Quantitative Research
O r i g i n a l A r t i c l e s
Authors alone are responsible for opinions expressed in the contribution and for its clearance through their federal health agency, if required.
M IL IT A R Y M E D IC IN E, 179, 10:1090,2014
P r e d i c t o r s o f A r m y N a t i o n a l G u a r d a n d R e s e r v e M e m b e r s ’ U s e o f V e t e r a n H e a l t h A d m in i s t r a t io n H e a l t h C a r e A f t e r D e m o b ili z in g
F r o m O E F / O I F D e p l o y m e n t
Alex H. S. Harris, PhD*; Cheng Chen, MA*; Beth A. Mohr, M S f; Rachel Sayko Adams, PhD, M P H f; Thomas V. Williams, PhD f; Mary Jo Larson, PhD, M P A f
ABSTRACT This study described rates and predictors o f Army National Guard and Army Reserve members’ enrollment in and utilization of Veteran Health Administration (VHA) services in the 365 days following demobiliza tion from an index deployment. We also explored regional and VHA facility variation in serving eligible members in their catchment areas. The sample included 125,434 Army National Guard and 48,423 Army Reserve members who demobilized after a deployment ending between FY 2008 and FY 2011. Demographic, geographic, deployment, and Military Health System eligibility were derived from Defense Enrollment Eligibility Reporting System and “Contingency Tracking System” data. The VHA National Patient Care Databases were used to ascertain VHA utiliza tion and status (e.g., enrollee, TRICARE). Logistic regression models were used to evaluate predictors o f VHA utilization as an enrollee in the year following demobilization. Of the study members demobilizing during the observa tion period, 56.9% of Army National Guard members and 45.7% of Army Reserve members utilized VHA as an enrollee within 12 months. Demographic, regional, health coverage, and deployment-related factors were associated with VHA enrollment and utilization, and significant variation by VHA facility was found. These findings can be useful in the design of specific outreach efforts to improve linkage from the Military Health System to the VHA.
INTRODUCTION Since September 11, 2001, more than 2.2 million members of the U.S. Armed Forces have served in the Operation Enduring Freedom (OEF) in Afghanistan and Operation Iraqi Freedom (OIF) in Iraq.1 The length and intensity of these operations, repeat deployments, as well as advancements in battlefield medicine, and increases in military members sur-
*Center for Innovation to Implementation (MPD: 152), Veterans Affairs Palo Alto Health Care System, 795 Willow Road, Menlo Park, CA 94025.
t Institute for Behavioral Health, Heller School for Social Policy and Man agement, Brandeis University, 415 South Street, Waltham, MA 02454-9110.
^Methods, Measures, and Analyses, Defense Health Cost Assessment and Program Evaluation, Department of Defense, Defense Health Agency, 7700 Arlington Boulevard, Suite 5101, Falls Church, VA 22042-5101.
This research has been conducted in compliance with all applicable federal regulations governing the protection of human subjects. Dr. Thomas V. Williams and Dr. Diana D. Jeffery are the DHA/DOD Government Pro ject Managers.
The opinions and assertions herein are those of the authors and do not necessarily reflect the view of the U.S. Department of Defense, Veterans Health Administration, or National Institutes of Health,
doi: 10.7205/MILMED-D-13-00521
viving with serious injuries, including traumatic brain injury, have placed tremendous demands on returning warriors, their families, and the health care systems of the Department of Defense (DoD) and Veterans Health Administration (VHA).2-5 Additionally, many service members return from deployments with ongoing psychological health problems, including post-traumatic stress disorder, depression, and sub stance use problems.6-10
Of the military members deployed to Iraq and Afghanistan as of 2010, roughly half have been members of the Army; with the Reserve Component (RC), specifically Army National Guard (ARNG) and Army Reservists (AR), com prising almost 44% of Army deployments.1 RC members receiving orders to deploy and activating under Title 10 authority are offered TRICARE health insurance coverage and free health care through the Military Health System (MHS) operated by the DoD, although co-pays may be required for services obtained outside the MHS. On return from deployment, RC members go through a requirement- based demobilization process that is designed to ensure that they get the services and assistance they need, including
1090 M ILITARY M ED ICIN E, Vol. 179, October 2014
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
medical, dental, and behavioral health assessments, assis tance in areas of identified need (e.g., vocational, finan cial, personal), offered time-limited MHS insurance (e.g., TRICARE Reserve Select), and information about their rights and benefits including those provided by VHA.
Combat veterans include members who served on active duty in an area of combat operations after 1998 and who were discharged under other than dishonorable conditions. The National Defense Authorization Act of 2008 (Public Law 110-181) entitles all combat veterans, including those in the RC, who meet minimum duty requirements up to 5 years of at least VHA Priority Group 6 status, which includes full access to VHA’s medical benefits package and free VHA services for conditions potentially related to service in a war zone. To rationalize the allocation of resources, VHA has established Priority Groups that are determined primarily by a member’s degree of service-related injury or disability, income, and other service characteristics (http://www.va.gov/healthbenefits/ resources/priority_groups.asp).
