Journal Logo


Analysis of Nonbattle Deaths Among U.S. Service Members in the Deployed Environment

Le, Tuan D. MD, DrPH∗,†; Gurney, Jennifer M. MD∗,‡; Akers, Kevin S. MD; Chung, Kevin K. MD§; Singh, Karan P. PhD; Wang, Heuy-Ching PhD||; Stackle, Mark E. MD; Pusateri, Anthony E. PhD||

Author Information
doi: 10.1097/SLA.0000000000005047


In recent conflicts, nonbattle injury (NBI) has accounted for approximately one-third of injuries requiring referral to a deployed hospital [role 3 medical treatment facility (R3MTF)].1 Leading causes of NBI include falls, motor vehicle accidents, and sports.1–3 As a subset of injury, NBI-related deaths (NBD) represent the most serious injuries. In the Vietnam War, NBD accounted for approximately 18% of deaths.4 During the Persian Gulf War (1990–1991), NBD accounted for approximately 50% of deaths.5,6 Advances in body armor and medical care have decreased the rates of death due to battle injury (BI).7 Understanding causes and trends in NBD may help guide efforts to develop specific preventive measures.

Previous analysis from the US Department of Defense Trauma Registry (DoDTR) demonstrated that NBI is a significant contributor to casualty rates, comprising approximately 34% of total casualties during combat operations in Iraq and Afghanistan.1 NBI, therefore, exerts a significant operational and clinical burden in the deployed setting. 1,8–15 Previous analyses examining the epidemiology of NBI suggest that many of the etiologies may be potentially preventable with increased safety measures and increased mechanistic understanding of these injuries.1,3

Our previous analysis of NBI was limited in that it could not account for NBI in which casualties died before reaching a R3MTF.1 To characterize the rates, etiologies, and trends of NBD during recent conflicts, we conducted a retrospective analysis utilizing 2 databases, the DoDTR and the Defense Casualty Analysis System (DCAS). The DCAS, although not as detailed as the DoDTR, accounted for all deaths, including those in the prehospital setting. Therefore, both databases were used to allow a more comprehensive examination of NBD. The objective of this comprehensive analysis is to: better quantify NBD; gain a better understanding of the factors and primary etiologies contributing to NBD, and characterize trends and forecast NDB rates. A better understanding of deaths from nonbattle etiologies will help inform primary prevention measures to decrease mortality from potentially preventable causes, while forecasted rates may serve as points of reference against which to assess future progress.


Study Population and Data Extraction

Following a protocol reviewed and approved by the U.S. Army Medical Research and Development Command Institutional Review Board (M-10563) with a waiver of informed consent, as allowed under 32 CFR 219.116(d), we conducted a retrospective analysis of prospectively collected data from the population of deployed U.S. service members (SM) with trauma sustained in the Iraq or Afghanistan Conflicts from January 1, 2003 to December 31, 2014. The 2 data sources used for this analysis were from the DCAS4 and the DoDTR managed by the joint trauma system.16 DCAS is an administrative database that aggregates monthly personnel and injury data, including prehospital injuries and mortality, but does not contain clinical details. The DoDTR includes clinical details for individual casualties arriving alive at R3MTFs, but until 2014, no information on prehospital injuries or mortality was included. A total of 59,799 casualties in Operation Enduring Freedom (OEF) and Operation Iraqi Freedom (OIF) extracted from DCAS was included in this analysis (Fig. 1A). Of 31,033 patients from DoDTR, 29,958 patients met study inclusion criteria described in Figure 1B. Patients with noninjury diseases or who were non-U.S. SMs were not considered in this study.

Study population and inclusion criteria. Figure depicts numbers of U.S. service members who sustained injuries and mortality involved in battle activities and nonbattle activities (NB) in the deployed environment during Operation Iraqi Freedom (OIF) and Operation New Dawn (OND) in Iraq and Operation Enduring Freedom (OEF) in Afghanistan from October 2001 to December 2014 reported by Defense Casualty Analysis System (DCAS) for all casualties (panel A) and recorded in DoDTR for patients who arrived at a role 3 medical treatment facility (panel B) and exclusion criteria. DoDTR indicates Department of Defense Trauma Registry.

