BG Margaret C. Wilmoth, USAR*; Andrea Linton, MS ; Richard Gromadzki, DSc ; Mary J. Larson, PhD, MPH ; Thomas V. Williams, PhD ; Jonathan Woodson, MD

MILITARY MEDICINE, 180, 1:53, 2015 Factors Associated With Psychiatric Evacuation Among Service Members Deployed to Operation Enduring Freedom and Operation Iraqi Freedom, January 2004 to September 2010 BG Margaret C. Wilmoth, USAR*; Andrea Linton, MS ; Richard Gromadzki, DSc ; Mary J. Larson, PhD, MPH ; Thomas V. Williams, PhD ; Jonathan Woodson, MD ABSTRACT Objectives: To calculate the annual rate of psychiatric evacuation of U.S. Service members out of Iraq and Afghanistan and identify risk factors for evacuation. Methods: Descriptive and regression analyses were performed using deployment records for Service members evacuated from January 2004 through September 2010 with a psychiatric diagnosis, and a 20% random sample of all other deployers (N = 364,047). Results: A total of 5,887 deployers psychiatrically evacuated, 3,951 (67%) of which evacuated on first deployment. The rate increased from 72.9 per 100,000 in 2004 to 196.9 per 100,000 in 2010. Evacuees were overrepresented in both combat and supporting duty assignments. In multivariate analysis, Army active duty had the highest odds of evacuation relative to Army National Guard (adjusted odds ratio [AOR] 0.852, 95% confidence interval [CI] 0.790 0.919), Army Reserve (AOR 0.825, 95% CI 0.740 0.919), and all other components. Accessions in 2005 had the highest risk (AOR 1.923, 95% CI 1.621 2.006) relative to pre-2001 accessions. Conclusions: Risk for psychiatric evacuation is highest among the Army Active Component. A strong link between multiple deployments or combat-related exposure and psychiatric evacuation is not apparent. Increased risk among post-2001 accessions suggests further review of changes in recruitment, training, and deployment policies and practices. INTRODUCTION Although advances in weapons technology, protective equipment, and battlefield triage have dramatically reduced troop fatalities relative to previous conflicts, 1 unit attrition remains a persistent challenge in Operation Enduring Freedom (OEF) and Operation Iraqi Freedom (OIF). From 2001 through September 2010, the fatalities from both operations totaled 5,707, with an additional 40,500 wounded in action. 2 There were an accompanying 62,087 aeromedical evacuations, approximately 80% of which resulted from nonbattle injury or disease. 3 From 2004 onward, a steady increase in the proportion of evacuations attributed to psychiatric disorders has been observed, as well as a reduced rate of psychiatric evacuees returning to theater after evacuation, relative to other medical evacuees. 3,4 *Byridine F. Lewis School of Nursing and Health Professions, Georgia State University, Atlanta, GA. Axiom Resource Management Inc., 2941 Fairview Park DDrive, Suite 900, Falls Church, VA 22042. Heller School for Social Policy and Management, Brandeis University, 415 South Street, MS 035, Waltham, MA 02454-9110. Defense Health Costing Analyses and Program Evaluation, Defense Health Agency, kassistant Secretary of Defense for Health Affairs, Department of Defense, 7700 Arlington Boulevard, Suite 5101, Falls Church, VA 22042-5101. Some findings from this study were presented at the 2013 Annual Research Meeting of Academy Health, as part of a panel presentation entitled, Deployment-Related Experiences Among U.S. Military Members: Challenges for Effective Assessment and Response to Psychological Injury, in Baltimore, MD. The opinions and assertions therein are those of the authors and do not necessarily reflect the view of the Department of Defense. doi: 10.7205/MILMED-D-14-00213 Evacuation is a significant and costly event for the evacuee, their unit, and the military. Psychiatric evacuees are generally accompanied by one or more nonmedical attendants, usually members of the same unit, which results in the loss of multiple unit members for every psychiatric evacuation that occurs. After returning home, the evacuee inevitably faces a lengthy recovery process likely to impact his family and professional life. If the evacuee is perceived to be emotionally unsuited for the rigors of military service, the confidence and trust of fellow service members may erode and the evacuees military careers may be adversely affected. The combined cost associated with the aeromedical evacuation event, loss of unit manpower, and the postdeployment health care burden placed on the DoD and Veteran s Affairs (VA) health care systems are difficult to estimate, but likely to be