Journal of Rehabilitation Research & Development (JRRD)

Quick Links

  • Health Programs
  • Protect your health
  • Learn more: A-Z Health
Veterans Crisis Line Badge
 

Volume 50 Number 8, 2013
   Pages 1047 — 1068

Healthcare utilization and costs of Veterans screened and assessed for traumatic brain injury

Kevin T. Stroupe, PhD;1–2* Bridget M. Smith, PhD;1–3 Timothy P. Hogan, PhD;4 Justin R. St. Andre, MA;5 Theresa Pape, DrPH, MA;1,6–7 Monica L. Steiner, MD;6 Eric Proescher, PsyD;8 Zhiping Huo, MS;1 Charlesnika T. Evans, MPH, PhD1,3,9

1Center for Management of Complex Chronic Care, Edward Hines Jr. Department of Veterans Affairs (VA) Hospital, Hines, IL; 2Program in Health Services Research, Stritch School of Medicine, Loyola University Chicago, Maywood, IL; 3Spinal Cord Injury Quality Enhancement Research Initiative (QUERI), Edward Hines Jr. VA Hospital, Hines, IL; 4Center for Health Quality, Outcomes and Economic Research & eHealth QUERI, National eHealth QUERI Coordinating Center, Edith Nourse Rogers Memorial Veterans Hospital, Bedford, MA; and Division of Health Informatics and Implementation Science, University of Massachusetts Medical School, Worcester, MA; 5Health Research & Educational Trust, Chicago, IL; 6Physical Medicine and Rehabilitation, Edward Hines Jr. VA Hospital, Hines, IL; 7Department of Physical Medicine and Rehabilitation, Feinberg School of Medicine, Northwestern University, Chicago, IL; 8Jesse Brown VA Medical Center, Chicago, IL; 9Center for Healthcare Studies, Feinberg School of Medicine, Northwestern University, Chicago, IL

Abstract — Approximately 15% of casualties in the Afghanistan (Operation Enduring Freedom [OEF]) and Iraq (Operation Iraqi Freedom [OIF]) conflicts received mild traumatic brain injury (TBI). To identify Veterans who may benefit from treatment, the Department of Veterans Affairs (VA) implemented a national clinical reminder in 2007 to screen for TBI. Veterans who screen positive are referred for a comprehensive TBI evaluation. We conducted a national retrospective study of OIF/OEF Veterans receiving care at VA facilities between 2007 and 2008. We examined the association of the TBI screen with healthcare costs over a 12 mo period following the initial evaluation. Of the Veterans, 164,438 met inclusion criteria: 31,627 screened positive, 118,545 screened negative, and 14,266 received no TBI screening. Total healthcare costs of Veterans who screened positive, screened negative, or had no TBI screening were $9,610, $5,184, and $3,399, respectively (p < 0.001). Understanding these healthcare utilization and cost patterns will assist policymakers to address the ongoing and future healthcare needs of these returning Veterans.

Key words: cost analysis, Department of Veterans Affairs, evaluation, head injury, healthcare costs, healthcare utilization, Operation Enduring Freedom, Operation Iraqi Freedom, traumatic brain injury, Veterans.

Abbreviations: CR = clinical reminder, DSS = Decision Support System, GLM = generalized linear model, NDE = National Data Extract, OEF = Operation Enduring Freedom, OIF = Operation Iraqi Freedom, OR = odds ratio, PNS = polytrauma network site, POC = point of contact, PRC = polytrauma rehabilitation center, PSCT = polytrauma support clinic team, PTSD = posttraumatic stress disorder, SCI = spinal cord injury, TBI = traumatic brain injury, VA = Department of Veteran Affairs.
*Address all correspondence to Kevin T. Stroupe, PhD; Center for Management of Complex Chronic Care, Edward Hines Jr. VA Hospital, 5000 South 5th Ave, (151H) Bldg 1B260, Hines, IL 60141; 708-202-3557; fax: 708-202-2316. Email: kevin.stroupe@va.gov
http://dx.doi.org/10.1682/JRRD.2012.06.0107
INTRODUCTION

