Up to 40% of veterans of the conflicts in Afghanistan (Operation Enduring Freedom [OEF]) and Iraq (Operation Iraqi Freedom [OIF]) who are enrolled in the Department of Veterans Affairs (VA) health care system screen positive for alcohol misuse (
1,
2), which covers the spectrum from risky drinking to alcohol use disorder (
3). Alcohol misuse affects health through direct harms on acute and chronic conditions and associations with self-care behaviors for conditions such as traumatic brain injury (TBI) (
4–
6).
Approximately 10%−15% of OEF/OIF veterans have sustained at least one TBI during their deployments (
7,
8), a significant percentage of whom report persisting postconcussive symptoms, even years postinjury (
9,
10). Most OEF/OIF patients with TBI have a comorbid psychiatric diagnosis (
11) or chronic pain (
9,
12). Such conditions may increase risk of alcohol misuse, although prior research yielded inconsistent findings on the relationship between TBI and alcohol misuse. Two studies have found higher rates of alcohol misuse among OEF/OIF military personnel with TBI compared with those without TBI (
13,
14). Other research has found no association between postdeployment alcohol misuse and TBI after adjustment for posttraumatic stress disorder (PTSD) symptoms (
15,
16).
Alcohol use is associated with numerous negative outcomes among patients with TBI and, even in limited quantities, may impair recovery from TBI (
7,
17,
18). TBI and alcohol misuse are each individually associated with changes in brain structure and reduced cognitive abilities (
19,
20). Alcohol misuse, including heavy episodic drinking (six or more drinks on one occasion) (
21), among younger veterans with TBI is concerning because heavy episodic drinking may adversely affect neurocognitive development (
22). Further, alcohol misuse can increase the risk of recurrent TBI (
23,
24). Emerging evidence suggests that TBI is associated with greater risk of developing an alcohol use disorder (
15,
16). These findings suggest that early detection and intervention targeting alcohol misuse among veterans with TBI may promote postinjury recovery and prevention of chronic alcohol problems.
Alcohol screening and brief intervention (BI) are recommended by the U.S. Preventive Services Task Force (
25) and deemed the third highest prevention priority for adults (
26). BIs as brief as five to 15 minutes consisting of advice to drink within recommended limits reduce alcohol use (
27). The VA implemented alcohol screening in 2003 and has since required annual administration of the Alcohol Use Disorders Identification Test alcohol consumption questions (AUDIT-C) to screen outpatients. In 2008, the VA implemented performance incentives and disseminated an electronic reminder to prompt providers to offer BI for outpatients who screen positive for alcohol misuse with AUDIT-C scores ≥5 (
28). Further, the
VA/DoD Clinical Practice Guideline for Management of Substance Use Disorders recommends that clinicians offer treatment referral to patients who have a diagnosis of or are at higher risk of (AUDIT-C score ≥8) an alcohol use disorder (
29). Among OEF/OIF veterans with alcohol misuse identified by screening, 70%−80% had BI and 20%−40% were referred for addiction treatment (
1,
2). However, rates of subsequent attendance at addiction specialty care are unknown.
Rates of alcohol misuse and TBI are high among OEF/OIF veterans, and in combination, these conditions are associated with poorer medical, neurobehavioral, and life satisfaction outcomes (
30). Although alcohol screening and BI for patients with TBI may reduce trauma recidivism and alcohol-related problems, results from a recent study among trauma centers suggest diminished effects among patients with TBI compared with those without TBI (
31). To our knowledge, no study has examined the prevalence of alcohol misuse and documented follow-up care for alcohol misuse—BI and addiction treatment—among OEF/OIF VA patients with and without TBI. This study used national VA health care utilization data to estimate the gender-specific prevalence of alcohol misuse and heavy episodic drinking among OEF/OIF VA patients with and without TBI. Among OEF/OIF VA patients with alcohol misuse, the study also estimated and compared the gender-specific prevalence of documented BI and addiction treatment attendance for those with and without TBI. This information could help identify gaps in the delivery of care for this high-risk population.
Methods
Data Source and Study Sample
This study used data from the VA Corporate Data Warehouse (CDW), a national data repository that includes clinical and administrative data for VA patients seen at all VA facilities, including patient-level data on demographic factors, clinical diagnoses, and alcohol screening and BI documented in the electronic medical record (EMR). Study approval was obtained from the VA Puget Sound Health Care System Institutional Review Board.
