Chores at Times of Fatal or Serious Injuries Associated with Tractor Overturns with and without Rollover Protection

This study describes chores when farmers were either fatally or seriously injured and required emergency medical treatment as a result of overturns of tractors with or without rollover protective structures (ROPS). Data from the 2002 Kentucky Farm Tractor Overturn Survey were used for this study. The data were collected by a telephone survey of a population-based random sample of 6063 (7.98%) of Kentucky’s 76,017 farm operators as listed in the Kentucky Agricultural Statistics Service database. Of farm operators interviewed, 551 (9.1%) reported 603 overturns and 5512 (90.9%) reported no overturns in the history of their farm, covering a period from 1925 to February 2002. Only the latest overturn was considered to improve recall accuracy. In addition, since the 1925 to 1959 time period had only 49 (8.1%) of the overturns reported, (14 farmers did not provide the year of most recent overturn); only data from the 1960 to 2002 period (approximately 41 years) were used. After making these adjustments, incidents evaluated included 25 cases (one fatal and four serious nonfatal injuries) that involved ROPS-equipped tractor overturns and 88 cases (24 fatal and 64 serious nonfatal injuries) that involved non-ROPS tractor overturns. Chores at highest risk for tractor overturns were identified for which educational and ROPS retrofit interventions could be emphasized. The highest frequency of overturn-related fatalities and nonfatal injuries were associated with hay harvesting, rotary mowing, and on-farm travel chores. These three chores represented 68.2% of fatal events and 50.0% of permanent and 56.6% of temporary disability overturn incidents. Tragically, in countries such as India and China with emerging mechanization, a large majority of tractors are produced without ROPS that can be expected to result in the same overturn-related epidemic of deaths experienced in highly mechanized countries, despite evidence of the protection provided by ROPS.


Introduction
In 1966, Knapp warned that implements added to the tractor can "radically" change its stability in different and unexpected ways [1].Unstable conditions may lead to tractor overturns that are responsible for thousands of deaths and serious injuries [2].While studies have addressed the overturn hazard related to tractor design, training, environmental conditions, including slope and terrain, and public policies, scant attention has been given to tractor use at the time of an overturn incident with one exception, roadway travel [3][4][5][6].To this day, tractor overturns remain the leading cause of death on farms in the United States and continue to merit attention regarding the dimensions of the hazard, including chores associated with the overturns [7].The history of success of rollover protective structures (ROPS) for reducing overturn-related deaths and serious injuries is well-documented [8].ROPS are known to reduce overturn-related fatalities by up to 98% [9].

The Non-ROPS Tractor Legacy
In 2004, Day et al. reported the success of a program in the state of Victoria in Australia to retrofit tractors with ROPS.Their program offered rebates for ROPS retrofits on tractors prior to a mandate that required that all tractors be equipped with ROPS [10].The program reduced the number of non-ROPS tractors from 17,420 in 1996 by 70% in 1998.With the rebates, the eventual prevalence of ROPS-equipped tractors reached 93% before the mandate became effective [11].In their study, Day et al. found that the most common decade for the manufacture of non-ROPS tractors was 1960 to 1969 [10].
In the United States, tractors dating back to the 1940s lacked a ROPS until 1985, when tractors were manufactured with a ROPS as standard equipment [12].Rules for ROPS on new tractors in eight European countries preceded this year ranging from 1959 in Sweden to 1978 in Spain; New Zealand required ROPS on new tractors in 1970 and six Australian states required them on new tractors in 1982 [8].While many of these countries required older tractors to be retrofitted with ROPS, other countries, such as the United States and Canada, have struggled with reducing tractor-related deaths from overturns regarding pre-1985 tractors that lack ROPS.
In 2010, Murphy et al. reported that the average age of farm tractors in use in the United States was more than 25 years, dating back before 1985 [13].Indeed, tractor longevity is the most critical issue regarding farm-related injuries in the United States and, more particularly, overturn-related deaths [14].From 1992-2007, as older tractors were retired in the United States (and Canada), ROPS-equipped tractors increased from a prevalence of 41% and 34% relative to the total tractor population, for crop and livestock farms, respectively, to 53% and 49%, respectively.The US has not yet reached the ROPS prevalence rate of 75% to 80% proven in Nordic countries to result in zero overturn deaths for farm tractors.However, there is more to a potential prevention strategy than the overall rate: Myers and Hendricks found that crop farmers and farmworkers experienced 1352 overturn-related deaths for the period, 1992 to 2007, at a rate of 7.8 deaths per 100,000 workers, whereas livestock farmers and farmworkers experienced 183 overturn-related deaths over the same period at a rate of 1.27 deaths per 100,000 workers.Thus, crop farming was 7.4-fold and 6.1-fold higher than livestock farming for fatality frequency and rate in 1992 and 2007, respectively.The authors claimed that more research was needed to identify the factors that continue to place crop farmers at a higher risk of overturn-related deaths [2].Moreover, older unprotected tractors are still in use for avocational chores on small acreages [15].Countries, such as India and China, have multiple small, widely-dispersed rural communities of farmers with less mechanized agricultural equipment.As a result many farm tractors lack ROPS and other safety features.This issue is examined in the Discussion section.