Traumatic brain injury (TBI) has emerged as the "signature wound" among U.S. troops since the conflicts in Afghanistan (Operation Enduring Freedom [OEF]) and Iraq (Operation Iraqi Freedom [OIF]) began in 2001 and 2003, respectively. Based on survey estimates, between approximately 15 and 20 percent of OIF/OEF Veterans may have had a TBI while deployed [1–3]. Mild TBI is a condition that can manifest with affective, somatic, and cognitive symptoms, including headaches; problems with sleep, balance, and/or memory; irritability; and sensitivity to light [4]. Although these symptoms usually resolve in a matter of hours, weeks, or months, in some cases they may persist for prolonged durations. To identify Veterans who may benefit from treatment and services, the U.S. Department of Veterans Affairs (VA) implemented a national clinical reminder (CR) in April 2007 to screen for TBI.

Veterans who screen positive are referred for a comprehensive TBI evaluation, and a diagnosis is made after completion of this TBI evaluation. Prior to implementation of the TBI screen in 2007, the VA estimated that in 2006, total annual costs of treating all VA users averaged $5,765 per patient [5]. For the VA to ensure that adequate resources are available for OIF/OEF Veterans, it is important to understand the healthcare utilization and costs of care for these Veterans following the TBI screen and evaluation. This article reports findings from a study of the annual healthcare utilization and costs for Veterans who complete the VA's TBI screen. Understanding these healthcare use and cost patterns will assist policymakers to address the ongoing and future healthcare needs of returning Veterans.

BACKGROUND

The TBI screen includes a series of questions to confirm OIF/OEF deployment and determine whether the Veteran has already been diagnosed with TBI during an OIF/OEF deployment [6]. For Veterans who confirm OIF/OEF deployment and do not have a prior diagnosis of TBI, the CR screening instrument proceeds using four sequential sets of questions [6]. If a Veteran responds positively to one or more possible answers in a section, he or she will proceed to the next section of the screening instrument; otherwise, the screen is negative and the screening process is complete [6].

The four sections of the screening instrument have questions about whether the Veteran had (1) exposure to events during deployment that may increase the risk of TBI (e.g., blast or explosion of improvised explosive device, rocket-propelled grenade, land mine, grenade, blow to the head), (2) symptoms immediately after the event (e.g., being dazed, confused, "seeing stars"), (3) new or worsening symptoms after the event (e.g., sensitivity to light, headaches), and (4) current symptoms (e.g., sensitivity to light, headaches). To screen positive, a Veteran must have an affirmative response to at least one question in each of the four sections of the screening instrument [6]. Consequently, a Veteran with a history of TBI but who does not have persistent symptoms would screen negative.

Since Veterans may respond positively to questions in the TBI screen because of the presence of symptoms related to other conditions, a positive screen does not indicate a definitive diagnosis of TBI. Most TBIs sustained are mild in nature; however, the screening is applied to everyone. As a result, this study mainly captures costs associated with those who potentially have mild TBI; however, it may also include Veterans with more severe TBI.

Accurate interpretation of data related to analysis of the TBI screening and evaluation program requires a thorough understanding of the development of documentation tools that capture the related clinical processes and subsequent limitations in the data. Data elements from the initial TBI screen were created at the time of process implementation in April 2007 and provide the means to identify the overall cohort of patients with a positive TBI screen. However, a Web-based template to assist providers with the comprehensive TBI evaluation did not become available until October 2007 [7], resulting in a 6 mo window during which TBI evaluation data could not be captured. While facilities were encouraged to use the standardized template for the comprehensive TBI evaluation, there was no automated means through which providers were required to enter TBI evaluation data prior to implementation of the Web-based template. Also, facilities may partner with TBI experts in the private sector to complete the TBI evaluation, and these non-VA providers would not have access to the standardized evaluation template in the VA computerized patient record system. These factors preclude using the TBI screening and evaluation data to totally assess completion rates for the overall population of patients with a positive TBI screen. Any analysis referencing the TBI evaluation data can only accurately detail results based on the cohort of patients for which the TBI evaluation data are available.

