TY - JOUR
AU1 - MSPH, Michael C. Lo,
AU2 - MSOH, Robert P. Giffin,
AU3 - CISSP, Kraig A. Pakulski, MEd,
AU4 - GCCRA, W. Sumner Davis, PhD, MPH, MDiv, MSc, GCMFS,
AU5 - MPH, Stephen A. Bernstein, MD,
AU6 - MEd, Daniel V. Wise,
AB - ABSTRACT Objective: The high-mobility multipurpose wheeled vehicle (HMMWV) is a light military tactical vehicle. During Operation Iraqi Freedom and Operation Enduring Freedom, the U.S. Army modified the HMMWV into a combat vehicle by adding vehicle armor, which made the vehicle more difficult to control and more likely to roll over. Consequently, reports of fatal rollover accidents involving up-armored HMMWVs began to accumulate during the up-armoring period (August 2003 to April 2005). Furthermore, the lack of occupant restraint use prevalent in a predominantly young, male, and enlisted military population compounded the injuries resulting from these accidents. In this retrospective case series analysis, we describe the characteristics of U.S. Army HMMWV rollover accidents, occupants, and injuries reported worldwide from fiscal year 1992 to 2013 based on reported occupant restraint use. Methods: We conducted all analyses using Microsoft Excel 2010 and SAS version 9.1. Because this analysis does not constitute human subjects research, no institutional review board review was required. First, we obtained U.S. Army HMMWV accident records from the U.S. Army Combat Readiness Center, and selected those records indicating a HMMWV rollover had occurred. Next, we successively deduplicated the records at the accident, vehicle, occupant, and injury levels for descriptive analysis of characteristics at each level. For each occupant position, we calculated relative, attributable, and population attributable risks of nonfatal and fatal injury based on reported occupant restraint use. Finally, we analyzed body part injured and nature of injury to characterize the injury patterns that HMMWV occupants in each position sustained based on restraint use. We performed a χ2 test of homogeneity to assess differences in injury patterns between restrained and unrestrained occupants. Results: A total of 819 U.S. Army HMMWV rollover accidents worldwide were reported from October 1991 through May 2013 involving 821 HMMWVs and 1,395 occupants (828 nonfatally injured, 151 fatally injured, and 416 noninjured). Thirty-five percent of more severe (class A and B) accidents involved the M1114 up-armored variant, whereas 32% of less severe (class C and D) accidents involved the M998 nonarmored variant. Unrestrained occupants were 20% more likely to be nonfatally injured and 5.6 times more likely to be fatally injured than were restrained occupants. Among unrestrained occupants, restraint use could have potentially saved 82% of lives lost. Among all occupants involved in a HMMWV rollover, an estimated 56% of fatalities could have been prevented by restraint use. Unrestrained drivers and vehicle commanders had greater than expected torso injuries, while restrained vehicle commanders and passengers had greater than expected upper extremity injuries. Unrestrained drivers had greater than expected fractures, whereas restrained drivers and vehicle commanders had greater than expected sprains/strains. Conclusion: While reporting bias may exist, nevertheless these results show that occupant restraint use confers substantial life-saving protection to HMMWV occupants in rollover accidents. Therefore, commanders, safety officers, and peers should continue to promote and enforce restraint use consistently during all Army ground operations and training involving HMMWVs. Doing so will save Soldiers' lives in rollover accidents during the remaining years of the HMMWV program. INTRODUCTION An icon of the 1990–1991 Persian Gulf War, the high-mobility multipurpose wheeled vehicle (HMMWV), or “Humvee,” is a light military tactical vehicle that replaced its equally iconic predecessor, the M151 jeep.1 As one of the U.S. military's most versatile ground vehicles, the HMMWV has served as a cargo/troop carrier, shelter carrier, weapons carrier, missile launcher, and ambulance since its debut in October 1985.1 While the Army and Marine Corps are the principal users of the HMMWV in the U.S. military, the Navy and Air Force have also used it on a limited basis.1 During Operation Iraqi Freedom (OIF) and Operation Enduring Freedom (OEF), the U.S. Army modified the HMMWV into a combat vehicle.1 This required hardening of the HMMWV cab, which the U.S. Army began in August 2003 and completed in April 2005 by increasing production of new up-armored HMMWVs and retrofitting existing HMMWVs already in theater with add-on armor kits.1 While reducing the HMMWV's vulnerability to weapons attacks and increasing occupant survivability, the added weight of the vehicle armor altered the HMMWV's center of gravity, destabilizing the vehicle and making it more difficult to control and more likely to roll over.1,2 Consequently, reports of fatal rollover accidents involving up-armored HMMWVs began to accumulate during this up-armoring period.2 During calendar years 2003 and 2004, 42% of all Soldiers injured in HMMWV accidents in OIF and OEF were injured in a rollover accident, 26% of whom were fatally injured, according to an unpublished 2005 technical report by the U.S. Army Center for Health Promotion and Preventive Medicine (M. Canham-Chervak, S. Canada, K. Hauret, J. Hadley, B. Jones, unpublished data, 2005). The lack of occupant restraint use prevalent in a predominantly young, male, and enlisted military population3 further compounded HMMWV rollover accident injury morbidity and mortality. In the 1995–2008 Department of Defense surveys of health-related behaviors among active duty military personnel, male Army respondents 25 years of age or younger consistently reported the lowest rates of using seat belts “always” or “nearly always” when driving or riding in a motor vehicle.4,–8 Soldiers who did not wear seat belts were three times more likely to die in a HMMWV rollover accident in OIF and OEF during calendar years 2003 and 2004 than Soldiers who wore seat belts (M. Canham-Chervak, S. Canada, K. Hauret, J. Hadley, B. Jones, unpublished data, 2005). Rollovers are especially hazardous to turret gunners, who are often unable to retract themselves within an overturning vehicle (especially if unrestrained), and are consequently either ejected or pinned underneath the vehicle.9 Although mine-resistant ambush-protected (MRAP) vehicles eventually replaced HMMWVs in combat roles in OIF and OEF, the Army National Guard still uses HMMWVs for missions providing homeland defense and logistical support.10 Thus, with approximately 150,000 HMMWVs remaining in the Army inventory, HMMWV rollover accidents will continue to be a concern until the new joint light tactical vehicle (JLTV) completely phases them out.10,11 As the HMMWV surpasses 30 years of continuous Army service, we take this opportunity to look back on its rollover accident, fatality, and nonfatal injury history, with the goal of preventing further rollover accident fatalities during the remaining years of the HMMWV program.10,11 To describe these accidents, fatalities, and nonfatal injuries, we analyzed a retrospective case series of U.S. Army HMMWV rollover accidents reported to the U.S. Army Combat Readiness Center (USACRC) at Fort Rucker, Alabama.12 U.S. Army accident investigators report ground accidents using Department of the Army (DA) Form 285 series forms to capture circumstantial information on the accident; property, materiel, and personnel involved; and the severity of damages and/or injuries sustained, if any.12 Using these accident reports, our objective is to describe the following: (1) HMMWV rollover accident characteristics, (2) HMMWV occupant characteristics involved in these accidents, and (3) the fatal and nonfatal injuries sustained by these occupants based on reported restraint use. We did not include in our analysis rollover events resulting from enemy action, such as detonation of enemy-deployed improvised explosive devices, because those are combat events, not accidents.12 MATERIALS AND METHODS Because this analysis does not constitute human subjects research, no institutional review board review was required. Under a data use agreement between the USACRC and the U.S. Army Aeromedical Research Laboratory, we downloaded anonymized records of U.S. Army HMMWV accidents occurring worldwide from fiscal year (FY) 1989 through 2013 