MILITARY MEDICINE, 180, 4:436, 2015 Mental Health Diagnoses and Attrition in Air Force Recruits LCDR Shawn M.S. Garcia, MC USN*; Lt Col Brian V. Ortman, USAF BSC ; Col Daniel G. Burnett, FS, USAF MC* ABSTRACT Introduction: Mental disorders effect military readiness. Evaluating the frequency and impact of mental health diagnoses (MHD) in recruits, the source of the military workforce, is key to identifying opportunities for screening and prevention. Objectives: This study assessed the relationship between MHD in the Air Force recruit population and time to discharge. Methods: A recruit cohort at Lackland Air Force Base was followed through Basic Military Training, technical school, and 14 months of service using data from Trainee Health Online Reporting System. Incidence rate of MHD was calculated. A risk ratio and attributable fraction were calculated for attrition comparing recruits with MHD to recruits receiving other diagnoses (non-mhd). A survival analysis was performed on recruits with MHD compared to those with non-mhd. Results: Incidence of MHD was 7.9%. A recruit with a MHD was 4.28 (95% CI = 4.04 4.54) times more likely to separate in the first 14 months of service as compared to a recruit with a non-mhd. Conclusions: Recruits with MHD were separated faster and more often when compared to non- MHD. This study increases visibility of mental health disorders in recruits as a step toward better identifying those at higher risk of attrition. INTRODUCTION Fourteen percent of the general U.S. population has been estimated to have a mental disorder that impacts their work or home life. 1 In a study of all military branches, 47% of personnel hospitalized with a mental disorder for the first time left military service within 6 months. This result was significantly different from the 12% attrition rate after hospitalization for medical diagnoses not related to mental health (MH). Moreover, 27% of those seen as outpatients for a mental disorder were discharged from military service within 6 months of mental health diagnoses (MHD) as compared to only 9% for individuals with other medical diagnoses. 2 Considering the impact MHD have on the active duty population, it is important to examine military accession, the process of bringing a person into the military. Once a MH disorder is identified, a decision must be made as to whether it is in the best interest of the individual and the military to graduate the recruit from training. In fiscal year (FY) 2007, Air Force (AF) recruits began their active duty service with 6 weeks (extended to 8½ weeks in November 2008) of Basic Military Training (BMT) at Lackland Air Force Base (AFB), Texas. During the first week of BMT, recruits took the Lackland Behavior Questionnaire administered through the Biographical Evaluation and Screening of Trainees program. 3 Based on these results, trainees could be referred for further MH evaluation. Trainees could also be referred for *Department of Preventive Medicine and Biometrics, Uniformed Services University, 4301 Jones Bridge Road, Bethesda, MD 20814. Public Health Flight, 633 AMDS/SGPM, 45 Nealy Avenue, Langley AFB, VA 23665. The views expressed are those of the authors and do not necessarily reflect the official views of the Uniformed Services University of the Health Sciences, the United States Navy, the United States Air Force, or the Department of Defense. doi: 10.7205/MILMED-D-14-00311 a mental evaluation via a medical provider, their commander, the chaplain s office, or through self-referral. All MH evaluations for AF trainees at Lackland AFB were performed by the Behavioral Analysis Service (BAS). From previous research, based on the results of the questionnaire and the other referral mechanisms mentioned above, the BAS saw 4 to 6% of the 30,000+ new enlisted recruits annually. The majority of referred basic trainees were referred during their first 2 weeks of training. The rate of referral decreased throughout the weeks of BMT with a very small minority referred in the last 2 weeks of training. 4 A1998study found the majority of trainees referred to BAS (65%) were deemed fit-for-duty from the MH perspective and returned to training. Of the remaining 35%, 31% had a MH disorder that warranted a recommendation for a routine discharge from military service. Those trainees who exhibited more severe psychological symptoms (2%) were recommended for immediate removal from the training environment with an expedited out-processing from the AF. Another 2% with the most severe symptoms were referred for inpatient psychiatric hospitalization. 5 It has previously been found that the best predictor of early military discharge was referral to a MH center, regardless of outcome. 6 Recognition and diagnosis of a MH disorder results in a decision of fit-for-duty or discharge from military service. For those who continue with training, their ability to complete BMT and continue on to technical training and beyond has a large impact on resource utility and force readiness. In a 2012 report on military attrition and morbidity, neurotic disorders, personality disorders, and other nonpsychotic mental disorders were the leading cause of hospital admissions within the first 1 and 2 years of service for active duty enlisted personnel. Within the first 2 years of accession, 10,070 hospital admissions were as a result of neurotic and personality disorders. The fourth leading cause of admissions 436