Study Definitions

NBI was defined as any injury not directly involving hostile actions or terrorist activity, including unintentional and/or self-inflicted injuries as well accidents occurring during the storage or transport of munitions or explosive ordnance.17,18 BI was defined as any injury occurring from battle-related activities or hostile actions. Mechanisms of injury were categorized as (1) explosion, including aerial bomb, explosively formed projectile, improvised explosive device (IED), IED-person borne, IED-vehicle borne, rocket propelled grenade, etc, (2) gunshot wound (GSW), (3) motor vehicle crash (MVC), (4) helicopter or airplane crash, and (5) other causes, such as sports, machinery/equipment, fall, chemical, electrical, fire/flame. Types of injury were categorized as penetrating, blunt, burn, and other/unknown. Injury severity was measured using the injury severity score (ISS) and the military injury severity score (mISS).17

The total number of SMs deployed who were exposed to risk of injury was not available in the DoDTR, which only captured those who have sustained traumatic injury. Therefore, as it was not possible to calculate person time incidence rates, we used a proportion of NBI or NBD as a surrogate. The proportion of NBI was defined as the number of NBI cases per 100 BI and NBI casualties.1 Proportion of NBD was defined as the number of NBD cases per 100 deaths caused by BI and NBI. The proportion of NBD was used as a primary outcome measure. NBD trend evaluation and forecasting were performed using the aggregate data from DCAS.

Statistical Analysis

Descriptive statistics characterized the demographics and injuries of the patients stratified by injury classification (battle vs nonbattle). Continuous variables were presented as mean and standard deviation (SD) or median and interquartile range (IQR) and were tested for difference using a t-test or Mann-Whitney test where appropriate. Categorical variables were expressed as frequencies and percentages and tested using the Chi-square or Fisher exact test. The proportion of NBD was compared among different strata of demographics and injury characteristic variables. The regression analysis of the proportion of NBD over time by calendar quarter was evaluated for linear trends. The changes in the proportion of NBD over time during the study period were examined using the weighted moving average (WMA), a common method sued for trend analysis. Time series analysis was accomplished with autoregressive integrated moving average (ARIMA), which is more appropriate for data with autocorrelation and better suited for evaluating trends and forecasting the NBD rate. The latter is the most suitable model to predict future NBD rates based on previously observed NBD. These projections may be used to help estimate the necessary supplies and interventions, or to help assess the effectiveness of the future prevention strategies.

The quarterly proportions of NBD from January 1, 2003 to December 31, 2014 were used for ARIMA model to estimate the proportion of NBD from 2003 to 2014 and forecast the proportion of NBD through 2025. Trend stationarity of NBD was examined using the autocorrelation function and tested by the Dickey-Fuller test.19 Autocorrelations at various lags with Ljung-Box Chi-square statistics were used to test for white noise or uncorrelated random variables with a mean of zero and no correlation between values at different time points.20 The values p, d, and q in the ARIMA model were selected to represent the number of autoregressive lags (p), level of differences (d), and moving average (q). Statistical significance was determined at the 2-sided P < 0.05 level. Data analyses were performed using SAS version 9.4 (SAS Institute, Inc., Cary, NC).


Patient Population and NBI Incidences

Between 2001 and 2014, there were 59,799 casualties recorded in DCAS. This includes 36,364 injured patients in OIF during 3/2003–8/2010 and 23,435 injured patients in OEF (39.2%) during 10/2001–12/2014 (Fig. 1A). Among SMs arriving at R3MTF recorded in DoDTR, 29,958 patients met the inclusion criteria (Fig. 1B) and are described in the Supplementary. Table 1, The majority of casualties were from OIF (61.2%) and from the US Army (74.0%). Most of the casualties were men (96.8%), young with a median age at injury of 24 years (IQR, 21–29 years), and white (70.4%). Overall, the most common primary mechanism of injury was explosion (51.5%). The leading injury types were blunt (54.3%) and penetrating (39.9%). The most common concomitant injuries were external (skin and soft tissue; 68.1%), extremities (49.4%), and head/neck (36.6%).