from the USACRC's Risk Management Information System (RMIS) into Microsoft Excel 2010 spreadsheets. RMIS is a database containing all U.S. Army accidents reported to the USACRC. After standardizing data elements with different formats to aggregate records across multiple years, we then analyzed the records using SAS version 9.1 (SAS Institute, Cary, NC). We applied two filters to obtain the final dataset used in our analysis: first, we selected records indicating a HMMWV was involved in the accident, and then we selected records indicating a rollover occurred in the accident. Next, we successively deduplicated the selected records at the accident, vehicle, occupant, and injury levels for descriptive analysis of characteristics at each level. In this manner, we obtained the deduplicated frequencies of rollover accidents, nonfatally injured occupants, and fatally injured occupants, and then plotted them by FY in a line chart using Microsoft Excel 2010. We analyzed accident and vehicle characteristics by accident class (A, B, C, or D). U.S. Army accident investigators categorize accidents based on severity as one of the following: class A, accident with property damage of $1 million (increased to $2 million in 2010) or more, or injury resulting in fatality or permanent total disability; class B, accident with property damage of $200,000 (increased to $500,000 in 2010) or more but less than $1 million (increased to $2 million in 2010), or injury resulting in permanent partial disability or inpatient hospitalizations of three or more personnel; class C, accident with property damage of $20,000 (increased to $50,000 in 2010) or more but less than $200,000 (increased to $500,000 in 2010), or injury resulting in lost work time; or class D, accident with property damage of $2,000 or more but less than $20,000 (increased to $50,000 in 2010), or injury resulting in restricted work activity.12 We identified occupant positions as driver, supervisory (vehicle commander), turret gunner, or passenger, and then analyzed occupant characteristics by occupant position within the vehicle. For each occupant position, we calculated the incidence rate per 1,000 accidents per year of being nonfatally or fatally injured in HMMWV rollover accidents occurring from October 1991 through May 2013 based on restraint use. We calculated risk ratios (RRs) with 95% confidence intervals (CIs) to determine the relative risk13 of being nonfatally or fatally injured based on restraint use. We also calculated the attributable risk (AR) and attributable risk percent (ARP)13 to determine the excess risk and percentage, respectively, of being nonfatally or fatally injured due to lack of restraint use among unrestrained occupants compared to restrained occupants. We then calculated the population attributable risk (PAR) and population attributable risk percent (PARP)13 to determine the excess risk and percentage, respectively, of being nonfatally or fatally injured due to lack of restraint use among unrestrained occupants compared to the entire analysis population of unrestrained and restrained occupants. Finally, we analyzed injuries using the body region injured and nature of injury categories of the Barell injury diagnosis matrix, developed by Barell et al.14 for analyzing injury data by cross-tabulating body region injured by nature of injury in a two-dimensional matrix. However, because of small cell sizes, when cross-tabulating body region injured by nature of injury, we analyzed instead the marginal totals of body region injured and nature of injury separately, stratified by occupant position and restraint use. We performed a χ2 test of homogeneity to assess differences in body region injured and nature of injury between restrained and unrestrained occupants in each position. As an intermediate step in calculating the χ2 test statistic, we calculated expected injury counts from the marginal totals of observed injury counts. RESULTS A total of 819 U.S. Army HMMWV rollover accidents worldwide were reported to the USACRC over 22 years from October 1991 through May 2013 involving 821 HMMWVs and 1,395 occupants (828 nonfatally injured, 151 fatally injured, and 416 noninjured). Figure 1 shows FY 2005 was the peak year for these accidents and the nonfatally and fatally injured occupants involved. This peak coincides with the completion of the U.S. Army's HMMWV up-armoring program in OIF in April 2005.1 FIGURE 1 View largeDownload slide U.S. Army HMMWV rollover accidents, nonfatally injured, and fatally injured, 1992–2013, by FY. FIGURE 1 View largeDownload slide U.S. Army HMMWV rollover accidents, nonfatally injured, and fatally injured, 1992–2013, by FY. Table I shows the characteristics of these rollover accidents by accident class. Fifty-two percent of class A accidents and 54% of class B accidents occurred in Iraq, whereas 51% of class C accidents and 49% of class D accidents occurred in the United States. Sixty-five percent of class A accidents and 56% of class B accidents occurred in combat areas of operation, while 45% of class C accidents and 43% of class D accidents were tactical training accidents. Thirty-five percent of both class A and class B accidents involved the M1114 up-armored variant, whereas 31% of class C accidents and 32% of class D accidents involved the M998 nonarmored variant. These findings distinguish more severe (classes A and B) rollover accidents from less severe (classes C and D) rollover accidents. TABLE I Characteristics of U.S. Army HMMWV Rollover Accidents, 1992–2013, by Accident Class Characteristic  Accident Class*  Total (%)  A (%)  B (%)  C (%)  D (%)  Total Accidents (%)  819 (100)  139 (17)  48 (6)  404 (49)  228 (28)  Country   United States  371 (45)  40 (29)  13 (27)  207 (51)  111 (49)   Iraq  285 (35)  73 (52)  26 (54)  124 (31)  62 (27)   Germany  45 (5)  4 (3)  3 (6)  19 (5)  19 (8)   South Korea  34 (4)  1 (1)  2 (4)  19 (5)  12 (5)   Afghanistan  29 (4)  15 (11)  1 (2)  12 (3)  1 (<1)   Yugoslavia/Serbia  22 (3)  1 (1)  1 (2)  7 (2)  13 (6)   Kuwait  16 (2)  4 (3)  2 (4)  7 (2)  3 (1)   Other  17 (2)  1 (1)  —  9 (2)  7 (3)  Accident Setting   Combat Accident  302 (37)  91 (65)  27 (56)  130 (32)  54 (24)   Tactical Training  317 (39)  29 (21)  8 (17)  182 (45)  98 (43)   Other Mission  200 (24)  19 (14)  13 (27)  92 (23)  76 (33)  HMMWV Variant   M998 (Utility)  228 (28)  21 (15)  7 (15)  126 (31)  74 (32)   M1114 (Expanded Capacity, Up-Armored)  182 (22)  49 (35)  17 (35)  73 (18)  43 (19)   M1151 (Enhanced Armament Carrier)  94 (11)  22 (16)  9 (19)  41 (10)  22 (10)   M1025 (Utility, Armament)  90 (11)  15 (11)  3 (6)  47 (12)  25 (11)   M1097 (Utility)  33 (4)  1 (1)  2 (4)  19 (5)  11 (5)   M997 (Ambulance, 4 litter)  32 (4)  2 (1)  1 (2)  21 (5)  8 (4)   M966 (Weapons Carrier)  23 (3)  6 (4)  —  9 (2)  8 (4)   M1037 (S280 Shelter Carrier w/o Winch)  18 (2)  6 (4)  1 (2)  8 (2)  3 (1)   M1038 (Cargo/Troop Carrier With Winch)  16 (2)  1 (1)  —  13 (3)  2 (1)   M1026 (Armament Carrier, Armored)  13 (2)  3 (2)  —  5 (1)  5 (2)   Other/Not Specified  92 (11)  14 (10)  8 (17)  43 (11)  27 (12)  Human Error   Yes  539 (66)  86 (62)  26 (54)  261 (65)  166 (73)   No  280 (34)  53 (38)  22 (46)  143 (35)  62 (27)  Environment Contributed   Yes  357 (44)  28 (20)  15 (31)  206 (51)  108 (47)   No  462 (56)  111 (80)  33 (69)  198 (49)  120 (53)  Materiel Failure   Yes  23 (3)  4 (3)  2 (4)  14 (3)  3 (1)   No  796 (97)  135 (97)  46 (96)  390 (97)  225 (99)  Total Occupants Not Injured (%)  416 (100)  48 (11)  28 (7)  166 (40)  174 (42)  Total Occupants Nonfatally Injured (%)  828 (100)  165 (20)  101 (12)  451 (55)  111 (13)  Total Fatalities (%)  151 (100)  151 (100)  —  —  —  Characteristic  Accident Class*  Total (%)  A (%)  B (%)  C (%)  D (%)  Total Accidents (%)  819 (100)  139 (17)  48 (6)  404 (49)  228 (28)  Country   United States  371 (45)  40 (29)  13 (27)  207 (51)  111 (49)   Iraq  285 (35)  73 (52)  26 (54)  124 (31)  62 (27)   Germany  45 (5)  4 (3)  3 (6)  19 (5)  19 (8)   South Korea  34 (4)  1 (1)  2 (4)  19 (5)  12 (5)   Afghanistan  29 (4)  15 (11)  1 (2)  12 (3)  1 (<1)   Yugoslavia/Serbia  22 (3)  1 (1)  1 (2)  7 (2)  13 (6)   Kuwait  16 (2)  4 (3)  2 (4)  7 (2)  3 (1)   Other  17 (2)  1 (1)  —  9 (2)  7 (3)  Accident Setting   Combat Accident  302 (37)  91 (65)  27 (56)  130 (32)  54 (24)   Tactical Training  317 (39)  29 (21)  8 (17)  182 (45)  98 (43)   Other Mission  200 (24)  19 (14)  13 (27)  92 (23)  76 (33)  HMMWV