Journal of Safety Research 53 (2015) 97–102

Contents lists available at ScienceDirect

Journal of Safety Research journal homepage: www.elsevier.com/locate/jsr

Nonfatal tractor-related injuries presenting to a state trauma system☆ Amanda R. Swanton, a,b,⁎ Tracy L. Young, a Kathy Leinenkugel, c James C. Torner, a,d Corinne Peek-Asa a,e,⁎⁎ a

Injury Prevention Research Center, University of Iowa, 2190 WL (Westlawn), Iowa City, IA 52242, USA Carver College of Medicine, University of Iowa, 375 Newton Road, Iowa City, IA 52242, USA c Iowa Department of Public Health, 321 E. 12th Street, Des Moines, IA 50319, USA d Department of Epidemiology, University of Iowa, 145 N. Riverside Drive, Iowa City, IA 52242, USA e Department of Occupational and Environmental Health, University of Iowa, 145 N. Riverside Drive, Iowa City, IA 52242, USA b

a r t i c l e

i n f o

Article history: Received 28 October 2014 Received in revised form 28 January 2015 Accepted 9 March 2015 Available online 20 March 2015 Keywords: Tractor Agriculture Farm equipment Injury surveillance Trauma registry

a b s t r a c t Purpose: To identify tractor-related injuries using data from a statewide trauma system, to characterize the mechanisms of nonfatal tractor-related injury, and to determine which injuries are associated with higher severity injury. Methods: A retrospective observational study was conducted using the Iowa State Trauma Registry to identify cases of nonfatal tractor-related injuries over an 11-year period from 2002 to 2012. Frequency of injury was reported by age, sex, severity, and nature. Injuries were classified by mechanism and a polytomous regression model was used to predict injury severity adjusting for sex and age. Results: Five-hundred thirteen nonfatal tractor-related injuries were identified with 18% classified as severe. Injuries were most frequent among males and among those ≥45 years of age. Rollovers were the most frequent mechanism of both total (25%) and severe injury (38%), although the frequency of injury mechanism varied by age. Falls were the next most frequent mechanism of injury (20%) but resulted in fewer high-severity injuries. Collision (adjOR = 1.89, 95% CI = 1.01–3.51), rollover (adjOR = 2.03, 95% CI = 1.21–3.40), and run over/rolled on (adjOR = 2.06, 95% CI = 1.17–3.62) injuries were significantly associated with higher injury severity. Advanced age was also a significant predictor of higher severity injury (adjOR = 1.82, 95% CI = 1.06–3.12). Summary: Mechanisms of nonfatal tractor-related injuries are heterogeneous, differ by age, and are associated with varying level of severity. Practical Applications: This work shows the burden of nonfatal tractor injuries on a rural state trauma system. These findings also demonstrate the heterogeneous nature of nonfatal tractor injuries and underscore the need for a multi-level approaches to injury prevention. © 2015 National Safety Council and Elsevier Ltd. All rights reserved.

1. Introduction Agriculture is among the most hazardous industries in the United States. Estimates suggest that for every 20 farmworkers, 1–2 will experience an injury per year (Gerberich et al., 1993; Hard, Myers, & Gerberich, 2002; Myers, 1997, 1998, 2001). Additionally, the fatality rate—nearly 23 deaths per 100,000 FTE workers—is also higher in the agriculture, forestry, and fishing sector than in any other sector (Statistics B of L, 2012). Tractors are known to be a frequent cause of agricultural injuries, accounting for 4–14% of nonfatal injuries (Gerberich et al., 1993; Myers, 1997, 1998, 2001; Pickett et al., 2001) and over one-third of fatal injuries (Hard et al., 2002). Tractors are also an integral ☆ This work was funded by the University of Iowa Great Plains Center for Agricultural Safety and Health (CDC/NIOSH U54 OH007548). ⁎ Correspondence to: A.R. Swanton, 145 N. Riverside Dr., S143 CPHB, Iowa City, IA 52252, USA. Tel.: +1 512 680 0754. ⁎⁎ Correspondence to: C. Peek-Asa, 145 N. Riverside Dr., S143 CPHB, Iowa City, IA 52252, USA. Tel.: +1 319 335 4895. E-mail addresses: [email protected] (A.R. Swanton), [email protected] (C. Peek-Asa).

http://dx.doi.org/10.1016/j.jsr.2015.03.002 0022-4375/© 2015 National Safety Council and Elsevier Ltd. All rights reserved.

part of farming, used frequently for a wide variety of activities; these activities may be directly related to farming operations (e.g., cultivating the land) but may also include other tasks around the farm (e.g., landscaping, transportation). Identification of activities that lead to tractor-related injuries could help reduce the frequency and severity of these injuries. Efforts to reduce tractor-related injury have mainly focused on preventing rollovers through the use of rollover protective structures (ROPS). Vehicle overturns often result in severe injury when the occupant is crushed or pinned under the heavy equipment; overturn events are responsible for a quarter of agricultural fatalities (Hard et al., 2002). The high mortality associated with tractor rollovers has prompted the development and promotion of ROPS, which are designed to keep the occupant in the driver compartment during a rollover. ROPS have been found to effectively prevent an overturning tractor from crushing the operator (Reynolds & Groves, 2000). While studies suggest that ROPS have successfully reduced fatality due to tractor rollovers (Myers & Hendricks, 2010), interventions for tractor-related injuries that occur through other mechanisms are needed, particularly those that address the high frequency of nonfatal tractor injuries.

