Recreational off-highway vehicle crashes resulting in victims being treated at a regional trauma center: mechanisms and contributing factors

Background Recreational off-highway vehicles (ROVs) have become increasingly popular in recent years; however, crash epidemiology is not well described. ROVs travel at least 30 mph, and unlike all-terrain vehicles, have a rollover protective structure (ROPS) and seat belts or a harness system for occupants. This study’s objective was to evaluate the demographics, mechanisms, injuries, and associated risk factors of ROV crashes. Methods A retrospective chart review was performed for patients of all ages with ROV-related injuries presenting to a Level 1 trauma center from 2004 to 2017. Cases were identified by ICD-9/10 codes and narrative searches. Person- and crash-related variables were examined in relation to injury outcomes including body area injured, injury severity score, and disposition (e.g. hospitalization, intensive care unit admission). Descriptive, bivariate (chi-square, Fishers exact test), and linear regression analyses were performed. Results Seventy-two patients with ROV-related injuries were identified. The number of injured patients increased over the study period (p < 0.01). Patients were 49% youth < 16 years old, 63% males, and 99% Caucasian. Half of the injured (51%) were passengers, with a higher proportion of youth being passengers (70%) as compared to adults (35%) (p < 0.01). Nearly one-third (30%) of crash victims < 16 years old were ROV drivers. Twenty-nine percent of all crashes occurred on roadways. Almost 40% of injured adults crashed at night, while all youth were injured during the day (p < 0.01). The primary crash mechanism was a rollover (67%). Only one patient was documented as being helmeted, and approximately one-fourth (24%) sustained head injuries and/or loss of consciousness. Other documented injuries included those to the face (20%), chest (22%), abdomen (11%), extremities (58%), and skin (51%). Over 90% of narratives were consistent with victims being unrestrained. Nearly three-fourths (74%) of victims were hospitalized and 26% required ICU care, one-half (53%) of these being children. Conclusions Although ROVs have ROPs, lack of helmet and safety belt use are reducing their benefit. Youth are a large proportion of those injured in ROV crashes, often while driving despite vehicle operation recommended only for those ≥16 years old. Increased public education is needed regarding proper safety measures while operating and riding ROVs.


Background
Recreational off-highway vehicles (ROVs) are off-road vehicles whose popularity has soared in recent years (U.S. Consumer Product Safety Commission, 2016). Consequently, ROV-related deaths and injuries have become an emerging safety issue, particularly among children (Linnaus et al., 2017;Richardson et al., 2018).
ROV types include sport models primarily for recreation, and multipurpose vehicles for both recreational and occupational use. All ROVs have maximum speeds of at least 30 miles per hour (mph), and most can travel at highway speeds (Wilson, 2015). A related off-road vehicle is the utility task vehicle (UTV). ROVs are often referred to as UTVs, but the latter have maximum speeds of < 25 mph (Wilson, 2015).
ROVs are designed solely for off-road use having low pressure tires with knobby treads, and a relatively narrow track and high clearance, which increases their risk of rollover. They also have a steering wheel for directional control, foot pedals for acceleration and braking, and are designed to carry more than one passenger (unlike most ATVs) with bench or bucket seating. ROVs have a rollover protective structure (ROPS) with seat belts or harness restraints for each seating position. In addition, in contrast to most ATVs that have all wheels traveling at the same speed due to a solid rear axle or locked differential, the differential in many ROVs may be selectively unlocked. This allows the outer wheels to rotate faster than the inner wheels when traveling around a corner, similar to a roadway vehicle, and makes the vehicle less likely to tear up turf such as lawns.
Decades of research have provided a wealth of information with regards to ATV crashes and injuries. In contrast, there are very few published studies describing ROV crash and injury epidemiology. Identifying the factors involved in ROV crashes and their associated injuries is vital for developing injury prevention strategies. The objective of this study was to evaluate ROV-related injuries among patients presenting to a Level 1 trauma center.