Variant   M998 (Utility)  228 (28)  21 (15)  7 (15)  126 (31)  74 (32)   M1114 (Expanded Capacity, Up-Armored)  182 (22)  49 (35)  17 (35)  73 (18)  43 (19)   M1151 (Enhanced Armament Carrier)  94 (11)  22 (16)  9 (19)  41 (10)  22 (10)   M1025 (Utility, Armament)  90 (11)  15 (11)  3 (6)  47 (12)  25 (11)   M1097 (Utility)  33 (4)  1 (1)  2 (4)  19 (5)  11 (5)   M997 (Ambulance, 4 litter)  32 (4)  2 (1)  1 (2)  21 (5)  8 (4)   M966 (Weapons Carrier)  23 (3)  6 (4)  —  9 (2)  8 (4)   M1037 (S280 Shelter Carrier w/o Winch)  18 (2)  6 (4)  1 (2)  8 (2)  3 (1)   M1038 (Cargo/Troop Carrier With Winch)  16 (2)  1 (1)  —  13 (3)  2 (1)   M1026 (Armament Carrier, Armored)  13 (2)  3 (2)  —  5 (1)  5 (2)   Other/Not Specified  92 (11)  14 (10)  8 (17)  43 (11)  27 (12)  Human Error   Yes  539 (66)  86 (62)  26 (54)  261 (65)  166 (73)   No  280 (34)  53 (38)  22 (46)  143 (35)  62 (27)  Environment Contributed   Yes  357 (44)  28 (20)  15 (31)  206 (51)  108 (47)   No  462 (56)  111 (80)  33 (69)  198 (49)  120 (53)  Materiel Failure   Yes  23 (3)  4 (3)  2 (4)  14 (3)  3 (1)   No  796 (97)  135 (97)  46 (96)  390 (97)  225 (99)  Total Occupants Not Injured (%)  416 (100)  48 (11)  28 (7)  166 (40)  174 (42)  Total Occupants Nonfatally Injured (%)  828 (100)  165 (20)  101 (12)  451 (55)  111 (13)  Total Fatalities (%)  151 (100)  151 (100)  —  —  —  * Defined as the following: class A, accident with property damage of $1 million (increased to $2 million in 2010) or more, or injury resulting in fatality or permanent total disability; class B, accident with property damage of $200,000 (increased to $500,000 in 2010) or more but less than $1 million (increased to $2 million in 2010), or injury resulting in permanent partial disability or inpatient hospitalizations of three or more personnel; class C, accident with property damage of $20,000 (increased to $50,000 in 2010) or more but less than $200,000 (increased to $500,000 in 2010), or injury resulting in lost work time; or class D, accident with property damage of $2,000 or more but less than $20,000 (increased to $50,000 in 2010), or injury resulting in restricted work activity.12 View Large TABLE I Characteristics of U.S. Army HMMWV Rollover Accidents, 1992–2013, by Accident Class Characteristic  Accident Class*  Total (%)  A (%)  B (%)  C (%)  D (%)  Total Accidents (%)  819 (100)  139 (17)  48 (6)  404 (49)  228 (28)  Country   United States  371 (45)  40 (29)  13 (27)  207 (51)  111 (49)   Iraq  285 (35)  73 (52)  26 (54)  124 (31)  62 (27)   Germany  45 (5)  4 (3)  3 (6)  19 (5)  19 (8)   South Korea  34 (4)  1 (1)  2 (4)  19 (5)  12 (5)   Afghanistan  29 (4)  15 (11)  1 (2)  12 (3)  1 (<1)   Yugoslavia/Serbia  22 (3)  1 (1)  1 (2)  7 (2)  13 (6)   Kuwait  16 (2)  4 (3)  2 (4)  7 (2)  3 (1)   Other  17 (2)  1 (1)  —  9 (2)  7 (3)  Accident Setting   Combat Accident  302 (37)  91 (65)  27 (56)  130 (32)  54 (24)   Tactical Training  317 (39)  29 (21)  8 (17)  182 (45)  98 (43)   Other Mission  200 (24)  19 (14)  13 (27)  92 (23)  76 (33)  HMMWV Variant   M998 (Utility)  228 (28)  21 (15)  7 (15)  126 (31)  74 (32)   M1114 (Expanded Capacity, Up-Armored)  182 (22)  49 (35)  17 (35)  73 (18)  43 (19)   M1151 (Enhanced Armament Carrier)  94 (11)  22 (16)  9 (19)  41 (10)  22 (10)   M1025 (Utility, Armament)  90 (11)  15 (11)  3 (6)  47 (12)  25 (11)   M1097 (Utility)  33 (4)  1 (1)  2 (4)  19 (5)  11 (5)   M997 (Ambulance, 4 litter)  32 (4)  2 (1)  1 (2)  21 (5)  8 (4)   M966 (Weapons Carrier)  23 (3)  6 (4)  —  9 (2)  8 (4)   M1037 (S280 Shelter Carrier w/o Winch)  18 (2)  6 (4)  1 (2)  8 (2)  3 (1)   M1038 (Cargo/Troop Carrier With Winch)  16 (2)  1 (1)  —  13 (3)  2 (1)   M1026 (Armament Carrier, Armored)  13 (2)  3 (2)  —  5 (1)  5 (2)   Other/Not Specified  92 (11)  14 (10)  8 (17)  43 (11)  27 (12)  Human Error   Yes  539 (66)  86 (62)  26 (54)  261 (65)  166 (73)   No  280 (34)  53 (38)  22 (46)  143 (35)  62 (27)  Environment Contributed   Yes  357 (44)  28 (20)  15 (31)  206 (51)  108 (47)   No  462 (56)  111 (80)  33 (69)  198 (49)  120 (53)  Materiel Failure   Yes  23 (3)  4 (3)  2 (4)  14 (3)  3 (1)   No  796 (97)  135 (97)  46 (96)  390 (97)  225 (99)  Total Occupants Not Injured (%)  416 (100)  48 (11)  28 (7)  166 (40)  174 (42)  Total Occupants Nonfatally Injured (%)  828 (100)  165 (20)  101 (12)  451 (55)  111 (13)  Total Fatalities (%)  151 (100)  151 (100)  —  —  —  Characteristic  Accident Class*  Total (%)  A (%)  B (%)  C (%)  D (%)  Total Accidents (%)  819 (100)  139 (17)  48 (6)  404 (49)  228 (28)  Country   United States  371 (45)  40 (29)  13 (27)  207 (51)  111 (49)   Iraq  285 (35)  73 (52)  26 (54)  124 (31)  62 (27)   Germany  45 (5)  4 (3)  3 (6)  19 (5)  19 (8)   South Korea  34 (4)  1 (1)  2 (4)  19 (5)  12 (5)   Afghanistan  29 (4)  15 (11)  1 (2)  12 (3)  1 (<1)   Yugoslavia/Serbia  22 (3)  1 (1)  1 (2)  7 (2)  13 (6)   Kuwait  16 (2)  4 (3)  2 (4)  7 (2)  3 (1)   Other  17 (2)  1 (1)  —  9 (2)  7 (3)  Accident Setting   Combat Accident  302 (37)  91 (65)  27 (56)  130 (32)  54 (24)   Tactical Training  317 (39)  29 (21)  8 (17)  182 (45)  98 (43)   Other Mission  200 (24)  19 (14)  13 (27)  92 (23)  76 (33)  HMMWV Variant   M998 (Utility)  228 (28)  21 (15)  7 (15)  126 (31)  74 (32)   M1114 (Expanded Capacity, Up-Armored)  182 (22)  49 (35)  17 (35)  73 (18)  43 (19)   M1151 (Enhanced Armament Carrier)  94 (11)  22 (16)  9 (19)  41 (10)  22 (10)   M1025 (Utility, Armament)  90 (11)  15 (11)  3 (6)  47 (12)  25 (11)   M1097 (Utility)  33 (4)  1 (1)  2 (4)  19 (5)  11 (5)   M997 (Ambulance, 4 litter)  32 (4)  2 (1)  1 (2)  21 (5)  8 (4)   M966 (Weapons Carrier)  23 (3)  6 (4)  —  9 (2)  8 (4)   M1037 (S280 Shelter Carrier w/o Winch)  18 (2)  6 (4)  1 (2)  8 (2)  3 (1)   M1038 (Cargo/Troop Carrier With Winch)  16 (2)  1 (1)  —  13 (3)  2 (1)   M1026 (Armament Carrier, Armored)  13 (2)  3 (2)  —  5 (1)  5 (2)   Other/Not Specified  92 (11)  14 (10)  8 (17)  43 (11)  27 (12)  Human Error   Yes  539 (66)  86 (62)  26 (54)  261 (65)  166 (73)   No  280 (34)  53 (38)  22 (46)  143 (35)  62 (27)  Environment Contributed   Yes  357 (44)  28 (20)  15 (31)  206 (51)  108 (47)   No  462 (56)  111 (80)  33 (69)  198 (49)  120 (53)  Materiel Failure   Yes  23 (3)  4 (3)  2 (4)  14 (3)  3 (1)   No  796 (97)  135 (97)  46 (96)  390 (97)  225 (99)  Total Occupants Not Injured (%)  416 (100)  48 (11)  28 (7)  166 (40)  174 (42)  Total Occupants Nonfatally Injured (%)  828 (100)  165 (20)  101 (12)  451 (55)  111 (13)  Total Fatalities (%)  151 (100)  151 (100)  —  —  —  * Defined as the following: class A, accident with property damage of $1 million (increased to $2 million in 2010) or more, or injury resulting in fatality or permanent total disability; class B, accident with property damage of $200,000 (increased to $500,000 in 2010) or more but less than $1 million (increased to $2 million in 2010), or injury resulting in permanent partial disability or inpatient hospitalizations of three or more personnel; class C, accident with property damage of $20,000 (increased to $50,000 in 2010) or more but less than $200,000 (increased to $500,000 in 2010), or injury resulting in lost work time; or class D, accident with property damage of $2,000 or more but less than $20,000 (increased to $50,000 in 2010), or injury resulting in restricted work activity.12 View Large Table II shows the characteristics of U.S. Army HMMWV occupants involved in rollover accidents by occupant position. Occupants were predominantly young (mean age = 25.5 years), male (87%), and enlisted (90%), although varying by position. Reported restraint use varied from a low of 33% among turret gunners to a high of 65% among drivers. Not surprisingly, unrestrained turret gunners had the highest ejection rate of all occupant positions (6/35 = 17%). Thirty-five percent of fatalities were turret gunners, the highest of any occupant position. TABLE II Characteristics of U.S. Army HMMWV Occupants in Rollover Accidents, 1992–2013, by Occupant Position Characteristic  Occupant Position  Total (%)  Driver (%)  Vehicle Commander (%)  Turret Gunner (%)  Passenger (%)  Total Occupants (%)  1,395 (100)  781 (56)  192 (14)  135 (10)  287 (21)  Age (Years)   Mean  25.5  24.4  29.1  25.0  26.6   Range  18–61  18–57  19–54  18–51  18–61  Sex   Male  1,210 (87)  662 (85)  182 (95)  129 (96)  237 (83)   Female  104 (7)  64 (8)  7 (4)  3 (2)  30 (10)   Unknown/Not Reported  81 (6)  55 (7)  3 (1)  3 (2)  20 (7)  Rank   Enlisted  1,256 (90)  714 (91)  168 (88)  130 (96)  244 (85)   Officer  54 (4)  15 (2)  20 (10)  2 (2)  17 (6)   Other  85 (6)  52 (7)  4 (2)  3 (2)  26 (9)  Restraints Used   Yes  794 (57)  511 (65)  114 (59)  44 (33)  125 (44)   No  298 (21)  132 (17)  54 (28)  33 (24)  79 (28)   Not Reported  303 (22)  138 (18)  24 (13)  58 (43)  83 (29)  Ejected   Restrained  21 (30)  9 (30)  2 (33)  6 (43)  4 (20)   Unrestrained  41 (59)  21 (70)  3 (50)  6 (43)  11 (55)   Not Reported  8 (11)  —  1 (17)  2 (14)  5 (25)  Not Injured (%)  416 (100)  334 (80)  41 (10)  7 (2)  34 (8)  Nonfatally Injured (%)  828 (100)  402 (49)  133 (16)  75 (9)  218 (26)  Fatalities (%)  151 (100)  45 (30)  18 (12)  53 (35)  35 (23)  Characteristic  Occupant Position  Total (%)  Driver (%)  Vehicle Commander (%)  Turret Gunner (%)  Passenger (%)  Total Occupants (%)  1,395 (100)  781 (56)  192 (14)  135 (10)  287 (21)  Age (Years)   Mean  25.5  24.4  29.1  25.0  26.6   Range  18–61  18–57  19–54  18–51  18–61  Sex   Male  1,210 (87)  662 (85)  182 (95)  129 (96)  237 (83)   Female  104 (7)  64 (8)  7 (4)  3 (2)  30 (10)   Unknown/Not Reported  81 (6)  55 (7)  3 (1)  3 (2)  20 (7)  Rank   Enlisted  1,256 (90)  714 (91)  168 (88)  130 (96)  244 (85)   Officer  54 (4)  15 (2)  20 (10)  2 (2)  17 (6)   Other  85 (6)  52 (7)  4 (2)  3 (2)  26 (9)  Restraints Used   Yes  794 (57)  511 (65)  114 (59)  44 (33)  125 (44)   No  298 (21)  132 (17)  54 (28)  33 (24)  79 (28)   Not Reported  303 (22)  138 (18)  24 (13)  58 (43)  83 (29)  Ejected   Restrained  21 (30)  9 (30)  2 (33)  6 (43)  4 (20)   Unrestrained  41 (59)  21 (70)  3 (50)  6 (43)  11 (55)   Not Reported  8 (11)  —  1 (17)  2 (14)  5 (25)  Not Injured (%)  416 (100)  334 (80)  41 (10)  7 (2)  34 (8)  Nonfatally Injured (%)  828 (100)  402 (49)  133 (16)  75 (9)  218 (26)  Fatalities (%)  151 (100)  45 (30)  18 (12)  53 (35)  35 (23)  View Large TABLE II Characteristics of U.S. Army HMMWV Occupants in Rollover Accidents, 1992–2013, by Occupant Position Characteristic  Occupant Position  Total (%)  Driver (%)  Vehicle Commander (%)  Turret Gunner (%)  Passenger (%)  Total Occupants (%)  1,395 (100)  781 (56)  192 (14)  135 (10)  287 (21)  Age (Years)   Mean  25.5  24.4  29.1  25.0  26.6   Range  18–61  18–57  19–54  18–51  18–61  Sex   Male  1,210 (87)  662 (85)  182 (95)  129 (96)  237 (83)   Female  104 (7)  64 (8)  7 (4)  3 (2)  30 (10)   Unknown/Not Reported  81 (6)  55 (7)  3 (1)  3 (2)  20 (7)  Rank   Enlisted  1,256 (90)  714 (91)  168 (88)  130 (96)  244 (85)   Officer  54 (4)  15 (2)  20 (10)  2 (2)  17 (6)   Other  85 (6)  52 (7)  4 (2)  3 (2)  26 (9)  Restraints Used   Yes  794 (57)  511 (65)  114 (59)  44 (33)  125 (44)   No  298 (21)  132 (17)  54 (28)  33 (24)  79 (28)   Not Reported  303 (22)  138 (18)  24 (13)  58 (43)  83 (29)  Ejected   Restrained  21 (30)  9 (30)  2 (33)  6 (43)  4 (20)   Unrestrained  41 (59)  21 (70)  3 (50)  6 (43)  11 (55)   Not Reported  8 (11)  —  1 (17)  2 (14)  5 (25)  Not Injured (%)  416 (100)  334 (80)  41 (10)  7 (2)  34 (8)  Nonfatally Injured (%)  828 (100)  402 (49)  133 (16)  75 (9)  218 (26)  Fatalities (%)  151 (100)  45 (30)  18 (12)  53 (35)  35 (23)  Characteristic  Occupant Position  Total (%)  Driver (%)  Vehicle Commander (%)  Turret Gunner (%)  Passenger (%)  Total Occupants (%)  1,395 (100)  781 (56)  192 (14)  135 (10)  287 (21)  Age (Years)   Mean  25.5  24.4  29.1  25.0  26.6   Range  18–61  18–57  19–54  18–51  18–61  Sex   Male  1,210 (87)  662 (85)  182 (95)  129 (96)  237 (83)   Female  104 (7)  64 (8)  7 (4)  3 (2)  30 (10)   Unknown/Not Reported  81 (6)  55 (7)  3 (1)  3 (2)  20 (7)  Rank   Enlisted  1,256 (90)  714 (91)  168 (88)  130 (96)  244 (85)   Officer  54 (4)  15 (2)  20 (10)  2 (2)  17 (6)   Other  85 (6)  52 (7)  4 (2)  3 (2)  26 (9)  Restraints Used   Yes  794 (57)  511 (65)  114 (59)  44 (33)  125 (44)   No  298 (21)  132 (17)  54 (28)  33 (24)  79 (28)   Not Reported  303 (22)  138 (18)  24 (13)  58 (43)  83 (29)  Ejected   Restrained  21 (30)  9 (30)  2 (33)  6 (43)  4 (20)   Unrestrained  41 (59)  21 (70)  3 (50)  6 (43)  11 (55)   Not Reported  8 (11)  —  1 (17)  2 (14)  5 (25)  Not Injured (%)  416 (100)  334 (80)  41 (10)  7 (2)  34 (8)  Nonfatally Injured (%)  828 (100)  402 (49)  133 (16)  75 (9)  218 (26)  Fatalities (%)  151 (100)  45 (30)  18 (12)  53 (35)  35 (23)  View Large Table III shows injury severity outcomes of U.S. Army HMMWV occupants injured in rollover accidents by occupant position and restraint use. U.S. Army accident investigators categorize injury severity outcomes in a continuum ranging from “No Injury” to “Fatality.”12 In general, in each successively greater injury severity outcome, the percentage of restrained occupants decreased while the percentage of unrestrained occupants increased. This trend was observed across all occupant positions. To show this trend more clearly, only occupants with restraint use or nonuse reported are shown. TABLE III Injury Severity Outcomes of U.S. Army HMMWV Occupants Injured in Rollover Accidents, 1992–2013, by Occupant Position and Reported Restraint Use Occupant Position  Injury Severity Outcome  Total (%)  No Injury (%)  First Aid (%)  Medical Treatment Beyond First Aid (%)  Workday of Restricted Activity (%)  Lost Workday Case (%)  Permanent Partial Disability (%)  Permanent Total Disability (%)  Fatality (%)  All Positions (N = 1,092)   Restrained  794 (73)  261 (84)  169 (82)  7 (100)  50 (70)  266 (70)  5 (36)  4 (50)  32 (32)   Unrestrained  298 (27)  48 (16)  36 (18)  —  21 (30)  113 (30)  9 (64)  4 (50)  67 (68)  Driver (n = 643)   Restrained  511 (79)  223 (89)  109 (85)  4 (100)  34 (83)  126 (70)  1 (100)  1 (33)  13 (36)   Unrestrained  132 (21)  28 (11)  19 (15)  —  7 (17)  53 (30)  —  2 (67)  23 (64)  Vehicle Commander (n = 168)   Restrained  114 (68)  23 (68)  24 (86)  2 (100)  8 (62)  48 (73)  1 (25)  2 (67)  6 (33)   Unrestrained  54 (32)  11 (32)  4 (14)  —  5 (38)  18 (27)  3 (75)  1 (33)  12 (67)  Turret Gunner (n = 77)   Restrained  44 (57)  2 (50)  3 (75)  —  1 (33)  29 (81)  3 (43)  —  6 (27)   Unrestrained  33 (43)  2 (50)  1 (25)  —  2 (67)  7 (19)  4 (57)  1 (100)  16 (73)  Passenger (n = 204)   Restrained  125 (61)  13 (65)  33 (73)  1 (100)  7 (50)  63 (64)  —  1 (100)  7 (30)   Unrestrained  79 (39)  7 (35)  12 (27)  —  7 (50)  35 (36)  2 (100)  —  16 (70)  Occupant Position  Injury Severity Outcome  Total (%)  No Injury (%)  First Aid (%)  Medical Treatment Beyond First Aid (%)  Workday of Restricted Activity (%)  Lost Workday Case (%)  Permanent Partial Disability (%)  Permanent Total Disability (%)  Fatality (%)  All Positions (N = 1,092)   Restrained  794 (73)  261 (84)  169 (82)  7 (100)  50 (70)  266 (70)  5 (36)  4 (50)  32 (32)   Unrestrained  298 (27)  48 (16)  36 (18)  —  21 (30)  113 (30)  9 (64)  4 (50)  67 (68)  Driver (n = 643)   Restrained  511 (79)  223 (89)  109 (85)  4 (100)  34 (83)  126 (70)  1 (100)  1 (33)  13 (36)   Unrestrained  132 (21)  28 (11)  19 (15)  —  7 (17)  53 (30)  —  2 (67)  23 (64)  Vehicle Commander (n = 168)   Restrained  114 (68)  23 (68)  24 (86)  2 (100)  8 (62)  48 (73)  1 (25)  2 (67)  6 (33)   Unrestrained  54 (32)  11 (32)  4 (14)  —  5 (38)  18 (27)  3 (75)  1 (33)  12 (67)  Turret Gunner (n = 77)   Restrained  44 (57)  2 (50)  3 (75)  —  1 (33)  29 (81)  3 (43)  —  6 (27)   Unrestrained  33 (43)  2 (50)  1 (25)  —  2 (67)  7 (19)  4 (57)  1 (100)  16 (73)  Passenger (n = 204)   Restrained  125 (61)  13 (65)  33 (73)  1 (100)  7 (50)  63 (64)  —  1 (100)  7 (30)   Unrestrained  79 (39)  7 (35)  12 (27)  —  7 (50)  35 (36)  2 (100)  —  16 (70)  Only occupants with restraint use or nonuse reported are shown. View Large TABLE III Injury Severity Outcomes of U.S. Army HMMWV Occupants Injured in Rollover Accidents, 1992–2013, by Occupant Position and Reported Restraint Use Occupant Position  Injury Severity Outcome  Total (%)  No Injury (%)  First Aid (%)  Medical Treatment Beyond First Aid (%)  Workday of Restricted Activity (%)  Lost Workday Case (%)  Permanent Partial Disability (%)  Permanent Total Disability (%)  Fatality (%)  All Positions (N = 1,092)   Restrained  794 (73)  261 (84)  169 (82)  7 (100)  50 (70)  266 (70)  5 (36)  4 (50)  32 (32)   Unrestrained  298 (27)  48 (16)  36 (18)  —  21 (30)  113 (30)  9 (64)  4 (50)  67 (68)  Driver (n = 643)   Restrained  511 (79)  223 (89)  109 (85)  4 (100)  34 (83)  126 (70)  1 (100)  1 (33)  13 (36)   Unrestrained  132 (21)  28 (11)  19 (15)  —  7 (17)  53 (30)  —  2 (67)  23 (64)  