DSM-IV Diagnosed Posttraumatic Stress Disorder in Women Veterans With and Without Military Sexual Trauma

Authors

  • Deborah Yaeger MD,

    1. Women's Comprehensive Healthcare Center, VA Greater Los Angeles Healthcare System, Los Angeles, CA, USA
    2. Department of Psychiatry and Biobehavioral Sciences, UCLA David Geffen School of Medicine, Los Angeles, CA, USA
    Search for more papers by this author
  • Naomi Himmelfarb PhD,

    1. Women's Comprehensive Healthcare Center, VA Greater Los Angeles Healthcare System, Los Angeles, CA, USA
    Search for more papers by this author
  • Alison Cammack BS,

    1. Department of Psychiatry and Human Behavior, University of California at Irvine, Irvine, CA, USA.
    Search for more papers by this author
  • Jim Mintz PhD

    1. Department of Psychiatry and Biobehavioral Sciences, UCLA David Geffen School of Medicine, Los Angeles, CA, USA
    Search for more papers by this author

  • Conflict of Interest: There was no outside funding for this study. None of the authors have financial interests in any company that might be interested in the study's results.

Address correspondence and requests for reprints to Dr. Yaeger: Women's Comprehensive Healthcare Center, Mailcode: OOAC/WC, 11301 Wilshire Boulevard, Los Angeles, CA 90073 (e-mail: Deborah.Yaeger@va.gov).

Abstract

BACKGROUND: This study compares rates of posttraumatic stress disorder (PTSD) in female veterans who had military sexual trauma (MST) with rates of PTSD in women veterans with all other types of trauma.

METHODS: Subjects were recruited at the Women's Comprehensive Healthcare Center when attending medical or psychiatric appointments or through a mailing; 230 women agreed and 196 completed the study. They completed questionnaires on health and military history, along with the Stressful Life Events Questionnaire (SLEQ). Those who met DSM-IV PTSD Criterion A completed the PTSD Symptom Scale-Interview (PSS-I) on which PTSD diagnoses were based.

RESULTS: Ninety-two percent reported at least 1 trauma. Forty-one percent had MST, alone or with other trauma, and 90% had other trauma, with or without MST. Overall, 43% of subjects with trauma had PTSD. Those with MST had higher rates of PTSD than those with other trauma. Sixty percent of those with MST had PTSD; 43% of subjects with other traumas (with or without MST) had PTSD. Military sexual trauma and other trauma both significantly predicted PTSD in regression analyses (P=.0001 and .02, respectively) but MST predicted it more strongly. Prior trauma did not contribute to the relationship between MST and PTSD.

DISCUSSION: Findings suggest that MST is common and that it is a trauma especially associated with PTSD.

Rape is a violent crime with long-lasting consequences that is often perpetrated against women. At particular risk are women working within male-dominated environments,1 such as the Armed Services, which, despite a recent influx of women, continue to be composed mainly of men.

National surveys found 13% to 30% of women veterans experienced rape during their military service.2–5 What is only beginning to be understood is the long-term impact that military sexual trauma (MST), such as rape, has on the lives of women veterans.

Posttraumatic stress disorder (PTSD) is one of the known consequences of rape. In fact, rape is the trauma most highly correlated with the development of this disorder.6 Posttraumatic stress disorder associated with rape is long lasting.7 Further, PTSD itself can cause impaired physical and social functioning.3,8 Most of what is known about rape and PTSD comes from the literature on civilian sexual trauma. Much less is known about the relationship between PTSD and MST. Previous studies examining the link between PTSD and MST have identified symptoms suggestive of PTSD or have created models connecting MST with the development of PTSD.9–11

There is only 1 study, to our knowledge, that examined rates of DSM-IV PTSD diagnoses in women veterans with MST. Suris et al.,12 using a sample of female Veterans Administration (VA) patients, compared rates of PTSD related to 2 types of civilian sexual trauma with PTSD rates related to MST. They found that MST was more frequently traumatizing than civilian assault. Thus, the literature indicates that MST is more predictive of PTSD than are other types of military trauma or civilian sexual trauma.

