Reproductive and other health outcomes in Iraq and Afghanistan women veterans using VA health care: association with mental health diagnoses

Over 220,000 women have served in Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF), and 14% of currently active military personnel are women ( Street, Vogt, & Dutra, 2009 ). Women represent not only the fastest growing population of new military recruits but also a rapidly increasing portion of patients who utilize Department of Veterans Affairs (VA) healthcare ( Friedman et al., 2011 ; Yano et al., 2010 ). Several studies have investigated the health of this population of returning women veterans, with a systematic review finding more studies were published on women veterans from 2004-2008 than in the preceding 25 years ( Bean-Mayberry et al., 2011 ). Understandably, the majority of studies on health conditions have focused on mental health, finding post-traumatic stress disorder (PTSD) and depression, including single and recurrent episodes of major depressive disorder, are common in returning women veterans ( Bean-Mayberry et al., 2011 ). Several studies of women veterans from prior eras have linked PTSD to increasing rates of physical symptoms and poorer physical health status, with more limited data on gender-specific conditions, such as abnormal pap smears, polycystic ovarian syndrome, and pelvic pain ( Araneta et al., 2004 ; Dobie et al., 2004 ; Frayne et al., 2004 ; Pierce, 1997 ; Wagner, Wolfe, Rotnitsky, Proctor, & Erickson, 2000 ). In addition, though returning women OEF/OIF veterans commonly access care for gender-specific services, few studies have examined connections between mental health disorders and gender-specific physical health problems in women veterans returning from the current conflicts ( Frayne et al., 2010 ; Haskell et al., 2011 ). More detailed study in this area may improve the provision of health services and treatment for this latest generation of women veterans (Maguen, Cohen et al., 2011). With the lack of distinct front lines in the current wars, an increasing number of women are exposed to the psychological and physical trauma of combat ( Zinzow, Grubaugh, Monnier, Suffoletta-Maierle, & Frueh, 2007 ). A recent study found that women serving in Iraq and Afghanistan were more likely to suffer a combat-related injury than women who served in the Gulf War (7% vs. 2%) and were also more likely to be exposed to death during combat (31% vs. 14%) ( Maguen, Luxton, Skopp, & Madden, 2011 ). Women also may experience non-combat trauma, such as sexual assault, during military service, which may be associated not only with PTSD, but also with poorer physical health ( Kimerling et al., 2010 ; Martin, Rosen, Durand, Knudson, & Stretch, 2000 ; Smith et al.; Suris & Lind, 2008 ). The women who have served in these recent conflicts are also primarily of reproductive age; thus reproductive health disorders, such as sexually transmitted infections, menstrual disorders, and infertility, and their association with combat-related mental health disorders are particularly important areas to investigate. In a survey of 1,206 women veterans seen at a single VA medical center, self report of cervical cancer, sexually transmitted infections, and chronic pelvic pain was significantly greater in women veterans who screened positive for PTSD ( Dobie et al., 2004 ). Focusing on women Iraq and Afghanistan veterans in an examination of national VA administrative data from 12,831 patients, Frayne and colleagues found women veterans with a diagnosis of PTSD had a significantly greater number of medical conditions than those without mental health conditions (7.0 vs. 4.5) ( Frayne et al., 2010 ). Most prior research on the association of mental and physical health problems in women veterans has focused on the effects of PTSD, but recent findings suggest that depression is also important to consider with women veterans. Several studies have now found that women returning from OEF/OIF are more likely to be diagnosed with depression than their male counterparts ( Adamson et al., 2008 ; Lapierre, Schwegler, & Labauve, 2007 ; Luxton, Skopp, & Maguen, 2010 ). Indeed, in a study of 329,049 OEF/OIF veterans seeking VA care, 23 % of women vs. 17% of men received a depression diagnosis ( Maguen, Ren, Bosch, Marmar, & Seal, 2010 ). Patients with comorbid PTSD and depression are also less likely to engage in or be retained in mental health treatment ( Bryant et al., 2008 ; McDonagh et al., 2005 ). Consequently, when evaluating the effect of mental health problems on physical health outcomes, it is important to consider not only the effects of PTSD but also the effects of depression as well as comorbid PTSD and depression on physical health outcomes. Therefore, the primary objective of this study was to expand upon prior work on the mental and physical health of women veterans to examine the associations of PTSD, depression, comorbid PTSD and depression, and other mental disorders with a broad range of reproductive and physical health outcomes using national-level data from all returning women OEF/OIF veterans who sought care at VA facilities over an extended time period. In evaluating depression, we included single and recurrent episodes of major depressive disorder as well as dysthymic disorder and depression not otherwise specified. In addition, we explored whether differences in demographics, military service characteristics, or utilization of primary care services explained these associations.