Vehicle Commander (n = 168)   Restrained  114 (68)  23 (68)  24 (86)  2 (100)  8 (62)  48 (73)  1 (25)  2 (67)  6 (33)   Unrestrained  54 (32)  11 (32)  4 (14)  —  5 (38)  18 (27)  3 (75)  1 (33)  12 (67)  Turret Gunner (n = 77)   Restrained  44 (57)  2 (50)  3 (75)  —  1 (33)  29 (81)  3 (43)  —  6 (27)   Unrestrained  33 (43)  2 (50)  1 (25)  —  2 (67)  7 (19)  4 (57)  1 (100)  16 (73)  Passenger (n = 204)   Restrained  125 (61)  13 (65)  33 (73)  1 (100)  7 (50)  63 (64)  —  1 (100)  7 (30)   Unrestrained  79 (39)  7 (35)  12 (27)  —  7 (50)  35 (36)  2 (100)  —  16 (70)  Occupant Position  Injury Severity Outcome  Total (%)  No Injury (%)  First Aid (%)  Medical Treatment Beyond First Aid (%)  Workday of Restricted Activity (%)  Lost Workday Case (%)  Permanent Partial Disability (%)  Permanent Total Disability (%)  Fatality (%)  All Positions (N = 1,092)   Restrained  794 (73)  261 (84)  169 (82)  7 (100)  50 (70)  266 (70)  5 (36)  4 (50)  32 (32)   Unrestrained  298 (27)  48 (16)  36 (18)  —  21 (30)  113 (30)  9 (64)  4 (50)  67 (68)  Driver (n = 643)   Restrained  511 (79)  223 (89)  109 (85)  4 (100)  34 (83)  126 (70)  1 (100)  1 (33)  13 (36)   Unrestrained  132 (21)  28 (11)  19 (15)  —  7 (17)  53 (30)  —  2 (67)  23 (64)  Vehicle Commander (n = 168)   Restrained  114 (68)  23 (68)  24 (86)  2 (100)  8 (62)  48 (73)  1 (25)  2 (67)  6 (33)   Unrestrained  54 (32)  11 (32)  4 (14)  —  5 (38)  18 (27)  3 (75)  1 (33)  12 (67)  Turret Gunner (n = 77)   Restrained  44 (57)  2 (50)  3 (75)  —  1 (33)  29 (81)  3 (43)  —  6 (27)   Unrestrained  33 (43)  2 (50)  1 (25)  —  2 (67)  7 (19)  4 (57)  1 (100)  16 (73)  Passenger (n = 204)   Restrained  125 (61)  13 (65)  33 (73)  1 (100)  7 (50)  63 (64)  —  1 (100)  7 (30)   Unrestrained  79 (39)  7 (35)  12 (27)  —  7 (50)  35 (36)  2 (100)  —  16 (70)  Only occupants with restraint use or nonuse reported are shown. View Large Table IV shows various rates and measures of relative risk of U.S. Army HMMWV occupants being injured in rollover accidents by occupant position. When aggregating all positions, unrestrained occupants were 20% more likely to be nonfatally injured than were restrained occupants (RR = 1.20, 95% CI = 1.11–1.31). However, when stratifying by position, unrestrained drivers were 35% more likely to be nonfatally injured than were restrained drivers (RR = 1.35, 95% CI = 1.17–1.54), but no such difference in relative risk was observed between unrestrained and restrained occupants in other positions. TABLE IV Rates (per 1,000 Accidents per Year) and Relative Risks of Nonfatally Injured or Fatality to U.S. Army HMMWV Occupants in Rollover Accidents, 1992–2013, by Occupant Position Injury Status  Measure  All Positions  Driver  Vehicle Commander  Turret Gunner  Passenger  Nonfatally Injured  Rate, Unrestrained  36.0  33.8  33.5  40.1  40.4  Rate, Restrained  29.9  25.1  35.8  43.1  40.4  Rate, Unrestrained and Restrained  31.3  26.7  35.2  42.1  40.4  RR (95% CI)  1.20 (1.11, 1.31)  1.35 (1.17, 1.54)  0.94 (0.76, 1.15)  0.93 (0.77, 1.13)  1.00 (0.90, 1.11)  AR (ARP)  6.1 (17.0%)  8.7 (25.7%)  —  —  —  PAR (PARP)  1.4 (4.6%)  1.6 (5.8%)  —  —  —  Fatality  Rate, Unrestrained  10.2  7.9  10.1  22.0  9.2  Rate, Restrained  1.8  1.2  2.4  6.2  2.5  Rate, Unrestrained and Restrained  4.1  2.5  4.9  13.0  5.1  RR (95% CI)  5.58 (3.74, 8.32)  6.85 (3.57, 13.15)  4.22 (1.67, 10.65)  3.56 (1.56, 8.09)  3.62 (1.56, 8.40)  AR (ARP)  8.4 (82.1%)  6.8 (85.4%)  7.7 (76.3%)  15.8 (71.9%)  6.7 (72.4%)  PAR (PARP)  2.3 (55.5%)  1.4 (54.6%)  2.5 (50.9%)  6.8 (52.3%)  2.6 (50.3%)  Injury Status  Measure  All Positions  Driver  Vehicle Commander  Turret Gunner  Passenger  Nonfatally Injured  Rate, Unrestrained  36.0  33.8  33.5  40.1  40.4  Rate, Restrained  29.9  25.1  35.8  43.1  40.4  Rate, Unrestrained and Restrained  31.3  26.7  35.2  42.1  40.4  RR (95% CI)  1.20 (1.11, 1.31)  1.35 (1.17, 1.54)  0.94 (0.76, 1.15)  0.93 (0.77, 1.13)  1.00 (0.90, 1.11)  AR (ARP)  6.1 (17.0%)  8.7 (25.7%)  —  —  —  PAR (PARP)  1.4 (4.6%)  1.6 (5.8%)  —  —  —  Fatality  Rate, Unrestrained  10.2  7.9  10.1  22.0  9.2  Rate, Restrained  1.8  1.2  2.4  6.2  2.5  Rate, Unrestrained and Restrained  4.1  2.5  4.9  13.0  5.1  RR (95% CI)  5.58 (3.74, 8.32)  6.85 (3.57, 13.15)  4.22 (1.67, 10.65)  3.56 (1.56, 8.09)  3.62 (1.56, 8.40)  AR (ARP)  8.4 (82.1%)  6.8 (85.4%)  7.7 (76.3%)  15.8 (71.9%)  6.7 (72.4%)  PAR (PARP)  2.3 (55.5%)  1.4 (54.6%)  2.5 (50.9%)  6.8 (52.3%)  2.6 (50.3%)  RR, risk ratio; CI, confidence interval; AR, attributable risk; ARP, attributable risk percent; PAR, population attributable risk; PARP, population attributable risk percent. View Large TABLE IV Rates (per 1,000 Accidents per Year) and Relative Risks of Nonfatally Injured or Fatality to U.S. Army HMMWV Occupants in Rollover Accidents, 1992–2013, by Occupant Position Injury Status  Measure  All Positions  Driver  Vehicle Commander  Turret Gunner  Passenger  Nonfatally Injured  Rate, Unrestrained  36.0  33.8  33.5  40.1  40.4  Rate, Restrained  29.9  25.1  35.8  43.1  40.4  Rate, Unrestrained and Restrained  31.3  26.7  35.2  42.1  40.4  RR (95% CI)  1.20 (1.11, 1.31)  1.35 (1.17, 1.54)  0.94 (0.76, 1.15)  0.93 (0.77, 1.13)  1.00 (0.90, 1.11)  AR (ARP)  6.1 (17.0%)  8.7 (25.7%)  —  —  —  PAR (PARP)  1.4 (4.6%)  1.6 (5.8%)  —  —  —  Fatality  Rate, Unrestrained  10.2  7.9  10.1  22.0  9.2  Rate, Restrained  1.8  1.2  2.4  6.2  2.5  Rate, Unrestrained and Restrained  4.1  2.5  4.9  13.0  5.1  RR (95% CI)  5.58 (3.74, 8.32)  6.85 (3.57, 13.15)  4.22 (1.67, 10.65)  3.56 (1.56, 8.09)  3.62 (1.56, 8.40)  AR (ARP)  8.4 (82.1%)  6.8 (85.4%)  7.7 (76.3%)  15.8 (71.9%)  6.7 (72.4%)  PAR (PARP)  2.3 (55.5%)  1.4 (54.6%)  2.5 (50.9%)  6.8 (52.3%)  2.6 (50.3%)  Injury Status  Measure  All Positions  Driver  Vehicle Commander  Turret Gunner  Passenger  Nonfatally Injured  Rate, Unrestrained  36.0  33.8  33.5  40.1  40.4  Rate, Restrained  29.9  25.1  35.8  43.1  40.4  Rate, Unrestrained and Restrained  31.3  26.7  35.2  42.1  40.4  RR (95% CI)  1.20 (1.11, 1.31)  1.35 (1.17, 1.54)  0.94 (0.76, 1.15)  0.93 (0.77, 1.13)  1.00 (0.90, 1.11)  AR (ARP)  6.1 (17.0%)  8.7 (25.7%)  —  —  —  PAR (PARP)  1.4 (4.6%)  1.6 (5.8%)  —  —  —  Fatality  Rate, Unrestrained  10.2  7.9  10.1  22.0  9.2  Rate, Restrained  1.8  1.2  2.4  6.2  2.5  Rate, Unrestrained and Restrained  4.1  2.5  4.9  13.0  5.1  RR (95% CI)  5.58 (3.74, 8.32)  6.85 (3.57, 13.15)  4.22 (1.67, 10.65)  3.56 (1.56, 8.09)  3.62 (1.56, 8.40)  AR (ARP)  8.4 (82.1%)  6.8 (85.4%)  7.7 (76.3%)  15.8 (71.9%)  6.7 (72.4%)  PAR (PARP)  2.3 (55.5%)  1.4 (54.6%)  2.5 (50.9%)  6.8 (52.3%)  2.6 (50.3%)  RR, risk ratio; CI, confidence interval; AR, attributable risk; ARP, attributable risk percent; PAR, population attributable risk; PARP, population attributable risk percent. View Large On the other hand, unrestrained occupants in each position were 3.56 to 6.85 times more likely to be fatally injured than were restrained occupants in the same position. When aggregating all positions, the difference between the unrestrained fatality rate per 1,000 accidents per year (10.2) and the restrained fatality rate per 1,000 accidents per year (1.8) was 8.4, which is the AR of fatalities among unrestrained occupants due to being unrestrained during a HMMWV rollover accident. This difference, expressed as an ARP, was 82% of the unrestrained fatality rate, meaning 82% of fatalities among unrestrained occupants were due to being unrestrained during a HMMWV rollover accident. Similarly, the difference between the unrestrained fatality rate and the population fatality rate was 2.3, which was the PAR of fatalities among the entire analysis population of unrestrained and restrained occupants due to being unrestrained during a HMMWV rollover accident. This difference, expressed as a PARP, was 56% of the population fatality rate, meaning 56% of fatalities in the entire analysis population were due to being unrestrained during a HMMWV rollover accident. When stratifying by position, unrestrained turret gunners had the greatest AR (15.8) of fatalities compared to restrained turret gunners, and the greatest PAR (6.8) of fatalities compared to both unrestrained and restrained turret gunners combined. However, unrestrained drivers had the greatest ARP (85.4%) of fatalities compared to restrained drivers, and the greatest PARP (54.6%) compared to both unrestrained and restrained drivers combined. Table V shows observed and expected injuries by