The present study is the first to compare the impact of MST versus that of all other kinds of traumas—life-threatening illnesses, accidents, or other sexual trauma—on the rate of current PTSD diagnoses. We hypothesize that MST will lead to higher rates of PTSD than all other traumas, civilian or military.

METHOD

Patient Population

Subjects for the study were recruited between December 2000 and December 2002, either from the Women's Comprehensive Healthcare Center at the VA West Los Angeles, or by letter inviting participation in the study. During the 2-year recruiting period, a total of 896 women were seen in the medicine and psychiatry clinics. Of these, nearly 25% (226) agreed to participate. Approximately 250 letters were sent out; only 4 recipients became subjects. A total of 196 women completed all questionnaires. About 98% of our subjects were recruited in the clinic, less than 2% of our subjects became participants as a result of the letter.

Measures

Demographics, Clinical Characteristics, and Military Service. Demographic characteristics, health behavior, childhood family functioning, and military service history were assessed with questions derived from the “VA Women's Health Project” questionnaire (an unpublished self-report measure, used by Skinner et al.13). We drew from 2 of the questionnaire's subsections, “Life Experiences” and “Remembering Your Military Service.” Additional questions were added to investigate possible dysfunctional childhood experiences and the use of cigarettes and other substances. We included questions regarding military service (e.g., branch, rank, years served) and demographic, employment, and income information.

Traumatic Life Events. Defining traumatic experiences and determining whether they met DSM-IV criteria for PTSD was a primary focus of this study. To assess stressful life experiences, we used the Stressful Life Events Questionnaire (SLEQ),14 a standardized instrument measuring exposure to all possible types of traumas, including sexual and physical assault, witnessing violence, combat trauma, illness, accidents, traumatic deaths, and natural disasters. Goodman et al.14 report that the SLEQ has good test-retest reliability (median κ=0.73), acceptable convergent validity (median κ=0.64), and good differentiation between criterion A-based traumatic and nontraumatic events. For this study, we amended the SLEQ to distinguish between subtypes of sexual trauma. These types included forced intercourse or anal sex, forced oral sex, forced insertion of objects, or threats of forced sex; exposure to any of these would be considered a trauma as defined by DSM-IV. These subtypes and their definitions were derived from the National Women's Study.15,16 We did not include verbal sexual harassment, as it did not qualify as a DSM-IV designated trauma. We used behavioral descriptors to make the meanings as clear as possible and to maximize the validity of the questions.

PTSD. To determine whether respondents suffered currently from PTSD, we used the PTSD Symptom Scale-Interview (PSS-I).17 The PSS-I is a semistructured standardized PTSD interview. It has 17 items, which assess the presence and severity of DSM-IV PTSD criteria. Scores reflect total PTSD severity and are summed from subscales of reexperiencing, numbing/avoidance, and arousal. Published reliability and validity data for the PSS-I show acceptable levels of internal consistency (0.65 to 0.71), test-retest reliability (0.66 to 0.77), and interrater reliability (0.93 to 0.95). It also had good concurrent validity, demonstrated by strong correlations with other measures of PTSD, depression, and anxiety.17

Procedure

All women presenting for treatment to the clinic saw fliers explaining the purpose of the study. A letter describing the study and inviting participation was mailed to female veterans in the Los Angeles area. This latter approach drew few responses, as many of the letters were returned because the addressee no longer lived there.

Patients expressing interest in the study either met with or called the research coordinator, a psychiatrist, or another of the researchers, who were psychologists or an internist. Prospective subjects were read a description of the study, had questions answered about enrollment and possible adverse consequences of participation, and were screened for eligibility. Veterans with dementia, psychosis, and suicidality were excluded. Those who agreed to participate were given an appointment at which they signed the informed consent. All were treated in accordance with institutional ethics guidelines. The VA Greater Los Angeles Healthcare System's Institutional Review Board approved the study.

Subjects first completed the VA Women's Health Project questionnaire and the SLEQ. They next met with a researcher, trained in the use of structured interviews, who reviewed responses with subjects to make certain that all stressors were noted correctly. Interviewers then asked subjects to specify when these traumas occurred in relation to their military service—before, during, or after—and at what age.