This study examines rates and predictors of RC members’ enrollment and utilization of VHA services through this entitlement. RC members are eligible to enroll in VHA immediately after demobilization, but regular Army mem bers cannot enroll in the VHA until they are discharged from military service, often years after deployments. Therefore, RC members, the focus of this study, must be considered separately from regular Army in understanding the predictors and timing of VHA enrollment and utilization.
As with other transitions or hand offs in health care, the transition from the DoD to civilian life and potential utiliza tion of VHA services is fraught with threats to continuity of care. Although “seamless transitions” between the DoD and VHA remain the stated goal, many factors continue to con tribute to suboptimal communication and care coordination between the two systems, including challenges of sharing medial record information, long wait times for determination of benefits, and geographic accessibility.1112
As a heterogeneous group, demobilized RC members may or may not return to civilian employment, regain private health insurance, or receive medical care through private programs. Some RC members enroll for and receive services from VHA as soon as they are eligible and others do not make the transi tion. Some members who do not enroll for and receive VHA services have simply chosen other good options for their health care needs. However, other members do not engage with the VHA system for less positive reasons including lack of knowl edge of benefits, frustration with the enrollment process, and perceptions of low-quality care.5,13,14 Also, stigma exists among combat veterans regarding treatment seeking in gen eral, particularly for mental health problems.15-17
Few studies have examined transitions from the DoD to VHA for OEF/OIF service members, and none have focused specifically on RC members. Copeland et al18 examined 994 service members (Active Duty, Guard, and Reserve) who were traumatically injured and medically discharged from
one inpatient DoD trauma treatment facility. The service members were followed to determine the rate, predictors, and patterns of subsequent VHA utilization, even though not all were discharged from the military. From this sample, 23% used VHA services as an enrollee in the 2-year observation period. Members wounded in action had longer transition times to VHA, whereas those with bums had shorter transi tion times. These data are difficult to interpret because not all members of the sample had been discharged from the mili tary and therefore might not have been eligible for VHA services as an enrollee.
In another study, Randall13 determined that the length of time to obtain VHA services at two VHA Medical Centers for 376 OEF/OIF combat service members who left active ser vice averaged 3.8 months. Respondents were from all branches and cited several factors that impeded them from enrolling in VHA sooner, including not knowing about bene fits, distance and transportation barriers, viewing help-seeking as a sign of weakness, and negative perceptions about the quality of care provided in VHA. Because all service members of the sample received VHA services, predictors of linkage to the VHA could not be examined.
The VHA Office of Public Health reported that between October 2001 and December 2012, 56% (n = 899,752) of the more than 1.6 million OEF/OIF/Operation New Dawn com bat veterans from all branches of the military who became eligible for VHA services eventually enrolled for and utilized VHA services.19 Of these, 688,414 were RC members of whom 55% utilized VHA services.19 However, little is known about the individual, geographic, or system factors that predict transitions from DoD to VHA care among RC members. Data on the predictors and locations of poor linkage are essential to the effective design and targeting of outreach and quality improvement efforts.
Thus, the purpose of this study of RC members was to describe rates and predictors of any utilization as an enrollee of VHA health care in the 365 days after the demobilization date following an index deployment ending in FY 2008-FY 2011. Although prior studies have estimated that on average half of eligible veterans eventually make use of VHA ser vices,19'20 the present study’s ability to track a large cohort of returning RC members and to examine predictors of linkage to the VHA within 1 year of their demobilization from deploy ment is unique. Further, we explored regional and VHA facil ity variability in enrolling and serving eligible veterans in their catchment areas. Information about patient and facility-level characteristics associated with linkage to the VHA after demo bilization can be used by the DoD and VHA to develop quality improvement efforts to design organizations and systems of care more likely to increase linkage to VHA facilities among returning OEF/OIF combat veterans.
METHODS This study was part of the Substance Use and Psychological Injury Combat study (SUPIC), a longitudinal, observational
MILITARY MEDICINE, Vol. 179, October 2014 1091
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
study of Army service members returning from deployment, funded by the National Institute on Drug Abuse and con ducted with the sponsorship from the Defense Health Agency/DoD. SUPIC is designed to study postdeployment health outcomes among Army members utilizing merged administrative data systems of the DoD and VHA. The SUPIC cohort is inclusive of Active Duty and RC members, but analyses are stratified by these components, and this study focuses only on those in the RCs. This study uses a prospective design to examine one key outcome associated with OEF/OIF deployment, successful linkage to the VHA to receive entitlement health care services after demobiliza tion. Members who demobilize from a combat deployment became eligible to enroll for VHA under Public Law 110- 181, the National Defense Authorization Act for Fiscal Year 2008 and previous authorizations.