within the first 2 years of accession was the category of other psychoses which added another 3,460 admissions to the scope of MH disorders impacting new service members. 7 For the entire active component of the U.S. Armed Forces in the year 2013, mental disorders accounted for more hospital bed days than any other disease category and for about 45% of all hospital bed days overall. 8 The purpose of this study was to determine the incidence of MHD in AF recruits during BMT through the first 14 months of service and to investigate the relationship of these diagnoses to attrition during the first 14 months of active duty service. This study also aimed to identify the most frequent MHD in this population and the most common reasons recruits with an MHD were separated. The primary research question was whether there was a difference in attrition rates in the first 14 months of service between those recruits who received a MHD as compared to those recruits who received medical diagnoses not related to MH during BMT. METHODS The FY 2007 cohort of recruits was followed from basic training through technical training and beyond up to 14 months of service. This time period was selected because it has been shown that an increased rate of attrition will occur in the first year of symptom onset and then return to baseline. Previous investigators hypothesized the reason for this was that those who left during the first year after diagnosis represented the more severe forms of MH disorders, whereas those with less severe disease who were able to recover in the initial 12 months after diagnosis no longer had the increased risk of separation. 9 The Trainee Health Online Reporting (THOR) System was used as the data source. This system was developed to be a source of centralized information on AF trainees. It integrated several databases to produce information that was pooled for epidemiologic studies with the goal of improving trainee health. THOR was a web-based system that was Health Insurance Portability and Accountability Act compliant. It integrated the following training and medical databases: the Airman s Database, medical hold (319 TRS), the medical data repository, Armed Forces Health Longitudinal Technology Application/Composite Health Care System I, Technical Training Management System, Medical Evaluation Board, Entry Level Separation, and Preventive Health Assessment/ Individual Medical Readiness. The designated MH diagnostic codes used to define a case were International Classification of Disease, Ninth Revision (ICD-9) codes 290-316 and the appropriate MHD V-codes, similar to diagnoses used by previous investigators. 2,10 A MHD case was defined as a trainee with a predefined ICD-9 or V-code in any diagnostic position in either an ambulatory or inpatient setting. V-codes described MH encounters that did not meet criteria for codes 290 to 316 because of lack of time to observe symptoms or other criteria, but did indicate a level of MHD concern by BAS personnel. A case was only counted once during the surveillance period of FY 2007. All cases were considered incident cases because they represented first diagnoses within the Military Health System. The comparison group consisted of those recruits who received only medical diagnoses not related to MHD while attending basic training. The population that received MHD had been referred for a MH evaluation through the Lackland Behavior Questionnaire, leadership referral, or through self-referral. The population that received non-mhd consisted of both wellvisits (e.g., routine physical examination) and sick visits (e.g., broken ankle). If a subject received both an MH and a non-mh diagnostic code, the individual was placed in the MHD group for the purposes of this study. An individual was censored (considered retained) in the analysis at 14 months (426 days) after the first day at BMT. The THOR data managers also pulled data on FY 2007 waivers from the waivers department at Lackland AFB. The primary endpoints of this study were to determine the incidence of MHD in the FY 2007 recruit population and compare the attrition rate of those FY 2007 recruits who received a MHD to those recruits that received