Veterans ages 18 and older were eligible for study inclusion if they received alcohol screening with the AUDIT-C questionnaire in 2012, had documented OEF/OIF service determined by an OEF/OIF patient record flag available through the VA Decision Support System, and used outpatient or inpatient VA services (one or more visits) in the year before the date of the AUDIT-C.
Outcome Measures
Alcohol screening.
Alcohol use was defined by scores on the AUDIT-C, a validated three-item screen for alcohol misuse (
32–
35). AUDIT-C items assess the frequency and quantity of typical drinking and the frequency of heavy episodic drinking. [A list of AUDIT-C items is included in an
online supplement to this article.] AUDIT-C scores range from 0 to 12, with higher scores indicating greater severity (
36,
37). VA providers are prompted by an electronic reminder in the EMR to administer the AUDIT-C to all patients annually. VA provides incentives for administering BI for patients with AUDIT-C scores ≥5 to minimize the clinical burden of providing BI for false-positive screens. Therefore, AUDIT-C scores ≥5 were used as our dichotomous measure of any alcohol misuse.
To facilitate interpretation of AUDIT-C scores, the scores were categorized into four risk groups (
37): nondrinkers (score of 0), low-level drinkers (score of 1–4), moderate alcohol misuse (score of 5–7), and severe alcohol misuse (score of 8–12). AUDIT-C item 3 was used to assess heavy episodic drinking, defined as reporting six or more drinks on a single occasion at least monthly.
Documented BI.
Since 2008, documentation of BI, consisting of alcohol-related advice and feedback on health consequences of drinking, has been facilitated by a national electronic reminder in VA’s EMR and is expected for outpatients who screen positive for alcohol misuse (AUDIT-C score ≥5) (
28,
38). We identified BI documented with the electronic reminder by using text data elements called “health factors” (HFs). HFs are generated when an electronic reminder is used to document VA care and can be extracted from the CDW. Although the BI electronic reminder is mandated nationally, facilities could edit HFs locally. We identified HFs reflecting advice or counseling regarding alcohol use in the 30 days before or after a positive screen and combined them into a single dichotomous measure of documented BI. Because the BI electronic reminder typically requires documentation of both advice and feedback, this measure was expected to be a good proxy for BI (
38). A period before the alcohol screen date was included to accommodate various alcohol screening strategies employed in VA clinics. For example, some clinics administer screening questionnaires in a paper-and-pencil format during patient visits and enter questionnaire data into the medical record at a later date.
Documented addiction treatment attendance.
Data on addiction treatment attendance in the VA was obtained by using outpatient clinic stop codes and defined as documented attendance at one or more visits in a VA addiction treatment clinic during the 90 days following the date of a positive screen. This measure did not include follow-up care for alcohol misuse in non-VA facilities.
Predictors and Covariates
TBI diagnostic status.
Patients were classified as having TBI if an appropriate
ICD-9-CM diagnosis was documented in the EMR between 365 days before and 30 days after the alcohol screen date. These
ICD-9-CM codes (310.2, 800.00–800.99, 801.00–801.99, 803.00–803.99, 804.00–804.99, 850.00–850.99, 851.00–854.19, 905.0, 907.0, 950.1–950.3, 959.01, 959.09, and V15.52) are used by the VA for TBI surveillance (
39,
40). TBIs that were diagnosed outside the VA and not documented in the VA EMR were not captured in this study. We included a 30-day window following the alcohol screen date to ensure inclusion of patients who screened positive on the TBI electronic reminder at the time of alcohol screening but needed to complete the VA-mandated TBI comprehensive evaluation to determine whether they met diagnostic criteria for TBI (
41).
Patient characteristics and medical facility.
Covariates associated with the prevalence of alcohol misuse and TBI (
42,
43) and available from the CDW were selected. Sociodemographic factors included age, gender, marital status, race, ethnicity, and VA disability compensation. Clinical characteristics included psychiatric and substance use disorders (current and in remission), tobacco use disorder, and pain diagnoses based on
ICD-9-CM codes documented between 365 days before and 30 days after the alcohol screen date. A variable noting the VA facility where screening occurred (of 138 medical facilities nationwide) was also included.