98

A.R. Swanton et al. / Journal of Safety Research 53 (2015) 97–102

While national data for fatalities are available through sources such as the Census for Fatal Occupational Injuries (CFOI), nonfatal tractorrelated injuries are more difficult to identify in current databases. Medical databases, such as those based on hospital or insurance data, rarely identify if an injury was agriculture-related. Even if occupational codes are available, they may undercount ties to agricultural activities as injuries also occur among family members (e.g., children) not formally employed on the farm. Occupational databases may exclude the reporting of farming operations, especially small ones, due to their exempt status under the Occupational Safety and Health Administration (OSHA). Furthermore, tractor injuries may not be reported to occupational databases if the tractor’s operation was not directly related to commodity production; for example, injuries sustained while doing property maintenance may be excluded. Additionally, such databases often do not have detail necessary to identify tractor involvement and may contain only limited information regarding injury characteristics. Data sources that allow researchers to identify and describe nonfatal tractor injuries are scant. While studies suggest that nonfatal injuries may occur at a rate exceeding 30 times that of fatal injuries (Hard et al., 2002; Myers, 1997, 1998, 2001), few studies have described the nature of nonfatal tractor injuries. The leading causes of nonfatal injury may differ from those contributing to fatalities. A survey of Midwestern farmers showed that 42% of nonfatal injuries occurred when mounting/dismounting the tractor rather than during operation (Lee et al., 1996). Likewise, another study of worker’s compensation in Colorado found that mounting/dismounting was a contributing factor in 20% of tractor-related injury claims; other leading contributing factors included performing tractor maintenance (15%) and having a body part hit the tractor (14%) (Douphrate, Rosecrance, Reynolds, Stallones, & Gilkey, 2009). Canadian surveillance data derived from hospital records identified that run overs were the most common mechanism of tractor-related injury (27%), followed by being pinned/struck (20%) and falls (18%) (Pickett et al., 2001). To our knowledge, no studies in the United States have utilized a statewide trauma registry to assess the causes of tractor-related injuries requiring emergency care. Using 11 years of data from a statewide trauma system, this study describes nonfatal tractor injuries and identifies which injury mechanisms were associated with the most severe injury. 2. Methods 2.1. Study data and sample The Iowa Trauma System, developed under the Iowa Trauma Care System Development Act of 1995, has been fully operational since 2001. In the statewide system, every acute care hospital is certified by the level of care they can provide. This rating ranges from Level 1, which provides the most advanced trauma care, to Level IV, which provides stabilization, transfer, and treatment for less severe injuries (Mellecker, Torner, Young, & Sparling, 2010). As part of the Trauma System, the Iowa Trauma Registry collects information using a standardized, system-wide software program on patients who meet trauma criteria from all Level I, II, and III hospitals as well as a sample of Level IV hospitals (Iowa Department of Public Health). Trauma centers submit pre-defined demographic, prehospital, injury, diagnostic, and outcome data for patients with injury ICD-9 codes (800.0–999.0) that are evaluated or treated at a their facility. The Trauma Registry collects additional information specifically on agricultural injuries, including a variable that designates if an injury is related to agriculture, and a brief narrative describing the injury. This study sample was identified using the Trauma Registry data from 2002 through 2012 by first selecting all Trauma Registry cases that were recorded as agriculture-related (n = 3490). Tractor-related injuries were then identified through review of the injury narrative. “Tractor-related” was defined as an injury due to or involving a tractor