METHODS
Study Design

The study included a national retrospective sample of OIF/OEF Veterans who received care at VA facilities in the United States between April 2007 (initiation of the mandatory TBI screen) and September 30, 2008. We examined the association between results of the TBI screen (positive, negative, or no screen) and patient characteristics, facility characteristics, healthcare utilization, and costs over a 12 mo period following an "index date." For Veterans who received the TBI screen, the index date was the date of the TBI screen. For Veterans who did not receive the TBI screen (e.g., due to not showing up for an appointment, refusal to take the screening), the index date was the date of their first VA healthcare utilization that occurred following separation from the military and after April 14, 2007.

Among Veterans who screened positive, we also examined the association of the TBI evaluation results with healthcare utilization and costs during the 12 mo following the index date. As noted previously, VA data systems may have not fully captured information on TBI evaluations. Therefore, the TBI evaluation results were only available for a subset of our sample of Veterans who screened positive for TBI.

National Population Study Sample

Veterans returning from OIF/OEF were included in this study if (1) they were a member of the VA OIF/OEF roster (described in the "Data Sources and Measures" section); (2) their military service separation date was after September 11, 2001, and before September 30, 2008; (3) they had a VA inpatient or outpatient visit between April 14, 2007, and September 30, 2008; (4) they indicated "yes" on the TBI screen that they had been deployed in Afghanistan and/or Iraq; and (5) they did not have a previous diagnosis of TBI. The final sample derived from this national population included 164,438 OIF/OEF Veterans.

Data Sources and Measures

This study utilized several national VA databases to obtain data about Veterans and their healthcare utilization and costs (Table 1). The VA's national OIF/OEF roster contains information on Veterans separated from OIF/OEF military service who have enrolled in VA healthcare. This roster is derived from the VA Health Eligibility Center enrollment file and the U.S. Department of Defense Defense Manpower Data Center database [8]. The OIF/OEF roster was used to identify all Veterans in our sample and to abstract their demographic data, including sex, race, ethnicity, marital status, branch of military service, and education.


Table 1. 

Results of each Veteran's TBI screen were abstracted from the VA National TBI Health Factors database, which is managed by the VA Office of Patient Care Services and derived from the VA's electronic health record. The elements of the TBI screen abstracted include the responses to the 4-item question set in the TBI screen, date of the TBI screen, and date of separation from the military. These abstracted data were used to create the index date for each Veteran as well as a variable indicating the number of days between separation from the military and the index date. The TBI evaluation results for each Veteran were abstracted from the Comprehensive TBI Evaluation database, which is also managed by the VA Office of Patient Care Services.

Additional clinical and demographic data for each Veteran were abstracted from the VA Medical SAS Inpatient and Outpatient data sets [9–10], including date of birth, most frequently occurring ZIP code of residence, sex, race, marital status, and comorbidities. Comorbid conditions were determined from diagnoses in VA medical SAS data sets during the 12 mo period prior to the index date. History of depression was based on diagnostic codes from fiscal years (i.e., October 1 to September 30) 1999 through 2008. Travel time in minutes to nearest VA facility was calculated using ZIP code and geographic information system software (ArcGIS version 9.3, Esri; Redlands, California).