The VA OEF/OIF Roster is a register of veterans who have been deployed to OEF/OIF military service, have separated from service, and have accessed VA healthcare ( Seal, Bertenthal, Miner, Sen, & Marmar, 2007 ). The OEF/OIF Roster includes information on veterans’ demographic characteristics as well as aspects of their military service, such as their rank and the branch of the military with which they served. Scrambled social security numbers were used to link OEF/OIF Roster data to VA clinical data contained in the VA National Patient Care Database (NPCD) through December 30, 2010. The VA NPCD includes data from outpatient and inpatient visits to any of the approximately 150 VA hospitals and over 900 VA clinics nationwide. The electronic record includes the date of the visit, a code designating the type of visit, and the diagnosis(es) associated with the visit classified using the International Classification of Diseases, Ninth Revision Clinical Modification (ICD-9-CM) codes. Data on the distance to and type of nearest VA medical facility based on each veteran's zip code were determined from the VA Planning Systems Support Group. Because our research questions focused on mental health diagnoses associated with OEF/OIF military service, we made exclusions to yield a study population of women OEF/OIF veterans who were new users of VA healthcare after the beginning of OEF or OIF. There were 78,689 women OEF/OIF veterans in the Roster with a VA encounter before the study end date of December 31, 2010. We excluded 7,185 women veterans with a first visit prior to October 7, 2001, the date of the US invasion of Afghanistan and the start of OEF. Our final study population consisted of 71,504 separated OEF/OIF women veterans who were new users of VA healthcare after October 7, 2001 and were followed through December 31, 2010. We used the OEF/OIF Roster to obtain dates of birth, sex, race/ethnicity, marital status, component type (Active Component versus National Guard or Reserve), military rank (enlisted versus officer), service branch (Army, Air Force, Marines, Navy/Coast Guard), and whether a veteran had been deployed more than once. Data were also obtained on the distance to and type of nearest VA facility (VA medical center versus VA community-based outpatient clinic). For this study, we defined mental health diagnoses as ICD-9-CM diagnostic codes 290.0-319.0, excluding codes for alcohol and drug use disorders, which correspond to mental health disorders in the Diagnostic and Statistical Manual-Fourth Edition (DSM-IV) ( APA, 2000 ). We used ICD-9-CM codes associated with VA outpatient or inpatient visits during the study period to categorize veterans into 5 groups: (1) those with no mental health diagnoses, (2) those with PTSD but not depression, (3) those with depression but not PTSD, (4) those with both depression and PTSD, and (5) those with other types of mental health diagnoses (not PTSD or depression). We chose these groupings because PTSD and depression are the most common MH diagnoses of OEF/OIF veterans enrolled in VA healthcare ( Seal et al., 2009 ). Specific codes for the mental health categories are presented in the Appendix in Table A1 . We used Clinical Classifications Software (CCS), a validated coding algorithm that groups ICD-9 codes into reproductive and physical health diagnostic categories to evaluate several broad outcomes categories, such as sexually transmitted infections and menstrual disorders. However, several outcomes of interest, including cervical dysplasia and sexual dysfunction are not represented as individual CCS categories. When CCS categories were not available, potential ICD-9 codes were reviewed independently by two VA internists (BC and KS) and an obstetrician/gynecologist (VJ) to create a coding algorithm. A complete list of codes used in this study and whether they are based on CCS categories is presented in Table A1 . Several categories that were initially examined were subsequently excluded due to low frequency of diagnostic codes, including HIV (prevalence = 0.03%), and concerns that these conditions may be under-coded as women would typically be referred outside of the VA for specialized treatment, including cervical cancer (prevalence = 0.3%) and spontaneous abortion/miscarriage (0.8%). We compared the proportion of veterans in each of the five mental health groups receiving the health outcome diagnoses using the Chi-square Test of Association with no mental health diagnoses as the reference category. We then created separate binary logistic regression models for each health outcome, using each of the five mental health groups as independent variables and adjusting for potential confounders