a non-mhd. The incidence of MHD was calculated as number diagnosed/ total recruit population. Relative risk and attributable risk fraction for attrition from a MHD was calculated in comparison to the non-mhd group using STATA 10.1. Survival analysis was used to model time until service separation in person-days. The cumulative proportion of recruits discharged over time was estimated using the Kaplan Meier method. Because the mean days to separation were significantly different between the genders using the independent sample t-test, the two diagnostic groups were stratified by gender and then compared between recruits with a MHD and those with a non-mhd using the log rank test. This part of the analysis was performed with SPSS 16.0 for Windows. Waiver status, defined as having any waiver on entrance into BMT, was explored using a two-sided Fisher s exact test to determine if there was a significant difference ( p < 0.05) in separation rates between those with MHD and those with only non-mhd. A Fisher s exact test was used because of the small cell size of some categories. Sample size calculation was conducted using nquery Advisor (R) 6.0 software (Statistical Solutions, Cork, Ireland). It assumed that 90% of recruits without a MHD would still be in the service after 1 year, versus 81% of recruits with a MHD (a hazard ratio of 2.0). Estimates showed that only 327 recruits per group (with and without a MHD) would be required to detect a significant difference between groups with 90% power and 5% two-sided significance level using a log rank test. An analyzable data set was constructed by the database managers at Lackland AFB, and all information was deidentified to protect subjects confidentiality before it was provided to the investigator. Each study subject had a unique study identification number and no information was received 437
TABLE I. FY 2007 Basic Military Training (BMT) Air Force Recruit Population Overview FY 2007 Male Female Total % BMT % All Diagnoses BMT Population 25,373 8,155 33,528 Any Diagnosis a 20,309 5,037 25,346 75.6% 100% Non-Mental Health 18,328 4,373 22,701 67.7% 89.6% Diagnosis b Mental Health Diagnosis c 1,981 664 2,645 7.9% 10.4% a All recruits who were seen by a provider for any reason during recruit training. b Any recruit who was seen by a provider for only well-visits or sick visits and did not receive a mental health diagnosis during recruit training. c Any recruit who received a mental health diagnosis, as defined in the methods section, during recruit training. TABLE II. Mental Disorder Diagnoses FY 2007 Air Force Recruits (Listed If Applied to >50 Recruits) Diagnosis Code Diagnosis Description Incidence ICD-9 300.0 Anxiety State NOS 65 (2.5%) ICD-9 311.0 Depressive Disorder NEC 51 (1.9%) V71.09 Observe for Other Suspected 1,435 (54.3%) Mental Conditions V62.2 Occupational Circumstances 52 (2.0%) V62.89 Psychological Stress NEC 568 (21.5%) V79.9 Screen for Mental Disorder/ 169 (6.4%) Not Otherwise Specified V62.84 Suicidal Ideation 119 (4.5%) that would allow linkage back to an individual. This study was approved by the Uniformed Services University Institutional Review Board. RESULTS The entire FY 2007 BMT population contained 33,528 recruits. Within that population, 75.6% were seen for a medical (MH or non-mh) reason at any time during BMT and therefore were included in this study. The incidence of MHD in the FY 2007 AF recruit population was 7.9%. Within the population of recruits with any diagnoses, the incidence was 10.4% (Table I). The top BAS diagnosis for this cohort was V71.09 Observe for other suspected mental conditions. The second highest was V62.89 Psychological Stress NEC. The only MH ICD-9 codes that were applied to more than 50 recruits were 300.0 Anxiety state NOS and 311.0 Depressive Disorder NEC (Table II). The 14-month attrition rate for those recruits who received a non-mhd was 10.0%, whereas the attrition rate for those who received a MHD was much higher at 42.7% ( p < 0.001). The attrition rate with a MHD was significantly different between males, 37.0%, and females, 60.2% ( p < 0.001). An AF recruit with a MHD was 4.28 (95% CI = 4.04 4.54) times more likely to separate in the first 14 months of service as compared to a recruit who had a non-mhd. The proportion of attrition that was because of the exposure of a MHD, as compared to those with a non-mhd, was 76.6%(95%CI= 75.2% 78.0%). In the recruit population who had received any diagnoses, 2.1% had waivers for entry into active duty service. When split into MHD and non-mhd, there was no statistical difference between the groups (2.4% MHD and 2.0% non- MHD; p > 0.05). However, if a waiver was granted to a recruit for entrance into the AF who then subsequently was diagnosed with a non-mhd, there was a significant relationship with separation status ( p < 0.01). When stratified by gender, the significant relationship remained for males ( p < 0.001) but not for females ( p > 0.05). There was no significant relationship between receiving a waiver for AF accession and separation for those with a MHD ( p > 0.05). In the first week of BMT, 63.5% of those who would receive a MHD in the first 14 months of service had been diagnosed. By week 6 of BMT, 97.4% had been diagnosed (Fig. 1). Seventy-one percent of all recruits who separated from military service with a MHD in their first 14 months of FIGURE 1. Histogram of week of initial mental health diagnosis for FY 2007 Air Force recruits. 438