occupant position, restraint use, and body region injured. Unrestrained drivers had greater than expected head/neck and torso injuries, while restrained drivers had fewer than expected of the same. Unrestrained vehicle commanders also had greater than expected torso injuries, but fewer than expected upper extremity injuries, while the reverse was true for restrained vehicle commanders. Restrained passengers had greater than expected upper extremity injuries, whereas unrestrained passengers had fewer than expected of the same. Differences in body region injured between restrained and unrestrained turret gunners were not statistically significant at the p = 0.05 level. TABLE V Observed and Expected Injuries by Occupant Position, Restraint Use, and Body Region Injured in U.S. Army HMMWV Rollover Accidents, 1992–2013 Occupant Position  Body Region Injured  Head/Neck  Spine/Back  Torso  Upper Extremity  Lower Extremity  Unspecified  Total  Driver*   Restrained    Observed (count)  114  18  66  81  61  283  623    Expected (count)  120  18  80  85  62  258   Unrestrained    Observed (count)  42  6  39  30  20  53  190    Expected (count)  36  6  25  26  19  78  Vehicle Commander**   Restrained    Observed (count)  34  11  15  36  16  40  152    Expected (count)  34  9  25  29  17  38   Unrestrained    Observed (count)  21  4  26  12  11  23  97    Expected (count)  21  6  16  19  10  25  Turret Gunner***   Restrained    Observed (count)  11  3  13  14  9  7  57    Expected (count)  13  3  13  10  6  12   Unrestrained    Observed (count)  16  2  13  7  3  17  58    Expected (count)  14  2  13  11  6  12  Passenger****   Restrained    Observed (count)  54  6  24  37  17  36  174    Expected (count)  54  9  25  29  21  36   Unrestrained    Observed (count)  36  9  18  11  19  25  118    Expected (count)  36  6  17  19  15  25  Occupant Position  Body Region Injured  Head/Neck  Spine/Back  Torso  Upper Extremity  Lower Extremity  Unspecified  Total  Driver*   Restrained    Observed (count)  114  18  66  81  61  283  623    Expected (count)  120  18  80  85  62  258   Unrestrained    Observed (count)  42  6  39  30  20  53  190    Expected (count)  36  6  25  26  19  78  Vehicle Commander**   Restrained    Observed (count)  34  11  15  36  16  40  152    Expected (count)  34  9  25  29  17  38   Unrestrained    Observed (count)  21  4  26  12  11  23  97    Expected (count)  21  6  16  19  10  25  Turret Gunner***   Restrained    Observed (count)  11  3  13  14  9  7  57    Expected (count)  13  3  13  10  6  12   Unrestrained    Observed (count)  16  2  13  7  3  17  58    Expected (count)  14  2  13  11  6  12  Passenger****   Restrained    Observed (count)  54  6  24  37  17  36  174    Expected (count)  54  9  25  29  21  36   Unrestrained    Observed (count)  36  9  18  11  19  25  118    Expected (count)  36  6  17  19  15  25  * χ2 = 24.0, df = 5, p = 0.0002. ** χ2= 15.4, df = 5, p = 0.009. *** χ2 = 10.6, df = 5, p = 0.06. **** χ2 = 10.9, df = 5, p = 0.05. View Large TABLE V Observed and Expected Injuries by Occupant Position, Restraint Use, and Body Region Injured in U.S. Army HMMWV Rollover Accidents, 1992–2013 Occupant Position  Body Region Injured  Head/Neck  Spine/Back  Torso  Upper Extremity  Lower Extremity  Unspecified  Total  Driver*   Restrained    Observed (count)  114  18  66  81  61  283  623    Expected (count)  120  18  80  85  62  258   Unrestrained    Observed (count)  42  6  39  30  20  53  190    Expected (count)  36  6  25  26  19  78  Vehicle Commander**   Restrained    Observed (count)  34  11  15  36  16  40  152    Expected (count)  34  9  25  29  17  38   Unrestrained    Observed (count)  21  4  26  12  11  23  97    Expected (count)  21  6  16  19  10  25  Turret Gunner***   Restrained    Observed (count)  11  3  13  14  9  7  57    Expected (count)  13  3  13  10  6  12   Unrestrained    Observed (count)  16  2  13  7  3  17  58    Expected (count)  14  2  13  11  6  12  Passenger****   Restrained    Observed (count)  54  6  24  37  17  36  174    Expected (count)  54  9  25  29  21  36   Unrestrained    Observed (count)  36  9  18  11  19  25  118    Expected (count)  36  6  17  19  15  25  Occupant Position  Body Region Injured  Head/Neck  Spine/Back  Torso  Upper Extremity  Lower Extremity  Unspecified  Total  Driver*   Restrained    Observed (count)  114  18  66  81  61  283  623    Expected (count)  120  18  80  85  62  258   Unrestrained    Observed (count)  42  6  39  30  20  53  190    Expected (count)  36  6  25  26  19  78  Vehicle Commander**   Restrained    Observed (count)  34  11  15  36  16  40  152    Expected (count)  34  9  25  29  17  38   Unrestrained    Observed (count)  21  4  26  12  11  23  97    Expected (count)  21  6  16  19  10  25  Turret Gunner***   Restrained    Observed (count)  11  3  13  14  9  7  57    Expected (count)  13  3  13  10  6  12   Unrestrained    Observed (count)  16  2  13  7  3  17  58    Expected (count)  14  2  13  11  6  12  Passenger****   Restrained    Observed (count)  54  6  24  37  17  36  174    Expected (count)  54  9  25  29  21  36   Unrestrained    Observed (count)  36  9  18  11  19  25  118    Expected (count)  36  6  17  19  15  25  * χ2 = 24.0, df = 5, p = 0.0002. ** χ2= 15.4, df = 5, p = 0.009. *** χ2 = 10.6, df = 5, p = 0.06. **** χ2 = 10.9, df = 5, p = 0.05. View Large Table VI shows observed and expected injuries by occupant position, restraint use, and nature of injury. Unrestrained drivers had greater than expected fractures, but fewer than expected sprains/strains, while the reverse was true for restrained drivers. Unrestrained vehicle commanders also had fewer than expected sprains/strains, whereas restrained vehicle commanders had greater than expected of the same. Differences in nature of injury between restrained and unrestrained turret gunners and passengers were not statistically significant at the p = 0.05 level. TABLE VI Observed and Expected Injuries by Occupant Position, Restraint Use, and Nature of Injury in U.S. Army HMMWV Rollover Accidents, 1992–2013 Occupant Position  Nature of Injury  Fracture  Internal  Open Wound  Contusion  Sprain/Strain  Unspecified  Total  Driver*   Restrained    Observed (count)  38  35  55  125  74  291  618    Expected (count)  60  37  59  128  66  268   Unrestrained    Observed (count)  40  13  21  42  12  58  186    Expected (count)  18  11  17  39  20  81  Vehicle Commander**   Restrained    Observed (count)  15  7  19  34  28  43  146    Expected (count)  19  10  19  34  20  44   Unrestrained    Observed (count)  17  10  13  20  5  30  95    Expected (count)  13  7  13  20  13  29  Turret Gunner***   Restrained    Observed (count)  15  4  7  10  5  11  52    Expected (count)  16  3  8  9  4  12   Unrestrained    Observed (count)  17  2  9  8  3  14  53    Expected (count)  16  3  8  9  4  13  Passenger****   Restrained    Observed (count)  28  17  28  35  17  46  171    Expected (count)  32  16  24  36  20  43   Unrestrained    Observed (cCount)  25  10  12  25  17  27  116    Expected (cCount)  21  11  16  24  14  30  Occupant Position  Nature of Injury  Fracture  Internal  Open Wound  Contusion  Sprain/Strain  Unspecified  Total  Driver*   Restrained    Observed (count)  38  35  55  125  74  291  618    Expected (count)  60  37  59  128  66  268   Unrestrained    Observed (count)  40  13  21  42  12  58  186    Expected (count)  18  11  17  39  20  81  Vehicle Commander**   Restrained    Observed (count)  15  7  19  34  28  43  146    Expected (count)  19  10  19  34  20  44   Unrestrained    Observed (count)  17  10  13  20  5  30  95    Expected (count)  13  7  13  20  13  29  Turret Gunner***   Restrained    Observed (count)  15  4  7  10  5  11  52    Expected (count)  16  3  8  9  4  12   Unrestrained    Observed (count)  17  2  9  8  3  14  53    Expected (count)  16  3  8  9  4  13  Passenger****   Restrained    Observed (count)  28  17  28  35  17  46  171    Expected (count)  32  16  24  36  20  43   Unrestrained    Observed (cCount)  25  10  12  25  17  27  116    Expected (cCount)  21  11  16  24  14  30  * χ2 = 48.8, df = 5, p < 0.0001. ** χ2 = 13.6, df = 5, p = 0.02. *** χ2 = 2.1, df = 5, p = 0.8. **** χ2 = 4.6, df = 5, p = 0.5. View Large TABLE VI Observed and Expected Injuries by Occupant Position, Restraint Use, and Nature of Injury in U.S. Army HMMWV Rollover Accidents, 1992–2013 Occupant Position  Nature of Injury  Fracture  Internal  Open Wound  Contusion  Sprain/Strain  Unspecified  Total  