Tractors and Chores
Tractors are considered utility vehicles on farms and are used for a variety of tasks.As an example, a 2016 issue of the magazine, Hobby Farms, listed chores for a tractor that demonstrate the range of farm-related tasks for which tractors are currently known to be useful [16].The Hobby Farms article suggests either vintage or new tractors for doing the chores listed in Table 1 along with the categories of chores used in the current paper.Few studies have addressed the task being performed with the exception of travel-related incidents on public roads, as noted earlier.Other studies that examined activities at the time of the overturn depended upon death certificate narratives and were limited by the small number of documented tasks during the overturn.These studies did not address non-fatal injuries associated with tractor overturns and did not use a population-based sample.
In a 1984 study of 1163 overturn-related deaths based upon death certificate narratives, McKnight found that 24 (2.0%) fatalities occurred while operating mowers and 39 (3.4%) deaths occurred when attempting to pull a tree, log, or other vehicle [17].In 12 cases, tractor front loader buckets that were loaded and/or raised contributed to the overturn.
Lehtola collected use data on 173 tractor incidents, 87 of which were fatal, from newspaper reports in Iowa for the years, 1988-1990 [18].She found that the highest number of incidents and fatalities occurred during the summer months of May, June, July, and August when tractors were used for mowing, hay harvesting, planting, cultivating, spraying, and routine chores.Of 90 tractor overturns recorded, 42% occurred on roadways, and 58% occurred in the field or the farmyard.Overturns were responsible for 58.6% of the fatalities, all of which involved tractors not equipped with a ROPS [19].
Lehtola et al. conducted another analysis of 136 tractor-related fatalities in Iowa for the years 1988-1992 based on newspaper reports [20].Of 77 overturn-related deaths in the cohort, 44 (59%) occurred in a field or farmyard, and 28 (37%) occurred on public roads in which no other vehicle was involved.Overturns in the field included carrying loads high with a loader (large round bales or soil), spraying weeds along fence lines, driving along a dead furrow, hay harvesting, and mowing of untilled ground and terraces.In this study, 41% of deaths regarding tractor overturns occurred on roadways, many when mowing ditches, while the remaining chores involved carrying loads high with a loader, spraying weeds along fence rows, herding cattle, driving along the edge of a dead-furrow, performing haying operations, or mowing untilled ground and terraces.
Lehtola and Rautiainen published a tractor safety facilitators' manual that identified chore-specific risk factors that contribute to tractor overturns [21].The risk factors included: roadway travel, rotary mower use, front-end-loader use, round bales hauled in front, and chores usually taking less than an hour each time.
In 1999, Bernhardt and Langley reported a North Carolina study that examined both death certificates and medical examiners' reports [22].They identified 344 deaths related to tractors for the period 1979-1988.Of these deaths, 56% were classified as an overturn event.The activities most closely aligned with an overturn event along with the number of events included the following: plowing, harrowing or discing (28), pulling something (28), driving on or beside a road (25), brush cutting (22), mowing (22), and dragging logs (16).
In 2011, DeGroot et al. used Canadian Agricultural Injury Surveillance Program data for the years 1990-2005 to examine fatal machine-related fatalities in agriculture.They found that an average of 21 deaths per year were associated with tractor overturns and reported the activity related to each death: transportation (40%), field work (19%), forestry (16%), towing (9%), farm yard work (6%), mowing (4%), and other (6%).They also reported contributing factors to the overturn that included dragging logs or implements (7%) and pulling stumps or trees (4%).Transportation included both public road and off-road transport [23].

Aims
We are responding to a need to identify factors that continue to place farmers and farmworkers at risk of tractor overturn-related deaths and serious injuries.We found no population-based study that specifically reported the type of tractor work or chore conducted when farmers were either fatally or seriously injured during overturns of tractors with or without a ROPS.This paper describes the chores involved in overturn-related fatal and nonfatal injuries to operators of tractors fitted with or without a ROPS.

Methods and Materials
This study used data from the 2002 Kentucky Farm Tractor Overturn Survey [24].This unique survey collected data from a large population-based random sample of farmers regarding the chore being performed when a tractor overturn occurred and included documentation of whether or not the tractor had a ROPS.The survey is unique in another way since it includes nonfatal injuries, as well as fatal injuries.Indeed, this survey has been used as a resource in other ways as recently as 2012 in which the relationship of terrain with tractor overturns was determined [25].We also describe below what is meant by a chore and chore classification terms.

The 2002 Kentucky Farm Tractor Overturn Survey
Tractor overturn cases, which resulted in farmers' fatal and nonfatal injuries that required emergency medical treatment (EMT)-defined as treatment by emergency response technicians or in a hospital emergency room-were extracted from data collected by the Kentucky Farm Tractor Overturn Survey conducted by the Kentucky field office for the US Department of Agriculture (USDA) National Agricultural Statistics Service (NASS).The survey data were collected from a 7.98% (n = 6063) population-based random sample of Kentucky farm operators who were interviewed by telephone.A 40-item survey collected information at the county level about each farm's history of overturns that included details about their single most recent tractor overturn event.The response rate was 79% [24].The sample was selected randomly from 76,017 farms statewide documented by USDA in 1997, which was stratified by six agricultural districts in Kentucky and farm size [24].
A total of 551 (9.1%) respondents reported 603 overturns for the period, 1925 to February 2002; whereas 5512 (90.9%) respondents reported no known overturn events in the history of their farm.To reduce respondent recall error, the analysis considered only the most recent overturns on the 551 farms, and since only 49 (8.1%) of the reported overturns occurred during the period, 1925-1959, we considered only those overturns that occurred over the period, 1960 to February 2002 [24,25].
As shown in Table 2, when we reduced the incidents to the most recent overturns reported, the total was 537 overturns in which 92 and 445 overturns occurred on ROPS-equipped and non-ROPS tractors, respectively (not included were 14 overturns for which the ROPS status was unknown).Five respondents did not report the type of injury regarding these incidents, reducing the number to 532 overturns, and to 89 overturns on ROPS-equipped and 443 on non-ROPS tractors.When we truncated the analytic period to the years 1960 to 2002, 49 previous non-ROPS tractor overturns were dropped from the analysis resulting in a study population of 483 overturns for which injury outcomes were known and, thus, reducing the number of non-ROPS incidents to 483 overturns [24,25].
Additionally, as shown in Table 2, we considered only fatal and serious nonfatal injuries.Accordingly, overturns of ROPS-equipped tractors were associated with one fatality and four nonfatal EMT injuries, and overturns of non-ROPS tractors were reported as the cause of 24 fatalities and 64 nonfatal EMT injuries [24].
The survey was conducted in two tiers in which the first tier asked general questions regarding the tractors on the farm and farmstead demographics, which included whether the farm entity or a member of the farm household had experienced an overturn.If so, they were asked a second tier of questions related to only the most recent tractor overturn [24].

Tractor Chore Descriptions
A chore is typically a routine job or task but includes atypical tasks by farmers who use the tractor as a source of power for traction (pulling or pushing loads) and hydraulic or mechanical power for implements.Informed by the chores identified in the introduction and terms used by farmer responses in our study, chores were categorized, first, for the fatal overturns of non-ROPS tractors.This set of chore names then was used to classify the type of chores for nonfatal injuries with a few additions and for fatal and nonfatal injuries associated with overturns of ROPS-equipped tractors.
Tractor chore descriptions were available for 23 of the 24 reported fatal events and from 443 non-ROPS tractor overturns.Chore description and eight other variables were collected for each event and placed into one of 10 non-overlapping chore categories in an Excel TM (Microsoft Corporation, Redmond, WA, USA) spreadsheet.Events that involved more than one possible chore category were classified based on the primary chore being performed.For example, an overturn that occurred while a farmer was "bush hogging" a ditch or on a sloped-bank at the side of a public roadway that bordered farm property was classified as rotary mowing as opposed to roadway travel or hay harvesting.
We also identified 64 nonfatal injury cases from non-ROPS tractor overturns that required EMT.The chore at the time of the overturn and the other data collected for each nonfatal event were entered into a spreadsheet in non-overlapping categories consistent with the procedure used for chore classification of fatal events described above.
From 92 overturns of ROPS-equipped tractors, we identified one fatality.In addition, we identified four serious nonfatal injuries related to overturns of ROPS-equipped tractors, none of which involved a permanent disabling injury.

Chores at Time of Fatal Non-ROPS Tractor Overturn
Descriptions of chores for 23 of the 24 fatal non-ROPS tractor overturns were distributed within the non-overlapping categories shown in Table 4.Of these 24 fatally injured operators, 11 received EMT, nine did not, and EMT was unknown for four individuals.Of the 24 operators, 14 died in ≤1 h.Of the five who were admitted to a hospital, one lived for 24 h (one day), one for 48 h (two days), two for 72 h (three days), and one for 336 h (14 days).Four operators, none of whom were admitted to a hospital, lived for an unknown period of time before their death.All 24 decedents were males.Twenty were family members (i.e., kin) of the survey respondents, and three were not.Family status

Chores at Time of Fatal Non-ROPS Tractor Overturn
Descriptions of chores for 23 of the 24 fatal non-ROPS tractor overturns were distributed within the non-overlapping categories shown in Table 4.Of these 24 fatally injured operators, 11 received EMT, nine did not, and EMT was unknown for four individuals.Of the 24 operators, 14 died in ≤1 h.Of the five who were admitted to a hospital, one lived for 24 h (one day), one for 48 h (two days), two for 72 h (three days), and one for 336 h (14 days).Four operators, none of whom were admitted to a hospital, lived for an unknown period of time before their death.All 24 decedents were males.Twenty were family members (i.e., kin) of the survey respondents, and three were not.Family status was unknown for one individual.Operator age at the time of death was known for 21 individuals and distributed as shown in Figure 2.   As shown in Figure 3, chores that were known for 23 of 24 fatal non-ROPS tractor overturns include rotary mowing = 6, farm travel = 6, road travel = 4, and hay harvesting = 3. Farm travel involves driving from one location to another on farm property as opposed to road travel.

Chores at Time of Serious Nonfatal Non-ROPS Tractor Overturns
A total of 64 nonfatal operator injuries from overturns of non-ROPS tractors received EMT.For 62 individuals, 59 (95.2%) were male and three (4.8%) were female.Chore categories, as reported at the time of overturn, and other selected variables were distributed as shown in Table 5. Fifty-seven of the injured operators were family members, and three were non-family members.Space does not permit a table that tallies values for all 19 variables for which data were collected.Chore category was known for 58 cases.As shown in Figure 4

Chores at Time of Serious Nonfatal Non-ROPS Tractor Overturns
A total of 64 nonfatal operator injuries from overturns of non-ROPS tractors received EMT.For 62 individuals, 59 (95.2%) were male and three (4.8%) were female.Chore categories, as reported at the time of overturn, and other selected variables were distributed as shown in Table 5. Fifty-seven of the injured operators were family members, and three were non-family members.Space does not permit a table that tallies values for all 19 variables for which data were collected.Chore category was known for 58 cases.As shown in Figure 4     Days in the hospital were strongly skewed to the right of the distribution as shown in Figure 5.About one-third of the farmers spent less than five days in a hospital, and about one-half of them spent less than 10 days in a hospital.Days hospitalized for 56 farmers totaled 1118 (M = 19.96,SD = 30.28).Days of farm work lost for 48 farmers totaled 6742 (M = 140.46,SD = 133.24).Ten were permanently disabled.
Operator age at the time of nonfatal, non-ROPS tractor overturns was known for 60 farmers as shown in Figure 6.Eleven farmers were reported to have sustained permanently disabling injuries that resulted from overturns of non-ROPS tractors.See Table 6.Six of the 11 recovered sufficiently to resume farm work.The reported disabilities for these six farmers included broken ankle, knee, rib, and shoulder bones, as well as torn muscles in shoulders and legs.
Five of the 11 farmers with permanent disabilities were never again able to perform farm work, and three were never able to do any other type of work.Of three fully-disabled farmers, one had a crushed chest; and one had injuries that required walking with a cane.The injury description for the third farmer was unknown.None of these fully-disabled farmers required special care typical of that provided by nursing homes.
Excluding eight farmers whose permanent disabilities prevented them from performing farm work or other work, the days of farm work lost were known for 48 farmers.These 48 farmers lost a total of at least 6742 farm work days with a mean of at least 140.46 days each (SD = 133.24)with a median of 90 days.Days of farm work lost are strongly skewed to the right, which suggests that non-ROPS tractor overturns result in high injury severity rates as shown in Figure 7. Eleven farmers were reported to have sustained permanently disabling injuries that resulted from overturns of non-ROPS tractors.See Table 6.Six of the 11 recovered sufficiently to resume farm work.The reported disabilities for these six farmers included broken ankle, knee, rib, and shoulder bones, as well as torn muscles in shoulders and legs.
Five of the 11 farmers with permanent disabilities were never again able to perform farm work, and three were never able to do any other type of work.Of three fully-disabled farmers, one had a crushed chest; and one had injuries that required walking with a cane.The injury description for the third farmer was unknown.None of these fully-disabled farmers required special care typical of that provided by nursing homes.
Excluding eight farmers whose permanent disabilities prevented them from performing farm work or other work, the days of farm work lost were known for 48 farmers.These 48 farmers lost a total of at least 6742 farm work days with a mean of at least 140.46 days each (SD = 133.24)with a median of 90 days.Days of farm work lost are strongly skewed to the right, which suggests that non-ROPS tractor overturns result in high injury severity rates as shown in Figure 7.

Injury Prevention Effectiveness of ROPS and Seatbelts
Of 24 overturns of ROPS-equipped tractors only one tractor operator died.He was not wearing a seatbelt when driving a ROPS-equipped tractor on a public road.He was ejected and crushed by the overturning tractor when he ran off the road while attempting to avoid a collision with an approaching motor vehicle.Myers et al. presented a separate analysis of seatbelt use regarding this cohort [26].
We identified only four of the 89 ROPS-equipped tractor overturns for which data were complete that produced operator injuries severe enough to require hospital admission [15].Details of these four ROPS-equipped tractor overturn events are summarized in Table 7.Three of the nonfatal injuries were related to chores of road travel, farm travel (hauling hay as feed), and rotary mowing (bush hogging) while one chore was unknown.All four overturns were to the side and involved violent overturns-one upside down and three with complete 360 • rolls-which resulted in neither death nor disabling injuries of the operators.None of these operators were known to be wearing seatbelts.It is worth noting that hospital stays regarding serious injuries associated with ROPS-equipped tractor overturns ranged from one to six days with a mean of 3.5 days, which is in sharp contrast with non-ROPS tractor overturn-related injuries that are strongly skewed toward much longer hospital stays as shown in Figure 5 with a mean of 20 days.Indeed, the length of hospital stays regarding ROPS-equipped tractor overturn injuries would reside at the low end of the non-ROPS tractor scale for hospital stays, rising up across two-thirds of the scale beyond six days to 180 days for a hospital stay, showing improved safety provided by ROPS.Improved safety of ROPS-equipped tractors was also shown by no permanent injury and many fewer days lost from work.

Discussion
No studies with survey data were conducted from a population-based random sample of this magnitude before or since the 2002 study documenting both fatal and nonfatal injuries associated with tractor overturns.Study limitations are noted below.Finally, the application of the study findings are discussed, and the prevalence of non-ROPS tractors is considered in some countries experiencing increasing agricultural machinery mechanization.

Limitations
Six limitations are described below: three regard the underreported number of fatal and nonfatal injuries, a fourth addresses the issue of exposure measurement, another discusses the current relevance of the 14-year-old study, and a sixth describes the generalizability of the results to a broader population.
First, the healthy worker effect may cause underreporting [27].Farmers' whose operations went out of business because of a severe or fatal tractor overturn injury were removed from the NASS state and national list of farm operations.During the approximate 41-year period spanned by the survey, the overturn events were those reported by healthy farmers, or by the family members of injured or deceased farmers' whose operations remained in business and were, therefore, included in the Kentucky NASS 2001 comprehensive farm list.Thus, data for nonfatal and fatal tractor overturn events from farm operations no longer in business were not counted.
Second, the number of overturn injuries and fatalities reported includes only each farm's most recent tractor overturn event of Kentucky's 75,780 farms at the time of the survey.One item at the beginning of the survey asked farmers to report the total number of all tractor overturns that had occurred in the history of their operation.A total of 551 farms reported 603 overturns during the history of their operation.However, to minimize the response burden for the farmers, and to help ensure accuracy of reporting, each farm operator was asked to report only the details for their most recent tractor overturn event.Thus, details for these additional 52 overturn events were not collected.
Third, the survey included only a 7.98% random sample of Kentucky farmers for an approximate 41-year period, As a consequence the total number of fatal and nonfatal tractor overturns within the Kentucky farm operator population over the period are 12.5 times greater than those identified by the survey.Even so, the survey results are robust and useful because they are based on a large statewide, population-based random sample stratified by farm size and USDA agricultural district.
Fourth, while this study addresses qualitative exposures (i.e., exposures to chores), quantitative exposure data would be useful for similar studies.These exposure data could better indicate priorities for educational or ROPS retrofit programs.However, the challenge for measuring exposure is problematic: (1) recall is a problem, especially considering the use of multiple tractors on a farm over a long period of time; and (2) tractor meters are a source of engine running time, but the meters do not differentiate between stationary and mobile operations, and the meters may become defective over time or by damage.
Fifth, this analysis is based on a study conducted 14 years ago in 2002, and the relevance of its results might be questioned.In a 2008 study, investigators found that non-ROPS prevalence on farm tractors in the United States decreased from 62% in 1993 to 54% in 2001, and to 49% in 2004.In 2004, there were 4 million tractors in the United States [28].A 2010 study found that, by 2006, the non-ROPS prevalence had dropped to 41%.Over the period, from 1992 to 2007, tractor overturns accounted for 1412 deaths, and for each 1% decrease in non-ROPS tractor prevalence, there was a decrease of 0.07 overturn-related deaths per 100,000 workers [2].The non-ROPS tractors reported in the current study are the same older tractors reported in the 2008 and 2010 studies.Nearly two million non-ROPS tractors manufactured prior to 1985 remain in service and a threat to life today [12,13].
Sixth, the current study was conducted in the state of Kentucky, so generalizability might be questioned.This data have been found to be generalizable to the non-ROPS problem in the high risk areas of the United States [24].Another study, using the same dataset, found that Kentucky and five other nearby states had the highest fatality rate regarding overturn-related deaths among all 50 states in the country.These six states have rugged terrain and small farms in common [29].Indeed, another study using the same data set affirmed that rough terrain was typical of small farms in the Appalachian Mountains' foothills that had a higher tractor overturn rate than large farms on relatively level ground [25].

