The use of all-terrain vehicles (ATVs) continues to grow in the USA. ATVs are defined as an off-road motorized vehicle with three or four low-pressure tires, a straddle seat, and handlebars. Compared with off-road motorcycles, ATVs are heavier and have high centers of gravity which create significant rollover risk. Engine sizes can reach 1000 cm3, allowing speeds of up to 80 miles per hour.

In the USA in 2005, there were 666 deaths attributable to ATV use, and an estimated 136 700 people received treatment for ATV-related injury in emergency departments (EDs). Children <16 years of age accounted for 22% of deaths and 30% of all ED visits.1 In 2004, an estimated 16 000 people were hospitalized due to ATV-related injuries, including over 4600 children <18 years of age.2 The economic impact of ATV deaths is immense, with direct and indirect costs estimated to be in excess of $3 billion in 2003.3 For people of all ages, ATV-related hospitalization charges exceeded US$370 million (£210 million; €270 million) in 2004, with ∼30% of patients hospitalized for ATV injuries having public insurance or no insurance.4 For children, ATV-related hospitalizations cost an estimated US$71 million (£40 million; €50 million) in 2004.4

Among motorcycle riders, helmets have been shown to reduce the risk of death by 42% and the risk of head injury by 69%.5 In a 1990 study, Rodgers6 assessed the effectiveness of helmets in reducing ATV-related deaths and head injuries using data from 1984–1985 and concluded that helmet use reduced the risk of death by 42% and head injury by 64%. In a comparison of morbidity between ATV and motorcycle injuries, Acosta and Rodriguez7 observed greater injury severity and a higher incidence of head and neck injuries among ATV patients than among motorcycle patients. As helmet use is generally more common among motorcycle riders than ATV riders, mainly because of mandatory motorcycle helmet laws in many states, this may account for the difference. Studies of ATV injuries often report low helmet use, in the range 8–50%.8^–15

The effectiveness of helmets in reducing ATV-related injuries has received minimal attention in recent years. The 1990 Rodgers study is dated, with 85% of ATVs in that study being of three-wheeled design, with a mean engine size of 200 cm3. Today’s four-wheeled ATVs often have much larger engines, weigh considerably more, and have different crash dynamics compared with three-wheelers. A more recent study of 109 ATV-related injuries at a pediatric trauma center reported greater injury severity and longer length of hospital stay for non-helmeted riders than helmeted riders, but low helmet usage (∼20%) limited the power of the analysis.16 Similarly, Murphy and Yanchar17 observed a strong relationship between unhelmeted ATV riding and increased risk of head and facial injuries. In a single-state study of pediatric ATV-related injuries, Gittelman and colleagues18 did not observe a difference in head or facial injuries between helmeted and unhelmeted children; however, this study relied on overall Injury Severity Score (ISS) rather than the more specific Abbreviated Injury Scale (AIS) by body region.

The purpose of this study was to compare injury patterns between helmeted and unhelmeted patients hospitalized due to ATV-related injuries. We examined whether lack of helmet use is associated with significant injuries to the head, neck, and face regions. We also assessed the likelihood of in-hospital death for helmeted versus unhelmeted ATV riders and the likelihood of arriving at the hospital with a depressed level of consciousness (ie, low Glasgow Coma Score (GCS)). Finally, we compared hospital utilization and procedures by helmet use.

METHODS

We obtained 2002–2006 data from the National Trauma Data Bank (NTDB), which is maintained by the American College of Surgeons. The NTDB contains information from over one million records voluntarily submitted from more than 400 designated trauma centers in the USA.19 Records in the NTDB include information on patient demographics, pre-existing conditions, injury characteristics including diagnoses and severity, prehospital care, ED care, outcomes, and financial data. The NTDB also includes hospital characteristics such as trauma designation level, bed size, teaching type, and public/private ownership. We included trauma patients of all ages who were hospitalized with an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) external cause-of-injury code (E-code) for an off-road motor vehicle accident. We included E821.0, E821.1, E821.8, and E821.9, as these correspond to injuries to the driver, a passenger, or another person injured in an off-road motor vehicle crash.220^–23 We excluded E821.3–E821.7, as these refer to passengers, occupants, or riders of motorcycles, streetcars, animals, or pedal cycles. Traumatic brain injury (TBI) was defined by either the presence of an ICD-9-CM diagnostic code of 800.0–801.9, 803.0–804.9, 850.0–854.1, or 959.01 per the Centers for Disease Control and Prevention framework or the presence of an AIS code between 113000 and 161100 for head injury, as defined by the Association for the Advancement of Automotive Medicine.2425 We restricted our study to patients admitted to the hospital and excluded patients discharged home from the ED. We excluded records with unknown or undocumented protective device use. We used 2002–2006 data, and our final 5-year dataset included 11 589 patients (mean slightly more than 2300 ATV-related trauma patients per year). Patients who were transferred from the hospital, either from the ED or after admission, were excluded, as we were unable to link records in the NTDB and because the NTDB does not typically include patients transferred from the ED. Including cases transferred after admission would artificially reduce the mortality for hospitals that transfer more patients. In addition, these patients may be included in the receiving hospital definitive care record.

