The Veterans’ Affairs medical system is the largest integrated provider of HIV care in the US. Up to 63% of HIV-infected veterans have at least one mental health diagnosis [1–3]. The co-occurrence of mental illness and HIV results in worse outcomes and higher healthcare costs .
The impact of depression among HIV-infected persons is well documented, affecting medication adherence , CD4 cell count , innate immune function , and mortality [8–11]. However, data regarding other psychiatric diagnoses are limited due to the low prevalence of some mental health disorders, under-recognition by HIV providers  and limited access to mental healthcare [13,14]. With combination antiretroviral therapy (cART) long-term HIV outcomes continue to improve . However, treatment outcomes in the setting of co-occurring mental illness are not well described.
For these analyses, the Veterans’ Affairs HIV Clinical Case Registry (CCR) was utilized to investigate the impact of mental health diagnoses on clinical outcomes, including the incident AIDS-defining illness (ADI) and death. We hypothesized that clinical outcomes in HIV-infected veterans with specific mental health diagnoses would be significantly worse when compared with individuals without mental health diagnoses.
The HIV CCR automatically populates from electronic Veterans’ Affairs medical records [16,17]. The CCR captures data on demographic, clinical, healthcare utilization, ICD-9-CM diagnostic codes, pharmacy and laboratory data, and mortality. The registry is updated daily to include all HIV-infected veterans identified at the local facility level.
We derived our cohort from 27 574 veterans registered in the HIV CCR between 1/1/2000 and 12/31/2006. To establish that patients were regular users of Veterans’ Affairs healthcare, we limited our cohort to patients with at least two outpatient visits during the observation period. To focus on new recipients of cART, we required that patients have 120-day period without any antiretroviral medication prescription before the receipt of cART, and did not have an undetectable HIV RNA in the 120 days prior to treatment initiation. Patients who had the outcome of interest prior to cART initiation were excluded. After applying these criteria, a sample of 9003 HIV-infected veterans with complete data was available.
Race, sex, and age at baseline were defined from data provided by CCR. The CD4 cell count and HIV RNA recorded prior to and most proximal to cohort entry were considered baseline values. A plasma HIV RNA value below 400 copies/ml was considered undetectable and assigned a value of 399 copies/ml for computational purposes. To adjust for potential confounding effects from other major medical comorbidities we utilized the validated, modified Charlson-Romano comorbidity index, calculated at cART initiation [18,19].
New cART regimen was defined as a combination of at least three antiretrovirals including either a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor (NNRTI) and two nucleoside reverse transcriptase inhibitors (NRTIs) or three NRTIs filled through the Veterans’ Affairs pharmacy concurrently. Baseline medication regimen was defined according to pharmacy data received in the first quarter of follow-up.
Diagnoses of schizophrenia, bipolar disorder, depression, anxiety, and substance use disorders were determined using inpatient and outpatient encounters (Table 1). We used one inpatient ICD-9-CM code or two outpatient ICD-9-CM codes to define each disorder to enhance diagnostic validity .
Adherence to cART was assessed using pharmacy data and expressed as percentage of days covered (PDC): total number of days of antiretroviral medication possession divided by the total number of days of observation . Healthcare utilization was measured as the number of outpatient visits per quarter.
Primary outcomes were death and incident ADI . Patients were followed until the development of a primary outcome, discontinuation of cART, or the last recorded clinic visit, whichever occurred first.
Death during follow-up was identified by date of death recorded in the CCR. Incident ADI after cART initiation was identified using ICD-9-CM diagnostic codes (Table 2).
Bivariate analyses for baseline data were conducted using χ2 and t-tests for categorical and continuous variables, respectively. Time-to-event analyses were performed using Cox-proportional hazard models. To account for time with a mental health diagnosis, we modeled the mental health diagnosis variables as time-dependent. Thus proportional hazards were not assumed. Modeling time-dependent variables fixed all covariates as occurring before the outcome event (e.g. mortality). Separate adjusted models were computed for each mental health disorder.
Models were adjusted for age, race, baseline CD4, baseline physical comorbidity, PDC, and healthcare utilization.