The type of facility where Veterans received care on their index date was categorized by the polytrauma and TBI care provided at the facility and by facility complexity. Polytrauma is defined as two or more injuries sustained in the same incident that affect multiple body parts or organ systems and result in physical, cognitive, psychological, or psychosocial impairments and functional disabilities. To meet the rehabilitation needs of patients with varying diagnoses and complexities, VA has four categories of facilities to provide polytrauma care. Polytrauma rehabilitation centers (PRCs) provide acute comprehensive medical and rehabilitation care for patients with complex and severe injuries. There are currently five PRCs, which serve as hubs for research and education related to polytrauma and TBI. There are 23 polytrauma network sites (PNSs) that have interdisciplinary treatment teams that manage the postacute sequelae of polytrauma and TBI and provide coordination for lifelong rehabilitation needs. Polytrauma support clinic teams (PSCTs) are established at 85 sites and have local rehabilitation teams who provide follow-up specialty care in consultation with regional and network specialists. The remaining 40 sites are polytrauma points of contact (POCs). These facilities refer Veterans for care according to their medical needs [11]. VA facilities were also categorized by complexity using VA complexity model classifications. The model uses several factors, including the total number of patients treated at the facility, the number and types of intensive care units, and the number of physician specialists used by the facility. This approach of grouping facilities was based on work in the private sector [12–13], which was revised to maximize relevance specifically to VA facilities [14].

Veteran Groups by Traumatic Brain Injury Screen and Traumatic Brain Injury Evaluation Results

All Veterans were categorized into TBI screening groups by whether they had no TBI screen, screened negative, or screened positive. Among those Veterans who screened positive and received the TBI evaluation, we categorized them into TBI evaluation results groups by whether the TBI evaluation results were positive or negative for TBI.

Measures of Healthcare Utilization and Cost

Healthcare utilization data for the 12 mo following the index date were obtained from the VA Decision Support System (DSS) National Data Extracts (NDEs) [15]. Outpatient utilization was categorized as primary care, rehabilitation care, polytrauma care, mental health care, other specialty care, and other outpatient care (e.g., ancillary care and home care) based on clinic codes in the VA DSS NDEs. We also computed the number of outpatient encounters and costs that Veterans had for each of these categories of care. Additionally, we assessed the costs on the day of the index visit and the costs on the day of the TBI evaluation. Because Veterans may visit more than one clinic while they are at a VA facility, Veterans may have had more than one encounter per facility visit. Outpatient pharmacy use was categorized as chronic medications, for which a Veteran received more than one 30 d supply, and acute medications, for which a Veteran received no more than one 30 d supply. Inpatient utilization included the proportion of Veterans who had hospital admissions for short-term medical or surgical, spinal cord injury (SCI), psychiatric, rehabilitation, and long-term care based on the care unit (i.e., VA bed section) and the proportion of patients with non-VA hospital care financed by VA from the VA Fee Basis databases [16]. We also determined the most frequently occurring inpatient admitting diagnoses for these Veterans in the VA Medical SAS Inpatient data sets.

We examined the direct costs of healthcare from the VA's (i.e., the payer/provider) perspective, where cost estimates reflect the VA's expenditures for each Veteran. Costs for outpatient care, outpatient pharmacy, and inpatient care provided by a VA facility were obtained from VA DSS NDEs [15]. The DSS extracts information from the VA's accounting and payroll system and combines it with workload information from patient care and administrative departments to produce cost estimates [15]. These databases contain estimates of personnel costs, including physicians, nurses, technicians, and other staff, as well as costs of supplies and other administrative and overhead expenses of inpatient stays and outpatient encounters. Pharmacy costs in DSS NDEs include the purchase price of the medication as well as dispensing and administrative and overhead costs [17]. Costs of non-VA inpatient hospitalizations financed by VA were obtained from the VA Fee Basis databases [16].

In summary, we examined total costs per patient, which consisted of total outpatient (primary care, rehabilitation care, polytrauma care, mental health care, other specialty care, other VA outpatient, and non-VA outpatient costs), total outpatient pharmacy (chronic and acute medication costs), and total inpatient (short-term medical or surgical, SCI, psychiatric, rehabilitation, long-term care, and non-VA Fee Basis care) costs during the 12 mo following the index date. Total outpatient costs also included costs on the day of the index visit and costs on the day of the TBI evaluation. All costs were adjusted to 2008 dollars using the Consumer Price Index.