FIGURE 2. Histogram of number of weeks to separation for FY 2007 Air Force recruits with a mental health diagnosis. active duty had been separated by week 6. The third week after entrance into BMT had the highest number of separations (23% of those to be separated with a MHD) (Fig. 2). The mean time in service for a recruit who received a MHD and attrited was 37.2 days and the mean time in service for a recruit that received a non-mhd and attrited was significantly longer at 46.7 days ( p < 0.001). Female recruits with a MHD had a mean number of days to separation that was significantly longer than the males: 41.4 versus 34.8 days ( p < 0.005) (Table III). The survival analyses were stratified by gender (Figs. 3 and 4) to account for the significantly higher attrition rate in females as compared to males. Both the female and male survival curves showed a significant difference between the retention in thosewithamhdversusthosewithanon-mhd(p < 0.001). One-quarter of the males with a MHD had been separated by 37 days, but one-quarter of females with a MHD were separated by only 24 days. Moreover, half of the females with a MHD had separated from the military by day 67, whereas only 37% of males had separated by the 14-month mark (Figs. 3 and 4). DISCUSSION This was a retrospective cohort study designed to describe and analyze the impact and scope that MHD have on the AF recruit population. The incidence of MHD was almost 8%, which accounted for approximately 10% of all diagnoses made in this 14-month cohort. Over 63% of MHD were made in the first week of BMT, which is consistent with past rates where the majority of recruits are referred for a MH evaluation in the first 2 weeks of training. This correlates with the administration of the Lackland Behavior Questionnaire, which screens the recruits for MH issues and may result in a BAS referral (and thus an ICD-9 code). 4 The two most frequent diagnoses recorded for this cohort were the V-codes observe for other suspected mental conditions and psychological stress NEC (accounting for 75% of all the MHD). The nonspecific nature of these diagnoses may indicate hesitancy on the part of the health care providers to assign these young recruits a specific non V-code MHD after such a short time of symptoms, knowing that diagnoses assigned during recruit training could impact future career options in the military. Because the initial diagnosis was also often made early on in training, the recruit may not have yet displayed all symptoms necessary to meet criteria for a more formal ICD-9 code diagnosis. Follow-on diagnoses on individual recruits, which were not captured by this study because of the focus on incidence, may have become more specific. The results of this study showed a clear increased risk of attrition for any recruit given a MHD. Additionally, BMT is a purposely stressful environment to identify those individuals who are not suited for military service. Looking at the attrition rate, it would appear those with MH issues are being TABLE III. Comparison of the Mean Days to Attrition for FY 2007 Air Force Recruits Using Independent Sample t-tests Days to Separation Mean Standard Deviation Standard Error p-value Male Mental Health Diagnosis 34.8 30.975 1.146 p < 0.001 Non-Mental Health Diagnosis 43.5 35.467 0.937 Female Mental Health Diagnosis 41.4 34.433 1.722 p < 0.001 Non-Mental Health Diagnosis 52.2 43.852 1.518 Total Mental Health Diagnosis 37.2 32.380 0.963 p < 0.001 Non-Mental Health Diagnosis 46.7 38.983 0.819 Mental Health Diagnosis Male 34.8 30.975 1.146 p < 0.005 Female 41.4 34.433 1.722 439