S tu d y S e t tin g
VHA is composed of over 1,700 sites of care, hierarchically organized into 140 major facilities in 21 Veterans Integrated Service Networks. VHA serves over 6 million veterans each year through integrated and comprehensive outpatient, resi dential, and inpatient services, including specialized pro grams for a host of issues of particular concern for Veterans (http://www.va.gov/health/programs/index.asp). Members of the RC become eligible to enroll after a qualifying index deployment (i.e., called to service by federal order, served the entire period of the order, released under Honorable or General Under Honorable conditions). Unlike other VHA utilization studies, SUPIC prospectively identifies a complete VHA eligible population by selecting a cohort through DoD records. The SUPIC deployment cohort is tracked longitudi nally to observe VHA enrollment, and utilization as an enrollee, in one of the VHA facilities in the United States.
S a m p le
As a part of SUPIC, this study used the Defense Enrollment Eligibility Reporting System, within the MHS Data Reposi tory, and the Contingency Tracking System of the Defense Manpower Data Center to identify RC members who had deployed to OEF/OIF countries (Iraq, Afghanistan, Qatar, and Kuwait), and who had demobilization dates associated with an index deployment end date between October 1, 2007 and September 30, 2011 (FY 2008-FY 2011). Among the one-third of members who had multiple deployments during the observation period, we chose an index deployment by selecting the first deployment ending in the study window that could be matched to a completed postdeployment health assessment (PDHA), or the first deployment ending in the study window for those without a completed PDHA. Service members were excluded if their deployments did not include OEF/OIF countries. Additional details about the sample development, matching of PDHAs to deployments, and selec tion of the index deployment are described in detail else
where.21 To be included in the analysis of linkage to a particular VHA facility, we further restricted the sample to those with a valid postdemobilization civilian zip code in the United States.
O u tc o m e
The outcome for this study was the receipt of any outpatient, inpatient, or residential care from a VHA facility as an enrollee (i.e., not TRICARE or other sharing arrangement) at least once during the 365 days after the index demobiliza tion date. We identified this utilization using the VHA admin istrative data including the VHA Enrollment File and National Patient Care Databases.
P r e d ic t o r V a ria b le s
Deployment record information was derived from the Con tingency Tracking System and demographic characteristics at the start and end of deployment and demobilization dates were derived from the Defense Enrollment Eligibility Reporting System. The following variables (measured at start of index deployment unless indicated) were used as predic tors of postdemobilization VHA utilization: age, gender, probable serious injury during deployment (defined as receiv ing inpatient services within a major MHS hospital after deployment), deployment again in the postindex year, postde ployment enrollment DoD provided health insurance (i.e., PRIME/TRICARE Reserve Select (TRS)) in the postindex year, number of deployments before the index deployment, length of index deployment in months, rank, race, marital status, region of residence at demobilization, and fiscal year of deployment end date (FY 2008-FY 2011). Other predictors derived from the VA Enrollment File and National Patient Care databases were pre- and postindex deployment VHA utilization as a nonenrollee (e.g., TRICARE) and preindex deployment VHA utilization as an enrollee (presumably eligi ble through demobilization from a previous deployment or active duty discharge). Driving time of the members’ postdemobilization residential zip code to the nearest VHA facility with primary care services was determined from the VHA Planning System Support group data. VHA Planning System Support Group data provide driving times for VHA patients within each 5-digit zip code to the nearest VHA facil ity with primary care services. Because postdemobilization zip codes for the entire sample were at the 3-digit level, we assigned the median driving time from distribution of VHA patients in the underlying 5-digit zip codes to all individuals in the sample with the same 3-digit stem.
S ta tis t ic a l A n a ly s e s
Mixed-effects logistic regression models were used to predict VHA utilization as an enrollee in the 365 days following the date of demobilization from the index deployment, with a random effect for VHA facility (N = 140) to account for the clustering of members within a facility. All candidate
1092 MILITARY MEDICINE, Vol. 179, October 2014
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
predictors were included and regression diagnostics, particu larly examination of variance inflation factors, were performed. For continuous variables such as age, assumptions regarding linearity were evaluated and found to be reasonable. Analyses were stratified by com ponent (i.e., ARNG and AR) because of the presence of significant interaction terms. Because o f the large size o f the sample, we emphasize the magnitude and confidence intervals o f effects rather than p values.
Describing VHA Facility-Level Variation in “Yield” From the Local Population To explore VHA facility-level variation in enrollm ent and utilization o f VHA services, we assigned each m em ber in the sample to one o f the 140 major VHA facilities with the shortest drive time by the abovem entioned method. Note that all o f VHA medical centers, clinics, and other settings
of care are organized into one o f these 140 organizational units. This is the level o f aggregation that is used for most system m onitoring and perform ance m easurement. Then, using the m ultivariate m ixed-effects regression models predicting utilization and controlling for other individual characteristics, we estim ated the proportion (95% confidence interval [Cl]) o f mem bers in each VHA facility’s catchment area that received VHA services as an enrollee. The purpose o f these analyses is to describe variation in facility-level yield from the local population o f eligible veterans, and to identify high and low outliers.
Protection of Human Subjects and Data Security To ensure protection o f human subjects, Brandeis University’s Com m ittee for Protection o f Human Subjects, the Insti tutional Review Boards o f the Stanford University and
T A B L E I. Characteristics of ARNG (N = 125,434) and Army Reserve Members (N = 48,423), Returning from Deployment FY 2008- FY 2011
Characteristic" National Guard N (%) Reserve N (%)
Female 10,887 (8.68) 7,316(15.11) Married 63,422 (50.56) 63,422 (50.56) Race/Ethnicity
Non-Hispanic White 95,155 (75.86) 95,155 (75.86) Non-Hispanic African-American 14,877 (11.86) 14,877 (11.86) Hispanic 9,963 (7.94) 9,963 (7.94) Asian or Pacific Islander 3,358 (2.68) 3,358 (2.68) American Indian/Alaskan Native 1,235 (0.98) 1,235 (0.98) Other 846 (0.67) 846 (0.67)
Received Preindex6 VHA Services as Enrollee 35,939 (28.65) 35,939 (28.65) Received Preindex VHA Services as Nonenrollee 11,708 (9.33) 11,708 (9.33) Received Postindex VHA Services as Nonenrollee 4,583 (3.65) 4,583 (3.65) Probable Serious Injury During Index Deployment 4,856 (3.87) 4,856 (3.87) Redeployed in the Postindex Year 2,411 (1.92) 2,411 (1.92) PRIME/TRSC After Index Deployment 50,364(40.15) 50,364(40.15) Rank
Enlisted, Junior 67,173 (53.55) 67,173 (53.55) Enlisted, Senior 45,773 (36.49) 45,773 (36.49) Officer, Junior 7,591 (6.05) 7,591 (6.05) Officer, Senior 2,800 (2.23) 2,800 (2.23) Warrant Officer 2,097(1.67) 2,097 (1.67)
Residence Region at Demobilization West 18,429(14.69) 18,429(14.69) Midwest 31,870(25.41) 31,870 (25.41) Northeast 21,029 (16.76) 21,029(16.76) South 54,106 (43.14) 54,106 (43.14)
Cohortrf 2008 54,106 (43.14) 8,231 (17.00) 2009 54,106(43.14) 11,531 (23.81) 2010 17,831 (14.22) 15,625 (32.27) 2011 27,723 (22.10) 13,036 (26.92)
Mean (SD) Mean (SD)
Age in Years 8,231 (17.00) 8,231 (17.00) Number of Deployments Before Index Deployment 11,531 (23.81) 11,531 (23.81) Length of Index Deployment in Months 15,625 (32.27) 15,625 (32.27) Drive Time (Minutes) to Nearest VHA Facility 13,036 (26.92) 13,036 (26.92)
“Measured at start o f index deployment unless indicated. '’Index refers to a deployment ending in FY 2008-FY 2011 and selected for analysis. “DoD provided health insurance. “'Cohort refers to fiscal year of index deployment end date.
MILITARY MEDICINE, Vol. 179, October 2014 1093
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
T A B L E II. Predictors of VHA Utilization as an Enrollee in the 365 days Postdemobilization for 48,423 National Reserve Members and 125,434 National Guard Members
Parameter" National Reserve OR* (95% Cl) National Guard OR* (95% Cl) Age in Years 1.011 (1.008, 1.014) 1.020(1.018. 1.022) Female 1.329(1.258, 1.404) 1.241 (1.188, 1.296) Preindex" VHA Services as Enrollee 4.210 (4.002,4.430) 2.737 (2.656, 2.820) Preindex VHA Services as Nonenrollee 1.558 (1.465, 1.656) 1.706(1.627, 1.788) Postindex VHA Services as Nonenrollee 2.096 (1.876,2.341) 1.857(1.727, 1.996) Probable Serious Injury During Index Deployment 1.282(1.159,1.417) 1.103 (1.033, 1.178) Deployed Again in the Postindex Year 0.574 (0.512,0.643) 0.604 (0.553, 0.659) PRIME/TRS^ Enrollment in the Postindex Year 0.963 (0.924, 1.005) 1.033 (1.006, 1.060) Number o f Deployments Before Index Deployment 0.809 (0.784, 0.834) 0.853 (0.835,0.872) Length of Index Deployment in Months 1.012(1.006, 1.018) 1.014 (1.010, 1.019) Drive Time (Minutes) to Nearest VHA Facility 0.997 (0.995,1.000) 0.997 (0.996, 0.999) Rank (Junior Enlisted as Reference)
Enlisted, Senior 0.778 (0.738, 0.820) 0.797 (0.771,0.823) Officer, Junior 0.581 (0.536, 0.629) 0.575 (0.545,0.606) Officer, Senior 0.481 (0.437,0.528) 0.471 (0.431,0.514) Warrant Officer 0.620 (0.525,0.733) 0.597 (0.542,0.658)
Race (Non-Hispanic White Reference) American Indian/Alaskan Native 0.947 (0.776,1.156) 0.854 (0.755, 0.966) Asian or Pacific Islander 0.959 (0.865, 1.063) 1.040(0.960, 1.126) Non-Hispanic African-American 1.092(1.030, 1.157) 1.103(1.061, 1.148) Hispanic 1.115(1.040, 1.195) 1.063 (1.008, 1.121) Other 0.906 (0.695,1.182) 1.004(0.868, 1.162)