Driver*   Restrained    Observed (count)  38  35  55  125  74  291  618    Expected (count)  60  37  59  128  66  268   Unrestrained    Observed (count)  40  13  21  42  12  58  186    Expected (count)  18  11  17  39  20  81  Vehicle Commander**   Restrained    Observed (count)  15  7  19  34  28  43  146    Expected (count)  19  10  19  34  20  44   Unrestrained    Observed (count)  17  10  13  20  5  30  95    Expected (count)  13  7  13  20  13  29  Turret Gunner***   Restrained    Observed (count)  15  4  7  10  5  11  52    Expected (count)  16  3  8  9  4  12   Unrestrained    Observed (count)  17  2  9  8  3  14  53    Expected (count)  16  3  8  9  4  13  Passenger****   Restrained    Observed (count)  28  17  28  35  17  46  171    Expected (count)  32  16  24  36  20  43   Unrestrained    Observed (cCount)  25  10  12  25  17  27  116    Expected (cCount)  21  11  16  24  14  30  Occupant Position  Nature of Injury  Fracture  Internal  Open Wound  Contusion  Sprain/Strain  Unspecified  Total  Driver*   Restrained    Observed (count)  38  35  55  125  74  291  618    Expected (count)  60  37  59  128  66  268   Unrestrained    Observed (count)  40  13  21  42  12  58  186    Expected (count)  18  11  17  39  20  81  Vehicle Commander**   Restrained    Observed (count)  15  7  19  34  28  43  146    Expected (count)  19  10  19  34  20  44   Unrestrained    Observed (count)  17  10  13  20  5  30  95    Expected (count)  13  7  13  20  13  29  Turret Gunner***   Restrained    Observed (count)  15  4  7  10  5  11  52    Expected (count)  16  3  8  9  4  12   Unrestrained    Observed (count)  17  2  9  8  3  14  53    Expected (count)  16  3  8  9  4  13  Passenger****   Restrained    Observed (count)  28  17  28  35  17  46  171    Expected (count)  32  16  24  36  20  43   Unrestrained    Observed (cCount)  25  10  12  25  17  27  116    Expected (cCount)  21  11  16  24  14  30  * χ2 = 48.8, df = 5, p < 0.0001. ** χ2 = 13.6, df = 5, p = 0.02. *** χ2 = 2.1, df = 5, p = 0.8. **** χ2 = 4.6, df = 5, p = 0.5. View Large DISCUSSION Trend analysis of U.S. Army HMMWV rollover accidents showed the risk of such rollovers did not become apparent until the period beginning in FY 2003 and ending in FY 2008, confirming reports of increased HMMWV rollover accidents during this period of increased OIF operational tempo.2,3 The drop in HMMWV rollover accidents occurring after FY 2005 coincided with the introduction of mine-resistant ambush-protected vehicles into OIF and OEF in FY 2008, which replaced HMWVs in combat roles.10 Although we have no evidence to prove this, increased driver experience, improved driver training, and changing mission profiles after FY 2005 could also have contributed to this drop. Analysis of HMMWV rollover accident characteristics showed one-half of all class A (fatal) rollover accidents occurred in Iraq, and one-third of class A rollover accidents involved the M1114 up-armored variant, confirming reports of fatal rollover accidents involving up-armored HMMWVs in OIF.2,3 Analysis of HMMWV occupant characteristics confirmed the population at risk for rollover accidents to be predominantly young, male, and enlisted.4,–8 Reported restraint use varied by occupant position, with turret gunners reporting the lowest percentage of restraint use and drivers reporting the highest. In this analysis, 28% of both vehicle commanders and passengers were reportedly unrestrained at the time of the rollover accident. This is of particular concern because vehicle commanders are responsible for enforcing the Army's restraint use requirement in motor vehicles15 and must therefore model this positive behavior themselves to their subordinates. Greater command and peer-to-peer enforcement are therefore necessary to increase occupant restraint use. However, these occupants could have been hindered from wearing restraints due to the bulky battle gear they were wearing.16 In 2006, a new personal restraint system for the M1114 variant was fielded that featured longer shoulder and lap straps to accommodate occupants wearing battle gear.16,17 However, since RMIS data do not indicate whether the older or newer personal restraint system was used, further research is needed to determine if this improvement has led to an increase in restraint use by M1114 occupants and consequently a reduction in injuries and fatalities. Restraint use is particularly challenging for turret gunners. In this analysis, 24% of turret gunners were reportedly unrestrained at the time of the rollover accident, and 43% reported neither restraint use nor nonuse. This lack of restraint use by turret gunners or their reticence to report not using them (if in fact they did not) are understandable, given that the gunner's restraint system initially fielded with the M1114 variant was cumbersome to use.16,18 By 2006, an improved gunner's restraint system was fielded featuring an adjustable tail strap to accommodate the gunner's height, a swivel to minimize tangling, and a quick-release push-button for rapid egress.16,–18 However, since RMIS data do not indicate whether the gunner's restraint system or improved gunner's restraint system was used, further research is needed to determine if these improvements have led to an increase in restraint use by M1114 turret gunners and consequently a reduction in injuries and fatalities. Stratified analysis of injury severity by restraint use showed the risk of being injured (both nonfatally or fatally) in a HMMWV rollover accident was greater for unrestrained occupants than for restrained occupants. Although the relative risk of nonfatal injury among unrestrained occupants was modestly higher (20%) than among restrained occupants, the relative risk of fatal injury was substantially higher. Unrestrained occupants from all positions combined were 5.6 times more likely to die in a HMMWV rollover accident than were restrained occupants. Among unrestrained occupants, restraint use could have potentially saved 82% of lives lost during the period analyzed (FY 1992 through 2013). Among all occupants involved in a HMMWV rollover during this period, an estimated 56% of fatalities could have been prevented by restraint use. Thus, these results unequivocally support the effectiveness of restraint use as a life-saving countermeasure in HMMWV rollover accidents, especially for turret gunners and drivers. However, these results could also reflect a possible reporting bias, as noted below. Stratified analysis of body region injured and nature of injury based on restraint use also showed differences in injury patterns between restrained and unrestrained occupants. Not surprisingly, unrestrained drivers and vehicle commanders had greater than expected torso injuries. This could be due to the unrestrained body striking against interior surfaces of the vehicle during a rollover. This could also explain the fact that unrestrained drivers had greater than expected fractures. However, restrained vehicle commanders and passengers had greater than expected upper extremity injuries. This could be due to flailing of the arms during a rollover while the rest of the body is restrained. Finally, restrained drivers and vehicle commanders had greater than expected sprains/strains. This could be due to hyperflexion of the head/neck, spine/back, or extremities during a rollover while the body is restrained. Because of small sample sizes of the injuries documented, differences in injury patterns between restrained and unrestrained turret gunners were not statistically significant at the p = 0.05 level. LIMITATIONS We note the impact of missing data on our analysis. Of 1,395 occupants involved in HMMWV rollover accidents during the period analyzed, 303 (22%) did not have restraint use or nonuse reported, ranging from 13% of vehicle commanders to 43% of turret gunners. These missing data could have skewed our analysis results. Furthermore, occupants could have underreported restraint nonuse, fearing the consequences of noncompliance with the Army's restraint use requirement in motor vehicles.15 Conversely, occupants could have overreported restraint use in less severe accidents than in accidents resulting in serious injury or fatality, possibly biasing these results in favor of restraint use. We note also the impact of missing data on our Barell matrix analysis of body region injured by nature of injury. Of 1,469 total injury diagnoses identified in this analysis, 33% did not indicate the