After reviewing the SLEQ with the subjects, interviewers determined which incidents met DSM-IV criteria for trauma (PTSD criterion A),18 which require there to be threat to life or physical integrity along with feelings of fear, horror, or helplessness. Using a scale of 0 to 10, with 10 being the most intense, subjects rated the intensity of their fear, horror, or helplessness at the time of the stressful event. If they rated the feelings as 8 or higher, we considered the event distressing enough to satisfy the DSM-IV criteria for trauma. If rated lower than 8, they were not considered traumas. While there is no specific measure in the literature for determining whether an event meets the level of distress required by DSM-IV, we wanted to assure a high threshold for an event to be labeled a trauma. The interviewer then reviewed with the subjects all the experiences that met trauma criteria and asked, “Which one bothers you the most now?” The participant selected her currently most distressing trauma, and the interviewer then assessed PTSD using the PSS-I. We assigned PTSD diagnoses on the basis of that test. These sessions lasted from about 45 minutes to 2 hours.

Training for the administration of the PSS-I was provided by a research psychologist with significant experience conducting this interview. All interviewers observed 2 administrations of this instrument by an experienced researcher, and were then observed for 2 administrations of the PSS-I. All but 1 interviewer was female.

During the course of the study, several patients were found to have untreated PTSD, depression, or anxiety and were referred for psychiatric treatment.

Statistical Methods

We reported rates of PTSD in subjects with military sexual trauma and with other trauma. In all analyses, subjects with both types of trauma were included in both groups. We conducted logistic regression analyses to determine the extent to which MST and other trauma were associated with PTSD. We also used logistic regression to control for the effects of demographic variables and earlier trauma on the relationship between MST and PTSD. Odds ratios were calculated to determine the extent to which MST and prior trauma increased one's risk of PTSD. Demographic data were compared with a both local and national random samples of female veterans. Statistical analyses were done using SPSS Version 13.0 for Windows.19

RESULTS

Characteristics of Respondents

Participants ranged in age from 22 to 88, with a mean age of 47.8 (SD=14.5). We had an ethnically diverse sample, and most subjects had moderate to low income. The majority were unmarried at the time of the study, and most were in the Army rather than other branches of service. Participants had spent an average of 4.9 years in active military service.

Our sample was similar to those in 2 other studies of women VA health care users. The first, a population-based telephone survey of women veterans in Southern California and Southern Nevada that assessed influences on VA health care use, was performed between March and September 2004.20 We used for comparison a subset of data from that sample that included women VA health care users from the Greater Los Angeles area. The second study was the 1999 Large Survey of Veterans, the largest survey ever of female veterans, that assessed the functional status of VA health care users.21 In the 3 samples, education, marital status, and age were all similar. Income level was similar in our sample and the one of the Greater Los Angeles area; these data were not available for the national sample. Our sample had a somewhat greater representation of minorities than the others (see Table 1).

Table 1. Descriptive Statistics of Demographics in the Current Study, a Greater Los Angeles (LA) Sample, and a National Sample of Women Veterans Receiving VA Care
VariableCurrent
Sample
LA Sample*National
Sample
(N=196)
(%)
(N=254)
(%)
(N=39,150)
(%)
  • *

    From a population-based telephone survey of female veterans in Southern California and Southern Nevada. Data are the subset of VA healthcare users from the Greater Los Angeles area. 20

  • From “1999 Large Survey of Veterans,”21 a national survey designed to assess the functioning of veterans using VA medical services.

  • Income information was not surveyed in this study (personal communication, Susan Frayne, October 25, 2005).

Age M=48 (SD=15) M=52(SD=7) M=51 (SD=17)
Racial ethnicity
 Caucasian406375
 African American392119
 Other22166
Educational level
 High school or less101025
 Some college546350
 College or more362725
Marital status
 Unmarried (divorced, separated, widowed, never married)786563
 Married/or living together223537
 Income $40,000 or less8478N/A

Prevalence of PTSD

In the following analyses, we made MST a dichotomous variable in which subjects who had MST, alone or in combination with other traumas, were considered positive for MST, and those who had other trauma without MST, or no trauma at all, were considered negative. The Other Trauma group was also dichotomous; those who had other trauma with or without MST were positive and those with either MST alone or no trauma were negative.