Married 0.962 (0.922, 1.005) 1.012(0.986, 1.037) Residence Region at Demobilization (West Reference)
Midwest 1.162 (0.987,1.369) 1.337 (1.107,1.616) Northeast 1.007 (0.843, 1.204) 1.107 (0.866, 1.383) South 0.981 (0.843, 1.142) 1.336(1.116, 1.599)
Cohort" (2008 Reference) 2009 1.193 (1.121,1.269) 1.104(1.057,1.152) 2010 1.023 (0.964, 1.086) 0.935 (0.899, 0.973) 2011 0.900 (0.847, 0.957) 0.767 (0.735,0.801)
"Measured at start of index deployment unless indicated. ^Adjusted for all characteristics shown. "Index refers to a deployment ending in FY 2008-FY 2011 and selected for analysis. *DoD provided health insurance. "Cohort refers to fiscal year of index deployment end date.
the VA Palo Alto Health Care System, and the Human Research Protection Program at the Office of the Assistant Secretary of Defense for Health Affairs/Defense Health Agency granted approval. The Defense Health Agency Pri vacy and Civil Liberties Office executed an annual data shar ing agreement.
RESULTS Descriptive statistics for the samples are presented in Table I. Of the 125,434 ARNG members who demobilized during the observation period and met other eligibility criteria, 71,322 (56.9%) utilized VHA as an enrollee within 12 months of their index demobilization date. Table II presents regression results for predictors of ARNG members’ VHA utilization as an enrollee. These results may be helpful in targeting and shaping enrollment efforts for increase enrollment and utilization of VHA services. Significant positive predictors were older age; female gender; receiving VHA services as an enrollee before the index deployment; receiving VHA services as a nonenrollee (e.g., TRICARE) before the index deployment; receiving VHA services as a nonenrollee (e.g.,
TRICARE) in the year after the index deployment; probable serious deployment-related injury; DoD provided health insurance (TRS or PRIME) selection or enrollment in the postindex year; length of index deployment; being Asian/ Pacific Islander, Hispanic, African-American, or “other” race/ethnicity compared to non-Hispanic white; residence after demobilization in the Midwest or South compared to the West region; and being a member of the FY 2009 cohort compared to the FY 2008 cohort.
Significant negative predictors of ARNG members VHA utilization as an enrollee were deployment again in the postdemobilization year, number of deployments before index deployment; drive time to nearest VHA facility; rank of Senior Enlisted, Junior Officer, Senior Officer or Warrant Officer compared to Junior Enlisted; being American Indian/ Alaskan native; and demobilization in 2010-2011 compared to 2008.
Of the 48,423 AR members who demobilized, 22,118 (45.7%) utilized VHA as an enrollee within 12 months of the index demobilization date. Table II presents regression results for predictors of AR members’ VHA utilization as an enrollee. Similar to ARNG, significant positive predictors
1094 MILITARY MEDICINE, Vol. 179, October 2014
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
Utilization of VHA Services as an Enrollee by National Guard Members in the Catchment Areas of 140 VHA Facilities
FIG UR E 1. Percent of National Guard Members with any VHA utilization as an enrollee in a VHA facility catchment area. Utilization measured in the year following member’s return from a FY 2008-FY 2011 deployment. Adjusted for characteristics shown in Table II.
were older age, female gender, receiving VHA services as an enrollee before the index deployment, receiving VHA ser vices as a nonenrollee (e.g., TRICARE) before the index deployment, receiving VHA services as a nonenrollee in the year after the index deployment, probable serious deployment- related injury, DoD provided health insurance (TRS or PRIME) selection or enrollment in the postindex year, length of index deployment, being Hispanic or African-American compared to non-Hispanic White, residence in the Midwest or South compared to the West region, and being a member of the FY 2009 cohort compared to the FY 2008 cohort.
Significant negative predictors of Army Reserve Members VHA utilization as an enrollee were deployment again in the postdemobilization year; number of deployments before index deployment; longer drive time to the nearest VHA facility; rank of Senior Enlisted, Junior Office, Senior Officer or War rant Officer compared to Junior Enlisted; being an American Indian or Alaskan Native compared to non-Hispanic White; and demobilization in 2010-2011 compared to 2008.
The percent (95% Cl) of ARNG and AR members in each VHA facility’s catchment area who received any VHA services as an enrollee are presented in Figures 1 and 2, respectively. The results are adjusted for the characteristics shown in Table II. As can be seen, substantial variation exists (as well as variation in the Cl widths driven by differences in the underlying sample sizes). For ARNG members, facility- level utilization in the year following the index deployment ranged from 31% to 89%. For AR members facility-level
utilization ranged from 27% to 81%. The percents for facili ties with nonoverlapping C l’s are significantly different.
DISCUSSION This descriptive analysis found that 56.9% of ARNG and 45.7% of AR members utilized VHA services as an enrollee within 365 days of demobilization from an index deployment to Iraq or Afghanistan. This implies that if linkage to the VHA does occur, most of it will occur within the first year of return from deployment. We plan future analysis to follow this cohort up to 3 years from deployment to determine if linkage is delayed for some members, and if so, what charac teristics are associated with delayed linkage. Perhaps the most significant finding, discussed in more detail below, was the substantial range in enrollment and utilization among the 140 VHA facilities, from as low as 25% to over 85%, even after adjusting for driving time, demographics, and service-related factors. Future research and quality improvement efforts with DoD and VHA should first strive to better understand this variation and the extent to which it is explained by factors such as the availability of non-VHA health care options, actual or perceived quality of VHA care, or the presence of specific education, outreach, and linkage activities. Additionally, rea sons underlying the higher VHA utilization among ARNG compared to AR should be explored.
We identified demographic, regional, health coverage, and deployment-related factors that were associated with
MILITARY MEDICINE, Vol. 179, October 2014 1095
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
Utilization of VHA Services as an Enrollee by National Reserve Members in the Catchment Areas of 140 VHA Facilities
FIG UR E 2. Percent of National Reserve Members with any VHA utilization as an enrollee in a VHA facility catchment area. Utilization measured in the year following member’s return from a FY 2008-FY 2011 deployment. Adjusted for characteristics shown in Table II.
VHA utilization as an enrollee. In general, these associations were very similar across the ARNG and AR samples. It is also noteworthy that female members were more likely to enroll and engage in VHA care compared to male members. This is an important finding given that women are the fastest growing segment of the veteran population, the number of women veterans using VHA has nearly doubled in the past decade, and VHA’s substantial efforts to meet the needs of
22women veterans. Also, members with pre- or postindex deployment contact
with VHA as a nonenrollee, or preindex contact as an enrollee, had increased odds of subsequently enrolling in VHA services after the index deployment. The only predictor that had substantially different predictive value between the samples was receiving preindex services at VHA as an enrollee, presumably related to a previous deployment. Although this experience was positively associated to postindex utilization as an enrollee in both ARNG and AR samples, the effect was substantially larger in the AR sample (odds ratio [OR] = 4.2 vs. 2.7). The drivers behind this dif ference are unknown. Therefore, previous contact with VHA may be interpreted as an “enabling factor,” that is, veterans who learn to navigate the process of making an appointment and arriving at a VHA have greater likelihood of doing so again compared to other veterans without VHA contact. This pattern of utilization predicting more utilization may also reflect some underlying medical need that was not measured. The one measure of medical need that was included in the
models, probable serious injury, was associated with a 16% increase in odds of VHA enrollment and utilization. It should be noted that all of these preindex utilization factors provided more opportunities to have contact with the active efforts of both systems to encourage VHA enrollment.
The negative relationship between number of preindex deployments and odds of VHA enrollment and utilization was unexpected, given the assumption that each deployment provides addition risk of medical problems and opportunities for VHA linkage. This result is perhaps driven by what has been termed the “healthy warrior effect,”23 which is analo gous to the epidemiological concept of the “healthy worker effect.” This apparent paradox refers to the fact that healthy people may experience more repeated deployments com pared to less healthy people, meaning that those with more previous deployments may have less medical need than with out previous deployments.
Military rank was associated with the odds of VHA enroll ment and utilization: junior enlisted were more likely than other ranks to link to VHA care. Lower rank is undoubtedly correlated with lower military income and perhaps general lower socioeconomic status, and thus these veterans may find seeking care at the VHA more attractive than private or other alternative health care settings. Race/ethnicity was also related to odds of VHA enrollment and utilization. African-Americans and Hispanics were somewhat more likely than non-Hispanic Caucasians to enroll in VHA. However, American Indian and Alaskan Natives ARNG members had 15% lower odds of
1096 MILITARY MEDICINE, Vol. 179, October 2014
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
enrollment compared to non-Hispanic Caucasians. This con cerning finding may be driven by American Indian and Alaskan Natives being eligible for other publically funded health care (e.g., Indian Health Service), but may also identify a possible disparity and target for improved outreach.
More in-depth research is needed to better understand why, for example, members from the West are substantially less likely to use VHA services than members from the Mid west or South. If these differences are driven by differences in outreach programs, education about available benefits, or perceptions about the quality of VHA care, then these can be addressed through targeted quality improvement programs. Also, we do not know why the utilization rates were lower in 2010 and 2011 compared to 2008. Hard economic conditions in 2008 may have restricted employment-related health insur ance access for returning members, highlighting the VHA role as an important health care safety net for veterans. All of these predictors of VHA linkage need to be better under stood as either understandable or anticipated, or problematic and an indicator of a disparity that requires further research and new remedies.
Although the most policy-relevant data from this study are perhaps the facility-level variations in linkage, these data alone do not explain their root causes or specific remedies. It is quite remarkable, however, that some VHA facilities are utilized by over 80% of eligible ARNG and AR members in their catchment areas and other are utilized by less than a quarter. Although our method of constructing catchment areas was ad hoc and probably crude, these data raise impor tant questions about potentially actionable factors that should be further examined. Our ongoing research is investigating whether facilities engaging more members are of higher qual ity, have fewer regional health care options, or have specific education and outreach programs. We are also exploring if identification of a behavioral health problem on one of the postdeployment screens is predictive of VHA enrollment. Future studies should explore what aspects of prior VHA utilization affect veterans’ experiences, impressions, and desires to use VHA care again. Finally, this study studied predictors of at least one VHA encounter, which is a low threshold. Knowing more about predictors of more extensive engagement (i.e., more than one visit) and how medical need relates to the type of care utilized would also be informative.
Several limitations should be noted about the current anal ysis. We have not yet accounted for other indicators of deployment-related medical needs. Also, although the binary outcome and logistic regression framework solved some ana lytic issues, particularly the failure of our data to meet the proportional hazards assumption of survival analyses, a time- to-event analysis could have addressed the timing of enroll ment in a more nuanced way and accounted for censorship related to postindex death and redeployments. Also, our con struction of VHA catchment areas based on median drive times between 3-digit zip codes was ad hoc and has not been validated. Results related to these catchment areas should be
interpreted in light of this uncertainty. Finally, the extent to which these findings generalize to members of the regular Army or other service branches is unknown.
In conclusion, this study is, to our knowledge, the first to describe the predictors of ARNG and AR members’ enroll ment and utilization of VHA services in the year after an index deployment. By identifying demographic and other predictors of enrollment and utilization of VHA services, and by describing the significant variation between VHA facilities in linking with ARNG and AR members in their catchment areas, this study provides a foundation for further research aimed at better understanding both positive and neg ative deviance, and developing quality improvement strate gies aimed at reducing disparities and maximizing veterans’ access to the benefits they have earned.
ACKNOWLEDGMENTS The Office of the Assistant Secretary of Defense for Health Affairs/Defense Health Agency (DHA) of the U.S. Department of Defense (DOD) provided access to these data. This study was funded by National Institute of Drug Abuse (NIDA) Grant No. R01DA030150 (PI: Dr. Larson).
REFERENCES 1. IOM (Institute of Medicine): Returning Home From Iraq and Afghanistan:
Assessment of Readjustment Needs of Veterans, Service Members, and Their Families. Washington, DC, National Academy of Sciences, 2013.
2. IOM (Institute of Medicine): Returning Home From Iraq and Afghanistan: Preliminary Assessment of Readjustment Needs of Military Personnel, Veterans, and Their Families. Washington, DC, National Academy of Sciences, 2010.
3. Hoge CW: Supporting resilience in a deployment context: Public health strategies and treatment of service members and veterans with combat- related mental health problems. In: Deployment Psychology: Evidence- Based Strategies to Promote Mental Health in the Military, pp 17-34. Edited by Adler AB, Bliese PD, Castro CA. Washington, DC, American Psychological Association, 2011.
4. Melcer T, W alker GJ, Galameau M, Belnap B, Konoske P: Midterm health and personnel outcomes of recent combat amputees. Mil Med 2010; 175(3): 147-54.
5. Tanielian T, Jaycox LH: Invisible Wounds of War: Psychological and Cognitive Injuries, Their Consequences, and Services to Assist Recov ery. Santa Monica, CA, RAND Corporation, 2008.
6. Larson MJ, Wooten NR, Adams RS, Merrick EL: Military combat deployments and substance use: review and future directions. J Soc W ork Pract Addict 2012; 12(1): 6—27.
7. Hoge CW, Auchterlonie JL, Milliken CS: Mental health problems, use of mental health services, and attrition from military service after returning from deployment to Iraq or Afghanistan. JAMA 2006; 295(9): 1023-32.