Approximately one-third of 29,958 casualties [n = 10,203 (34.1%)] arriving alive to the R3MTF were NBI (Supplementary Table 1, The proportion of NBI was higher in females than males (63.2% vs 33.1%; P < 0.0001), higher in black patients (52.1%) than in Hispanic patients (36.6%), and White patients (34.2%); P < 0.0001; Supplementary Table 2, A higher proportion of NBI was observed in Operation New Dawn (OND) (70.9%) and OIF (36.3%) than in OEF (29.0%) (P < 0.0001), and in Air Force (66.3%) than Navy (48.3%), Army (34.7%), and Marine Corps (25.7%); P < 0.0001. A higher proportion of NBIs was observed in those involved in MVC (87.5%) and helicopter or airplane crash (63.3%) compared to those involved in explosion (1.5%) and GSW (16.4%; P < 0.0001). The most common body regions associated with NBI etiologies were the extremity (54.2%), followed by external, including the skin and soft tissues (47.7%), and head/neck (25.0%), while the most common injured body regions caused by BI were external (78.6%), followed by extremity (46.9%) and head/neck (42.6%) (Supplementary Table 1, The median ISS and mISS were 5 (IQR, 2–10) and 5 (2–13), respectively, and NBI was less severe than BI, with a median ISS of 4 (IQR, 1–5) versus 6 (IQR, 2–14) and a median mISS of 4 (IQR, 1–5) versus 6 (IQR, 2–18), respectively (Supplementary Table 1,; all P < 0.0001).

Proportion of Nonbattle Death (NBD)

Between the 2 available data sources there were significant differences in the overall case fatality rates (CFRs) recorded during the study period. In DCAS, the CFR was 11.3% (6745 of 59,799); NBD accounted for 21.2% (n = 1431) (Fig. 1A). Among patients included in DoDTR, overall mortality was 6.0% (1788 of 29,958) with NBD accounting for 11.5% (n = 206). NBD proportions in DCAS were similar for OIF (21.1%) and OEF (21.4%) (Fig. 1A). In the DODTR, NBD proportion was slightly higher in OIF (12.4%) than in OEF (9.8%) but there was not a statistically significant difference (P = 0.10; Fig. 1B). Comparative analyses demonstrated a higher proportion of NBD in OND and OIF than in OEF (22.2% vs 12.3% vs 9.8%, respectively; P < 0.0001), in female versus male (27.3% vs 11.2%; P < 0.0001), in Air Force versus Army (18.8% vs 12.5%) and lowest in Marine and Navy SMs (8.0% and 7.7%, respectively) (Supplementary Table 2, The proportion of NBD was highest in Operation Iraqi Resolve and OND (Supplementary Figure 1, No significant differences in the proportions of NBD were found according to race or ethnicity, which ranged from 9.4% to 17.2% (P = 0.58). Although explosion was the major cause of battle death (63% all deaths), nonbattle GSW (33.5%) and MVC (32.5%) were major causes of NBD (Table 1). Being injured in the head/neck, external injuries including the skin and soft tissues, and thoracic injury occurring during NB activities accounted for 50.0%, 40.8%, and 34.0% of all NBD, respectively, (Supplementary Table 1, Nonsurvivors were associated with a higher median ISS and mISS, 25 (IQR, 9–33) and 25 (9–26), respectively whether caused by BI or NBI (Supplementary Table 1,; all P > 0.3).

TABLE 1 - Causes of Nonbattle Injury and Nonbattle Deaths
Rank Causes of NBI No. (%) Causes of NBD No. (%)
1 Fall 2178 (21.4) Gunshot wound 69 (33.5)
2 Motor vehicle crash 1921 (18.8) Motor vehicle crash 67 (32.5)
3 Machinery/equipment 1283 (12.6) Helicopter crash 21 (10.2)
4 Blunt objects 1107 (10.7) Inhalation injury 8 (3.9)
5 Gunshot wounds 728 (7.1) Machinery/equipment 6 (2.9)
6 Sport 697 (6.8) Submersion/drowning 6 (2.9)
7 Knife/other sharp object 549 (5.4) Fall 5 (2.4)
8 Crush 322 (3.2) Fire/fame 4 (1.9)
9 Fire/flame 251 (2.5) Crush 3 (1.5)
10 Altered ROM 205 (2.0) Electrical 3 (1.5)
11 Helicopter crash 150 (1.5) Explosive-NOS 3 (1.5)
12 Altercation/fight 117 (1.1) Hand grenade 2 (1.0)
13 Explosive-NOS 87 (0.9) Knife/other sharp object 2 (1.0)
14 Electrical 86 (0.8) Plane crash 2 (1.0)
15 Chemical 71 (0.7) UXO 2 (1.0)
Other 451 (4.5) Other 3 (1.5)
Due to rounding, totals may not be 100%.
Any nonsport related sprain/strain/dislocation.
If mortar rounds are being loaded, unloaded, or transported, they are classified as UXO and NBI.
If mortar rounds are being loaded, unloaded, or transported, they are classified as UXO and NBI.NBI indicates nonbattle injuries; NBD, nonbattle deaths; ROM, range of motion; UXO, unexploded ordnance.