Analysis

Analyses were conducted using SAS version 9.1 (SAS Institute Inc; Cary, North Carolina) and STATA SE version 11.0 (StataCorp LP; College Station, Texas). Veteran and facility characteristics, prior to the index date, were compared between the three TBI screening groups and the two TBI evaluation results groups using chi-square tests.

Healthcare utilization and costs were compared between the three TBI screening groups and the two TBI evaluation results groups using analysis of variance and chi-square tests. We also reported the 10 most frequently occurring inpatient admitting diagnoses for Veterans in the three TBI screening groups and the two TBI evaluation results groups.

To investigate the association of the TBI screening groups and TBI evaluation groups with healthcare utilization and costs, we used multivariable regression analyses, controlling for Veterans– demographic, clinical, and other factors previously described. Specifically, we used multivariable logistic regression to examine the probability of hospital admission during the 12 mo period following the index date. Because the numbers of outpatient visits in the 12 mo period after the index date were nonnegative integers, we used negative binomial count models for our multivariable analyses of outpatient visits. The negative binomial model is a count data model that allows for overdispersion where the conditional variance exceeds the conditional mean. We used hierarchical logistic and negative binomial models to adjust for the correlation of patients within VA facility.

To examine the association of the three TBI screening groups and two TBI evaluation results groups with total direct healthcare costs after controlling for other factors, we used generalized linear models (GLMs) [18]. The GLM includes a distribution function that describes the expected distribution of the costs and a link function that describes the scale on which the variables in the model are related to costs [19]. We used gamma distribution with a log link based on results from a modified Park test and a Box-Cox test [20]. The GLM analyses estimated robust (i.e., sandwich) standard errors to accommodate nonindependence within VA facilities.

RESULTS
Sample Description

Of the 164,438 Veterans meeting inclusion criteria for this study, 14,266 (9%) had no TBI screen, 118,545 (72%) screened negative, and 31,627 (19%) screened positive. Among those who screened positive and received the TBI evaluation, 6,045 (46%) tested positive and 7,231 (54%) tested negative. As noted previously, because the TBI evaluation information was not fully captured in VA data systems, we do not have any information about the TBI evaluation for some Veterans in our sample who screened positive on the TBI screen.

Table 2 shows Veteran and facility characteristics by TBI screening groups and TBI evaluation results groups. Among the differences between Veterans who were screened versus were not screened, Veterans who were not screened were less likely to be white, were more likely to be of Hispanic ethnicity, were less likely to be married, were less likely to be ≥35 yr old, were less likely to have a service-connected condition, and had a longer duration from military separation date to index date. Moreover, among the differences between Veterans who screened negative versus screened positive, Veterans who screened negative were less likely to be male, more likely to be black, more likely to be ≥35 yr old, less likely to have a service-connected condition, and had a longer duration from military separation date to index date.


Table 2. 
No TBI Screen
(
n = 14, 266)
TBI Screen (n = 150,172)
p-Value*
Negative Screen
(
n = 118,545)
86.9†‡
85.6–?**
94.1††
????**
12.3† ‡‡
43.7†‡ ‡‡
35
33.8†‡ ‡‡
35.5–?**
41.2††
5.8
5.6–?**
5.5†‡
5.0–?**
5.4†‡
4.4–?**
0.9
0.9??
0.6??
15.4†‡ ‡‡
13.2–?**
26.7††
11.6‡ ‡‡
11.3–?**
60
725.5 ± 454.6†‡ ‡‡
605.9 ± 487.8–?**
23.5††
50.1‡ ‡‡
27.7† ‡‡
67.6†‡ ‡‡
65.6??**
64.4††
1.8??**
0.6††
*p-value is for comparison across five groups.
p < 0.05 for comparison between No TBI Screen (n = 14,266) vs Negative Screen (n = 118,545).
p < 0.05 for comparison between No TBI Screen (n = 14,266) vs Positive Screen (n = 31,627).
??p < 0.05 for comparison between Negative Screen (n = 118,545) vs Positive Screen (n = 31,627).
??p < 0.05 for comparison between Negative Screen (n = 118,545) vs Negative TBI Evaluation (n = 7,231).
**p < 0.05 for comparison between Negative Screen (n = 118,545) vs Positive TBI Evaluation (n = 6,045).
††p < 0.05 for comparison between Negative TBI Evaluation (n = 7,231) vs Positive TBI Evaluation (n = 6,045).
‡‡p < 0.05 for comparison between No TBI Screen (n = 14,266) vs TBI Screen (n = 150,172).
Traumatic Brain Injury Screening and Healthcare Utilization and Costs