We assessed for the presence of 16 types of trauma on the SLEQ. Ninety-two percent (181) had at least 1 trauma; 8% (15) had none. Prevalence rates of MST, other trauma, and PTSD are shown in Table 2. Most notably, 41% (80) had MST with or without other trauma. Ninety percent (177) had other trauma either alone or in combination with MST.

Table 2. Prevalence of Trauma Among the Overall Sample (n=196), and Prevalence of PTSD, by Trauma Type
 Prevalence of TraumaPrevalence of PTSD
NumberPercentageNumberPercentage
No trauma158
MST only42125
MST with other trauma76394762
MST with or without other trauma80414860
Other trauma only101522929
Other trauma with or without MST177907643
Any trauma181927743

Forty-three percent (77) of subjects with any kind of trauma had PTSD. Sixty percent (48) of subjects in the MST group and 43% (76) of those in the Other Trauma group had PTSD. The rate of PTSD was therefore higher for women in the MST group than the Other Trauma group. Furthermore, among subjects with PTSD, MST was most often identified as the most distressing trauma at the time of the interview (29% or 22 subjects).

Relationship Between Trauma and PTSD

The correlation between the MST group and Other Trauma group (r=.13, P=.07) reflected a weak relationship. We conducted a logistic regression analysis in which PTSD was regressed on MST and Other Trauma. Both MST (Wald χ2=20.3, df=1, P=.0001) and Other Trauma (Wald χ2=5.4, df=1, P=.02) significantly predicted PTSD, but MST predicted it more strongly. This finding is notable because the number of women positive for MST (80) was less than half of those positive for Other Trauma (177), yet the relationship of the MST group with PTSD was stronger.

Controlling for the Effects of Other Variables

In the literature, several demographic variables have been associated with MST, including age at the time of the study, age at entry into the service, education, work status, marital status, and years on active duty.4,22,23 We also considered race and branch of service, as these variables could reasonably relate to MST, too. To rule out the possibility that the strength of the relationship between MST and PTSD was being significantly affected by these variables, we performed a logistic regression analysis in which MST predicted PTSD, with demographic variables entered as covariates. Despite the inclusion of the demographic variables, the relationship between MST and PTSD remained highly significant (Wald χ2=22.1, df=1, P<.0001).

To rule out the possibility that prior trauma, rather than MST, was accounting for the high rates of PTSD in women with MST, we calculated a logistic regression analysis in which MST and presence of trauma before the military predicted PTSD. We found that MST was a significant predictor (Wald χ2=22.4, df=1, P<.001) while prior trauma was not (Wald χ2=0.44, df=1, P=.509). The odds ratios indicated that subjects with MST were almost four and a half times more likely than subjects without MST to develop PTSD (odds ratio [OR]=4.4, 95% confidence interval [CI]=[2.4 to 8.2]), while having prior trauma did not significantly increase the risk of PTSD (OR=1.3, 95% CI=[0.63 to 2.5]).

DISCUSSION

Our findings suggest that MST is a particularly distressing trauma. Women with MST had higher rates of PTSD than those with other traumas; 60% of subjects with MST had PTSD compared with 43% of women in the Other Trauma group. The rate of PTSD in women with MST was therefore about 40% higher than in women with other traumas. Further, subjects identified MST most often as the most distressing trauma at the time of the interview, no matter what other traumas they had experienced.

While our data are consistent with other research findings showing that women veterans have high rates of MST, we found a somewhat higher rate of MST than that reported in many previous studies.2–4 One explanation is the increased resources for treatment offered by the VA. In 1992, Congress authorized counseling for MST; subsequently, the VA mandated universal screening and treatment. Many of the studies citing lower rates of women with MST were published closer to the inception of these initiatives. Additionally, increased enrollment by women in the military has led, subsequently, to an increasing percentage of women veterans, many of whom may have experienced MST. It may be that the cumulative efforts of more than 10 years of recruitment, by the military and by VA providers, have resulted in a higher number of women with MST seeking VA services today.