FIGURE 3. Cumulative retention of FY 2007 male Air Force recruits with a mental health diagnosis compared to male Air Force recruits with a nonmental health diagnosis. separated efficiently. Key questions remain including whether or not these recruits were truly unsuited or if the symptoms potentially were the result of a transient adjustment disorder in a young, naïve individual, which could have been addressed with adequate MH resources/interventions. Regarding the primary research question, the overall attrition rate over the 14-month duration of the study for recruits who received a MHD was 43%, which is over four times higher than the attrition rate of those recruits who received a non-mhd. A female recruit with any diagnosis had over twice (24.4% vs. 10.6%) the attrition rate of a male recruit, and this effect remained when the diagnoses were stratified by MHD and non-mhd. The mean days to separation for those recruits who had received a MHD and would separate within the 14-month period was consistently different between genders. The female mean number of days to separation in all categories indicated that females on average took longer to be separated than males regardless of diagnosis. Because of a concern for effect modification, the survival analysis was performed stratified by gender. Survival analysis showed females with a MHD were more likely to be discharged and many were discharged sooner as compared to males with MHD. The greater mean time to discharge for females can be explained by their continued high rates of attrition later in the surveillance period, weeks after the male survival curve had leveled off. Prior studies have identified the demographic risk factors of age, gender, race, marital status, and education level. 5,11 Motivation toward military service, disciplinary history, and factors associated with MH history as related to attrition have also been examined. 6,11 This study found that having obtained a waiver for entrance into military service was not associated with attrition following a MHD. However, a significant relationship was found for males with a non-mhd waiver and subsequent attrition. These results could partially be as a result of the smaller numbers present in the MH and FIGURE 4. Cumulative retention of FY 2007 female Air Force recruits with a mental health diagnosis compared to male Air Force recruits with a nonmental health diagnosis. 440
female non-mh diagnosis groups. Evaluation of other variables such as physical fitness, other medical diagnoses, and Armed Forces Qualification Test scores, all of which hypothetically could have a relationship with the mental condition of the trainee, is a goal for future research with the hope of creating a multivariate model for risk factors predictive of attrition when a recruit has a MHD. The strengths of this study were the large numbers of recruits included in the analysis, the military relevance of the data, and the ability to retrospectively follow a cohort forward in time. One of the weaknesses was the inability to get the entire BMT population data for the 14-month period. Only those recruits who had been seen by a health care provider were available for data extraction. Because the attrition rate for those with any diagnosis would be expected to be higher than the attrition rate of the baseline population, using those with a non-mhd as the comparison group likely biased the results of this study toward the null hypothesis of no difference between diagnostic groups. CONCLUSIONS MH disorders have a substantial impact on recruit populations. Increased efforts in the arena of recruit MH could both improve the morbidity of mental illness for the recruits and help training commands reduce their attrition rates. First, validated screening tools to identify MH risk factors predictive of future attrition and disability could be used at recruitment before the beginning of basic training to identify many of those who are unsuited for military service. Active surveillance of MH status should be initiated at all recruiting sites considering the large impact MH disease has on the military. The Lackland Behavior Questionnaire is a good screening tool to be duplicated by other services. This study further demonstrates its effectiveness in identifying recruits with potential MH disorders, which facilitates their early evaluation. 12 Second, increased resources for effective evaluation and treatment may improve the outcome for individuals diagnosed with a mental disorder and simultaneously reduce their attrition rates. Research should be done to determine the most effective assessment tools and therapies which not only reduce attrition in basic training, but also through a military career. Throughout the Department of Defense, behavioral health screening is integrated into the pre and postdeployment health assessments to screen active duty members for behavioral health concerns. 13 However, there is no consistent standard for screening within the recruit and early technical training time frame. The importance of MH screening early in a military career is further emphasized in a cross-sectional study on the U.S. Army population, which found 76.6% of the soldiers reporting a MHD on their survey also reported a pre-enlistment age at onset. 