or tractor part, excluding those injuries caused exclusively by rear attachments. For the agricultural injuries, a query was performed to find the word “tractor” in the narrative (n = 545). To avoid missing cases based on spelling errors, the cause of injury narrative was manually reviewed for all motor vehicle-related injuries (ICD-9 810.0–825.9), other road traffic injuries (ICD-9 825.0–829.9), and injuries caused by machinery (ICD-9 919.0–919.9) yielding 2 additional tractor-related injuries. Thus, 547 cases were assessed to determine whether they met the case definition. Of the assessed injuries, 4 were excluded because the tractor did not contribute to the injury (e.g., individual suffered a non-tractor-related injury and returned to their tractor for transport), and 8 injuries were excluded because the injury was due to a rear attachment (e.g., rakes, sprayers, etc.). Since our focus was nonfatal tractor injuries, those patients that were confirmed dead prior to hospital discharge (n = 22) were also excluded. In total, 513 nonfatal, tractorrelated injuries were identified and included in the analysis. 2.2. Study variables Injuries were categorized into eight mechanisms based on the External Cause of Injury code and a review of the injury narrative. The mechanisms were collision, rollover, run over, fall/jump/ejection, caught/pinched/cut, struck, fire/chemical burn, and other/unclassified. When multiple injury mechanisms applied, a semihierarchical structure was used, with collisions as the first priority, non-collisions rollovers or tips as the second priority, and a person being run over as the final priority. For example, a tractor that collided with another vehicle and subsequently rolled over would be coded as a “collision,” as opposed to “rollover;” in our sample, 7 injuries were classified as collisions instead of rollovers based on this hierarchy. Similarly, several rollovers (n = 16) and run overs (n = 18) could have been classified as fall/jump/ejection. All other injuries were classified into the most appropriate of the remaining categories without further hierarchy. Injuries that were not well-described by the defined mechanisms or for which the cause of injury was insufficient to determine the mechanism were classified as “other/unclassified.” Due to the heterogeneity of injury causes, each mechanism was divided into two or more injury subtypes. Injury severity was available in the Iowa Trauma Registry as quantified by the Injury Severity Score (ISS) (Baker, O’Neill, Haddon, & Long, 1974). ISS is cumulative index that ranges from 1 to 75 and incorporates information from the three most severely injured body regions. Scores N15 are considered severe injuries. The nature of injury was classified according to the Barell Matrix classification scheme using the ICD-9 code of the primary injury (Barell et al., 2002); all categories containing b10 injuries were collapsed into the category “Other.” 2.3. Analysis Frequency tables were generated to show the number and proportion of injuries by age, sex, severity, and nature of injury. Injuries were also tabulated by mechanism of injury, both overall and stratified by the age of the injured person. Injury severity was represented graphically to show the relative frequency by injury mechanism. Injury subtypes were also cross-tabulated with injury severity to provide more detailed information about specific types of injuries within each mechanism category. In order to determine which mechanisms led to the most severe injuries, a multivariable polytomous logistic regression model was fit with injury mechanism category as the main predictor and injury severity as the outcome. The outcome variable was categorized into three ordered levels: minor injuries (ISS 1–8), moderate injuries (ISS 8–15), and severe injuries (ISS N15). Because tractor injuries were unequally observed by sex and age, these variables were included in the model as two-level and four-level categorical covariates, respectively. Adjusted

A.R. Swanton et al. / Journal of Safety Research 53 (2015) 97–102

odds ratios with 95% confidence intervals were constructed to describe the associations.

99

Table 2 Most frequent causes of nonfatal tractor-related injuries, by age category. n

3. Results 3.1. Injury characteristics Five-hundred thirteen nonfatal tractor injuries were reported to the Iowa Trauma Registry from 2002 to 2012. Patients had a median hospital stay of 3 days (range: 0–68 days) with 23% of patients requiring at least one day in the intensive care unit (median: 3 days, range: 1–34 days). Males (93%) had nonfatal tractor injuries far more frequently than females (Table 1). The majority of injuries occurred in adults 25 and older (88%) with adults aged 45–64 years accounting for 37% of injuries and those over 65 years accounting for 34%. Most of the injuries were of minor or moderate severity with only 92 injuries (18%) classified as severe. Fractures accounted for nearly half of all observed injuries seeking hospital care, followed by internal injuries (13%) and open wounds (10%).

3.2. Injury mechanisms by age Overall, the most common mechanisms of nonfatal tractor-related injuries were rollover (n = 127, 25%), fall/jump/ejection (n = 104, 20%), and run over (n = 85, 17%); burns were the least frequent cause of nonfatal tractor injuries (Table 2). However, the frequencies of the injury mechanisms were different when stratified by age. Among children, fall/ jump/ejection (n = 11, 33%) and run over (n = 8, 24%) were the most common mechanisms of injury. Among young adults, rollover (n = 8, 27%) was most frequent followed by 5 cases of each collision, run over, and caught/pinched/cut. Among adults aged 25–44 years, collision (n = 22, 26%) and rollover (n = 15, 18%) were the most frequent mechanisms. Lastly, for both older adults and seniors, the most frequent mechanisms of injury were rollover and fall/jump/ejection. Fire/chemical burn was the least frequently observed mechanism of injury for all age-groups.

Table 1 Characteristics of nonfatal tractor-related injury cases identified from the Iowa State Trauma Registry from 2002 to 2012.

Total Age ≤14 15–24 25–44 45–64 ≥65 Sex M F Missing Injury Severity Score (ISS) ISS 1–8 ISS 9–15 ISS N15 Missing Nature of injury Fracture Internal Open Wound Contusion Burn Crush Sprain Dislocation Other Missing

n

%

513

100%

33 30 84 191 175

6.4% 5.9% 16.4% 37.2% 34.1%

477 35 1

93.0% 6.8% 0.2%

273 139 92 9

53.2% 27.1% 17.9% 1.8%

262 67 52 50 18 16 15 10 16 7

51.1% 13.1% 10.1% 9.8% 3.5% 3.1% 2.9% 2.0% 3.1% 1.4%

Children (b15 years) Fall/jump/ejection Run over Caught/pinched/cut Struck Rollover Collision Other/unclassified Young adults (15–24 years) Rollover Collision Run over Caught/pinched/cut Struck Fall/jump/ejection Adults (25–44 years) Collision Rollover Fall/Jump/Ejection Caught/Pinched/Cut Run Over Struck Fire/Chemical Burn Older Adults (45–64 years) Rollover Fall/Jump/Ejection Collision Run Over Struck Caught/Pinched/Cut Fire/Chemical Burn Other/Unclassified Seniors (65+ years) Rollover Fall/jump/ejection Run over Struck Caught/pinched/cut Collision Fire/chemical burn Other/unclassified Total Rollover Fall/jump/ejection Run over Collision Struck Caught/pinched/cut Fire/chemical burn Other/unclassified

33 11 8 6 4 2 1 1 30 8 5 5 5 4 3 84 22 15 13 13 10 9 2 191 54 39 24 24 22 17 8 3 175 49 38 37 17 13 12 6 3 513 127 104 85 64 56 54 16 7

% 33.3% 24.2% 18.2% 12.1% 6.1% 3.0% 3.0% 26.7% 16.7% 16.7% 16.7% 13.3% 10.0% 26.2% 17.9% 15.5% 15.5% 11.9% 10.7% 2.4% 28.3% 20.4% 12.6% 12.6% 11.5% 8.9% 4.2% 1.6% 28.0% 21.7% 21.1% 9.7% 7.4% 6.9% 3.4% 1.7% 24.8% 20.3% 16.6% 12.5% 10.9% 10.5% 3.1% 1.4%

3.3. Cause of injury and injury severity Of the 513 nonfatal injuries identified, 92 (18%) were severe. Certain injury mechanisms had disproportionately high rates of severe injury (Fig. 1, Table 3). Rollover had the highest proportion of severe injuries compared to the other mechanisms with over one-quarter (27%) of rollovers classified as severe (Table 3). The frequent occurrence of rollover as a mechanism and its high-severity rate made rollovers the most frequent cause of severe nonfatal injury; 38% of all severe injuries were due to rollovers. As shown in Table 3, several injury narratives stated that the tractor occupant attempted to jump from the vehicle during the rollover in 10 cases, which still resulted in moderate or severe injury in 50% of the cases. Run overs were the third most frequent mechanism of injury over all, but the second most frequent mechanism of severe injury (Fig. 1). Run overs had a severity rate similar to that of rollover with onequarter of injury events classified as high severity. Although the majority run over events affected individuals that had been working outside the vehicle, cases in which the injured person fell/jumped/was ejected

100

A.R. Swanton et al. / Journal of Safety Research 53 (2015) 97–102

pinched/cut had the lowest severity rate with only 5.6% of events (n = 3) resulting in severe injury.

Collision Rollover

3.4. Predictors of injury severity

Run Over Fall/Jump/Ejection Caught/Pinched/Cut Severe

Struck

Moderate Fire/Chemical Burn

Minor

Other/Unclassified

Unknown 0

50

100

150

Frequency Fig. 1. Mechanisms of nonfatal tractor-related injuries identified from the Iowa State Trauma Registry (2002–2012), sub-categorized by Injury Severity Score (ISS).

from the tractor prior to being run over tended to be high severity (Table 3). Collisions with motor vehicles led to severe injury in 17% of cases and were the third most frequent mechanism of severe injury. Of the 61 motor vehicle collisions, 48% (n = 30) were tractor operators, while 31% (n = 19) were occupants of the other vehicle; the identity of the injured victim was unclear in the remaining 21% (n = 13) of cases. Only 3 cases of collisions not involving another motorized vehicle were identified, none of which resulted in severe injury (Table 3). All remaining injury mechanisms led to less than 10 severe injuries during this 11-year period. Although fall/jump/ejection was the second most frequent mechanism of injury overall, fall/jump/ejection only led to severe injuries in 9% (n = 9) of cases. All of the severe injuries in fall/jump/ejection were the result of falling or being thrown from the tractor; in the absence of a rollover, no severe injuries resulted from jumping from the tractor (Table 3). Of all mechanisms, caught/

Three injury mechanisms were significantly associated with increasing injury severity. Run over injuries were most strongly associated with increasing injury severity resulting in 2.1 (95% CI = 1.2–3.6) times the odds of being in a higher severity category compared with fall/jump/ejection injuries (Table 4). Similarly, rollovers were also associated with a two-fold increase in odds of higher injury severity (adjOR = 2.0, 95% CI = 1.2–3.4). Collisions were also marginally associated with higher severity injury (adjOR = 1.8, 95% CI = 1.0–3.5). On the other hand, caught/pinched/cut injuries were associated with a marginally decreased odds of having a high injury severity (adjOR = 0.5, 95% CI = 0.3–1.1). The elderly (age 65+) had 1.8 (95% CI = 1.1–3.1) times the odds of suffering a higher severity injury compared to adults aged 25–44 years. Those 45–64 years of age also had a marginally significant increase in odds of severe injury with a 1.6-fold increase in the odds of higher injury severity (95% CI = 1.0–2.8). The magnitude of association was highest for those aged 15–24 years who experienced a two-fold increase in odds of higher severity injury, but this result did not reach statistical significance. Children appeared to be associated with a decreased risk of high-severity injury; however, this also did not reach statistical significance (adjOR = 0.5, 95% CI = 0.2–1.4). Sex also showed no significant association with injury severity. 4. Discussion This is among the first studies to quantify nonfatal, tractor-related injuries using a statewide trauma registry. The Iowa State Trauma Registry is unique in that it allows for identification of tractor-related injuries from prospectively collected data at a statewide level. Prior methods of surveying nonfatal tractor injury in the United States have relied on retrospectively collected survey data, including the Regional

Table 3 Mechanisms of nonfatal tractor-related injuries with detailed injury subtypes, by Injury Severity Score (ISS). Total

ISS 1–8 n

ISS N 15

ISS 9–15

Injury category

n

% column

% row

Collision MV collision Non-motor vehicle collision Rollover Non-collision rollover/tip Jumped from tractor during rollover Run over Run over Fall/thrown and run over fall/jump/ejection (non-run over/rollover/collision) Fall/Ejection from tractor Jumped from tractor Caught/pinched/cut Caught/pinched/crushed in tractor/part Cut/Lacerated Caught in power take off Struck Hit by load/debris while in/on tractor Struck against tractor/part Struck by tractor/part Fire/chemical burn Fire/chemical exposure from tractor Other fire Other/unclassified Total

64 61 3 128 118 10 84 66 18 104

100.0% 95.3% 4.7% 100.0% 92.2% 7.8% 100.0% 78.6% 21.4% 100.0%

28 26 2 57 52 5 37 31 6 62

43.8% 42.6% 66.7% 44.5% 44.1% 50.0% 44.0% 47.0% 33.3% 59.6%

98 6 54 37 11 6 56 20 16 20 16 13 3 7 513

94.2% 5.8% 100.0% 68.5% 20.4% 11.1% 100.0% 35.7% 28.6% 35.7% 100.0% 81.3% 18.8% 100.0% 100.0%

59 3 40 27 11 2 35 9 12 14 10 9 1 4 273

60.2% 50.0% 74.1% 73.0% 100.0% 33.3% 62.5% 45.0% 75.0% 70.0% 62.5% 69.2% 33.3% 57.1% 53.2%

n

Missing ISS

% row

n

% row

n

% row

24 23 1 32 28 4 24 19 5 32

37.5% 37.7% 33.3% 25.0% 23.7% 40.0% 28.6% 28.8% 27.8% 30.8%

11 11 0 35 34 1 21 15 6 9

17.2% 18.0% – 27.3% 28.8% 10.0% 25.0% 22.7% 33.3% 8.7%

1 1 0 4 4 0 2 1 1 1

1.6% 1.6% – 3.1% 3.4% – 2.4% 1.5% 5.6% 1.0%

30 2 10 7 0 3 12 5 4 3 3 2 1 2 139

30.6% 33.3% 18.5% 18.9% 0.0% 50.0% 21.4% 25.0% 25.0% 15.0% 18.8% 15.4% 33.3% 28.6% 27.1%

9 0 3 2 0 1 9 6 0 3 3 2 1 1 92

9.2% – 5.6% 5.4% – 16.7% 16.1% 30.0% – 15.0% 18.8% 15.4% 33.3% 14.3% 17.9%

0 1 1 1 0 0 0 0 0 0 0 0 0 0 9

– 16.7% 1.9% 2.7% – – – – – – – – – – 1.8%

A.R. Swanton et al. / Journal of Safety Research 53 (2015) 97–102 Table 4 Multivariable polytomous regression model of Injury Severity Score (ISS) category.

Sex Male Female Age (years) ≤14 15–24 25–44 45–64 ≥65 Injury mechanism Collision Rollover Run over Fall/jump/ejection Caught/pinched/cut Struck Fire/chemical burn Other/unclassified

Adj OR

(95% CI)

0.70 ref

(0.36,1.37)

0.52 1.99 ref 1.63 1.82

(0.20,1.37) (0.76,4.55)

1.88 2.03 2.06 ref 0.53 1.03 1.03 1.21

(1.01,3.51) (1.21,3.40) (1.17,3.62)

(0.96,2.76) (1.06,3.12)

(0.25,1.13) (0.53,1.99) (0.36,2.94) (0.26,5.52)

Rural Injury Study-I (RRIS-I) (Myers, 1997, 1998, 2001) and the Traumatic Injury Surveillance of Farmers (TISF) survey (Lee et al., 1996), which both included items on tractor injury. Other studies have used hospital records (Gelberg, Struttmann, & London, 1999; Hartling, Pickett, Dorland, & Brison, 1997; Pickett & Brison, 1995; Pickett et al., 2001) and insurance claims (Douphrate et al., 2009) to identify tractor-related injuries from clinical data, but those conducted in the United States have been limited to smaller geographic areas or single insurance providers. While countries, such as Canada (Pickett et al., 2001), have developed state and national surveillance systems using hospital data, such programs do not currently exist in the United States. The ability to identify tractor-related injuries within the context of an established statewide trauma system represents a step towards more population-based surveillance methods. Although our study found that fractures were the most frequent injury type, data from previous studies relying on self-reported or claimsbased injury identification have been dominated by soft tissue injuries. According to the RRIS-I (Myers, 1997, 1998, 2001) and TISF (Lee et al., 1996) surveys, nonfatal tractor injuries were 17–41% bruises, 6–30% cuts, and 5–36% sprains/strains; combined, these three injury types accounted for approximately 60% of the injuries observed in each study. A study of insurance claims from a worker’s compensation program had similar findings (Douphrate et al., 2009). Whereas these previous studies found that fractures were 10–21% of tractor-related injuries (Douphrate et al., 2009; Lee et al., 1996; Myers, 1997, 1998, 2001), fractures accounted for nearly half of the tractor-related injuries in the Iowa Trauma Registry. We found very few isolated soft tissue injuries, likely because our cases were from a trauma registry and required hospital care, and were likely more severe than injuries identified through surveys. The literature has not consistently identified which mechanisms of nonfatal tractor injury are most frequent. Our results showed that rollovers (25%) were the most common cause of nonfatal injury, followed by falls (20%) and run overs (16%). Run overs and rollovers were identified as mechanisms that commonly led to hospitalization in a Canadian study using discharge summaries from a single province, but falls accounted for only 8% of injuries (Pickett & Brison, 1995). A national Canadian study conducted several years later showed tractor-related injuries were more commonly caused by run overs (27%), being struck by the tractor (20%), falls (18%), and rollovers (11%) (Marlenga, Pickett, & Berg, 2001). In contrast, a sub-analysis of 764 tractor injuries from the RRIS-I survey identified only 3 rollovers leading to injury (Lee et al., 1996). Again, the differences observed in the proportions of injury mechanisms may be due to differences in severity between the samples and may be further complicated by differences in categorization schemes. Another possibility is that increasing prevalence of ROPS

101

leading to decreased rollover fatalities (Myers & Hendricks, 2010) has led to an accompanying rise in the nonfatal rollover injuries compared to 20 years ago. Whatever the cause, our results show that not only are nonfatal rollover injuries frequent among those seeking care at hospitals but demonstrate a significantly increased risk of higher severity. The second most commonly identified mechanism of nonfatal tractor injury was falling/jumping/being ejected from a tractor. National estimates from RRIS-I consistently identified falls among the leading mechanisms of nonfatal tractor-related injuries leading to lost-time from work (Myers, 1997, 1998, 2001). In the Iowa Trauma Registry, fall/jump/ejection injuries accounted for nearly one-fifth of all the nonfatal tractor-related injuries. Although the Iowa Trauma Registry did not collect specific information regarding the activity at the time of injury, other studies have suggested that 20–42% of nonfatal tractor-related injuries occur while mounting/dismounting the vehicle (Douphrate et al., 2009; Lee et al., 1996). Engineering modifications to facilitate safe entry to and exit from the tractor have the potential to reduce falls and related injuries, and further studies examining mounting devices, such as stairs and railings, are warranted. Although falls represent a relatively low injury severity subtype, the frequent occurrence of falls among most agegroups plus the potential for risk modification make falls an excellent candidate for future injury prevention measures. We also found that injury mechanism differed by age. Rollovers were the primary injury type for all age categories except children and adults aged 25–44 years. Children were more likely to be involved in falls or to be run over, which may be consistent with their roles as extra passengers or as fieldworkers outside of the tractor (Marlenga et al., 2001; Pickett et al., 2005; Smith, Scherzer, Buckley, Haley, & Shields, 2004). Adults aged 25–44 years involved in nonfatal tractorrelated injuries were most frequently involved in collisions, which could suggest that these individuals are the primary transporters of farm goods on roadways. Fall/jump/ejection events were among the top three mechanisms of injury in all age categories, except for young adults. While rollovers are the dominant cause of fatal tractor-related injuries, the multivariable model for injury severity shows that rollovers are not unique in their association with high-severity nonfatal injuries. In fact, multiple mechanisms, including rollovers, run overs, and collisions, each demonstrated an approximately 2-fold increase in odds of high-severity injury compared to fall/jump/ejection. This suggests that a multipronged approach to preventing tractor injuries is needed as strategies aimed solely at reducing the rate of tractor rollovers would fail to prevent the 62% of high-severity injuries by other mechanisms. We also found increasing age to be highly predictive of increased injury severity. Lee et al. have previously shown that the rates of tractorrelated injuries differ by age with the lowest rates of injury among children aged 10–19 years and the highest rates of injury among those 40– 49 years of age (Lee et al., 1996). Although those in their 40s have high rates of injury, our results suggest that these individuals are more likely to experience less severe injuries. Instead, our results indicated that older adults had greater odds of a more severe injury. This result is consistent with other work that has shown that farmworkers greater than 55 years of age are more likely to have restricted work duties and be hospitalized than younger workers for agricultural injuries (Myers, Layne, & Marsh, 2009), which could be related to the physiologic effects of aging and reduced resilience (Sterling, O’Connor, & Bonadies, 2001). Additionally, although not statistically significant, our results suggest that young adults aged 15–24 years may be more likely to suffer higher severity injury, perhaps resulting from greater hazard due to inexperience or increasingly risky activities. 4.1. Limitations Although these data provide important information regarding the occurrence of nonfatal tractor-related injuries, this study has several limitations. First, since the sample was derived from trauma registry

102

A.R. Swanton et al. / Journal of Safety Research 53 (2015) 97–102

data, only those who were hospitalized for their injuries were accessible for analysis; this sampling method was more likely to capture severe injuries. The hospitals participating in the Iowa Trauma Registry include all levels with those at Levels I–III required to submit trauma patient data. Thus, this sample may also over-represent those requiring transfer to higher level facilities due to the need for surgical or specialty care. However, these data do represent the burden of tractor-related injury on a state trauma system and may be useful when considering policy making and resource allocation. A second limitation of this study is that these data are presented as frequencies rather than rates. Accurate data for population exposure to tractors are not available, particularly since tractors are used both for commodity production by farmworkers, but also by other farm occupants for non-production-related tasks. The lack of exposure data limits our ability to present both aggregate and age-specific rates of nonfatal tractor-related injury. Rates have been presented by previous studies in the context of large surveys, but it is currently infeasible to conduct these surveys at a statewide level. Finally, the Iowa Trauma Registry lacks consistent descriptive data about the activity at the time of injury and what safety features were available or used. Such information would be useful at identifying new strategies for reducing injury and quantify the adoption safety devices/practices. Future work is needed to devise and evaluate interventions to prevent tractor-related injury. 5. Conclusion Nonfatal tractor-related injuries reported to a statewide trauma registry show that several mechanisms frequently lead to injury, and that these injury mechanisms vary by age and severity. Rollovers still represent the most common cause. Intervention approaches that consider nonfatal injuries can help reduce the burden of these events on individuals, their families, and the healthcare system. Successful approaches will likely need to consider tractor design, the interaction of the tractor with the roadway and farm environment, as well as individual behavior. References Baker, S., O’Neill, B., Haddon, W., & Long, W. (1974). The Injury Severity Score: A method for describing patients with multiple injuries and evaluating emergency care. Journal of Trauma and Acute Care Surgery, 14(3), 187–196. Barell, V., Aharonson-Daniel, L., Fingerhut, L. A., Mackenzie, E. J., Ziv, A., Boyko, V., Abargel, A., Avitzour, M., & Heruti, R. (2002). An introduction to the Barell body region by nature of injury diagnosis matrix. Injury Prevention, 8(2), 91–96. Douphrate, D. I., Rosecrance, J. C., Reynolds, S. J., Stallones, L., & Gilkey, D. P. (2009). Tractor-related injuries: An analysis of workers’ compensation data. Journal of Agromedicine, 14(2), 198–205. Gelberg, K. H., Struttmann, T. W., & London, M. A. (1999). A comparison of agricultural injuries between the young and elderly: New York and Kentucky5(518), 73–81. Gerberich, S. G., Gibson, R. W., French, L. R., Carr, P., Renier, C. M., Gunderson, P. D., Martin, F., True, J. A., Shutske, J., & Brademeyer, K. (1993). The Regional Rural Injury Study-I (RRIS-I): A population based effort. Minneapolis, MN: University of Minnesota. Available from: NTIS, Springfield, VA; PB94–134848. Sponsored by the Centers for Disease Control and Prevention. Hard, D. L., Myers, J. R., & Gerberich, S. G. (2002). Traumatic injuries in agriculture, 51–65. Hartling, L., Pickett, W., Dorland, J., & Brison, R. J. (1997). Hosptial costs associated with agricultural machinery injuries in Ontario509, 502–509. Iowa Department of Public Health (h). Trauma Program. Available at: http://www.idph. state.ia.us/ems/trauma.asp (Accessed September 8, 2014). Lee, T. Y., Gerberich, S. G., Gibson, R. W., Carr, W. P., Shutske, J., & Renier, C. M. (1996). A population-based study of tractor-related injuries: Regional rural injury study-I (RRIS-I). Journal of Occupational and Environmental Medicine, 38(8), 782–793. Marlenga, B., Pickett, W., & Berg, R. L. (2001). Assignment of work involving farm tractors to children on North American farms. American Journal of Industrial Medicine, 40(1), 15–22.