Our finding that those with a history of MST have higher rates of current PTSD than do women with other types of trauma reflects an even stronger link than previously reported. Suris et al.12 found that 42% of their sample with military sexual assault had PTSD, compared with 60% in ours. The reasons for this difference are unclear; samples and procedures were similar in these 2 studies. This discrepancy highlights the need for larger, national samples to be studied to reduce the possible effects of geographic or other sampling differences.

However, in spite of the differences in prevalence rates, these 2 studies were consistent in their finding that MST is associated with higher rates of PTSD. This finding raises the question of why there is such a strong association between MST and PTSD. One explanation is that rape and interpersonal violence have been found in civilian studies to cause the highest rates of PTSD, and there are factors unique to the military that may increase that risk. Examples of such factors include fear of negative career consequences if the assault becomes known, having to continue to see or work with the perpetrator,24 and having limited access to outside support.

Limitations of the Study

This study has several limitations. First, generalizability may be limited by our use of a cross-sectional design and a VA clinical sample from 1 geographical area. A second factor limiting generalizability is self-selection. Because subjects were told of the study's focus on trauma, their decisions to participate may have been influenced by their victim status. It is possible that women with more trauma either refused participation because of concerns about recalling traumatic events, or that they enrolled more often because of their personal interest in this subject. In either case, getting a representative sample for this kind of research may be difficult for this reason. Further, in this study, those who dropped out did so before giving their trauma histories. Therefore, there is no way to determine if dropout rates are related to level of trauma.

A third limitation was that we did not control for status of service-connection claims, a variable that could have affected reported PTSD symptomatology. Veterans seeking service connection might tend to overstate PTSD symptoms to improve their chances of getting compensation. Finally, the study is retrospective and the data are therefore dependent on memory, which can be faulty. In spite of these shortcomings, retrospective designs using clinical samples are standard in the literature on MST.

Implications of the Study

Posttraumatic stress disorder symptoms in women veterans have been associated with physical and psychologic problems, as well as functional impairment.8,25–27 Our finding that rates of PTSD are high in women with MST argues for screening all women veterans for MST and PTSD. Doing so will require educating medical professionals in how to assess women for trauma exposure and its sequelae. Further, there is some evidence that PTSD-related health care costs exceed those of patients without this disorder.28–30 If true, treating PTSD early may decrease medical costs while improving physical and mental health outcomes in women veterans. Therefore, our findings have implications for the design and implementation of health services for female veterans.

Voices of Women Veterans (continued)

PRIDE IN SERVING

“I went in the Navy as a journalist. What a wonderful experience it was for a young woman who had many goals. My family could not afford to send me to college. I became in that short time structured, principled, and organized, although I believe some of the qualities were instilled by my parents. The Navy reinforced strong decision-making, self-worth, and unparalleled leadership skills. I have always been proud to be a veteran with a profound love of country.”

“I loved being in the military. I spent [many] years' active duty as a chaplain's assistant. I enjoyed serving the military families and healing their spiritual needs. I enjoyed serving those who served. It made me well-rounded and able to adapt to any situation and fit in any culture.”

“While I was on active duty, women were expected to perform to the same standards as the men. I never had anyone treat me differently. During my years of active duty, I learned more and was allowed to do more than you could do in the civilian life.”

Acknowledgments

We would like to thank Donna L. Washington, MD, MPH, for use of a subset of her data for comparison purposes. These data came from a study funded by the Department of Veterans Affairs, Health Services Research and Development Service (#GEN-00-082).

We would also like to thank Ned Rodriguez, PhD, for his assistance in designing and implementing this study, Caroline Goldzweig, MD, and Jack Mearns, PhD, for their helpful comments, and Sun S. Hwang, MS, for her statistical assistance.

There was no grant support for this study. Donations from the VA's Women Veteran Coordinator and from the Jewish War Veterans enabled us to give a small payment to each participating veteran.

Ancillary