Our main dependent variables were in-hospital mortality, documented TBI, ISS, and significant injury to the face or neck region. The primary independent variable was lack of helmet use (coded 1 if helmet = no, 0 if helmet = yes).

We included variables in multivariable analytical models based on an a priori assessment of clinical relevance. Indicator variables were created for hypotension (systolic blood pressure <90 mm Hg in ED) and receipt of a blood transfusion. Trauma patients who are hypotensive and/or receive blood transfusion tend to have higher mortality than normotensive patients and those not needing blood products; therefore we include these control variables in our models.2627 We also included gender, age, and geographic region (eg, Midwest, Northeast, South, and West). The NTDB includes the ISS for each record and the AIS score and code for each injury diagnosis. The overall ISS ranges from 1 (minor) to 75 (nearly always fatal).28 The AIS is an anatomically based system that classifies individual injuries by body region on a 6-point ordinal severity scale ranging from AIS 1 (minor) to AIS 6 (fatal). From the detailed 6-digit AIS codes, we created an indicator variable for the presence of a severe TBI, defined as AIS 4–6 injury to the head.

Univariate and bivariate analyses were completed using t tests for continuous variables and χ² testing for categorical variables. To identify significant associations between helmet use and in-hospital outcomes, we used logistic regression, with in-hospital death as the dependent variable and the independent and control variables described above. We performed separate logistic regressions with the following dependent variables: (1) any TBI, (2) severe TBI (AIS⩾4), and (3) moderate to severe injury (AIS⩾3) to each body region (head, face, neck, thorax, abdomen, spine, lower extremity, upper extremity). We adjusted for clustering on hospital identifier to account for correlation of individual outcomes within hospitals using the cluster option in Stata 10.0/MP. Odds ratios are presented for explanatory variables, with p values and confidence intervals.

RESULTS

We identified 11 589 hospitalizations for ATV injuries, including 248 (2.1%) patients who died during their hospital stay. About half of the hospitalized patients were wearing helmets at the time of injury. Table 1 presents relative frequencies of patient and injury characteristics and their relationship to helmet use. Overall, men represented 80.3% of ATV hospitalizations. Helmet use was less common among women, with 40.2% of women wearing helmets compared with 53.1% of men. Compared with helmeted patients, unhelmeted patients were slightly older (mean 27.7 vs 25.0 years, p<0.001) and more likely to be of a non-white race. Helmet use was most common in the Northeast (90.7%) and least common in the South (38.2%). The ISS and GCS were, on average, slightly worse for unhelmeted riders. Unhelmeted patients were significantly more likely to experience a TBI of any severity (40.3% vs 31.1%, p<.001) and nearly three times more likely to experience a severe TBI (AIS⩾4). No differences were observed for injuries to the thorax, abdomen, spine, or extremities.

Table 2 presents hospital treatment and outcomes for helmeted versus unhelmeted patients. There was no difference in the likelihood of receiving a CT scan of the head. However, unhelmeted patients were more likely to undergo a neurosurgical procedure (eg, craniotomy, skull procedure) compared with the helmeted group (OR = 2.60, p<0.001). Unhelmeted patients were more likely to require admission to the intensive care unit (OR = 1.44, p<0.001), and, if admitted, stayed longer on average (4.6 vs 3.8 days, p<0.001). The length of hospital stay tended to be longer for unhelmeted patients (4.8 vs 4.2 days, p<0.001), and they were more likely to be discharged to a post-acute care facility (ie, rehabilitation center, skilled nursing facility, nursing home). Compared with patients who had worn helmets, those who had not were nearly three times more likely to die during their hospital stay (OR 2.8, p<0.001).

Table 3 presents the results of logistic regression models evaluating the odds of receiving significant injuries (AIS⩾3) to each body region. Compared with helmeted persons, unhelmeted ATV riders were significantly more likely to receive a diagnosis of a TBI of any severity (OR = 1.62, 95% CI 1.49 to 1.76, p<0.001). Unhelmeted patients were three times more likely to have a moderate to severe TBI (OR = 2.99, 95% CI 2.30 to 3.89, p<0.001). Significant facial injuries were nearly twice as likely in unhelmeted ATV riders (OR = 1.94, 95% CI 1.28 to 2.94, p = 0.002), as were significant neck injuries (OR = 3.53, 95% CI 1.26 to 9.91, p = 0.017). Helmet use was not associated with any significant differences in injury to other body regions.

We also assessed the likelihood of arriving in the ED with a GCS of 3–8 and the likelihood of experiencing a severe TBI (AIS⩾4), controlling for patient characteristics. We found that unhelmeted ATV riders were more than twice as likely to have depressed levels of consciousness (GCS<9) than were helmeted riders (OR = 2.39, 95% CI 1.89 to 3.01, p<0.001). Unhelmeted ATV riders were also more than three times more likely to receive a severe TBI (OR = 3.19, 95% CI 2.39 to 4.25, p<0.001).