All analyses were performed using SAS software version 9.2 (SAS Institute, Cary, North Carolina, USA), with α = 0.05, using two-tailed tests. Cox-proportional hazard models were computed using the PROC PHREG procedure.
The study was approved by the Institutional Review Board for the St. Louis Veterans’ Affairs Medical Center. Individual informed consent was not required due to the retrospective nature of these analyses and use of a de-identified dataset.
The cohort was predominantly male and African American (Table 3). Median follow-up was 3 years [interquartile range (IQR) 1.75–3.0]. Median age at initiation of cART was 47 (IQR 42–53) years. During the follow-up period, 930 deaths and 712 ADIs occurred. The most common ADIs were Pneumocystis jiroveci pneumonia, Candida esophagitis, and Mycobacterium avium complex infection.
Greater than two-thirds of veterans (69%) in this cohort had at least one mental health diagnosis. The most common mental health disorders were illicit drug use disorders (50%), depression (41%), alcohol use disorder (30%), anxiety disorders (18%), bipolar disorder (7%), and schizophrenia (6%). Among persons with mental health diagnoses, 53% had more than two mental health diagnoses.
In bivariate analyses, veterans with a mental health diagnosis were more likely to be female, have HCV co-infection, higher comorbidity index, more outpatient visits, and lower cART adherence rates (Table 3).
A substantial proportion (44%) of the HIV-infected veterans initiated cART at a CD4+ cell count below 350 cells/μl. Forty-three percentage of HIV-infected veterans were commenced on a cART regimen that included NNRTI, 22% on a boosted protease inhibitor, 18% on a triple NRTI, and 17% on unboosted protease inhibitor. At the end of the follow-up period, 43% remained on the initial regimen, 28% switched regimens, and 28% discontinued cART altogether. The presence of a mental health diagnosis predicted regimen switching or discontinuation (P < 0.0001).
Overall, the median PDC was 72% for the entire cohort, and was significantly lower for veterans with a mental health diagnosis (69 vs. 77%, P < 0.0001). Patients had a median of 4.5 visits per quarter in the entire cohort; 3.0 among those without mental health diagnosis vs. 5.7 among those with a mental health diagnosis (P < 0.0001). There were further differences when specific mental health diagnoses were considered: patients with anxiety disorders had 8.1 visits per quarter vs. 5.0 among those without. The majority of visits were general medical outpatient clinic visits across the cohort.
In unadjusted survival analyses, having a mental health diagnosis was associated with a 1.25 times higher hazard of all-cause mortality. This association remained significant after adjusting for age, race, baseline CD4 cell count and comorbidity, PDC, and healthcare utilization (Table 4). Specifically, veterans with schizophrenia, bipolar and substance use disorders had 1.40, 1.32, and 1.23 times higher hazard for death, respectively, compared to veterans without these diagnoses in the adjusted analysis.
Anxiety disorder was associated with a statistically significant protective effect on all-cause mortality in adjusted analysis: veterans with an anxiety disorder had 20% reduced hazard for death compared to veterans without this diagnosis.
In unadjusted survival analyses, the presence of a mental health diagnosis was associated with significantly higher hazard for ADI. Veterans with any mental health diagnosis during the follow-up period were 1.25 times more likely to develop an ADI compared to veterans without a mental health diagnosis. Among specific mental health diagnoses, only substance use disorder was significantly associated with ADI, with a hazard ratio of 1.19 in the adjusted analysis (Table 4).
In this study, HIV-infected veterans had a high prevalence (69%) of mental health diagnoses, substantially higher compared to estimates derived from the general population , and from a nonveteran HIV-infected population . Affective and substance use disorders were the most common categories. Our results confirm the higher prevalence of mental health diagnoses among HIV-infected veterans [2,25].
All-cause mortality was higher among HIV-infected veterans with schizophrenia, bipolar and substance use disorders. In the general population, persons with schizophrenia and bipolar disorder experience higher mortality due to both medical and nonmedical causes [26–29]. To our knowledge, this is the first study describing the increased risk of mortality associated with these disorders among HIV-infected persons.
Contrary to published studies from the early cART era [30–32], we found no increase in mortality associated with depression or anxiety disorders. Interestingly, comorbid anxiety disorders had protective effect on mortality. We found no effect of depression on incidence of AIDS-defining illness. These findings may be explained by a favorable influence of anxiety disorders on health behaviors observed even after the adjustment for healthcare utilization. We believe that depression and anxiety disorders had a different impact on outcomes due to factors unique to the Veterans’ Affairs healthcare system, including routine symptom monitoring, universal access to mental health services including collaborative care models, counseling, pharmacotherapy and comorbid substance use programs [6,33]. Nevertheless management of severe mental illnesses (SMI) remains a challenge requiring more complex care models.
AIDS-defining illnesses were more frequent among veterans with a mental health diagnosis, specifically those with substance use disorders likely due to suboptimal medication adherence demonstrated in this study, and commonly reported in this patient category [31,34].
This study was limited by reliance on ICD-9-CM codes. We optimized the accuracy of coding as others have, by requiring one inpatient or multiple outpatient codes . Veterans’ Affairs administrative data generally have high reliability for mental health conditions, with more than 96% agreement reported between ICD-9-CM code and medical record . Veterans’ Affairs national databases and ICD-9-CM codes have been utilized extensively for clinical epidemiology and outcomes research [36–40]. Another limitation of this study is incomplete mortality data in CCR. Though adherence was not directly measured, pharmacy data serve as a reliable surrogate marker . Due to male predominance and demographic and behavioral differences between veteran and nonveteran populations, our results may not be generalizable to nonveteran populations.
As long-term survival among patients with HIV continues to improve , increased attention is geared towards management of comorbidities [25,41,42], and comorbid mental health conditions, particularly SMI and substance use, remain challenging. Our finding that depression was not associated with negative outcomes is notable. For persons with a diagnosis of depression engaged in mental healthcare, outcomes are significantly better than for persons with undiagnosed or undertreated depression. The access to mental health services within the Veterans’ Affairs may partially explain our findings. Nevertheless, the association of SMI and increased mortality is independent and deserves further investigation. Our results reflect ‘the real world’ of clinical practice, in which challenging patients with complex comorbidities are much more common than in the clinical trial setting. Even in the Veterans’ Affairs medical system, in which comprehensive mental health services are universally available and easily accessible, HIV-infected veterans with SMI remain vulnerable to unfavorable outcomes .
All authors equally contributed to this manuscript.
Conflicts of interest
The study was funded by VISN 15 Career Development Award.
D.N. has received a Career Development Award from the VA VISN 15 network.
J.F.S. is supported by a Career Development Award from the VA HSRD.
J.R.M. is currently supported by a Merit Award from VA HSRD.
E.T.O has served as a consultant, on speaker's bureau, or on an advisory board for the following companies: Gilead, Bristol Myers Squibb, Glaxo-Smith-Kline, Tibotec, Merck, Monogram Sciences and Boehringer Ingelheim. He also has received research support from the following companies: Gilead, Glaxo-Smith-Kline, and Tibotec as well as the NIH and CDC.
For the remaining authors none were declared.
1. Lyketsos CG, Hutton H, Fishman M, Schwartz J, Treisman GJ. Psychiatric morbidity on entry to an HIV primary care clinic
2. Fuller BE, Loftis JM, Rodriguez VL, McQuesten MJ, Hauser P. Psychiatric and substance use disorders comorbidities in veterans with hepatitis C virus and HIV coinfection
. Curr Opin Psychiatry
3. Backus LI, Boothroyd D, Deyton LR. HIV, hepatitis C and HIV/hepatitis C virus co-infection in vulnerable populations
2005; 19 (Suppl 3):S13–S19.
4. Rothbard AB, Metraux S, Blank MB. Cost of care for medicaid recipients with serious mental illness and HIV infection or AIDS
. Psychiatric Serv
5. Gordillo V, del Amo J, Soriano V, Gonzalez-Lahoz J. