Prevalence of Mental Health Problems Among Iraq and Afghanistan Veterans Who Have and Have Not Received VA Services

Potential study participants consisted of all individuals who had been deployed for OEF/OIF, who became eligible for VA services between summer 2004 and summer 2009, and who had a New York address at the time they left active duty. Individuals who had since reenlisted in the military or been called back into active duty were also considered eligible for the study. Potential participants were identified from a list of names and addresses of all individuals who became eligible for VA services. The list was obtained via a release-of-names-and-addresses (RONA) request as part of a larger project to assess the needs of veterans residing in New York State (7). For the study reported here, a sample of 7,400 veterans was randomly selected for contact from the complete list of approximately 45,000 individuals.

The mixed-mode survey included Web-based and computer-assisted telephone interviewing (CATI) modes. All sampled individuals were initially mailed an invitation to participate that contained instructions for accessing the Web-based survey. The invitation was accompanied by a letter of support from the New York State Division of Veterans Affairs. The contact information included a substantial number of inaccurate mailing addresses and did not include telephone numbers. When possible, invalid addresses were replaced and land-line phone numbers were obtained from commercial databases. Individuals who did not respond via the Web and who had a valid telephone number were subsequently called, screened for study eligibility, and asked to participate in a CATI or Web survey. Of the 958 persons screened by phone, 47% (N=454) were eligible; the main reasons for ineligibility were not having been deployed to Iraq or Afghanistan (36%, N=348) or no longer living in New York (16%, N=156). All respondents (N=913) were provided $30 for participating. The survey was fielded between August and October 2010.

The recruitment of participants via phone and mail is an unconventional sampling procedure that yields two sampling frames, a dual-mode sampling frame and a Web-only sampling frame (that is, those without a telephone number). The dual-mode sampling frame, which comprised 2,536 individuals with a valid telephone number, had an estimated response rate of 53%, as defined (definition RR3) by the American Association for Public Opinion Research (8). A total of 728 respondents were recruited from this sampling frame. The Web-only sampling frame, for which study eligibility was unknown, provided the remainder of participants (N=185). This study was approved by the RAND Human Subjects Protection Committee. All respondents provided informed consent to participate.

Participants reported a range of sociodemographic and service characteristics. Utilization of VHA services was assessed by asking participants to indicate whether they had used VA health care since leaving the military. PTSD symptoms were assessed with the PTSD Checklist (PCL), an instrument that includes 17 symptom items keyed directly to DSM-IV. The PCL has been used to study posttraumatic distress in various military samples (9,10). On a 5-point scale, respondents rate the extent to which they have been bothered by symptoms during the past 30 days. Symptoms were considered present if respondents had been at least moderately bothered (3). Probable diagnoses of PTSD were derived following the cluster scoring method (11), which corresponds to the DSM-IV criteria for PTSD and has been shown to have high specificity and sensitivity (12).

Major depression was assessed with the Patient Health Questionnaire–8, a well-validated, widely used brief screening measure (13,14). Items correspond to DSM-IV criteria for major depressive disorder, except for thoughts of suicide. Responses indicate on a 4-point (0–3) scale the frequency of symptoms experienced in the past four weeks. Probable depression was indicated by a total score of 10 or above, which yields a sensitivity of .99 and a specificity of .92 (13).

Most of the 913 respondents were male (N=814, 89%), white (N=661, 73%), married (N=482, 53%), and age ≥35 years (N=492, 54%) and had no college degree (N=612, 67%) and no children under 18 living at home (N=503, 55%). Most respondents had been enlisted personnel (N=744, 82%), had experienced only a single deployment (N=540, 60%), had returned from their most recent deployment more than two years ago (N=668, 75%), and reported that their most recent deployment lasted for at least seven months (N=656, 73%). Slightly over half were in the Army at the time of their most recent deployment (N=497, 55%), and just under a fifth were in the Navy (N=123, 14%), Air Force (N=149, 17%), or Marine Corps (N=130, 14%). Slightly over half reported having received VHA services (N=537, 59%). On a wide range of sociodemographic characteristics, the sample closely resembled the national population of military personnel previously deployed for OEF/OIF, as determined by comparison with data from the 2008 Contingency Tracking System Deployment File and the Work Experience File from the Defense Manpower Data Center.

Rates of current probable PTSD and depression in the entire sample were both 16% (for PTSD, N=144, 95% confidence interval [CI]=13%–18%; for depression, N=143, CI=13%–18%). About a fifth of the 913 respondents (N=200, 22%, CI=19%–25%) met criteria for one or both conditions. A total of 144 respondents (16%) had probable PTSD; of these, 87 (60%) also met criteria for probable depression.

Table 1 shows rates of probable PTSD and depression by whether or not respondents reported use of VHA services. For both disorders, these rates were approximately three times greater among respondents who had received VHA care than among those who had not.

Overall, the rates of mental health problems observed in this sample of OEF/OIF veterans residing in New York State closely resemble those observed in an earlier national study of OEF/OIF veterans and service members (1). Both our study and the previous national study attempted to obtain a representative sample of those who had previously been deployed to Iraq or Afghanistan, regardless of their current military status. In addition, the estimates of probable PTSD and depression among those who had received VHA care are very similar to those reported in a census based on VA data that documented prevalence estimates of 22% for PTSD and 17% for depression among OEF/OIF veterans served by the VHA between 2002 and 2008 (3).

Our study is the first to examine rates of mental health problems among OEF/OIF veterans who had not received VHA services, a subgroup that constitutes approximately half of the OEF/OIF veteran population. Rates of probable PTSD and depression among veterans who had received VHA services were approximately three times greater than rates among veterans who had not received such services. This helps to explain the relatively large gap in prevalence estimates across studies that sample from the military and those that sample from the VHA (5). These findings suggest that estimates based on VHA administrative records are not particularly useful for documenting the scale of the public health problem, although they are likely useful for internal VHA planning and resource allocation decisions.

Although we found that PTSD and depression were more prevalent among veterans who had sought VHA services than among their peers who had not, a substantial proportion of the latter group (11%) had PTSD or depression. Given the size of the previously deployed force (15) and the low rate of VHA enrollment (6), this proportion represents a significant number of veterans whose treatment needs are not being met by the VHA. Considerable improvement in the health status of veterans may be achievable if they can be connected with high-quality services. Additional research is needed to identify the barriers to care for these individuals and the extent to which they are accessing services through the private health care system.

The primary limitations of this study pertain to the sampling procedures. Many of the addresses obtained through the RONA request were out of date, which impeded our ability to contact potential participants. Because individuals in the Web-only sampling frame self-screened, it is unclear how many met inclusion criteria for the study. The findings are from a sample of New York State veterans, and the extent to which they are representative of the larger population of OEF/OIF veterans is not known. However, the prevalence estimates from this study are generally similar to those of other studies. In particular, in our study the prevalence rates of PTSD and depression among veterans who had received VHA services are almost identical to those documented in research on veterans accessing VHA services (3).

This study found relatively high rates of PTSD and depression among OEF/OIF veterans. However, rates of mental health problems varied substantially across segments of this population, with rates of mental health problems approximately three times higher among those who had sought VA services than those who had not. Although rates were lower among those not engaged in the VHA system, many of these individuals may benefit from VHA treatment and services. To improve health outcomes for these veterans, further research may be needed on barriers to treatment for this group, along with better outreach.

The original data collection was funded by the New York State Health Foundation. Data analysis and manuscript preparation were funded by grant R01MH87657 from the National Institute of Mental Health. The authors gratefully acknowledge Robin Beckman, M.P.H., for her programming support throughout the project.

The authors report no competing interests.