Etiologies of NBD

Primary causes and type of injury were different between NBI and NBD (Table 1). The most common etiologies of NBD were GSW (33.5%), MVC (32.5%), helicopter crash (10.2%), flame/inhalation injury (5.8%), misused machinery/equipment (2.9%), and submersion/drowning (2.9%) (Table 1). Among 196 of 206 NBD with documented manner intent, self-inflicted accounted for 21.9% (n = 43), in which GSW accounted for 71.7% of all self-inflicted deaths among documented manner of intent (n = 60); while unintentional manner of intent was 23.2% (n = 16) (data not shown). DCAS data indicate that approximately 60% and 25% of NBD were accidental and self-inflicted, respectively, in both OIF and OEF (Fig. 2A and B).

Distribution of causes of NBD by military service branches. Figure depicts the distribution of common possible causes involved nonbattle deaths (NBD) by military service branch during Operation Iraqi Freedom (OIF; panel A) and Operation Enduring Freedom (OEF; panel B).

Trend in NBD and Forecasting

The quarterly time plots of NBD cases from January 2003 to December 2014 show a generally declining trend in the numbers of NBD in OIF and OEF (Fig. 3A and B). Trends in proportion of NBD using WMA method show an initial decrease from 35% in 2003 to 21% in 2006 in OIF and from 82% in 2002 to 21% in 2011 in OEF, followed by consistent annual rates of 21% for both OIF and OEF (Fig. 3A and B). For OIF and OEF combined, the overall WMA mean (±SD) for the proportion of NBD was 23.6% (±5.7%) (Fig. 3C). However, the decrease in the proportion of NBD over time by quarter was not significant according to regression analysis (P = 0.84; Fig. 3C).

Trends in proportion of nonbattle deaths. Figure depicts the frequency, proportion of nonbattle death (NBD), trend in NBD quarterly from 2001 to 2014. Line graphs depict the frequency of NBD and weighted moving average (WMA) trends in NBD (red lines) in Iraq War [Operation Iraqi Freedom (OIF); Panel A], in Afghanistan War [Operation Enduring Freedom (OEF); panel B], and NBD overall in combined OIF and OEF (panel C) with a fitted regression line for trend in NBD in 2003–2014 (black line).

In the time series analysis, the plot of time series shows the original data of proportion of NBD (d = 0) were stationary with the mean (±SD) of 23.8% (±2.2%) (Supplementary Figure 2A, The autocorrelation function plot suggests that the change in proportion of NBD was highly correlated with lag 1 (p = 1), and the partial autocorrelation function plot suggests that the best model for moving average uses lag 0 (Supplementary Figure 2B and C, The best ARIMA model that is suitable for estimation was ARIMA (1,0,0). Assuming stable battlefield injury risk conditions, the proportion of NBDs from 2015 to 2025 would be predicted to be 24.0% (±2.7%) (Fig. 4).

Time series analysis with ARIMA (1,0,0) model with forecasting to year 2025. Figure depicts the time series analysis with ARIMA (1,0,0) model with forecasting to year 2025. ARIMA indicates autoregressive integrated moving average.


This is the first study to describe the high incidence of NBD during OEF/OIF. Overall, we found that NBD accounted for approximately one-fifth of all deaths and this NBD rate was consistent over time and predictable. Although NBI has been recognized as a significant burden of injury with resultant attrition from the combat zone, NBD has much longer-term impacts on the military and SM families. This analysis expands on our previous work to quantify and forecast injuries in which we determined that NBI accounts for 34% of injuries seen at R3MTF in Iraq and Afghanistan and almost 12% of all deaths in deployed SMs.1

The DoDTR, which is the primary source for clinical combat casualty care data, underestimates the NBD because patients who died before arrival at the R3MTF were not captured in this registry until 2014 and those that died before receiving clinical care (classified as killed in action) or who were treated at a lower level of care and returned to duty were not captured at all. To address this limitation of the clinical registry, we used the DCAS data, which may provide a more accurate estimate of NBD rates during military deployment. The current analysis demonstrates that approximately 1 in 5 deaths was due to NBI, and may have a preventable component. The significant proportion of NBD classified as being the result of self-inflicted injuries, suggests significant roles for predeployment screening for existing behavioral health conditions that may place SMs at highest risk in the deployed environment and for mental health support in the deployed environment. Analysis also revealed that the DoDTR accounted for only about 15% of all NBD, suggesting that a more comprehensive approach to this database may be warranted.