Application
This study provided detailed information about the frequency, type, and extent of injuries associated with non-ROPS and ROPS-equipped tractor overturns across an array of chores.Data about these chores and other variables provide information about EMT received, hospital admission rates, days in the hospital, days of work lost, and disability outcomes and their duration.These data are useful in at least three ways.
First, the data can be used to provide more accurate estimates of the costs of tractor overturn injury outcomes as related to the chore performed, as well as for calculating cost effectiveness estimates for preventing injuries.For example, differentiating the potential cost savings of injuries prevented for crop-type chores (e.g., mowing) relative to livestock-type chores (e.g., fence work) would help to counter the averaging effect in cost analyses for ROPS retrofit interventions.As described earlier, ROPS-equipped tractors are less prevalent on crop farms than livestock farms [2].Thus, a cost savings case could potentially be made for targeted education and ROPS retrofit programs and retrofitting tractors used for crop farming where the overturn-related injury risk is higher and the potential benefit is higher.
Second, the results can be used to design educational materials and interventions.These can be used to inform farmers and others who have a stake in preventing tractor-related injuries about the circumstances and details of chores that result in tractor overturn injuries and methods for their prevention.
Third, this study reinforces the need to encourage retrofitting older tractors with ROPS.Four overturns involving ROPS-equipped tractors resulted in nonfatal hospital admissions-however, none resulted in permanent disabilities.

Non-ROPS Tractors in Countries with Emerging Mechanization
Economically-challenged small farms and rough terrain, including unpaved roadways, are present in Kentucky [30] as well as in other countries such as China and India.While farm mechanization has come late to many countries, there is emerging interest regarding tractor overturning incidents and safety [31].Tractors are coming into increased use in these countries, and they typically lack ROPS protection.India and China have numerous tractor manufacturers.Major tractor manufacturers in India include Escorts, Crossword Agro Industries, Mahindra Gujurat Tractor Limited, Punjab Tractors Ltd. (Swaraj), and MARS Farm Equipment Ltd.In India, tractors are a common mode of road transport in rural areas and often involve multiple riders per tractor [32].Based upon product lines viewed at the webpages of these companies, none of the models shown were equipped with ROPS.China has more than 60 companies that manufacture tractors, which vary from 1960s technology to modern designs.Some of these companies have partnered with international companies, such as John Deere and AGCO [33].One company that manufactures tractors in China, Wuzheng Agricultural Equipment Co., shows 94 tractor models on its website, 38% with cabs (whether ROPS outfitted or not), 2% fitted with a roll bar, and 60% with no ROPS [34].
The lack of ROPS on these tractors can be expected to result in the same tragedies that continues to be experienced in some highly mechanized countries despite the knowledge of the protection provided by ROPS.Regarding chores, studies need to be conducted in these countries to better understand risks associated with tractor overturns.While a tractor is likely the only motorized vehicle operating on a farm, its broader use as a utility vehicle (e.g., people transport) needs to be better understood so as to guide priorities for education and, more importantly, fitting the tractors with rollover protection.