8. Milliken CS, Auchterlonie JL, Hoge CW: Longitudinal assessment of mental health problems among active and reserve component soldiers returning from the Iraq war. JAMA 2007; 298(18): 2141-8.
9. Adams RS, Larson MJ, Conrigan JD, Horgan CM, Williams TV: Fre quent binge drinking after combat-acquired traumatic brain injury among active duty military personnel with a past year combat deploy ment. J Head Trauma Rehabil 2012; 27(5): 349-60.
10. Bray RM, Pemberton MR, Hourani LL, et al: 2008 Department of Defense Survey of Health Related Behaviors Among Active Duty Military Personnel. RTI/10940-FR. Durham, NC, RTI International, Research Triangle Institute, 2009.
MILITARY MEDICINE, Vol. 179, October 2014 1097
Predictors o f Army National Guard and Reserve Members Linkage to the VHA
11. Department of Defense, Department of Veterans Affairs, Department of Health and Human Services: Interagency Task Force on Military and Veterans Mental Health: 2013 Interim Report. Available at http://www .whitehouse.gov/sites/default/files/uploads/2013_interim_report_of_the_ interagency_task_force_on_military_and_veterans_mental_health.pdf; accessed February 19, 2014.
12. IOM (Institute o f Medicine): Returning Home from Iraq and Afghanistan: Readjustment Needs of Veterans, Service Members, and Their Families. Washington, DC, National Academy of Sciences, 2013.
13. Randall MJ: Gap analysis: transition of health care from Department of Defense to Department of Veterans Affairs. Mil Med 2012; 177(1): 11- 6.
14. Bumam MA, Meredith LS, Tanielian T, Jaycox LH: Mental health care for Iraq and Afghanistan W ar Veterans. Health Aff 2009; 28(3): 771-82.
15. Mittal D, Drummond KL, Blevins D, Curran G, Corrigan P, Sullivan G: Stigma associated with PTSD: perceptions of treatment seeking combat veterans. Psychiatr Rehabil 2013; 36(2): 86-92.
16. Pietrzak RH, Johnson DC, Goldstein MB, Malley JC, Southwick SM: Perceived stigma and barriers to mental health care utilization among OEF-OIF veterans. Psychiatr Serv 2009; 60(8): 1118-22.
17. Rosen CS, Greenbaum MA, Fitt JE, Laffaye C, Norris VA, Kimerling R: Stigma, help-seeking attitudes, and use of psychotherapy in veterans with diagnoses of posttraumatic stress disorder. J Nerv Ment Dis 2011; 199(11): 879-85.
18. Copeland LA, Zeber JE, Bingham MO, et al: Transition from military to VHA care: psychiatric health services for Iraq/Afghanistan combat- wounded. J Affect Disord 2011; 130(1-2): 226-30.
19. Epidemiology Program P-DHG, Office of Public Health Veterans Health Administration: Analysis of VA Health Care Utilization among Operation Enduring Freedom (OEF), Operation Iraqi Freedom (OIF), and Operation New Dawn (OND) Veterans, from 1st Qtr FY 2002 through 1st Qtr FY 2013. Washington, DC, Office of Public Health Veterans Health A dm inistration, 2013. Available at http://www .publichealth.va.gov/docs/epidemioiogy/healthcare-utilization-report- fy2013-qtrl.pdf; accessed February 19, 2014.
20. VHA Office of Public Health and Environmental Hazards: Analysis of VA Health Care U tilization among US Global W ar on Terrorism (GWOT) Veterans: Operation Enduring Freedom, Operation Iraqi Free dom. Washington, DC, VHA Office of Public Health and Environmental Hazards, 2009. Available at http://www.networkofcare.org/library/ GWOT_4th%20Qtr%20FY08%20HCU.pdf; accessed February 19, 2014.
21. Larson M, Adams R, Mohr B, et al: Rationale and methods of the Substance Use and Psychological Injury Combat Study (SUPIC): a longi tudinal study of army service members returning from deployment in FY 2008-2011. Subst Use Misuse 2013; 48(10): 863-79.
22. Frayne SM PC, Friedman SA, Saechao F, et al: Sourcebook: Women Veterans in the Veterans Health Administration. Volume 2: Sociodemo graphics and Use of VHA and Non-VA Care (Fee). Washington, DC, W omen’s Health Evaluation Initiative, W omen’s Health Services, Vet erans Health Administration, Department of Veterans Affairs, 2012. Available at http://www.womenshealth.va.gov/WOMENSHEALTH/ docs/SourcebookVo!2_508c_FINAL.pdf; accessed February 19, 2014.
23. Haley RW: Point: bias from the “healthy-warrior effect” and unequal follow-up in three government studies of health effects of the Gulf War. Am J Epidemiol 1998; 148(4): 315-23.
1098 MILITARY MEDICINE, Vol. 179, October 2014
Copyright of Military Medicine is the property of AMSUS and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.