The limitations of the DoDTR with respect to capturing prehospital deaths are highlighted when data are compared to the DCAS system. The identification of differing causes of death between NBI and BI, underscores the need for ongoing mortality reviews in trauma registries to provide information on the specific etiology of death, which can lead to potential opportunities for intervention. Comprehensive mortality analyses to understand the etiology of death from trauma can only be performed if assessments of prehospital deaths are included in trauma registries.

The DoDTR data has previously been used to examine NBI. The DoDTR captured 206 NBD from 2003 to 2014 and accounted for 11.5% of all SM mortality.1 Until 2014, to be captured for inclusion into the DoDTR, a casualty had to be admitted to a R3MTF. R3MTFs are the highest level of care on the battlefield, so if a casualty arrived at a more austere level of care (eg, Role 2 Forward Surgical Team) and died before being transferred to R3MTF, they would not be included in the DoDTR. After 2014, this inclusion criterion changed and the DoDTR began to capture all injured SMs, although the lack of available medical documentation from far forward medical units poses a significant challenge. So, although NBD accounts of 11.5% (n = 206) of all U.S. SM mortality that arrived at the R3MTFs (n = 1788), it likely underestimates the true incidence of NBD, compared to NBD rates reported in DCAS.

In this study, the primary causes of NBD were GSW and MVC, accounting for two-thirds of all NBD. Both of these mechanisms of injuries may have potentially preventable components. Of the 69 NBD caused by GSW, 75.4% of those (n = 52) were to the head. Suicide has been reported to be the second leading cause of death in the U.S. military.21 Suicide risk is higher in deployed SM and has surpassed the U.S. general population since 2008,22–24 and GSW from firearms was the primary cause of death. Although these data cannot discriminate accidental from intentional GSW, our finding that 71.7% of self-inflicted deaths was due to GSW is consistent with previous findings for military suicides.25 Predeployment screening and targeted efforts towards those in the combat zone have the potential to decrease the risk of suicide while deployed, and may also empower SMs with skills that could decrease suicide risk postdeployment.25

Visual inspection of Figure 3 reveals that, although the ARIMA model confirms a stable proportion of nonbattle related deaths, there may be periods when the proportion was higher and more variable. Early during OEF and late during OIF were periods when the extent of combat was lower, as reflected by relatively lower numbers of overall deaths.4 During these periods, the proportion of deaths attributable to NBI accounted for over 50% in some years.4 These proportions are similar to the 52% reported for the conflict in the Persian Gulf War in the early 1990's, when combat operations represented a relatively small proportion of the overall deployment period.5 These levels of NBD may reflect a relatively stable absolute level of NBD that may approach levels encountered during training operations.

The present findings that 2 of the top 3 causes of NBDs were aircraft crashes and ground vehicular accidents are consistent with the leading causes of death in DoD training accidents.26,27 Considering that these incidents occur in the noncombat training environment, it would be expected that the same issues may occur in the deployed environment. Individual US military units track injury and death due to training accidents but there is currently no DoD-wide clinical registry to assess and study NBD in the nondeployed environment. As combat operations are decreasing, turning focus to understanding NBI and NBD in the training environment and expanding the DoDTR to include all training centers could facilitate an improved understanding of NBI and NBD, and ideally lead to increased prevention in training environments, and in conflict environments. Although the DoD places a strong emphasis on training and maintenance, it has been recognized that additional measures in these areas and others, such as equipment modifications, may be needed.26,27 Whether the deployed environment may place extra strain on equipment or additional stressors on vehicle operators or crews, making them more susceptible to accidents, cannot be determined from the present study.

Our trend analysis demonstrated that the trend in proportion of NBD did not decrease over time, stabilizing at approximately 21% in OIF after 2006, and OEF after 2011. Despite changes over time in the number of SMs at risk, and regardless of the statistical method used, WMA demonstrates an overall NBD rate of 24%. Assuming stable battlefield conditions the ARIMA model for time-series analysis predicts that the proportion of NBD from 2015 through 2025 or later would remain at approximately 24%.