substantial. There is ample literature regarding the association of postdeployment psychiatric outcomes with combat exposure 5 8 and noncombat deployment to OEF/OIF. 9,10 Fewer studies exist, however, that explore the factors associated with psychiatric evacuation from OEF/OIF. Ferrier-Auerbach et al 11 identified combat-related injury and exposure to explosive blast as significant predictors of emotional distress during deployment, but an outcome of evacuation was not addressed. Findings of higher frequency of postdeployment mental health concerns reported among National Guard/ Reserve members relative to active duty (AD) personnel, 12,13 have fostered a hypothesis that the rise in psychiatric evacuations may be the result of deploying more National Guard/ Reserve units relative to AD units over time. 4 The cumulative psychological effect of multiple deployments has also been MILITARY MEDICINE, Vol. 180, January 2015 53

suggested as a contributing factor. 4 Clear evidentiary support for these notions, however, is lacking. This article aims to provide a comprehensive examination of the risk factors associated with psychiatric evacuation among U.S. armed force members deployed to OEF/OIF. We present the trend in annual evacuations as a count per 100,000 deployers, and address the question of whether the service member s service branch/component group or repeat deployments are associated with risk of evacuation. METHODOLOGY Study Design A retrospective analysis was performed using administrative and operational Department of Defense (DoD) data from January 2004 to September 2010. Clinical information on evacuations was obtained from the TRANSCOM Regulating and Command and Control Evacuation System (TRAC2ES). The Contingency Tracking System (CTS) was the data source for deployment location, start date, and demographic information. All personnel with psychiatric evacuation (n = 5,887) and a 20% random sample of all other deployed personnel (n = 364,047) were selected for inclusion in the study sample. Analysis was restricted to personnel with deployments in support of OEF/OIF during the period January 2004 through January 2010. The definition of psychiatric evacuation was aeromedical transport out of the theater of operation (most commonly to Germany) and a primary or secondary diagnosis of mental disorder, using International Disease Classification, 9th revision, Clinical Modification (ICD9-CM) diagnosis codes in the range from 290 to 319. This study was approved by the Assistant Secretary of Defense for Health Affairs/TRICARE Management Activity (OASD HA/TMA), Human Subjects in Research Protection Office under 32 CFR 219.101(b)(4). Statistical Analysis The primary outcome was psychiatric evacuation. The annual evacuation rate is expressed as a rate per 100,000 deployers where the denominator is the number of Service members who were deployed at any time during that year. For the 34 service members who experienced more than one psychiatric evacuation, only the first evacuation was examined. Sample weights were applied to adjust the 20% sample to the total deployed population. The independent variables included a concatenation of the Service branch (Army, Navy, Air Force (AF), Marines, and Coast Guard) and component (AD, Reserve, National Guard); total number of deployments, personal characteristics (gender, age group, race, education, marital status, and number of dependents) and military characteristics (year of accession into military service, primary occupation code assigned at accession); and the deployment location (OEF, OIF). These characteristics were measured at the start of the first deployment as the analysis was conducted at the person-level. Descriptive analysis of the first psychiatric evacuation includes the number of prior deployments, total months deployed (inclusive of all deployments), and event outcome. Binary logistic regression analysis was used to estimate the odds of psychiatric evacuation associated with each personal and deployment characteristic. The variables for occupational code, rank, and pay grade were omitted from the model on the basis of collinearity with Service branch, and variables for age and education, respectively. Statistical significance was assessed at a 95% confidence level. All analyses were performed using SPSS version 19.0. RESULTS A total of 5,887 of the 1,796,687 deployed service members (32.8 per 100,000 deployers) experienced one or more psychiatric evacuations from either OEF or OIF during the study period. The annual psychiatric evacuation rate increased from 