Identification of ROV crashes
The ICD codes include crashes of both ATVs and ROVs. To identify the latter, narratives were searched for the terms "ROV" and "UTV," vehicle makes/models, and mention of any of the following: steering wheel, foot pedals, seat belt, harness, front seat/back seat, more than 4 wheels, rollover protection structure (ROPS), roll cage, or roll bar.

Person-related variables
Demographic variables included patient sex, age, race, seating position (operator, passenger), and helmet and/ or seatbelt use. Ages were grouped as < 16 (youth) and > 16 years (adult). Other person-related variables were whether the operator tested positive or negative for alcohol use. Drug/medication use was defined as operators testing positive for illicit drugs (e.g. amphetamines) or reporting use of prescription or over the counter (OTC) medications.
For helmet use, the narrative was searched for the word "helmet." For seatbelts, the narrative was searched using variations on the word "seatbelt" (8 cases). In addition, there were 52 cases where the narrative was consistent with the victim being unrestrained, e.g. the mechanism included being ejected/falling off the vehicle or the victim was found outside/under the vehicle. The narrative was also used to determine whether there was adult supervision of youth operators.

Crash-related variables
Crash-related variables included the year, the season (coded as Winter (Dec-Feb), Spring (Mar-May), Summer (Jun-Aug), and Fall (Sep-Nov)), and the day of the week grouped as Weekday (Mon-Fri) or Weekend (Sat-Sun). Crash location was coded as on roadways or off-road based on location-related variables. Light conditions were coded as Day, Dawn (30 min before sunrise), Dusk (30 min before sunset), and Night. For bivariate analyses, comparisons were made for Day vs. Night.
The mechanism was coded as a collision with another motor vehicle or an object, or as a non-collision event, either an ejection/fall from the vehicle or a rollover. For bivariate comparisons, crashes were categorized as collision or non-collision events. Narratives were used to determine whether the victim was trapped/pinned under the vehicle. Crashes were also coded as whether they involved multiple victims. Vehicle use was a trauma record coded variable termed work-related. Estimated speeds were extracted from patient notes and grouped as < 20 mph and > 20 mph.
The maximum abbreviated injury severity (MAIS) score was used to determine whether there was an injury to the head, face, chest, abdomen, extremities, and/or external (skin). MAIS values of 0 were coded as "No" and > 0 as "Yes." A major trauma was defined as an overall injury severity score (ISS) of > 15. Other injuryrelated characteristics included whether the patient was admitted to the hospital, admitted to the intensive care unit (ICU), was endotracheally intubated and ventilated, and/or died because of their injuries.

Data analysis
Descriptive (frequency) and bivariate (chi-square, Fishers Exact test) analyses were performed using IBM SPSS Statistics. Missing data were not included in analysis. Statistical significance for bivariate analysis was defined as a two-sided p-value of < 0.05. Linear regression analysis of the data was performed to determine the trendline of the number of ROV victims over time. The β (slope), r 2 (correlation coefficient) and p-value were determined.

Person-related characteristics
From 2004 to 2017, 72 patients from 67 crashes presented to the trauma center with injuries as an operator or passenger in an ROV. Patient ages ranged from 4 to 88 years with 49% being youth < 16 years old, and 64% were ≤ 25 years old (Table 1). Almost two-thirds (63%) of patients were male, and 99% were Caucasian. Approximately half (48%) were operators and half (52%) passengers. Less than 10% were belted. Among those adults with alcohol test results (n = 23), more than twofifths (43%) tested positive. Only one patient (a 17-yearold male) was documented as wearing a helmet at the time of the crash. Where supervision of child operators < 16 years old was documented (7/10, 70%), two cases had adult supervision and five were unsupervised.

Crashes
The number of patients increased over the study period ( Fig. 1). The highest proportions of victims were in ROV crashes occurring during the summer months and on weekdays (Table 2). Nearly 30% of crashes were on public roads. Around one in five occurred at night. Twothirds of victims were injured in a rollover. The second most common mechanism was an ejection/fall from the vehicle. Crash narratives indicated almost two-fifths of victims had at least some part of their body pinned under the ROV, including the ROPS. Four patients were injured while working. Where speed was documented (36/67, 64%), 61% were injured in crashes at speeds of < 20 mph.

