Depression and posttraumatic stress disorder (PTSD) are common among veterans who served in Afghanistan (Operating Enduring Freedom; OEF) or Iraq (Operation Iraqi Freedom; OIF). Positive screens range from 11.4% to 14.7% for depression and from 9.4% to 9.9% for PTSD (
1,
2). Among OEF/OIF veterans receiving Veterans Health Administration (VHA) care, 48% have at least one diagnosed mental health problem (
3,
4). In one VHA network, depression and PTSD were reported to be the second and third most common diagnoses among OEF/OIF veterans; pain was the most frequently diagnosed condition (
5). Therefore, access to high-quality, evidence-based mental health care is essential.
Rural residence can be a barrier to general medical and mental health care among veterans (
6–
8). Analyses of data from fiscal year 2004 for veterans receiving VHA care for PTSD, depression, or anxiety found that veterans in urban areas were almost twice as likely as those in rural areas to receive four or more or eight or more psychotherapy sessions (p<.001) in the 12 months after diagnosis (
9). Since 2004, VHA has made significant efforts to improve identification and treatment of depression and PTSD and to ensure access to mental health care for rural veterans. In addition to implementing screening for depression and PTSD, the VHA Primary Care–Mental Health Integration Initiative provides mental health services within primary care (
10,
11). Since fiscal year 2005, VHA has created 536 new specialty mental health positions in VHA community-based outpatient clinics (CBOCs) nationwide (
12). A CBOC is a VHA-operated or VHA-contracted health care facility that is geographically separate from the Department of Veterans Affairs medical center (VAMC). VHA also began requiring a mental health care provider on every home-based primary care team and has mandated that VAMCs and very large CBOCs have specialized PTSD programs (
9,
10).
Using a retrospective, cohort design, this study, which was conducted in a single VHA network, examined differences in depression and PTSD screening, diagnosis, and treatment among OEF/OIF veterans in rural and urban areas who received care in fiscal year 2008–2009. During the study period, veterans had at least two visits within this network. The network is geographically large and covers all or part of five states.
Methods
Sample
The data were extracted from a VHA network data warehouse (NDW), a relational database containing information from the VHA electronic health record from all ten VAMCs and 46 CBOCs in the network. The extract included demographic and service priority data, pharmacy and service use records, and screening results. The NDW was selected because, at the time of the study, screening data were not available in national VHA repositories. This network was selected because it had cared for approximately 89,000 OEF/OIF veterans, more than any other VHA network (
15).
Inclusion criteria were as follows: no record of service use within the network before January 1, 2008; at least one visit to a VHA “nexus” clinic between January 1, 2008, and March 17, 2009; and use of a VHA facility in the network at least once in the 12 months after the index visit. The resulting cohort used VHA care in that network at least twice during the study time frame. Screening data were not recorded in the NDW until January 2008, so veterans who received care before that time were excluded. In fiscal year 2008, the VHA defined nexus clinics as all primary care, general medicine, cardiology, diabetes, pulmonary, hypertension, and mental health clinics. The date of the first nexus clinic visit was designated as the index date. These criteria resulted in a cohort of veterans who, on the basis of VHA policy, should have received depression and PTSD screening at the index visit and who had a follow-up visit in that network for a total of at least two visits.
The study was approved by the Central Arkansas Veterans Healthcare System Institutional Review Board.
Measures
Demographic information.
Age at the index visit was included in all models as a continuous variable. Race was categorized into three groups (Caucasian, non-Caucasian, and unknown), with Caucasian as the reference group. Marital status was dichotomized as married versus not married. Rural and urban residence was determined on the basis of residential zip code and the Rural-Urban Commuting Area (RUCA) classification system (
16). On the basis of RUCA category B, veterans were divided into three groups: urban areas, large rural towns, and small or isolated rural towns (
16), with the third category serving as the comparator.
Service priority category.
VHA classifies veterans into eight service priority categories that account for the extent to which the veteran has a service-connected disability and that determine the veteran’s copayment for inpatient and outpatient care and for prescriptions. Categories range from priority group 1, which includes veterans who are ≥50% disabled as a result of a service-connected injury, to priority group 8, which includes veterans who either have no service-connected disability or have a disability that is not compensable. Veterans in categories 1 through 4 were coded high-priority status; groups 5 through 8 were coded as non–high-priority status.
Screening and diagnosis.
Veterans were screened for depression with the two-item Patient Health Questionnaire (PHQ-2). The PHQ-2 is scored from 0 to 6; a score of 3 or more was considered a positive screen. Ideally, individuals who screen positive on the PHQ-2 receive additional screening with the PHQ-9. However, most facilities reported only the PHQ-2 results; therefore, the results of the PHQ-2, reported as positive or negative, were used in all analyses. The Primary Care–PTSD (PC-PTSD) checklist was used to screen for PTSD. This four-item screen is considered positive if a veteran answers yes to three items. The NDW recorded the PC-PTSD results as either positive or negative. Veterans were classified as having depression if there was at least one VHA health care encounter in the 90 days after the index date with an ICD-9-CM code of 2962–2969, 2980, 3004, 30112, 3090, 3091, 30928, or 3110. Veterans were classified as having PTSD if they had at least one VHA health care encounter in the 90 days after the index date with an ICD-9-CM code of 30981.
Treatment variables
The index visit was coded as occurring at a CBOC or VAMC. Psychotherapy was defined as a health care encounter with any of the following current procedural terminology codes in the 90 days after the date of diagnosis: 90804, 90806, 90808, 90810, 90812, 90814, 90845–90847, 90849, 90853, 90857, 90875, 90876, and 96152–96155. Psychotherapy was further categorized as one to three visits versus four or more visits in the 90 days after diagnosis. Pharmacotherapy for depression was defined as at least one antidepressant prescription filled within 90 days after diagnosis; pharmacotherapy for PTSD was defined as at least one psychotropic prescription filled within 90 days after diagnosis.
Analyses
Chi square and t tests were used for bivariate analyses of each variable with RUCA categories. To control for confounding effects of case-mix variables on the relationship between rural residence and outcomes, we fit logistic regression models for each disorder (depression and PTSD). Veterans who had diagnoses of both depression and PTSD were included in both sets of analyses. Covariates in all models included age, gender, race-ethnicity, marital status, service priority category, and index treatment location (VAMC or CBOC). Rather than excluding veterans who screened negative, models for diagnosis and treatment also included a variable for screening results for each disorder because a substantial proportion of veterans who screened negative received a depression or PTSD diagnosis. This variable was not significant and thus is not included in tables. All models controlled for clustering of patients within facilities by including facility code as a repeated measure.
Results
The cohort included 4,782 OEF/OIF veterans who used VHA care at least twice during the study period (January 1, 2008, through March 17, 2009). The mean±SD age was 31.0±8.9 years (range 18–64); 88% (N=4,202) were male. About two-thirds (67%, N=3,212) were Caucasian, 23% (N=1,099) were non-Caucasian, and race was unknown for 10% (N=467). Half of the sample (50%, N=2,414) were not married; most (79%, N=3,766) were in the non–high-priority service category. Most veterans (74%, N=3,553) lived in urban areas, 13% (N=621) lived in large rural towns, and 13% (N=608) lived in small towns or rural areas. Almost half of the index visits (47%, N=2,266) occurred at a CBOC. Most veterans (82%, N=3,941) were screened for both conditions; only 13% (N=618) did not receive either screening. A few were screened only for depression (3%, N=135) or only for PTSD (2%, N=88).
Among the 4,076 veterans (85%) who were screened for depression, 21% (N=840) screened positive. In the full cohort (N=4,782), regardless of screening status, 27% (N=1,267) were diagnosed as having depression. Among the 84% (N=4,029) who were screened for PTSD, 34% (N=1,375) screened positive. In the full cohort (N=4,782), regardless of screening status, 27% (N=1,280) were diagnosed as having PTSD. It was possible for a patient to screen positive for PTSD but not to meet diagnostic criteria for PTSD on examination by a qualified clinician (
17).
Among the veterans with a depression diagnosis, 85% (N=1,077) received either medication or psychotherapy in the 90 days after the diagnosis: 74% (N=938) received at least one antidepressant prescription, and 50% (N=632) had at least one psychotherapy session. Among veterans who had one or more psychotherapy visits, only 24% (N=153) had four or more sessions during that time (mean visits=3.3±6.4).
Among the veterans with a diagnosis of PTSD, 88% (N=1,124) received medication or had at least one psychotherapy visit in the 90 days after diagnosis; 79% (N=1,016) received at least one psychotropic prescription. Selective serotonin reuptake inhibitors were the most commonly prescribed medications. Over half (53%, N=677) had one or more psychotherapy visits in the 90 days after diagnosis. Among those with one or more psychotherapy visits, 25% (N=168) received four or more visits (mean visits=3.0±4.8).
In unadjusted analyses (
Table 1), screening for depression and PTSD were higher for veterans in urban areas compared with those in small or isolated rural towns. For veterans with depression or with PTSD, use of pharmacotherapy and receipt of at least one psychotherapy visit was more common among veterans in small or isolated rural towns than among those in either urban areas or large rural towns. No significant differences between rural and urban areas were noted for number of psychotherapy sessions for either depression or PTSD.
In multivariable analyses (
Table 2), no significant differences between rural and urban areas were found in the odds of depression screening or diagnosis. Among those diagnosed as having depression, veterans in urban areas had significantly lower odds than those in small or isolated rural towns of receiving an antidepressant in the 90 days after diagnosis (odds ratio [OR]=.56, p<.05) and of receiving psychotherapy (OR=.61, p<.05).