This section presents comparisons of the characteristics and outcomes among the TBI screening groups: the 14,266 who were not screened, the 118,545 who screened negative, and the 31,627 who screened positive. Veterans who screened positive on the TBI screen had more healthcare utilization during the 12 mo period following the index date than patients who screened negative or with no TBI screen (Table 3). Veterans who screened positive averaged 3.3 primary care visits compared with 2.6 visits for Veterans who screened negative and 1.6 visits for Veterans with no TBI screen (p < 0.001). Moreover, Veterans who screened positive, negative, or had no TBI screening averaged 1.8, 0.8, and 0.4 rehabilitation care visits (p < 0.001); 1.6, 0.1, and 0.1 polytrauma care visits (p < 0.001); 7.2, 3.3, and 1.9 mental health visits (p < 0.001); and 3.4, 2.5, and 1.6 specialty care visits (p < 0.001), respectively (Figure 1(a)). A similar relationship existed for acute and mental health inpatient care. The most common inpatient admitting diagnosis among all three TBI screening groups was posttraumatic stress disorder (PTSD), accounting for 32.7, 17.7, and 11.4 percent of all admissions among Veterans who screened positive, screened negative, or had no TBI screening, respectively, followed by alcohol dependence accounting for 4.2, 3.9, and 3.5 percent of admissions, respectively (Table 4). Moreover, Veterans who screened positive received more medications from VA pharmacies than Veterans who screened negative or had no screening.


Table 3. 
No TBI Screen
(n = 14,266)
p-Value*
Negative Screen
(n = 118,545)
Positive Screen
(n = 31,627)
1.6†‡??
2.6??
0.4†‡??
0.8??
0.1†??
0.1??
1.9†‡??
3.3??
1.6†‡??
2.5??
5.7†‡??
9.9??
0.5†‡??
0.8??
0.6†‡??
0.9??
7.0†‡??
2.0†‡??
2.5??
No TBI Screen
(n = 14,266)
p-Value*
Negative Screen
(n = 118,545)
Positive Screen
(n = 31,627)
1.4†‡??
2.0??
0.2??
0.3??
0.2†‡??
0.5??
1.1??
1.1??
1.1‡??
1.0??
2,732†‡??
349†‡??
403??
0†‡??
19??
448†‡??
687??
71†‡??
129??
36†??
33??
446†‡??
782??
603†‡??
909??
616†‡??
980??
162†‡??
241??
217†‡??
330??
No TBI Screen
(n = 14,266)
p-Value*
Negative Screen
(n = 118,545)
Positive Screen
(n = 31,627)
450†‡??
671??
174†‡??
265??
6??
97†‡??
168??
6??
51†‡??
111??
75??
80??
3,399†‡??
*p-value is for comparison of means across three groups.
p < 0.05 for comparison between No TBI Screen (n = 14,266) vs TBI Screen (n = 150,172).
p < 0.05 for comparison between No TBI Screen (n = 14,266) vs Negative Screen (n = 118,545).
??p < 0.05 for comparison between No TBI Screen (n = 14,266) vs Positive Screen (n = 31,627).
??p < 0.05 for comparison between Negative Screen (n = 118,545) vs Positive Screen (n = 31,627).
**Number of chronic medications (i.e., medications with more than one 30 d supply) counts number of 30 d equivalent prescriptions filled by VA pharmacies (i.e., one 90 d prescription is counted as three 30 d equivalent prescriptions). Number of acute medications (i.e., medications with no more than one 30 d supply) counts number of prescriptions filled by VA pharmacies.
††Costs of index evaluation and TBI evaluation included all outpatient costs that occurred on day of those visits.