Although it is intuitive that NBI and NBD should have a preventable component, there has been relatively little research documenting their causes and mechanisms, which in turn could inform preventative strategies. Research efforts to date have focused primarily on the incidence, prevention and treatment of BI; however, NBI has carried a significant impact on military operations, including increasing cost of care, depleting medical resources, and decreasing military forces, and its etiologies are potentially preventable.8–12 Previous research has focused primarily on the incidence and prevention of NBI in regards to the anatomic region injured, aggregate rates of NBI and diseases that contribute to SM attrition from the combat zone.28,2911,30–34 Improving our understanding of the mechanisms of injury, situational influences, etiologies and trends of NBD may help inform medical leaders and operational commanders of this risk, increase awareness amongst the medical community, inform research priorities for materiel or solutions, and focus the improvement efforts of military safety centers


Our study has some limitations. First, the accuracy of the data in a retrospective study depends on complete and accurate documentation, so results should be interpreted cautiously. Second, the DoDTR does not capture all data, particularly prehospital deaths; thus, a selection bias exists in the DoDTR data favoring less seriously injured SM who are more likely to survive to reach the R3MTF. This is likely reflected in the fact that the NBD rate recorded in DCAS data (21%) is nearly twice that recorded in DoDTR data (12%). DCAS data were used to create a more comprehensive picture but, given that DCAS data are captured only for administrative/tracking/reporting purposes, the data are limited by not being clinically validated. In this study we mitigated the above issues by analyzing both data sources, DoDTR for individual-level analysis and DCAS for aggregate-level analysis. Although multiple efforts are underway to improve the fidelity of the mortality data to inform and improve the DoD trauma system, there is currently no readily available data source that captures all of the mortality data from the recent conflicts. Third, the DoDTR does not include all traumatically injured SM who were treated at a lower level of care and returned to duty. There is a significant discrepancy in the number of traumatically injured SMs in the DODTR compared to DCAS; the DoDTR captured about 52% of those injured when compared to DCAS. However, a strength of the DoDTR is that data were collected prospectively by dedicated personnel on all patients presenting at the R3MTF and validated using available medical records.


Death from noncombat causes is common in the deployed setting, having caused approximately 1 in 5 deaths among U.S. SMs deployed to Iraq and Afghanistan. The majority of these deaths were attributable to GSW or vehicular accidents, and 1 in 5 NBD resulted from self-inflicted injuries. The death rate was consistent and was predicted to remain at this level under equivalent conditions. The findings from this study may aid military leadership in developing targeted safety interventions to reduce NBI and deaths to the minimum possible.


The authors acknowledge the Department of Defense Trauma Registry (DoDTR) and Patient Administration Systems and Biostatistics Activity (PASBA) for providing data for this study.