Outcomes
Around 10% of patients had an abnormal GCS score < 15 (Table 3). Additionally, six had severe brain injuries (GCS < 8), and 19 suffered a loss of consciousness. Based on MAIS scores, one-fourth of victims suffered a head injury and over half suffered an injury to the extremities and/or externally to the skin. Major trauma (ISS > 15) was observed in 13% of victims. Almost three-quarters of patients were admitted to the hospital, 26% were admitted to the ICU, and 10% were ventilated. About half of those admitted to the ICU (48%) had an ICU length   The sum of n for a variable may not equal the total Group N due to missing data of stay of three or more days. Three patients died (two adults, one youth 6 years old).

Males versus females
Comparisons by sex showed females were a higher proportion of youth victims < 16 years old than they were of those ≥16 years (p = 0.03) (Table 4). Additionally, relative to males, females were a smaller percentage of patients injured on weekends (p = 0.02). No other sex-dependent differences in variables were observed. The patterns and severity of injuries for males and females were also similar (Table 5), with about one-quarter suffering head injuries. Two-thirds of females and almost 80% of males were hospitalized for their injuries. Other results for sexbased comparisons included the observation that 90% of victims testing positive for alcohol were male. Three of the four persons suffering work-related injuries were also male.

Youth versus adults
A higher percentage of youth were passenger victims compared to those > 16 years (p < 0.01) (Table 4). Still, 30% of youth victims were operators of the vehicle. Additionally, compared to 39% of adult patients being injured at night, no victims < 16 years old were in nighttime crashes (p < 0.01). Seat belt use by both youth and adults was very low. Similarly to adults, youth were most often injured in the summer and during the week. More than a quarter of youth were injured in a roadway crash, and over 80% of their crashes involved a non-collision event. About a third of youth victims were pinned by the vehicle, and two-thirds of crashes occurred at estimated speeds ≤20 mph. Injury patterns among youth were similar to adults (Table 5). A quarter of youth suffered injuries to the head and/or face, with one-half or more having extremity or external (skin) injuries. Two-thirds of children were hospitalized. All occupational-related ROV crashes (n = 4) and patients testing positive for alcohol (n = 10) were adults.

Other comparisons
Except for the seating differences by age, no differences in crash characteristics and outcomes were observed when comparing operators and passengers. There was a higher percentage of victims testing positive for alcohol on weekends (6/8, 75%), as compared to weekdays (4/15, 27%) (p = 0.04). All eight victims with an abnormal GCS were unrestrained, were hospitalized in the ICU, and seven of the eight were mechanically ventilated. Those pinned by the vehicle were more commonly ventilated (5/26, 19%) than those who were not (1/45, 2%) (p = 0.02). Pinned victims were also more commonly admitted (23/26, 88%) than those not trapped by the vehicle (29/45, 64%) (p < 0.05).

Discussion
Our study characterized the demographics, crash mechanisms, and clinical outcomes for ROV crash victims Based on maximum abbreviated injury severity (MAIS) score; 0 = no injury, > 0 = injury presenting to a Level 1 trauma center over a 14-year study period. Overall, we found youth < 16 years old were approximately half (49%) of all victims, rollovers were the major crash mechanism, and personal protective equipment (PPE) was rarely used. Clinical outcomes from these crashes were also relatively severe, with the majority of patients being admitted to the hospital, including 26% to the ICU.

Recreational off-highway vehicles
Over the last decade, healthcare providers have noticed an increasing number of ROV-related injuries (Linnaus et al., 2017;Richardson et al., 2018). Between January 2003 and August 2016, 942 ROV crashes involving 665 fatalities and 843 injuries were reported by the U.S. Consumer Product Safety Commission (CPSC) (U.S. Consumer Product Safety Commission, 2016). Another  . Finally, researchers at a pediatric trauma center studying off-road vehicle crashes  (Linnaus et al., 2017).