Mellecker, M., Torner, J. C., Young, T., & Sparling, K. (2010). Iowa trauma system ten year report. , 1–22 (Available at: http://www.public-health.uiowa.edu/iprc/resources/IATrauma-System_Ten-Year-Report.pdf. Accessed August 19, 2014). Myers, J. (1997). Injuries among Farm Workers in the United States, 1993. DHHS (NIOSH) Publication No. 97–115. Cincinnati, OH: National Institute for Occupational Safety and Health (Available at: http://www.cdc.gov/niosh/docs/97-115/. Accessed August 19, 2014). Myers, J. (1998). Injuries among Farm Workers in the United States, 1994. DHHS (NIOSH) Publication No. 98–153. Cincinnati, OH: National Institute for Occupational Safety and Health. Myers, J. (2001). Injuries among Farm Workers in the United States, 1995. DHHS (NIOSH) Publication No. 2001-153. Cincinnati, OH: National Institute for Occupational Safety and Health (Available at: http://www.cdc.gov/niosh/docs/2001-153/. Accessed August 19, 2014). Myers, J. R., & Hendricks, K. J. (2010). Agricultural tractor overturn deaths: Assessment of trends and risk factors. American Journal of Industrial Medicine, 53, 662–672. Myers, J. R., Layne, L. a, & Marsh, S. M. (2009). Injuries and fatalities to U.S. farmers and farm workers 55 years and older. American Journal of Industrial Medicine, 52(3), 185–194. Pickett, W., & Brison, R. (1995). Tractor-related injuries in Ontario. Canadian Journal of Public Health. Revue Canadienne de Sante Publique, 86(4), 13–16. Pickett, W., Brison, R. J., Berg, R. L., Zentner, J., Linneman, J., & Marlenga, B. (2005). Pediatric farm injuries involving non-working children injured by a farm work hazard: Five priorities for primary prevention. Injury Prevention, 11, 6–11. Pickett, W., Hartling, L., Dimich-Ward, H., Guernsey, J. R., Hagel, L., Voaklander, D. C., & Brison, R. J. (2001). Surveillance of hospitalized farm injuries in Canada. Injury Prevention, 7(2), 123–128. Reynolds, S. J., & Groves, W. (2000). Effectiveness of roll-over protective structures in reducing farm tractor fatalities. American Journal of Preventive Medicine, 18(4S), 63–69. Smith, G. A., Scherzer, D. J., Buckley, J. W., Haley, K. J., & Shields, B. J. (2004). Pediatric farmrelated injuries: A series of 96 hospitalized patients. Clinical Pediatrics, 43, 335–342. Statistics B of L (2012). Census of Fatal Occupational Injuries: 2012 Chart Package. (Washington, DC; Available at: http://www.bls.gov/iif/oshcfoi1.htm#charts. Accessed August 12, 2014). Sterling, D. A., O’Connor, J. A., & Bonadies, J. (2001). Geriatric falls: Injury severity is high and disproportionate to mechanism. Journal of Trauma, 50(1), 116–119. Amanda Swanton is an MD/PhD candidate pursuing her doctoral work in Translational Biomedicine Program at the University of Iowa (UI). She received a BA in Physics and Russian from Grinnell College before joining the UI Medical Scientist Training Program (MSTP) in 2010. She is also an Occupational Injury Prevention fellow through the Heartland Center for Occupational Health and Safety and currently works on agricultural injury surveillance and evaluating time to care among farmworkers. Tracy L. Young, MS, is an injury epidemiologist at the UI Injury Prevention Research Center where she provides expertise in data management and analysis. She received her MS degree from the University of Iowa. Her interests include acute trauma care, including road traffic injuries, falls, and injuries. She also acts as a consultant for the Iowa Statewide Trauma System facilitating database maintenance as well as trauma system evaluation. Kathy Leinenkugel, MPA, is a community health consultant and co-investigator of the Occupational Health & Safety Program at the Iowa Department of Public Health. She received her MPA from the University of Nebraska. Her work focuses on occupational safety at multiple levels including surveillance, education, program implementation, and evaluation and spans numerous content areas. She also serves on the Board of Directors for Iowa’s Center for Agricultural Safety and Health (I-CASH), which focuses on agricultural-related injury surveillance, awareness, and outreach. James C. Torner, PhD, is the professor and head of the UI Department of Epidemiology and Senior Associate Director of the Institute for Clinical and Translational Science. Dr. Torner an MS in Biostatistics and a PhD in Epidemiology from the University of Iowa. Dr. Torner’s interests include injury and trauma systems evaluation, stroke and cerebrovascular disorders, traumatic brain injury, epidemiological methods, and translational research. Dr. Torner is also the team leader of the IPRC Acute Care Research Team and primary liaison with the Iowa Statewide Trauma System on the Trauma System Advisory Council and oversees the Iowa Trauma Registry. Corinne Peek-Asa, PhD, is the Associate Dean for Research of the University of Iowa, College of Public Health, a Professor of Occupational and Environmental Health, and Director of the CDC-funded Injury Prevention Research Center. Dr. Peek-Asa received her PhD in Epidemiology from the University of California, Los Angeles. Her work focuses on the implementation and evaluation of programs and policies to prevent acute traumatic injuries and violence. She has current studies in the areas of teen driving safety, bicycle safety, workplace violence, intimate partner violence, bullying prevention, and international road traffic safety.