72.7 per 100,000 deployers in 2004 to 196.9 in 2010, with a spike in 2007 (192.2) (Fig. 1). Data collection terminated as of September 2010, so the 2010 rate is based on only 9 months of data. Characteristics Associated With Evacuation Personal characteristics of psychiatric evacuees are compared to other deployers in Table I. Relative to nonevacuees, a disproportionally higher number of evacuees were women, aged 17 to 24 years, and white. Also overrepresented among the evacuee population were those with no more than a high school diploma, never married, and with one or two dependents. Military characteristics of evacuees and other deployers at the time of first deployment are compared in Table II. The psychiatric evacuee population was overrepresented among Army components, junior enlisted personnel, and accessions into military service between 2003 and 2008. Overrepresentation of evacuees is present among combat occupations, as well as occupations generally considered to provide supporting functions. Notably, occupations FIGURE 1. Annual psychiatric evacuation rate per 100,000 deployed personnel, Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF), January 2004 through September 2010. 54 MILITARY MEDICINE, Vol. 180, January 2015

TABLE I. Demographic Characteristics of Psychiatric Evacuees and Other Deployers, OEF/OIF, January 2004 Through September 2010 Evacuees Other Deployers Characteristics at First Deployment N* n % n % p Value Total 1,796,687 5,887 100.0 1,790,800 100.0 Gender Male 1,591,783 5,013 85.2 1,586,770 88.6 <0.01 Female 204,904 874 14.8 204,030 11.4 <0.01 Age Group (Years) 17 24 806,085 3,260 55.4 802,825 44.8 <0.01 25 34 591,974 1,804 30.6 590,170 33.0 <0.01 35 44 313,706 646 11.0 313,060 17.5 <0.01 45+ 84,922 177 3.0 84,745 4.7 <0.01 Race/Ethnicity White 1,179,595 4,115 69.9 117,5480 65.6 <0.01 Black 272,473 818 13.9 271,655 15.2 0.006 Hispanic 190,437 542 9.2 189,895 10.6 0.001 Asian 75,876 191 3.2 75,685 4.2 <0.01 American Indian 29,061 81 1.4 28,980 1.6 0.141 Other 492 9 0.2 8,360 0.5 <0.01 Education High School Diploma or Less 132,3649 4,934 83.8 1,318,715 73.6 <0.01 Less Than 4 Years of College 169,295 365 6.2 168,930 9.4 <0.01 Bachelor s Degree 201,994 349 5.9 201,645 11.3 <0.01 Advanced Degree 70,115 115 2.0 70,000 3.9 <0.01 Marital Status Single, Never Married 858,068 2,893 49.1 855,175 47.8 0.033 Married 865,964 2,744 46.6 863,220 48.2 0.015 Other 72,098 248 4.2 71,850 4.0 0.434 Number of Dependents No Dependents 778,612 2,577 43.8 776,035 43.3 0.497 1 2 Dependents 622,701 2,186 37.1 620,515 34.7 <0.01 3 or More Dependents 393,394 1,114 18.9 392,280 21.9 <0.01 *Weighted to reflect 100% of the deployed population. A value of unknown was coded for race/ethnicity (n = 40,876), education (n = 31,634), marital status (n = 557), and dependents (n = 1,980). requiring advanced skills or training, such as special forces and pilots, are significantly underrepresented among the psychiatric evacuees. Deployment characteristics of psychiatric evacuees are presented in Table III. The majority (n = 3,951, 67.1%) evacuated from their first deployment, and 93.5% overall evacuated by the second deployment. Approximately one third (32.9%) of evacuees accrued 12 or more months of total deployment time before evacuation, the remaining two thirds evacuated with less than 1 year of total deployment time. Together, primary diagnoses of depression (24.9%) adjustment disorders (14.4%), and post-traumatic stress disorder (PTSD) (10.5%), accounted for nearly half of psychiatric evacuations. Psychotic diagnoses accounted for approximately 18%. Only a small fraction (n = 218, 3.7%) of psychiatric evacuees were evacuated for a medical condition with a secondary psychiatric diagnosis only. Depression accounted for significantly more evacuations among women (23.9%) than men (30.4%) while men were more frequently evacuated than women for PTSD (11.3% and 5.7%, respectively) and postconcussion syndrome (6.0% and 1.4%, respectively). All other evacuation diagnoses occurred at a similar rate for men and women (data not shown). Disposition of Evacuation Person-level outcomes among the psychiatric evacuees are also presented in Table III. The majority (n = 4,754, 80.8%) of evacuees did not return to theater nor had a subsequent deployment by end of data collection. A total of 592 evacuees (10.1%) returned to theater and completed their deployment, but did not subsequently deploy, while 89 evacuees (1.5%) returned to theater and deployed again at a later time. A small number of evacuees (n = 21, 0.4%) returned to