Demographics
Almost two-thirds of patients in our study were males. This proportion is similar to the 2016 CPSC report (68%), the 2018 CPSC/FARS study (71%), and the newspaper article study (70%) (U.S. Consumer Product Safety Commission, 2016; Richardson et al., 2018;Jennissen et al., 2016). In contrast, the pediatric trauma center study reported males comprised only 55% of victims (Linnaus et al., 2017). All of these studies, including ours, found a lower proportion of males than is commonly seen for ATVs, where they are typically > 85% of victims (Denning et al., 2013a;Denning et al., 2013b;Denning et al., 2014;. ROVs are designed for multiple riders. Consistent with this, we found about half of those injured were passengers. The newspaper article study reported passengers were 37% of victims . In contrast to ROVs, almost all ATVs are designed for an operator only, and previous studies found passengers were about 15-22% of ATV-related injuries and deaths Denning et al., 2013a;Denning et al., 2013b;Denning et al., 2014;. These ATV studies found injured females were more commonly passengers than were males. This was also true for ROV crashes in the study using newspaper articles: males, 75% operators, 25% passengers; females, 43% operators, 57% passengers . We saw a similar pattern in our study (Table 4), but the differences did not reach statistical significance.
Children and adolescents appear to be a particularly vulnerable riding population. Nearly one-half (49%) of the ROV crash victims in our study were < 16 years old. This was similar to the newspaper article study showing youth were 44% of ROV crash victims . In both cases, youth were a considerably higher percentage of victims than that typically reported for ATVs (24-29%) Denning et al., 2014;Denning & Jennissen, 2018). A significantly higher proportion of injured children were passengers (70%) compared to adults (35%) in our study. Still, 30% of child crash victims were drivers of the ROV at the time of their injury. Children < 16 years old were drivers in 21% of crashes in reports collected by the CPSC where the age of the driver was known, and 14% of ROV drivers killed on public roads in the FARS database (Richardson et al., 2018).
Manufacturers warn consumers in the owner's manuals (Kawasaki Heavy Industries, 2012; Polaris Industries, n.d.-b) and with vehicle decals that adult-size ROVs should not be driven by anyone < 16 years old. Manufacturers also recommend that child passengers be at least 12 years old. They state all riders should be able to sit with their backs against the seat with both feet flat on the floor and grip the passenger hand hold (Kawasaki Heavy Industries L, 2012;Polaris Industries, n.d.-b). Unfortunately, many families do not adhere to these recommendations. Sometimes parents will put children in car seats or booster seats in ROVs, but these products have not been tested for use in these vehicles and manufacturers provide no guidance. Greater efforts are needed to educate consumers on the dangers of ROVs and provide safety information on their proper use.

Crash mechanism
Rollovers were the primary crash mechanism in our study. Other studies have found the majority of ROV crashes U.S. Consumer Product Safety Commission, 2014), like ATV crashes (Denning et al., 2014;Unni et al., 2012;Humphries et al., 2006), were rollovers. In addition, the CPSC study determined 68% were lateral rollovers with over one-half occurring during a turn. About 90% of the most severely injured had been in a rollover, and among rollover crashes for which terrain was known, more than half occurred on flat surfaces. The lateral stability and vehicle handling characteristics of ROVs have been a concern (U.S. Consumer Product Safety Commission, 2016;U.S. Consumer Product Safety Commission, 2014), and efforts to improve their safety engineering and design is very much needed.