by adding them as covariates in the logistic regression models. These included age, race, marital status, military component type, rank, branch, multiple deployments, distance to and type of nearest VA facility. All health outcomes were treated as binary variables (received the diagnosis vs. did not receive the diagnosis), including the composite sexually transmitted infections variable, which was coded as positive if the person received any of the diagnoses in this CCS category. We also conducted additional analyses to evaluate whether differences in the prevalence of substance use disorders explained the associations between mental health diagnoses and health outcomes. To examine this, we repeated the adjusted logistic regressions adding a covariate indicating whether the Veteran had been diagnosed with an alcohol or drug use disorder (see Table A1 for ICD-9 codes used to define substance use disorders). Finally, to explore potential ascertainment bias from increased use of medical services by women veterans with mental health disorders, we ran alternate logistic regression models which included the potential confounders described above (demographics, military service characteristics, distance to and type of nearest VA facility) plus each veteran's number of primary care encounters entered as a continuous covariate. All analyses were performed using SAS software (version 9.2, Cary, NC)

Baseline characteristics of the study population by mental health status are presented in Table 1 . The majority of women veterans (56.4%) did not receive a mental health diagnosis. Six percent received a diagnosis of PTSD without depression, 13.6 % received a diagnosis of depression without PTSD, 15.4% had diagnostic codes for both PTSD and depression, and 8.5% received another type of mental health diagnosis. The mean age of women veterans was 28.5-29.5 years in all categories. Due to the large sample size, statistically significant differences existed across mental health categories for most sociodemographic and military service variables. Practically, demographics were generally similar across mental health categories, though women with mental health disorders were particularly more likely to be non-officers and to have served in the Army. Table 2 shows the unadjusted diagnostic prevalence of each health outcome by mental health category (unadjusted odds ratios are presented in Table A2 ). Compared to women without mental health disorders, those with mental health disorders had significantly greater prevalences of all outcomes including sexually transmitted infections, other infections such as urinary tract infections, pain-related conditions such as dysmenorrhea, and other diagnoses such as infertility. For example, the diagnostic prevalence of urinary tract infections ranged from a low of 6% in the group without mental health diagnoses to 19% in those with diagnoses for both PTSD and depression. Findings were similar after adjusting for demographics, military service characteristics, and distance to/type of nearest VA medical center ( Table 3 ). In general, women veterans with mental health disorders had two to four times the odds of receiving these reproductive and physical health diagnoses versus women veterans without mental health diagnoses, and prevalences were highest in women with comorbid PTSD and depression. This increased risk was particularly notable for pain-related conditions. For example, the odds of receiving a diagnosis of female genital pain were 4.88 times greater in women with PTSD and depression than in those with no mental health disorders. The most striking difference was in sexual dysfunction, which was a relatively rare outcome, but women diagnosed with mental health disorders had 6-10 times the odds of receiving this diagnosis than those without mental health disorders. Additional adjustment for alcohol and drug use disorders led to a small reduction in the association of mental health diagnoses and the health outcomes, with the magnitudes of the odds ratios generally being reduced by less than 10% ( Table A3 ). Adjustment for VA primary care utilization did attenuate the magnitude of the odds ratios. The differences in some of the less common sexually transmitted infections were no longer significant and the odds ratio for the composite sexually transmitted infections category in those with the highest risk (comorbid PTSD and depression) was reduced from 2.92 to 1.52 ( Table A4 ). Despite the expected decrease in the magnitude of the odds ratios with this adjustment, we still found women veterans with mental health disorders had significantly higher risk of receiving diagnoses of the majority of conditions.