theater and experienced a second psychiatric evacuation from the same deployment. A total 433 psychiatric evacuees (7.4%) did not return theater, but deployed again later during the data collection period. Likelihood of Psychiatric Evacuation The results of the multivariate logistic regression analysis on psychiatric evacuation are presented in Table IV. Relative to Army AD (reference group), all other components had a reduced odds of psychiatric evacuation. Army National Guard had reduced odds of evacuation (adjusted odds ratio [AOR] 0.852, 95% confidence interval [CI] 0.790 0.919), as did Army Reserve (AOR 0.825, 95% CI 0.740 0.919), Navy MILITARY MEDICINE, Vol. 180, January 2015 55

TABLE II. Military Characteristics of Psychiatric Evacuees and Other Deployers, OEF/OIF, January 2004 Through September 2010 Evacuees Other Deployers Characteristics at First Deployment N* n % n % p Value Total 1,796,687 5,887 100.0 1,790,800 100.0 Service Branch/Component Active Duty Army 603,335 3,375 57.3 599,960 33.5 <0.01 Active Duty Navy 260,204 234 4.0 259,970 14.5 <0.01 Active Duty Air Force 239,463 233 4.0 239,230 13.4 <0.01 Active Duty Marines 197,985 470 8.0 197,515 11.0 <0.01 Reserve Army 109,727 402 6.8 109,325 6.1 0.021 Reserve Navy 29,643 73 1.2 29,570 1.7 0.013 Reserve Air Force 25,065 10 0.2 25,055 1.4 <0.01 Reserve Marines 29,330 45 0.8 29,285 1.6 <0.01 National Guard Army 252,745 1,025 17.4 251,720 14.1 <0.01 National Guard Air Force 45,205 20 0.3 45,185 2.5 <0.01 Grade Junior Enlisted 984,152 4,392 74.6 979,760 54.7 <0.01 Senior Enlisted 571,449 1,154 19.6 570,295 31.8 <0.01 Junior Officer 143,024 214 3.6 142,810 8.0 <0.01 Senior Officer 98,062 127 2.2 97,935 5.5 <0.01 Accession Year Before 2001 843,809 2,039 34.6 841,770 47.0 <0.01 2001 118,621 371 6.3 118,250 6.6 0.353 2002 133,160 455 7.7 132,705 7.4 0.352 2003 137,600 535 9.1 137,065 7.7 <0.01 2004 125,107 507 8.6 124,600 7.0 <0.01 2005 116,670 595 10.1 116,075 6.5 <0.01 2006 125,232 602 10.2 124,630 7.0 <0.01 2007 96,540 445 7.6 96,095 5.4 <0.01 2008 73,502 282 4.8 73,220 4.1 0.007 2009 25,576 56 1.0 25,520 1.4 0.002 2010 870 0 0.0 870 0.0 0.091 DoD Primary Occupational Code Infantry 197,490 980 16.6 196,510 11.0 <0.01 Special Forces 7,444 9 0.2 7,435 0.4 0.002 Artillery and Gunnery 50,991 231 3.9 50,760 2.8 <0.01 Pilots and Other Air Crew 115,806 101 1.7 115,705 6.5 <0.01 Boatswain and Other Seamen 23,384 39 0.7 23,345 1.3 <0.01 Combat Engineers, Armor, and Amphibious 61,087 302 5.1 60,785 3.4 <0.01 Motor Vehicle Operators 68,017 477 8.1 67,540 3.8 <0.01 Auto, Aircraft, Armament Maintenance and Repair 247,044 564 9.6 246,480 13.8 <0.01 Construction, Facility Maintenance,and Utilities 203,250 565 9.6 202,685 11.3 <0.01 Medical, Dental, and Veterinary Personnel 98,354 479 8.1 97,875 5.5 <0.01 Scientists, Lawyers, Chaplins 49,199 114 1.9 49,085 2.7 <0.01 Admin, Personnel, Legal, and Accounting 117,819 309 5.2 117,510 6.6 <0.01 Food, Laundry, and Auxillary Labor 42,865 185 3.1 42,680 2.4 <0.01 *Weighted to reflect 100% of deployed population. No psychiatric evacuations occurred among the 2,890 Active and 1,095 Reserve Coast Guard members deployed during the study period. A total of 10,609 deployers had null values for occupation code. No significant differences were observed among administrative and logistics personnel, law enforcement and security personnel, combat communications and intelligence personnel, undesignated general or excecutive officers, or student/cadets. Reserve (AOR 0.585, 95% CI 0.461 0.742), and Marine Corps AD (AOR = 0.390, 95% CI 0.353 0.430). The remaining Service components had further reduced odds of psychiatric evacuation. Regarding our second hypothesis, the odds of evacuation increased with a second deployment (AOR 1.248, 95% CI 1.174 1.326) and third deployment (AOR 1.381, 95% CI 1.250 1.526, while the odds decreased for those deploying more than 4 times (AOR 0.400, 95% CI 0.220 0.725). First deployment to OEF was associated with a lower likelihood (AOR 0.814, 95% CI 0.752 0.880) of evacuation. Other demographic risk factors for psychiatric evacuation included female gender (AOR 1.551, CI 1.438 1.673). Increasing age group was protective for psychiatric evacuation, as was increasing education level. Being married (AOR 1.207, CI 1.109 1.313) or previously married (AOR 1.323, CI 1.145 1.529) had slightly elevated odds of evacuation relative to being single, as did having one or two dependents 56 MILITARY MEDICINE, Vol. 180, January 2015