Veterans in urban areas were less likely than those in small or isolated rural towns to be diagnosed as having PTSD (OR=.79, p<.05) (
Table 3). Among veterans diagnosed as having PTSD, those in urban areas had significantly lower odds than those living in small or isolated rural towns of receiving psychotropic medication (OR=.52, p<.01) or of having a psychotherapy visit (OR=.61, p<.05). There were no significant differences in receipt of four or more psychotherapy sessions.
As veterans’ age increased by one year, the odds of being diagnosed as having depression decreased by 2% (OR=.98, p<.001). For each one-year increase in age, veterans had a 3% lower odds of being diagnosed as having PTSD (OR=.97, p<.001), 2% greater odds of receiving psychotherapy for PTSD (OR=1.02, p<.05), and 3% greater odds of receiving four or more psychotherapy sessions for PTSD (OR=1.03, p<.05). Compared with male veterans, female veterans were more likely to be diagnosed as having depression (OR=1.35, p<.01), and less likely to be diagnosed as having PTSD (OR=.62, p<.001). Non-Caucasian veterans were less likely than Caucasian veterans to be screened for depression (OR=.74, p<.01) and less likely to receive psychotropic medications in the 90 days after a PTSD diagnosis (OR=.71, p<.05).
Veterans receiving care in a CBOC were more likely than those at a VAMC to be screened for depression (OR=1.69, p<.05) and were more likely to receive antidepressant medication in the 90 days after a depression diagnosis (OR=1.39, p<.05). Veterans who were non–high-priority status were less likely than those with high-priority status to be diagnosed as having PTSD (OR=.57, p<.001).
Discussion
In this study of OEF/OIF veterans in a VHA network who used care at least twice, rural veterans were not less likely than those in urban areas to receive depression or PTSD screening, diagnosis, or treatment. This may reflect VHA’s efforts to increase care to rural veterans (
10–
14). In these analyses, veterans in urban areas were less likely to receive a diagnosis of and be treated for depression and PTSD—a finding that suggests a difference opposite to that reported previously. The fairly high screening rates for depression and PTSD for all veterans in the study suggest that this VHA network has effectively implemented screening programs for these diagnoses. The fairly high rates of pharmacotherapy and psychotherapy suggest that OEF/OIF veterans with depression or PTSD who used VHA care in this network were likely to receive at least some treatment. Although the psychotherapy initiation rates are promising (about 50%), the low number of psychotherapy sessions over a 90-day period is concerning, particularly because this cohort consisted of veterans who were known users of the VHA system. The VHA–Department of Defense clinical practice guideline for depression treatment recommends 16–20 sessions of cognitive-behavioral therapy or interpersonal therapy over 16 weeks (
18). Other authors have identified eight or more or four or more psychotherapy sessions over 12 months as minimally adequate treatment (
9). Analyses of this cohort of veterans with at least two visits in one VHA network found that among those who received psychotherapy, the mean number of visits in the 90-day period after diagnosis was 3.3±6.4 for depression and 3.0±4.8 for PTSD. Four or more psychotherapy visits were received by only 24.2% of veterans after a diagnosis of depression and by 24.8% after a diagnosis of PTSD.
The racial and gender differences in diagnosis and treatment of depression and PTSD are consistent with those in other published studies (
4,
18–
21). This suggests that the findings may be generalizable to VHA care for OEF/OIF veterans outside this network who have used VHA services at least twice.
The conclusions should be viewed with some limitations in mind. Although the findings are useful in understanding rural-urban differences among veterans using VHA care, they cannot be extrapolated to veterans who had no follow-up visit. All veterans in this cohort had at least a baseline and a follow-up visit. The proportion of veterans excluded because they lacked a follow-up visit was similar for each RUCA category (urban, 29.8%; large rural city or town, 32.5%; and small or isolated rural town, 28.7%). Therefore, the sample selection criteria probably did not account for the findings. This analysis is relevant to understanding care for OEF/OIF veterans who are known to be engaged in VHA care, but findings are not generalizable to veterans who may use VHA care once with no follow-up visits. The screening rates in these analyses are slightly lower than those reported for this network by the VA Office of Information and Analytics (OIA), which monitors VHA screening. During the second quarter of fiscal year 2009—the end of the study period—OIA estimated an overall screening rate of approximately 96% in this network, compared with 82.5% in the analysis reported here. This study used NDW data, whereas OIA estimated the rates from chart abstraction, which may identify screening activities that are not documented in the NDW. Given the large numbers of OEF/OIF veterans treated in this network (
14), these data provide useful findings for veterans engaged in VHA care. Future work should examine these issues using national-level VHA data. Finally, these analyses are subject to the limitations of all analyses utilizing secondary data. It is possible that a veteran screened positive for depression or PTSD and received a diagnosis but the record was not coded accordingly and thus the data from that veteran would not have been included in this study. Given the VHA emphasis on depression and PTSD screening, diagnosis, and treatment, this is likely to be a rare occurrence and is unlikely to account for study findings.