body part injured and 35% did not indicate the nature of injury. These missing data diminished both individual cell sizes and the number of populated cells within the Barell matrix, making meaningful stratified analyses difficult, if not impossible. Consequently, we were able to compare only the marginal totals of body region injured and nature of injury separately, stratified by occupant position and restraint use. Because of small cell sizes and the paucity of populated cells, we chose not to show our cross-tabulations of body region injured by nature of injury for each occupant position stratified by restraint use. In addition, these injury diagnoses were not coded according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM), as specified in the Barell matrix,14 nor were they coded using an injury scale such as the Abbreviated Injury Scale,19 making analysis of injury severity impossible. Finally, our analysis was limited by the lack of an exposure metric (vehicle miles traveled) over which the incidence of HMMWV rollover accidents, fatalities, and nonfatal injuries could be normalized. This lack of exposure data is a limitation common to all military motor vehicle crash injury studies reviewed by Krahl et al.20 CONCLUSION Given the limitations described above, nevertheless we were able to describe U.S. Army HMMWV rollover accidents, fatalities, and nonfatal injuries during 22 years of Army service. We found restraint use conferred substantial life-saving protection to HMMWV occupants in rollover accidents. Therefore, commanders, safety officers, and peers should continue to promote and enforce restraint use consistently during all Army ground operations and training involving HMMWVs and other military motor vehicles equipped with occupant restraints. Specific recommendations to leverage these findings into positive action include the following: (1) engage commanders and safety officers Army-wide in enhanced outreach to educate Soldiers during safety briefings, emphasizing the potential life-saving benefit of restraint use and the consequences of nonuse, (2) retrofit existing seatbelts with seatbelt extension kits developed by the program manager, light tactical vehicles, or replace them with improved seatbelt assemblies with longer straps to accommodate occupants wearing battle gear,21 and (3) enforce mandatory restraint use by making Soldiers more accountable to each other during training and operations. Implementation of these measures will save Soldiers' lives in rollover accidents during the remaining years of the HMMWV program. ACKNOWLEDGMENTS The authors thank the following individuals for critically reviewing this manuscript: Amy Barrett, BA, Laulima Government Solutions, LLC; and COL Mark McPherson, MC, USA, formerly with the U.S. Army Combat Readiness Center and currently with the U.S. Army School of Aviation Medicine. The authors are supported by Joint Trauma Analysis and Prevention of Injury in Combat (JTAPIC) program funds. Earlier versions of this analysis were previously presented in USAARL report no. 2013-13, “Injuries to Occupants of U.S. Army High Mobility Multipurpose Wheeled Vehicles, 1989–2007,” and in a poster presented at the 2008 Force Health Protection Conference, Albuquerque, NM. This analysis was completed under contract number W81XWH-14-D-0012 funded by the U.S. Army Medical Research Acquisition Activity (USAMRAA) in support of the Joint Trauma Analysis and Prevention of Injury in Combat (JTAPIC) program. REFERENCES 1. Zaloga SJ, Johnson H HMMWV Humvee 1980–2005: US Army Tactical Vehicle . Oxford, UK, Osprey Publishing, 2006. 2. Carollo R, Wagner M Deadly price paid for Humvee armor used to protect soldiers, Dayton Daily News , 2006. Available at https://www.highbeam.com/doc/1P2-11969245.html; accessed August 6, 2016. 3. Rossen LM, Pollack KM, Canham-Chervak M, Canada S, Baker SP Motor vehicle crashes among active duty U.S. Army personnel, 1999 to 2006. Mil Med  2011; 176( 9): 1019– 26. Google Scholar CrossRef Search ADS PubMed  4. Bray RM, Kroutil LA, Wheeless SC, et al.   1995 Department of Defense survey of health related behaviors among military personnel . Research Triangle Park, NC, Research Triangle Institute, 1995. Available at http://www.dtic.mil/dtic/tr/fulltext/u2/a317298.pdf; accessed August 6, 2016. 5. Bray RM, Sanchez RP, Ornstein ML, Lentine D, Vincus AA HIGHLIGHTS 1998 Department of Defense survey of health related behaviors among military personnel . Research Triangle Park, NC, Research Triangle Institute, 1999. Available at http://www.dtic.mil/cgi-bin/GetTRDoc?Location=U2&doc=GetTRDoc.pdf&AD=ADA361901; accessed August 6, 2016. 6. Bray RM, Hourani LL, Rae KL, et al.   2002 Department of Defense survey of health related behaviors among military personnel . Research Triangle Park, NC, RTI International, 2003. Available at http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA431566; accessed August 6, 2016. 7. Bray RM, Hourani LL, Rae Olmsted KL, et al.   2005 Department of Defense survey of health related behaviors among active duty military personnel . Research Triangle Park, NC, RTI International, 2006. Available at http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA465678; accessed August 6, 2016. 8. Bray RM, Pemberton MR, Hourani LL, et al.   2008 Department of Defense survey of health related behaviors among active duty military personnel . Research Triangle Park, NC, RTI International, 2009. Available at https://www.documentcloud.org/documents/694941-2008-department-of-defense-survey-of-health.html; accessed August 6, 2016. 9. Peik SM, Pollack KM, Canham-Chervak M, Hauret KG, Baker SP Injuries to deployed US Army Soldiers involved in HMMWV crashes, 2002–2006. Mil Med  2012; 177( 8): 963– 9. Google Scholar CrossRef Search ADS PubMed  10. Erwin S Army says Humvees no longer to be used in combat roles. National Defense , 2012. Available at http://www.nationaldefensemagazine.org/blog/lists/posts/post.aspx?ID=745; accessed August 6, 2016. 11. Jean GV Improvements to discontinued Humvees may last another 20 years . National Defense, 2010. Available at http://www.nationaldefensemagazine.org/ARCHIVE/2010/MAY/Pages/ImprovedArmyHumveesMayLastAnother20Years.aspx; accessed August 6, 2016. 12. Department of the Army Pamphlet 385-40, Army Accident Investigations and Reporting . Available at https://armypubs.army.mil/Search/ePubsSearch/ePubsSearchDownloadPage.aspx?docID=0902c85180010e81; accessed August 6, 2016. 13. Friedman GD Primer of Epidemiology , Ed 5, New York, NY, McGraw-Hill, 2004. 14. Barell V, Aharonson-Daniel L, Fingerhut LA, et al.   An introduction to the Barell body region by nature of injury diagnosis matrix. Inj Prev  2002; 8: 91– 6. Google Scholar CrossRef Search ADS PubMed  15. Department of the Army Regulation 385-10, The Army Safety Program . Available at https://armypubs.army.mil/Search/ePubsSearch/ePubsSearchDownloadPage.aspx?docID=0902c85180010d83; accessed August 6, 2016. 16. Olive-Drab.com HMMWV safety improvements . Available at http://olive-drab.com/od_mvg_hmmwv_safety_improvements.php; accessed October 8, 2016. 17. United States Army Tank-automotive and Armaments Command M1114 UAHMMWV Combined Safety Smart Card, Version 2.0 . Warren, MI, United States Army Tank-automotive and Armaments Command, 2006. Available at http://8tharmy.korea.army.mil/safety/Toolbox/resources/11/M1114CombinedSafetyCard.pdf; accessed October 8, 2016. 18. Department of the Army Up-armored HMMWV gunner's restraint system: restraints keep gunners safe! PS The Preventive Maintenance Monthly  2007; 657: 14– 6. 19. Gennarelli TA, Wodzin E AIS 2005: a contemporary injury scale. Injury  2006; 37: 1083– 91. Google Scholar CrossRef Search ADS PubMed  20. Krahl PL, Jankosky CJ, Thomas RJ, Hooper TI Systematic review of military motor vehicle crash-related injuries. Am J Prev Med  2010; 38: 189– 96. Google Scholar CrossRef Search ADS   21. United States Army Combat Readiness Center Get “belted” in that HMMWV! Countermeasure  2003; 24( 6): 18. Reprint & Copyright © Association of Military Surgeons of the U.S.
TI - High-Mobility Multipurpose Wheeled Vehicle Rollover Accidents and Injuries to U.S. Army Soldiers by Reported Occupant Restraint Use, 1992–2013
JF - Military Medicine
DO - 10.7205/MILMED-D-16-00318
DA - 2017-05-01
UR - https://www.deepdyve.com/lp/oxford-university-press/high-mobility-multipurpose-wheeled-vehicle-rollover-accidents-and-9y8DMuw4G3
SP - e1782
EP - e1791
VL - 182
IS - 5
DP - DeepDyve
ER -