TABLE III. Characteristics of Deployment, Evacuation Diagnoses, and Evacuation Outcomes, Among Psychiatric Evacuees on First Evacuation, OEF/OIF, January 2004 Through September 2010 Evacuee Characteristics at First Psychiatric Evacuation n % Cumulative % Total Number of Psychiatric Evacuees 5,887 100.0 100.0 Deployment Number for First Evacuation 1 Deployment 3,951 67.1 67.1 2 Deployments 1,553 26.4 93.5 3 Deployments 333 5.7 99.2 4+ Deployments 50 0.8 100.0 Total Deployed Time Up To First Evacuation 0 2 Months 411 7.0 7.0 2 4 Months 683 11.6 18.6 4 6 Months 790 13.4 32.0 6 8 Months 803 13.6 45.6 8 10 Months 679 11.5 57.1 10 12 Months 565 9.6 66.7 >12 Months 1,935 32.9 99.6 Missing 21 0.4 100.0 Principal Diagnosis at First Evacuation Depression 1,465 24.9 25.8 Adjustment Disorders 848 14.4 39.3 Other or Unspecified Psychotic Disorder 630 10.7 50.0 PTSD 619 10.5 60.5 Serious and Persistent Mental Illness 454 7.7 68.2 Postconcussion Syndrome 314 5.3 73.6 Al Other Psychiatric Diagnoses 1,342 22.8 96.3 Psychiatric Diagnosis Secondary to Medical Diagnosis 218 3.7 100.0 Outcome of First Evacuation Returned and Completed; No Subsequent Deployment 592 10.1 10.1 Returned and Reevacuated 21 0.4 10.4 Returned and Completed; Subsequently Deployed 89 1.5 11.9 Not Returned; Subsequently Deployed 433 7.4 19.0 No Return to Theater or Subsequent Deployment 4,754 80.8 100.0 Serious and persistent mental illness includes schizophrenic disorders, bipolar affective disorder, and other manic depressive psychosis. (AOR 1.144, 95% CI 1.058 1.243) or three to four dependents (AOR 1.142, 95% CI 1.025 1.272) relative to no dependents. Relative to service members identified as white, service members identified as black had reduced odds (AOR 0.785, 95% 0.726 0.848), as did Hispanic (AOR 0.757, 95% CI 0.691 0.830), and Asian service members (AOR 0.792, 95% CI 0.683 0.917). Relative to accession into military service before 2001, there appeared to be dose-response associated with each year between 2001 and 2007 with the highest relative risk of psychiatric evacuation (AOR 1.923, CI 1.621 2.006) among 2005 accessions. DISCUSSION This report is unique in that it is one of the few to compare psychiatric evacuees to the total deployed population. Previous reports compared psychiatric evacuees to other medical evacuees. The method used in this study thus provides a rate of psychiatric evacuation based on the total deployed force, providing a more accurate accounting of the total numbers and rate of psychiatric evacuees for a nearly 7-year period of war. The annual psychiatric evacuation rate increased during the study period from 72.7 per 100,000 deployers in 2004 to 196.9 per 100,000 deployers in 2010. This is consistent with previous reports of increased absolute frequency of psychiatric evacuation over the same time period. To put this in context, psychiatric evacuations represented approximately 11% of all aeromedical evacuations during this period. 3 Though psychiatric evacuation occurred among all Services and components, members of the Army represented 48% of the deployed force but disproportionately represented 82% of psychiatric evacuees. Among Army members, AD members had the highest risk of psychiatric evacuation, compared with Reserve and Guard personnel, whose likelihood was reduced by approximately 20%. These findings contradict prior suggestions that increasing reliance on Reserve or Guard units was responsible for the rise in the frequency of psychiatric evacuations. 4,13 While the odds of psychiatric evacuation increased slightly up to the third deployment, it is important to remember that two-thirds of psychiatric evacuations during this period were among those on first deployment. Our observations are consistent with prior research that demonstrated a bias inherent in studying repeat deployments known as the Healthy Warrior effect. 14 The healthy warrior effect describes the phenomenon that there is selection of the fittest for the first and each subsequent deployment. Those that do not fare well mentally or physically during a deployment are MILITARY MEDICINE, Vol. 180, January 2015 57