Data Analyses
All analyses were performed separately for men and women, because gender is associated with alcohol misuse. Descriptive statistics were generated for demographic and clinical characteristics of OEF/OIF patients with and without TBI. Unadjusted rates of alcohol use (AUDIT-C categories) and heavy episodic drinking were estimated among patients with and without TBI. Logistic regression was used to estimate age-specific (<30, 30–39, and ≥40) prevalence (95% confidence intervals [CIs]) of alcohol misuse (AUDIT-C score of ≥5), severe alcohol misuse (AUDIT-C score of ≥8), and heavy episodic drinking among patients with and without TBI. Unadjusted and adjusted logistic regression models including patients who screened positive for any alcohol misuse (AUDIT-C ≥5) and severe misuse (AUDIT-C ≥8) were used to compare documented receipt of BI and addiction treatment for patients with and without TBI. Covariates included in adjusted models were age, race, ethnicity, marital status, VA disability compensation, substance use and psychiatric disorders, and pain diagnosis. Medical facility was included as a grouping variable to account for intragroup correlation at the facility level.
Results
Patient Characteristics
We identified 358,147 OEF/OIF VA patients (13.2% were women) eligible for this study, and of those 8.4% (N=30,197) had a documented TBI diagnosis (3.9% of women and 9.1% of men). Most women and men with TBI were white and non-Hispanic (
Table 1). The most common psychiatric disorders among women and men with TBI were PTSD and depression, and over 80% of women and men with TBI had a pain diagnosis.
Compared with patients without TBI, those with TBI were younger (age <30) and more likely to have ≥50% VA disability and psychiatric disorders and pain diagnoses (p<.01 for all). Women and men with TBI were approximately twice as likely as those without TBI to have an alcohol use disorder (p<.01) or a drug use disorder (p<.01). Most patients screened negative for alcohol misuse (AUDIT-C ≤5), regardless of TBI status. Rates of abstinence (AUDIT-C score of 0) were higher among women and men with TBI compared with those without TBI (p<.01).
Prevalence of Alcohol Misuse and Heavy Episodic Drinking
The overall prevalence of alcohol misuse (AUDIT-C ≥5) among women with TBI was 6.8% (CI=5.8–8.1) compared with 5.6% (CI=5.4–5.8) among women without TBI (p<.05). Among men, rates of alcohol misuse for those with and without TBI were 20.3% (CI=19.9–20.8) and 16.4% (CI=16.3–16.6), respectively (p<.001).
Age-specific estimates of alcohol misuse among women did not differ between those with and without TBI in any age group (
Table 2). Estimates of heavy episodic drinking were significantly higher only among women ages 30–39 with TBI compared with those without TBI (p<.05). Among men, age-specific estimates of alcohol misuse and heavy episodic drinking were greater among those with TBI compared with those without TBI in each age group (p<.001). The prevalence of heavy episodic drinking and alcohol misuse was highest among both women and men younger than 30 with TBI.
Prevalence of Documented BI and Addiction Treatment
Table 3 presents unadjusted rates of documented BI and addiction treatment attendance. Among women with moderate and severe alcohol misuse, the unadjusted prevalence of BI was between 73% and 80% for those with and without TBI and did not significantly differ between groups. Similarly, in adjusted analyses (data not shown), rates of BI did not significantly differ between women with and without TBI. Among men with moderate and severe alcohol misuse, the unadjusted prevalence of BI was between 78% and 80% for those with and without TBI and did not significantly differ between groups. Consistent with findings for women, adjusted rates of BI for men with TBI and men without TBI did not significantly differ (data not shown).
Among both women and men, the highest rate of addiction treatment attendance was for those with TBI who screened positive for severe alcohol misuse; 36.6% of women and 20.0% of men had documented addiction treatment. Adjusted estimates of addiction treatment attendance were similar to unadjusted estimates, with higher rates of attendance observed for men and women with TBI compared with men (p<.001) and women (p<.05) without TBI (data not shown).
Discussion
This study used national VA health care utilization data to compare the prevalence of alcohol misuse, documented BI, and VA addiction treatment among OEF/OIF veterans with and without TBI. Our results indicate that alcohol misuse was common among OEF/OIF veterans with TBI, especially men, with 20% of men and 7% of women with TBI screening positive for alcohol misuse. The risk of screening positive for alcohol misuse was highest among younger men and women (age <30). The prevalence of documented BI for OEF/OIF men and women who screened positive was high (70%−80%), regardless of TBI status. However, for men and women who screened positive, rates of documented VA addiction treatment were higher among TBI patients compared with those without TBI. Among those with severe alcohol misuse and TBI, approximately one in five men and one in three women attended one or more addiction treatment visits in the 90 days after their alcohol screen.