Personal protective equipment
A major contributor to severe injury in ROV crashes is riding unrestrained. Only 8% of victims presenting to our trauma center appeared to be restrained, very similar to the 6% of victims documented as being restrained in the newspaper article study . In that same study, nearly three-quarters of those who died had been ejected from the vehicle . In reports using national data, only 14% of ROV occupants were known to be belted, and half to threequarters of victims who were injured or killed were not wearing a seat belt or harness system (Richardson et al., 2018;U.S. Consumer Product Safety Commission, 2014). The vast majority of these victims were partially or fully ejected from the vehicle. Among pediatric patients with ROV-related injuries, 70% were not restrained (Linnaus et al., 2017).
We found 37% of the victims had been pinned by the vehicle. Other studies have shown about one-half or more of those injured were struck or pinned U.S. Consumer Product Safety Commission, 2014). Nearly 30% of victims in reports collected by the CPSC were pinned by the vehicle, usually by the ROPS structure (Richardson et al., 2018). ROVs weigh up to a 1000 pounds. They can transmit significant forces to occupants who are ejected from the vehicle and cannot be easily lifted off a pinned victim. One of the most important factors in decreasing ROV-related injuries and deaths is for occupants to always wear the seat belt or harness system for each and every ride, and there should be no passengers who cannot be properly restrained.
Manufacturers and safety experts recommend the use of helmets when riding in ROVs. Only one patient seen at our trauma center was documented as wearing a helmet, and one-fourth of victims suffered a head injury. In a nine-state study of ROV crashes, three (3%) were reported as being helmeted , and < 15% of pediatric victims injured in ROV crashes were wearing helmets, both significantly lower percentages than helmet use on ATVs and dirt bikes (Linnaus et al., 2017). Only 2% of victims in the CPSC database were documented as being helmeted (Richardson et al., 2018).

Nighttime driving
Despite ROVs being equipped with headlights, two-fifths of adults were injured in nighttime crashes. This is similar to the proportion of all ROV crashes occurring in compromised light conditions in the newspaper article study . Appreciating terrain changes and identifying obstacles is inevitably more difficult at night. Nighttime ROV driving should be done with extreme caution and avoided if possible. Adult alcohol use is also likely to be playing a factor in many adult ROV crashes at night.

Roadways
Like ATVs, ROVs are designed for off-road use only and manufacturers strongly warn against roadway riding (Specialty Vehicle Institute of America, n.d.; Recreational Off-Highway Vehicle Association, n.d.-a). For both vehicles, the dangers of riding on the road are largely related to their fundamental vehicle design Denning & Jennissen, 2018). Despite the warnings, almost 30% of crashes in our study occurred on public roadways. This proportion is slightly lower than that found in other studies. Specifically, one-half of crashes in the newspaper articles from nine states  and 43% of crashes in the CPSC database (Richardson et al., 2018) were on the road. Interestingly, four-fifths of ROV crashes reported in newspapers from nine states  and 82% of fatal ROV crashes in the FARS (Richardson et al., 2018) occurring on public roads did not involve another motor vehicle. In other words, the vast majority were single vehicle crashes. Similarly, ATV crash studies have shown that over two-thirds of deaths and threefourths of injuries on public roadways do not involve another motorized vehicle (Denning et al., 2013a;Denning et al., 2013b). Because of their design, off-road vehicles including ROVs can have unpredictable interactions with roadway surfaces and should not be driven on public roads.

Limitations
Our study was retrospective, had a relatively limited sample size, and involves a single trauma center, which limits its generalizability to other populations. As potential cases for our study were determined by ICD coding, it is possible some ROV cases were not identified due to miscoding and/or missing narrative cues. In addition, narratives where patients were clearly ejected/fell from the vehicle were interpreted as their being unbelted. However, they could also represent improper belt use. Moreover, the need to identify ROVs and code other variables from narratives may introduce documentation bias. Finally, ED populations do not include all fatalities as those occurring pre-hospital are not included in the trauma database. Despite these limitations, ROV-related crashes and injuries are an emerging safety issue about which little has been published, and our findings contribute to this small body of knowledge.

Conclusions
Although ROVs have rollover protective structures, lack of adherence to manufacturer safety recommendations including helmet and safety belt use and refraining from driving on public roadways is likely reducing their benefit. Almost all ROVs are designed to be driven by those ≥16 years old, and manufacturers recommend passengers be 12 years or older. Our studies support the hypothesis that youth are a vulnerable ROV riding population. Laws as outlined in the Recreational Off-Highway Vehicle Association's model legislation (Recreational Off-Highway Vehicle Association, n.d.-b) should be adopted and strictly enforced. Increased public education and anticipatory guidance by healthcare providers on ROV safety are also greatly needed. Given the increasing popularity and prevalence of these vehicles, ROV crashes and injuries should be considered an emerging public health concern and multi-pronged strategies should be developed to prevent these serious injuries.