1. Le TD, Gurney JM, Nnamani NS, et al. A 12-year analysis of nonbattle injury among US service members deployed to Iraq and Afghanistan. JAMA Surg 2018; 153:800–807.
2. Hauret KG, Pacha L, Taylor BJ, et al. Surveillance of disease and nonbattle injuries during US army operations in Afghanistan and Iraq. US Army Med Dep J 2016. 15–23.
3. Patel AA, Hauret KG, Taylor BJ, et al. Non-battle injuries among U.S. army soldiers deployed to Afghanistan and Iraq, 2001-2013. J Safety Res 2017; 60:29–34.
4. Defense Casualty Analysis System. Conflict Casualties. Available at: Accessed December 23, 2019
5. Writer JV, DeFraites RF, Brundage JF. Comparative mortality among US military personnel in the Persian Gulf region and worldwide during Operations Desert Shield and Desert Storm. JAMA 1996; 275:118–121.
6. Writer JV, DeFraites RF, Keep LW. Non-battle injury casualties during the Persian Gulf War and other deployments. Am J Prev Med 2000; 18: (3 Suppl): 64–70.
7. Jadick R, Hayden T. On Call in Hell: A Doctor's Iraq War Story. New York, NY: NAL Caliber; 2007.
8. Jones BH, Canham-Chervak M, Canada S, et al. Medical surveillance of injuries in the U.S. military descriptive epidemiology and recommendations for improvement. Am J Prev Med 2010; 38: (1 Suppl): S42–S60.
9. Smith GS, Dannenberg AL, Amoroso PJ. Hospitalization due to injuries in the military. Evaluation of current data and recommendations on their use for injury prevention. Am J Prev Med 2000; 18: (3 Suppl): 41–53.
10. Sanders JW, Putnam SD, Frankart C, et al. Impact of illness and non-combat injury during operations Iraqi freedom and enduring freedom (Afghanistan). Am J Trop Med Hyg 2005; 73:713–719.
11. Eaton M, Marshall SW, Fujimoto S, et al. Review of non-battle injuries in air force personnel deployed in support of operation enduring freedom and operation Iraqi freedom. Mil Med 2011; 176:1007–1014.
12. Withers BG, Erickson RL, Petruccelli BP, et al. Preventing disease and non-battle injury in deployed units. Mil Med 1994; 159:39–43.
13. Blood CG, Jolly R. Comparisons of disease and nonbattle injury incidence across various military operations. Mil Med 1995; 160:258–263.
14. Garfield RM, Neugut AI. Epidemiologic analysis of warfare. A historical review. JAMA 1991; 266:688–692.
15. Wojcik BE, Hassell LH, Humphrey RJ, et al. A disease and non-battle injury model based on Persian Gulf War admission rates. Am J Ind Med 2004; 45:549–557.
16. United States Department of Defense Joint Trauma System Center of Excellence for Trauma. Available at: Accessed December 31, 2020.
17. Le TD, Orman JA, Stockinger ZT, et al. The military injury severity score (mISS): a better predictor of combat mortality than injury severity score (ISS). J Trauma Acute Care Surg 2016; 81:114–121.
18. DoD Trauma Registry Data Dictionary. Version 3.2, Build 44 - Apr 2018, Updated Nov 2019. Available at: Accessed December 31, 2020.
19. Dickey DAF, Wayne A. Distribution of the estimators for autoregressive time series with a unit root. J Am Stat Assoc 1979; 74:427–431.
20. Ljung G, Box G. On a measure of lack of fit in time series models. Biometrika 1978; 65:297–303.
21. Hoyert DL, Xu J. Deaths: preliminary data for 2011. Natl Vital Stat Rep 2012; 61:1–51.
22. Eaton KM, Messer SC, Garvey Wilson AL, et al. Strengthening the validity of population-based suicide rate comparisons: an illustration using U.S. military and civilian data. Suicide Life Threat Behav 2006; 36:182–191.
23. Schoenbaum M, Kessler RC, Gilman SE, et al. Predictors of suicide and accident death in the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS): results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). JAMA Psychiatry 2014; 71:493–503.
24. Hoge CW, Castro CA. Preventing suicides in US service members and veterans: concerns after a decade of war. JAMA 2012; 308:671–672.
25. Uniformed Services University. Center for Development Psychology. Suicide in the Military. Military Suicide Web site. Published 2019. Available at: Accessed September 8, 2020.
26. Remfer K. Army Vehicle Rollover Deaths Spiked This Fiscal Year. Available at: Accessed December 31, 2020.
27. Ritchie IL. Military Injuries and Deaths Off the Battlefield Are Increasing. Available at: Accessed December 31, 2020.
28. Hollingsworth AC, Bowley DM, Lundy JB. The operative burden of general surgical disease and non-battle injury in a deployed military treatment facility in Afghanistan. Mil Med 2016; 181:1065–1068.
29. Wilson CD. Disease and non-battle injury (DNBI) have always caused much greater morbidity. J R Army Med Corps 2013; 159:135.
30. Madson AQ, Tucker D, Aden J, et al. Non-battle craniomaxillofacial injuries from U.S. military operations. J Craniomaxillofac Surg 2013; 41:816–820.
31. Hauret KG, Taylor BJ, Clemmons NS, et al. Frequency and causes of nonbattle injuries air evacuated from operations Iraqi freedom and enduring freedom, U.S. army, 2001-2006. Am J Prev Med 2010; 38: (1 Suppl): S94–S107.
32. Belmont PJ Jr, Goodman GP, Waterman B, et al. Disease and nonbattle injuries sustained by a U.S. Army Brigade Combat Team during operation Iraqi freedom. Mil Med 2010; 175:469–476.
33. Kauvar DS, Cancio LC, Wolf SE, et al. Comparison of combat and non-combat burns from ongoing U.S. military operations. J Surg Res 2006; 132:195–200.
34. Skeehan CD, Tribble DR, Sanders JW, et al. Nonbattle injury among deployed troops: an epidemiologic study. Mil Med 2009; 174:1256–1262.

nonbattle death; nonbattle death in deployed environment; nonbattle injury

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.