TABLE IV. Logistic Regression Model of Probability of Psychiatric Evacuation by Deployer Characteristics at First Deployment, OEF/OIF, January 2004 Through September 2010 Likelihood of Psychiatric Evacuation (N = 1,796,687) Characteristics at First Deployment AOR 95% CI Component/Service (Reference: Army Active Duty) Active Duty Navy 0.164 0.144 0.188 Active Duty Air Force 0.180 0.157 0.206 Active Duty Marines 0.390 0.353 0.430 Reserve Army 0.825 0.740 0.919 Reserve Navy 0.585 0.461 0.742 Reserve Air Force 0.097 0.052 0.180 Reserve Marines 0.300 0.223 0.404 National Guard Army 0.852 0.790 0.919 National Guard Air Force 0.120 0.077 0.188 Female Gender (Reference: Male Gender) 1.551 1.438 1.673 Age Group (Years) (Reference: 17 24 Years) 25 34 0.935 0.870 1.006 35 44 0.756 0.674 0.848 45+ 0.823 0.689 0.983 Race/Ethnicity (Reference: White Race) Black 0.785 0.726 0.848 Hispanic 0.757 0.691 0.830 Asian 0.792 0.683 0.917 American Indian 0.943 0.754 1.179 Other 0.743 0.384 1.439 Education (Reference: High School Diploma or Less) Less Than 4 Years of College 0.737 0.659 0.825 Bachelor s Degree 0.504 0.450 0.565 Advanced Degree 0.590 0.485 0.717 Marital Status (Reference: Single, Never Married) Married 1.207 1.109 1.313 Other 1.323 1.145 1.529 Dependents (Reference: No Dependents) 1 2 Dependents 1.144 1.054 1.243 3 or More Dependents 1.142 1.025 1.272 Year of Accession (Reference: Before 2001) 2001 1.188 1.053 1.340 2002 1.292 1.154 1.447 2003 1.488 1.335 1.658 2004 1.535 1.374 1.715 2005 1.923 1.621 2.006 2006 1.803 1.568 1.985 2007 1.764 1.280 1.676 2008 1.474 0.645 1.696 2009 0.250 0.645 1.121 Deployed to OEF (Reference: Deployed to OIF) 0.814 0.752 0.880 Number of Deployments (Reference: 1 Deployment) 2 Deployments 1.248 1.174 1.326 3 Deployments 1.381 1.250 1.526 4 Deployments 1.058 0.832 1.344 5 or More Deployments 0.400 0.220 0.725 C-statistic = 0.715. more likely to attrite from the military or be otherwise less likely to be deployed again. That the small group of service members who had four or more deployments were less likely to be psychiatrically evacuated should not be interpreted as a protective effect of repeat deployments rather they may be a group with specialized training, or specific personal characteristics or occupational assignments that better enabled them to endure the rigors of deployment. Observed deployment length also varied widely in this study population, but information regarding the scheduled length of deployment was not available. Further study is needed to identify the threshold deployment length beyond which the risk of psychiatric evacuation increases. It seems natural to assume that psychiatric evacuation may be more likely among those who engage in fire-fights or patrol areas susceptible to blasts and explosions, but our 58 MILITARY MEDICINE, Vol. 180, January 2015

findings indicate that other occupations also are associated with evacuation risk. Given the primarily ground-based operations in OEF/OIF, a lower likelihood of psychiatric evacuation rates among Naval and Air Force personnel, relative to Army personnel is not surprising. However, it is well known that Marines engage the enemy as frequently or more frequently than most Army units, and we found lower odds among the Marines. Likewise, evacuee overrepresentation among infantry, artillery, and motor vehicle operators is not consistent with overrepresentation also observed among medical, food, laundry, or auxiliary labor personnel or the underrepresentation observed for special forces personnel. Prior research by Sutker et al, 15 found Operation Desert Storm veterans with qualities of personal commitment, sense of control, problem-focused coping skills, and satisfaction with social support were strongly related to psychological resilience and resistance to the negative effects of deployment. Possibly the selective recruiting and screening, and the advanced training for special forces and the Marines, both selects individuals more suited for deployment duties and better prepares them for stresses likely to be encountered. Additional factors beyond the combat-related deployment experiences, and not available in this study may also be contributing to the observed evacuation patterns. Other research has found that preexisting psychiatric illness, 16 predeployment training, 17 morale, 18 and separation from family 19 have been significantly associated with deployment-related stress or adverse postdeployment outcomes. Perhaps these factors, largely unmeasured in this study, were a source of differences between Service or occupation-based groups. We observed elevated evacuation risk among certain demographic groups. Women were 50% more likely to experience a psychiatric evacuation than men despite ineligibility for assignment to combat-specific occupations during this period. 20 This finding is consistent with the other reports of increased likelihood of depression or generalized distress among women during and following OEF/OIF deployment, 9,11 as well as women in the civilian population nationwide. 