In table 4, we present the results of the logistic regression modeling in-hospital mortality for ATV injuries. Compared with helmeted patients, unhelmeted patients were much more likely to die (OR = 2.58, 95% CI 1.79 to 3.71, p<0.001). Female patients were less likely to die irrespective of helmet use. Injury severity, hypotension, and blood transfusion were all significantly associated with increased odds of an in-hospital death. We also performed the same logistic regression, limiting patients to children (age 0–19). As with adults, children not wearing helmets were much more likely to die than helmeted children (OR = 2.32, 95% CI 1.23 to 4.37, p = 0.005).

DISCUSSION

Despite the evidence in support of helmet use for motorcycle riders, lack of acceptance and highly variable, poorly enforced policy intervention has resulted in low helmet use among ATV riders. In 2006, 26 states required young ATV riders to wear helmets; 13 of these states also required helmet use for adults, and 15 of the 26 states required helmets to be worn by both drivers and passengers. Two of the state laws pertained only to three-wheeled ATVs.2930 To the best of our knowledge, this is the first study of helmet use and ATV injuries using the NTDB. The inclusion of physiological data and anatomical injury profiling in the NTDB allows more comprehensive adjustment for potential confounders. After adjusting for patient characteristics, we found a strong relationship between not wearing a helmet and significant injuries to the head, neck, and face regions. As one might expect, we did not observe differences in injuries to other body regions not protected by a helmet. Our results are contrary to the smaller study by Gittelman and colleagues18 that did not address severity of injury by body region, but instead relied on the overall ISS and the presence of any injury regardless of severity to a body region. With 11 589 ATV-injured patients, our study is well powered to address differences in significant injuries between helmeted and unhelmeted riders. Our findings are consistent with the work of Acosta and Rodriguez,7 who identified significant morbidity in ATV crashes and found greater likelihood of head injury in ATV crashes than in motorcycle crashes.

In this study, nearly 75% of people who died from ATV injuries had a diagnosis of a TBI. Although slightly over half of patients were wearing helmets, nearly three-quarters of the patients who died were not helmeted. After adjustment for overall injury severity to reflect multisystem injuries, unhelmeted ATV riders were more than 2.5 times more likely to die. This is consistent with a previous study by Helmkamp31 that showed that states without ATV safety requirements, such as helmet laws, have ATV-related death rates that are twice as high as states with ATV safety measures. Our study provides strong evidence in support of the requirement for helmet use for all ATV riders and for active enforcement of these regulations.

There are several limitations to our study. Firstly, we relied on AIS codes to control for TBI. Misclassification of TBI severity is possible through incomplete or inaccurate injury coding. Also, we cannot rule out the potential for measurement error in our dependent variable (mortality). Although we believe that our model adequately adjusts for patient and hospital characteristics, we cannot rule out the possibility of unmeasured covariates that could influence our results. We also do not have data on ATV riders who may have died at the scene, who may have walked away from the crash because of helmet usage, or who were treated for mild injuries at non-trauma centers. However, we believe that any effect of this omission would bias our results towards not demonstrating helmet effectiveness. We also limited our study to hospitalizations with documented protective device use/non-use which could potentially bias our results.32 As patients with missing or unknown protective device use were similar in age, gender, ISS, GCS, and mortality to our sample, we believe that excluding these cases would not significantly bias our results. We also did not have information about helmet type (eg, full vs open face), whether helmets were appropriately fastened, or the speed of the ATV at the time of the crash, all of which would be informative. Finally, ATV-related hospitalizations may have been undercounted, as noted in previous studies, because the ICD9-CM E-code 821 identified only non-traffic ATV injuries and not those that occurred in traffic or on paved surfaces. In fact, E821 may have identified some “other off-road” cases (eg, those related to golf carts, go carts, dirt bikes and dune buggies), but we believe that there were very few of these cases identified based on the 94% ATV specificity noted by the Consumer Product Safety Commission from their 1990–1998 fatality records. Although the NTDB provides useful data on more severely injured patients, a national injury surveillance system that captures both hospitalizations and protective device use is needed.

IMPLICATIONS FOR PREVENTION

ATVs are associated with a significant and increasing number of deaths and serious disabilities for both children and adults. Helmet use appears to substantially decrease the mortality and morbidity associated with ATV-related crashes. Helmet use is, however, abysmally low among current ATV users. Increasing the proportion of ATV riders who wear helmets would be an important step to reducing TBIs and other significant injuries to the head, face, and neck areas. Enforceable state-level policy to promote helmet use among ATV riders appears to be not only warranted, but critical, to significantly reduce the burden of ATV-related injuries.

• Download figure
• Open in new tab
• Download powerpoint