In this study of women OEF/OIF veterans who were new users to VA care over a 9-year period following the start of the current combat operations, we found that those with mental health diagnoses were significantly more likely to receive diagnoses of sexually transmitted infections, cervical dysplasia, dysmenorrhea and gynecologic pain syndromes, and other reproductive and gynecologic health conditions than women veterans without mental health diagnoses. This was generally not due to differences in demographics, military service characteristics, drug and alcohol use disorders, or primary care utilization. Women with comorbid PTSD and depression diagnoses were at highest risk in all outcome categories. Our findings extend prior literature demonstrating associations between mental and physical health problems by focusing on specific reproductive health outcomes relevant to the growing number of women veterans returning from the current wars. A recent updated systematic review found an increasing number of studies on women veterans’ health in the last several years ( Bean-Mayberry et al., 2011 ). The majority of studies focused on the mental health of women veterans as well as access to and quality of care, but some included physical health outcomes. In a random sample of 30,865 women using VA care, those with PTSD or depression reported a greater number of medical conditions and had significantly worse physical health status ( Frayne et al., 2004 ). A survey of Iraq veterans 1 year after deployment found those with PTSD reported poorer general health and had a greater number of sick call visits and somatic symptoms, but the study population included only 80 women ( Hoge, Terhakopian, Castro, Messer, & Engel, 2007 ). Several other prior studies have focused on specific reproductive health conditions. In a study of 251 women deployed to Iraq, 21% reported gynecological health problems during deployment and nearly half were not up to date on cervical cancer screening based on current guidelines ( Thomson & Nielsen, 2006 ). Studies have also demonstrated that rates of sexually transmitted infections, including gonorrhea and chlamydia, are higher in female military recruits and personnel than in the general population, though these studies did not examine the connection between mental health disorders and these sexually transmitted infections ( Aldous et al., 2011 ; Boyer, Pollack, Becnel, & Shafer, 2008 ; Gaydos et al., 1998 ; Jordan, Lee, Nowak, Johns, & Gaydos, 2011 ). In a study of VA administrative data from 12,831 women Iraq and Afghanistan veterans seeking care in fiscal year 2006-2007, women with a PTSD diagnosis also had significantly higher risk of receiving diagnoses of urinary tract infections, pelvic inflammatory conditions, and menstrual disorders than those without mental health conditions ( Frayne et al., 2010 ). However, rates of cervical dysplasia, sexually transmitted infections, infertility, or sexual dysfunction were not significantly different. In the current study, we expand on this important prior work by studying a larger sample of women OEF/OIF veterans and by examining additional gender-specific and reproductive health outcomes. Interestingly, we did find significant differences by mental health status for these outcomes, which may be due to our increased power to detect these relatively rare diagnoses in the larger sample. Several mechanisms may explain the associations between mental and reproductive health conditions ( Meyers et al., 2008 ; von Sadovszky, Ryan-Wenger, Germann, Evans, & Fortney, 2008 ). One possibility is that increased sexual risk taking behaviors could lead to conditions such as sexually transmitted infections. Upon returning home from deployments, exposure to combat, which is also associated with higher rates of mental health problems, especially PTSD, has been found to be associated with increased risk taking and perceived invincibility, which may include engaging in high risk sexual behaviors ( Killgore et al., 2008 ). In addition, women with a history of trauma, particularly sexual trauma, may avoid preventive health measures such as cervical cancer screening. In a survey of women veterans seen for pelvic examinations at the VA, Weitlauf and colleagues found that women with a history of sexual violence and PTSD reported higher levels of fear and embarrassment related to the pelvic examination and were also more likely to endorse beliefs that the examination was unnecessary or unsafe ( Weitlauf et al., 2010 ). A survey and chart review of 999 women seen in VA care also found that women with a history of sexual assault were more likely to have abnormal cervical cytology (Sadler, Mengeling, Syrop, Torner, & Booth). Finally, increased alcohol and drug use that can co-occur with mental health problems may decrease inhibition and increase sexual risk-taking behavior ( Breslau, Davis, & Schultz, 2003 ; Jacobsen, Southwick, & Kosten, 2001 ; Tate, Norman, McQuaid, & Brown, 2007 ). Substance use may also directly contribute to other reproductive health problems. In our study, we found comorbid substance use disorders were diagnosed in 19% of those with PTSD and 17% of those with depression diagnoses. Additional adjustment for substance use disorders led to a small reduction in the association of mental health and reproductive health outcomes. However, lack of universal screening and stigma surrounding reporting of drug and alcohol use problems may have led to under-coding of substance use in this sample and therefore may have decreased our ability to examine the true effect of these comorbid conditions ( Seal et al., 2011 ). In addition to these potential behavioral mechanisms, the symptoms of mental health disorders themselves may affect women's reproductive health. Emotional numbing, one of the symptoms of PTSD is associated with poorer sexual functioning ( Nunnink, Goldwaser, Afari, Nievergelt, & Baker, 2010 ). Those with PTSD may be less likely to engage in sexual activity due to avoidance or numbing, or conversely, sexual activity may be increased in an attempt to overcome numbing, resulting in frequent and/or high risk sexual behaviors. Major depressive disorder is also associated with sexual dysfunction, and this may be mediated through a combination of biologic factors, such as changes in neurotransmitter and hormone levels, and psychosocial factors, such as social isolation and relationship difficulties ( Clayton, 2010 ; Fabre & Smith, 2012 ). Pharmacotherapy for mental health disorders may also have side effects, such as sexual dysfunction ( Hirschfeld, 2003 ; Opbroek et al., 2002 ). Additional biological factors associated with mental health diagnoses could also influence risk of disease in this population. For example, women OEF/OIF veterans with mental health disorders have higher rates of metabolic problems including obesity, which increase risk for polycystic ovarian syndrome ( Cohen, Marmar, Ren, Bertenthal, & Seal, 2009 ). Polycystic ovarian syndrome could, in turn, be linked to conditions such as infertility. PTSD and depression have also been associated with inflammation and immune dysfunction, which could increase risk for infections and pain syndromes ( Duivis et al., 2011 ; Gill, Saligan, Woods, & Page, 2009 ; Kendall-Tackett, 2007 ; von Kanel et al., 2007 ). Finally, mental health disorders are associated with decreased treatment adherence, which could directly worsen physical health problems or their known risk factors ( Gehi, Haas, Pipkin, & Whooley, 2005 ; Zen, Whooley, Zhao, & Cohen, 2012 ). Strengths of this study include the large number of women veterans and the examination of multiple gender-specific health diagnoses. However, our findings should be interpreted in light of several important limitations. This study relied on administrative data, and therefore the mental and physical health conditions we examined could be misclassified. In addition, our results are from a treatment seeking population enrolled in VA care and may not generalize to women veterans seen in other settings. Prior studies have found veterans who use VA care have lower incomes, worse health status and poorer social support than veterans who use other systems of care ( Mengeling, Sadler, Torner, & Booth, 2011 ; Yano et al., 2010 ). Still, as large numbers of women veterans are enrolling in VA care, this represents an important population. As statistical methods such as Chi-square tests are affected by sample size, our large sample size could have led to relatively small differences being considered statistically significant, though we have tried to focus on differences that would be clinically meaningful. In addition, this study design cannot establish timing of the development of mental health and physical disorders or causality, and reverse causality is possible as several of the health outcomes examined could cause mental health problems ( Dhaliwal, Gupta, Gopalan, & Kulhara, 2004 ; Mindel & Marks, 2005 ). Finally, we did not have information on traumatic event history and did not examine military sexual trauma screening results in this study. Despite these limitations, our study highlights several avenues for further research and clinical care. The VA conducts annual screening for mental health conditions, including PTSD, depression, and alcohol use disorders. The results of this study underscore the importance of recognition and treatment of mental health disorders in this population. However, veterans seen in other care settings may not receive such comprehensive assessment, and in fact may not even be asked about military service. It is important that primary care providers in all settings are aware of women's military service history. In addition to prompting screening for mental health disorders, primary care or other providers should evaluate risk behaviors and counsel patients about safe sexual practices. Providers can also counsel patients about the connection between their mental health and acute or chronic physical problems. Beyond providing pscyhoeducation, providers can refer patients to psychological services for treatment with evidence based therapies ( Cloitre, Koenen, Cohen, & Han, 2002 ; Resick, Williams, Suvak, Monson, & Gradus, 2011 ). The work of prior studies also emphasizes the need to improve engagement in preventative screening, particularly for those with a history of trauma who may have discomfort with routine gynecologic examinations. In recognition of the growing number of women veterans, the VA has made a comprehensive effort to study women veteran's health concerns, evaluate their preferences and needs for services, and improve availability and quality of gender-specific care ( Bean-Mayberry et al., 2010 ; Mengeling et al., 2011 ; Yano et al., 2006 ; Yano et al., 2010 ). Our finding that women veterans with any mental health disorder, regardless of whether it is PTSD, depression, or another type, have a greater prevalence of several important health outcomes provides support for the VA's efforts to increase collaboration among mental health and primary care providers. It also lends support to the VA's strategic plan to improve women's health care, which incorporates initiatives in mental health, primary care, and reproductive health services ( Department of Veterans Affairs, 2010 ). Adoption of these VA recommendations has led to improved satisfaction among women veterans ( Washington, Bean-Mayberry, Mitchell, Riopelle, & Yano, 2011 ). As the VA continues its mission to improve healthcare for women veterans ( Yano et al., 2010 ), an increased understanding of the connection between women's mental and physical health outcomes can facilitate identification of high-risk groups that may benefit from additional evaluation and development of more integrative approaches to mental and physical health care for women veterans.