21 Unmeasured sources of stress may include sexual harassment within the unit, personal victimization of sexual assault by unit member or enemy, and concerns about family and finances have been noted as significant sources of stress specific to deployed women. 22 24 We also observed a differential risk associated with race/ethnicity whereby whites had a 30% higher likelihood of psychiatric evacuation than other groups, whereas a higher likelihood was reported among minority groups during the earlier years of the war. 13 There may be other unmeasured characteristics associated with gender or race/ethnicity groups, or a change in the way evacuation procedures are carried out that affects these groups differently. Finally, we observed a steady increase in likelihood of psychiatric evacuation among individuals who accessed into military service since 2001. The highest likelihood of evacuation among 2005 accessions is consistent with the peak in the evacuation rate observed in 2007, when the 2005 recruits would be have completed a full year of training before deploying. The absolute number of evacuations and Service members who deployed also peaked in 2007, primarily as a result of a major troop surge in Iraq. The demands placed on the military services to maintain the troop strength necessary for nearly a decade of sustained conflict in multiple theaters has been substantial. Policies and procedures regarding recruitment, training, and deployment operations were adapted to accommodate this demand. These findings suggest that a review of policy or procedural changes is warranted to ensure a full understanding of their implications for Service member readiness and fitness for deployment. LIMITATIONS The findings from this study are subject to certain limitations. First, the mental health status of the Service member before deployment was not controlled. Some of the diagnoses assigned at time of evacuation would appear to represent preexisting or emergent conditions (such as serious and persistent mental illness). Variation in the length of deployment, which was typically 12 months or more for Army members, but generally much less for Navy and Air Force was also not controlled. There may have also been changes in evacuation criteria, thresholds, procedures, or availability of in-theater mental health resources during the study period that may have confounded our results. Finally, this analysis was limited to administrative data. Specifically, there was no information on direct combat exposure, measures of unit morale or confidence in leadership, or the extent to which combat-related stress management techniques were applied. These are topics for future research if the data become available. CONCLUSIONS Though a relatively rare event, the spike in 2007 and the trend toward increasing rate of psychiatric evacuations over time adds to growing concerns regarding adverse psychological consequences of military deployments. This study found that psychiatric evacuations are concentrated in the Army, and unlike a priori hypotheses, higher among AD troops than National Guard and Army Reserve Soldiers. We also found evidence that one third of evacuations occurred among personnel with more than 12 cumulative months of deployment and that, while most evacuations occurred on first deployments, there was increased risk associated with second and third deployment. Further study is warranted to isolate and mitigate the factors associated with this costly contributor to unit attrition. ACKNOWLEDGMENTS The authors acknowledge Dr. Angelia Eick-Cost of the Armed Forces Health Surveillance Center and Ms. Melissa Fraine, Force Health Protection and Readiness (FHP&R) for their invaluable assistance in the compilation of the MILITARY MEDICINE, Vol. 180, January 2015 59

dataset used in this study. This study was funded solely by the Office of the Deputy Assistant Secretary of Defense for FHP&R/U.S. Department of Defense. All authors disclose that they hold no affiliations with organizations that may incur a financial benefit as a result of this work being published. REFERENCES 1. Lee R: The History Guy: Casualties From America s Wars. Available at http://www.historyguy.com/american_war_casualties.html; accessed May 21, 2012. 2. Defense Manpower Data Center: DoD Military Casualty Information. Available at http://siadapp.dmdc.osd.mil/personnel/casualty/castop.htm; accessed May 21, 2012. 3. Armed Forces Health Surveillance Center: Causes of Medical Evacuations from Operation Iraqi Freedom (OIF), Operation New Dawn (OND), and Operation Enduring Freedom (OEF), Active and Reserve Component, U.S. Armed Forces, October 2001 September 2010. Monthly Surveillance Medical Report (MSMR) 2011; 18(2): 2 7. 