Figure 1. Healthcare utilization and cost 1 yr following index date by traumatic brain injury screening group. (a) Outpatient visits. (b) Healthcare costs.

Figure 1.

Healthcare utilization and cost 1 yr following index date by traumatic brain injury screening group. (a) Outpatient visits. (b) Healthcare costs.

Click Image to Enlarge. View as PowerPoint Slide


Table 4. 
No TBI
Screen
Negative
TBI Screen
Positive
TBI Screen
Negative TBI Evaluation
Positive TBI Evaluation
722.10: Displacement of thoracic or lumbar intervertebral disc without myelopathy

Mean total healthcare costs per patient during the 12 mo period following the index date were nearly double for Veterans who screened positive ($9,610) compared with the mean cost of Veterans who screened negative ($5,184) and nearly three times the costs of Veterans with no TBI screening ($3,399) (p < 0.001) (Table 3, Figure 1(b)). Outpatient costs comprised approximately 80 percent of total healthcare costs. Among Veterans who screened positive, screened negative, or had no TBI screening, mean costs per patient for the index evaluation were $449, $403, and $349, respectively (p < 0.001), and total outpatient costs per patient were $7,746, $4,182, and $2,732, respectively (p < 0.001). There was a similar relationship between the TBI screening categories and each category of outpatient care. Inpatient costs comprised approximately 14 percent of total costs: $1,353 for Veterans who screened positive, $671 for Veterans who screened negative, and $450 for Veterans with no TBI screening (p < 0.001).

After adjusting for patient and facility characteristics, there continued to be an association between TBI screening group and healthcare utilization and costs (Table 5). During the 12 mo period after the index date, there were 1.63 times more primary care, 1.67 times more rehabilitation, 1.53 times more polytrauma, 1.49 times more mental health, and 1.54 times more other specialty care outpatient visits for Veterans who screened negative than for Veterans with no TBI screening. Over this period, total healthcare costs from GLM analysis were $5,576 higher for Veterans who screened positive than for Veterans with no TBI screening after adjusting for the effects of the other factors in the model. Healthcare utilization and costs were also higher for Veterans who screened positive than for Veterans with no TBI screening controlling for other factors. It should be noted that the estimates from the multivariable analyses do not represent the differences seen in the typical Veteran but instead represent what the differences would be if all of these other factors in the model are held constant. The adjusted results isolate the effects of various factors, while the unadjusted results might be a better indicator of the typical effect seen across the overall population of VA users. The significant association of TBI screening results groups with utilization and costs in the multivariable analyses highlights that there remain unmeasured differences between these groups that are related to these outcomes.


Table 5. 
Primary
Care IRR
Polytrauma
Care IRR
1.33
1.63
1.67
1.53
1.49
1.54
2.21
2.04
3.93
2.87
2.14
0.78
0.73
1.03
1.18
0.94
0.81
906
1.12
0.92
1.13
0.88
0.94
0.91
1.03
456??
0.48
0.88
0.81
0.89
0.84
0.91
1,728
1.11
1.06
1.06
1.14
16
0.83
0.95
1.07
433
1.04
1.07
1.04
1.06
1.10
1.09
1.19
1.11
1.10
1.21
516
35
0.81
1.21
1.49
1.14
1.62
968
0.83
0.96
0.94
339
0.60
0.89
0.93
0.82
0.93
1,063
0.59
0.87
0.74
1,097
1.21
1.11
1.31
785
1.02
1.20
1.03
509
3.20
1.07
1.32
0.93
1.26
1.16
1.07
1.43
3.22
1.22
1.87
1.41
1.08
1.16
1.09
1.74
1.10
1.59
1.12
1.07
1.50
1.11