Consistent with prior studies of OEF/OIF veterans (
11,
44), this study found disproportionately high rates of concomitant pain and psychiatric and substance use disorders among men and women with TBI compared with those without TBI. Because most TBI patients had a PTSD or pain diagnosis, the higher estimates of VA addiction treatment for those with TBI may reflect more severe psychiatric comorbidity and greater use of VA health care services more generally. Given the small number of OEF/OIF women with both TBI and alcohol misuse, results for this group should be interpreted with caution. Nonetheless, this is the first study to report rates of VA addiction treatment attendance after alcohol screening among OEF/OIF veterans with and without TBI. Future research should assess these patients’ engagement in VA addiction treatment through documentation of subsequent visits over a longer period.
Early recognition and management of alcohol misuse among young veterans with TBI are critical given research suggesting that alcohol use may impair neurologic recovery and magnify cognitive deficits among TBI patients (
42). Although this study showed that alcohol use was common among OEF/OIF patients with TBI, it is important to note that 41% of women and 31% of men with TBI were abstinent from alcohol in the prior year. Although alcohol use may initially decline after injury, many people resume drinking at later periods (
4,
45). Examining changes in drinking in the context of TBI is beyond the scope of this study; however, the notably high rates of abstinence are encouraging. Studies are needed to evaluate how changes in drinking, including resolution of alcohol misuse, affect veterans’ recovery from TBI.
Although it is critical that veterans with TBI and alcohol misuse receive timely assessment and symptom management, their treatment needs are poorly understood. It is unclear whether alcohol screening and BI methods need to be adapted to the cognitive deficits associated with TBI (
46,
47). One study found that the effectiveness of an alcohol BI to reduce hazardous drinking among trauma patients was significantly lower among patients with TBI than among those without TBI (
31). Because previous evaluations of alcohol screening and BI have systematically neglected patients with more severe TBI symptoms, further research is needed to assess TBI patients’ responsiveness to alcohol BI (
48). Prospective studies are also needed to better understand short- and long-term health outcomes among patients with alcohol misuse and TBI and psychiatric comorbidity, including suicide risk. In the general population, the risk of death by suicide for individuals with both TBI and alcohol misuse has been reported as being four times higher than among individuals without these conditions (
49,
50). Information on TBI and alcohol misuse and their association with psychiatric comorbidity and suicide risk among OEF/OIF women and men could inform delivery of targeted screening and intervention strategies for this important group of veterans.
This study had some limitations. The study used alcohol data for screening and follow-up care for alcohol misuse documented in VA patients’ EMR. Prior research suggests that patients report higher rates of alcohol misuse on confidential surveys (
9), and thus our results may underestimate the prevalence of alcohol misuse. Results of our analyses relied on documentation of BI without any assessment of the fidelity or quality of the intervention. Because medical facilities could modify HFs, we may not have identified all BI-related HFs; furthermore, some BIs may have been documented without HFs. In addition, this study relied on VA data, and thus any screening and follow-up interventions that occurred in non-VA care were not considered in the analyses. Nevertheless, this study describes alcohol screening and BI documented in patients’ EMRs and available to clinicians, and we expect most BIs were documented with electronic reminders that generate HFs to ensure that medical facilities met performance measure requirements. Our analyses did not take into account either the time between TBI and alcohol screening or levels of TBI severity and their association with alcohol misuse. Because this study only captured
ICD-9 TBI diagnoses documented in the VA EMR within one year of the alcohol screen date, our results may have underidentified the prevalence of TBI.
Conclusions
Findings indicate that alcohol misuse and psychiatric problems were prevalent among OEF/OIF VA patients with TBI. The risk of alcohol misuse was found to be particularly high among younger women and men. It is encouraging that most patients who screened positive for alcohol misuse had documented BI, regardless of a TBI diagnosis. A significant minority of patients who screened positive did not have BI, and most patients with severe alcohol misuse did not attend VA addiction treatment in the 90 days following screening. Timely recognition and management of alcohol misuse among OEF/OIF veterans with TBI are critical to support reintegration and recovery and reduce the health burden of these conditions.