4. Cohen SP, Brown C, Kurihara C, Plunkett A, Nguyen C, Strassels SA: Diagnoses and factors associated with medical evacuation and return to duty for Service members participating in Operation Iraqi Freedom or Operation Enduring Freedom: a prospective cohort study. Lancet 2010; 375: 301 9. 5. Pietrzak RH, Whealin JM, Stotzer RL, Goldstein MB, Southwick SM: An examination of the relation between combat experiences and combatrelated posttraumatic stress disorder in a sample of Connecticut OEF-OIF Veterans. J Psychiatr Res 2011; 45(12): 1579 84. 6. Killgore WD, Cotting DI, Thomas JL, et al: Post-combat invincibility: violent combat experiences are associated with increased risk-taking propensity following deployment. J Psychiatr Res 2008; 42(13): 1112 21. 7. Thomas JL, Wilk JE, Riviere LA, McGurk D, Castro CA, Hoge CW: Prevalence of mental health problems and functional impairment among active component and National Guard soldiers 3 and 12 months following combat in Iraq. Arch Gen Psychiatry 2010; 67(6): 614 23. 8. Bell NS, Hunt PR, Harford TC, Kay A: Deployment to a combat zone and other risk factors for mental health-related disability discharge from the U.S. Army: 1994 2007. J Trauma Stress 2011; 24(1): 34 43. 9. Shen YC, Arkes J, Williams TV: Effects of Iraq/Afghanistan deployments on major depression and substance use disorder: analysis of Active Duty Personnel in the US Military. Am J Public Health 2012; 102: S80 7. 10. Peterson AL, Wong V, Haynes MF, Bush AC, Schillerstrom JE: Documented combat-related mental health problems in military noncombatants. J Trauma Stress 2010; 23(6): 674 81. 11. Ferrier-Auerbach AG, Erbes CR, Polusny MA, Rath CM, Sponheim SR: Predictors of emotional distress reported by soldiers in the combat zone. J Psychiatr Res 2010; 44(7): 470 6. 12. Milliken CS, Auchterlonie JL, Hoge CW: Longitudinal assessment of mental health problems among active and reserve component soldiers returning from the Iraq war. JAMA 2007; 298: 2141 8. Available at http://jama.jamanetwork.com/article.aspx?doi=10.1001/jama.298.18.2141; accessed May 23, 2012. 13. Rundell JR: Demographics of and diagnoses in Operation Enduring Freedom and Operation Iraqi Freedom personnel who were psychiatrically evacuated from the theater of operations. Gen Hosp Psychiatry 2006; 28: 352 6. 14. Haley R: Point. Bias from the Healthy-Warrior Effect and Unequal Follow-up in Three Government Studies of Health Effects of the Gulf War. Am J Epidemiol 1998; 148(4): 315 23. Available at http://aje.oxfordjournals.org/content/148/4/315.long; accessed August 26, 2013. 15. Sutker PB, Uddo M, Brailey K, Allain AN: War zone trauma and stressrelated symptoms in Operation Desert Shield/Storm (ODS) returnees. J Soc Issues 1993; 49: 33 49. 16. Crain JA, Larson GE, Highfill-McRoy RM, Schmied EA: Postcombat outcomes among Marines with preexisting mental diagnoses. J Trauma Stress 2011; 24(6): 671 9. 17. Hourani LL, Council CL, Hubal RC, Strange LB: Approaches to the primary prevention of posttraumatic stress disorder in the military: a review of the stress control literature. Mil Med 2011; 176(7): 721 30. 18. Jones N, Seddon R, Fear NT, McAllister P, Wessely S, Greenberg N: Leadership, cohesion, morale, and the mental health of UK Armed Forces in Afghanistan. Psychiatry 2012; 75(1): 49 59. 19. Hinojosa R, Hinojosa MS, Högnäs RS: Problems with veteran-family communication during operation enduring freedom/operation Iraqi freedom military deployment. Mil Med 2012; 177(2): 191 7. 20. National Defense Authorization Act for Fiscal Year 1994 (P.L. 103-160, Nov. 30, 1993). Available at http://www.gpo.gov/fdsys/pkg/bills- 103hr2401eh/pdf/BILLS-103hr2401eh.pdf; accessed June 17, 2014. 21. Grossbard JR, Lehavot K, Hoerster KD, Jakupcak M, Seal KH, Simpson TL: Relationships among veteran status, gender, and key health indicators in a national young adult sample. Psychiatr Serv 2013; 64(6): 547 53. 22. Dutra L, Grubbs K, Greene C, et al: Women at war: implications for mental health. J Trauma Dissociation 2011; 12(1): 25 37. 23. Mattocks KM, Haskell SG, Krebs EE, Justice AC, Yano EM, Brandt C: Women at war: understanding how women veterans cope with combat and military sexual trauma. Soc Sci Med 2012; 74(4): 537 45. 24. Gibbons SW, Barnett SD, Hickling EJ, Herbig-Wall PL, Watts DD: Stress, coping, and mental health-seeking behaviors: gender differences in OEF/OIF health care providers. J Trauma Stress 2012; 25(1): 115 9. 60 MILITARY MEDICINE, Vol. 180, January 2015