14 Finally, the findings of this study indicate that female recruits are at a higher risk of attrition. It has been established in the military literature that female AF recruits are more frequently referred for MH evaluations than males. 5 Research has shown that there is a higher rate of previous sexual abuse in enlisting females, which may indicate a specific need for focused counseling and treatment to increase the rates of female retention. 5,11 There is also research to suggest that women show higher rates of internalizing MH disorders, such as depression and dysthymia, whereas men show higher rates of externalizing disorders, such as antisocial personality and substance use disorders. 15 Targeted studies should be done focusing on the risk factors for female attrition and reasons why females with MHD are being separated at rates greater than males. Results of these studies might lead to evidence-based prevention and treatment efforts that focus on gender specific psychological needs. In summary, MH assessments should be an effective, efficient, and normal part of military life, starting with accession. Additionally, increased education at all levels of military leadership is critical to reduce the stigma associated with MH disorders and increase the probability that those experiencing MH symptoms will seek and receive treatment promptly. 13 ACKNOWLEDGMENTS The authors would like to thank Cara Olsen, DrPH, Assistant Professor, Department of Preventive Medicine and Biometrics, Uniformed Services University, for her expertise as a biostatistical consultant in determining the appropriate statistical methods to use for this analysis and for her assistance in analyzing the data. 441
APPENDIX A: IRB APPROVAL LETTER 442
APPENDIX B: NOTICE OF PROJECT APPROVAL 443
REFERENCES 1. Kessler RC, Chiu WT, Demler O, Merikangas KR, Walters EE: Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry 2005; 62(6): 617 27. 2. Hoge CW, Lesikar SE, Guevara R, et al: Mental disorders among U.S. military personnel in the 1990s: association with high levels of health care utilization and early military attrition. Am J Psychiatry 2002; 159(9): 1576 83. 3. Biographical Evaluation and Screening of Trainees (BEST) Program. Air Education and Training Instruction 40 105. Air Education and Training Command, United States Air Force, October 15, 2013. Available at http://static.e-publishing.af.mil/production/1/aetc/publication/ aetci40-105/aetci40-105.pdf; accessed October 10, 2014. 4. Englert DR, Hunter CL, Sweeney BJ: Mental health evaluations of U.S. Air Force basic military training and technical training students. Mil Med 2003; 168(11): 904 10. 5. Cigrang JA, Carbone EG, Todd S, Fiedler E: Mental health attrition from Air Force basic military training. Mil Med 1998; 163(12): 834 8. 6. Retzlaff P, Deatherage T: Air Force mental health consultation: a sixyear retention follow-up. Mil Med 1993; 158(5): 338 40. 7. Accession Medical Standards Analysis & Research Activity: Attrition and Morbidity Data for 2011 Accessions. Annual Report 2012. Silver Spring, MD, Walter Reed Army Institute of Research, 2012. Available at http://www.amsara.amedd.army.mil; accessed March 28, 2014. 8. Armed Forces Health Surveillance Center: Absolute and relative morbidity burdens attributable to various illnesses and injuries, U.S. Armed Forces, 2013. MSMR 2014; 21(4): 2 7. 9. Creamer M, Carboon I, Forbes AB, et al: Psychiatric disorder and separation from military service: a 10-year retrospective study. Am J Psychiatry 2006; 163(4): 733 4. 10. Hoge CW, Toboni HE, Messer SC, Bell N, Amoroso P, Orman DT: The occupational burden of mental disorders in the U.S. military: psychiatric hospitalizations, involuntary separations, and disability. Am J Psychiatry 2005; 162(3): 585 91. 11. Carbone EG, Cigrang JA, Todd SL, Fieldler ER: Predicting outcome of military basic training for individuals referred for psychological evaluation. J Pers Assess 1999; 72(2): 256 65. 12. Garb HN, Wood JM, Schneider K, Baker M, Travis W: Suitability screening during basic military training. Mil Psychol 2013; 25(1): 82 91. 13. American Psychological Association: Presidential Task Force on Military Deployment Services for Youth, Families, and Service Members. The Psychological Needs of U.S Military Service Members and Their Families: A Preliminary Report. February 2007. Available at http://www.ptsd.ne.gov/publications/military-deployment-task-force-report.pdf; accessed March 28, 2014. 14. Kessler RC, Heeringa SG, Stein MB,et al: Thirty-day prevalence of DSM-IV mental disorders among nondeployed soldiers in the US Army: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS). JAMA Psychiatry 2014; 71(5): 504 13. 15. Eaton NR, Keyes KM, Krueger RF, et al: An invariant dimensional liability model of gender differences in mental disorder prevalence: evidence from a national sample. J Abnorm Psychol. 2012; 121(1): 282 8. 444