Conclusions
This is the first population-based random sample study of tractor chores during operator fatal and nonfatal overturns of non-ROPS and ROPS-equipped tractors.Furthermore, the data used in this study reflect the problem of non-ROPS tractors and the chores for which they are used.These older tractors persist in being used for farm chores, and our study provides a lens into the type of chores that represent overturn hazards.Taking the perspective of tractor use in farm chores reinforces the need for ROPS retrofits on non-ROPS tractors.In addition, high hazard chores have been identified for which educational and ROPS retrofit interventions can be targeted, especially for travel on the farm, rotary mowing, and hay harvesting.These three chores represent 68.2% of fatal events and 50.0% of permanent and 56.6% of temporary disability overturn incidents.In addition, we found that the hospital residence time of people with nonfatal injuries related to overturns of non-ROPS tractors greatly exceeded the residence time of those patients with nonfatal injuries that resulted from ROPS-equipped tractor overturns.Despite the knowledge of the protection provided by ROPS, an implication of this study concerns countries with emerging mechanization where most tractors are produced without ROPS.The lack of ROPS on tractors in these countries can be expected to result in the same overturn tragedies experienced by the highly-mechanized countries.

Figure 1 .
Figure 1.Number of fatal, permanent, or temporarily disabling injuries associated with overturns of tractors that lacked rollover protective structures in Kentucky,1960-2002, n = 88.

Figure 1 .
Figure 1.Number of fatal, permanent, or temporarily disabling injuries associated with overturns of tractors that lacked rollover protective structures in Kentucky,1960-2002, n = 88.

Figure 2 .
Figure 2. Age of operators killed in an overturn of non-ROPS tractors in Kentucky, 1960-2002, n = 21.Figure 2. Age of operators killed in an overturn of non-ROPS tractors in Kentucky, 1960-2002, n = 21.

Figure 2 .
Figure 2. Age of operators killed in an overturn of non-ROPS tractors in Kentucky, 1960-2002, n = 21.Figure 2. Age of operators killed in an overturn of non-ROPS tractors in Kentucky, 1960-2002, n = 21.

Figure 5 .
Figure 5. Days in hospital for operators injured as a result of non-ROPS tractor overturns, n = 56.Figure 5. Days in hospital for operators injured as a result of non-ROPS tractor overturns, n = 56.

Figure 5 .
Figure 5. Days in hospital for operators injured as a result of non-ROPS tractor overturns, n = 56.Figure 5. Days in hospital for operators injured as a result of non-ROPS tractor overturns, n = 56.

Figure 5 .
Figure 5. Days in hospital for operators injured as a result of non-ROPS tractor overturns, n = 56.

Table 1 .
Tractor chores listed in 2016 and associated chore categories used later in this article.

Table 2 .
Background data collected about overturns from the 2002 Kentucky Farm Tractor Overturn Survey (6063 farms).

Table 3 .
Top five chores associated with fatal and nonfatal injury overturns of tractors without rollover protective structures.

Table 3 .
Top five chores associated with fatal and nonfatal injury overturns of tractors without rollover protective structures.

Table 4 .
Details of fatal non-ROPS tractor overturns by chore category, n = 24.
* EMT: emergency medical treatment; ‡ all deaths >1 h required hospital admittance; ** Farm travel refers to driving the tractor from one location to another between tasks; † This is a term used by farmers.Safety 2016, 2, 18 of 17

Table 5 .
Chore category and selected variables for overturns of non-ROPS nonfatal injuries, n = 64.

Table 5 .
Chore category and selected variables for overturns of non-ROPS nonfatal injuries, n = 64.

Table 6 .
Details of nonfatal non-ROPS tractor overturns by chore category, n = 64.Chore Category No.

Table 7 .
Chore category, overturn type, injury outcome, and operator age and gender for overturns of four ROPS-equipped tractors.