system (FRMS), describes the policy and procedures that reflect a safety culture.⁵⁷ This includes education for workers, methods of measuring fatigue levels and reporting, investigating and recording incidents attributable to fatigue. To facilitate this approach, processes for evaluating fatigue-related incidents, undertaking interventions, and evaluating interventions are required.⁴⁶ A focus on safety culture and the effective implementation of FRMS have been shown to improve safety outcomes among mining,^35,42 transport,⁴³ and agriculture,²⁷ with these strategies being employed where HoS limitations are not desirable or feasible.⁴⁶

While there are numerous guides and fatigue management resource materials available that farmers may avail themselves of through various agricultural departments, commodity sector organisations or occupational health services (including online), we are unaware of any evidence-based programs that have been implemented and formally evaluated within an agricultural context internationally. In the case of Australian agriculture, we also believe there are no such interventions currently in progress. This highlights a significant gap in our knowledge and thereby reduces the potential for effective preventive approaches.

In the Australian context, employers are legally required to eliminate or minimise the risk of fatigue, so far as is reasonably practicable.⁵⁸ In general terms, this could include establishing relevant FRMS within agricultural businesses, as examples this may be inclusive of – rostering, monitoring fatigue signs, job rotation (where feasible), ensuring sufficient sleep periods and use of new fatigue detection technologies as they are developed.

The limitations of this study include the lack of a consistent definition of “fatigue” in the literature, thereby restricting the ability to synthesise conclusions and understand the true nature of the problem. Furthermore, ethical restrictions in regard to studying fatigue and injury limit the possibility for conducting more robust research e.g. controlled studies with fatigue inducement. The studies in this review are exclusively cross-sectional or case study designs, thereby restricting the ability to draw causal conclusions and increasing the potential for confounders such as sleep quality and stress. Furthermore, data for farmers and farm workers have not been disaggregated and while the impact of fatigue and its contribution to subsequent injury in these groups are not fully understood, there may well be important differences between these cohorts. Due to the nature of the studies conducted, much of the data is self-reported, with few studies seeking to use validated sleep and fatigue scales resulting in potential for response bias. As this is not a systematic review, no detailed assessment of the overall quality of the studies has been undertaken (as previously noted, they are predominantly cross-sectional assessments). However, the consistency of the findings does support the generalisability of the results even for the agricultural studies, which were lower in number (n = 10). Finally, limiting the relevance of the conclusions made by this review is the scarcity of the literature in the agriculture sector, with none having been conducted in the Australian agricultural sector. This also brings into question the practicality of interventions used in other industries within agriculture. Interventions such as HoS limitations, log books, enforced rest breaks and vacation leave, may not be desired, practical, feasible, or enforceable in the agricultural sector. Similar issues have also been identified within the USA, where such “best practices” have been highlighted as being incompatible with work in the agricultural sector, necessitating a greater focus on risk mitigation.¹³

Based on what has been revealed in this review and the limitations outlined, future studies should focus on defining fatigue, then assessing the direct empirical impacts that fatigue has on agricultural workers in relation to injury. Due to the complex nature of agricultural work and the unique demands of the industry, consultation with members of the agricultural sector is imperative to assess potential efficacious and practical solutions to the problem of fatigue. This should then be followed by well-designed and robustly evaluated intervention studies.

There is unanimous agreement that fatigue contributes to occupational injury in industries comparable to agriculture, however there is a scarcity of the literature specific to agriculture in general and Australia specifically. This limits the relevance of conclusions that can be drawn and brings into question the practicality of interventions used by other industries for agriculture and potentially a greater focus on risk mitigation.¹³ It is contended that the unique demands, conditions, and culture of the Australian agriculture sector are likely to contribute to fatigue generation and injury. Whilst other sectors have introduced some effective strategies, these may or may not be suitable for agriculture. Therefore, potential solutions require consultation with and enthusiastic adoption from the agricultural sector and suggest that international collaboration among countries that have somewhat similar farming practices, may be feasible to address this issue.

Disclosure statement

There are no issues to disclose.

Additional information

Funding

No funding was received for this study.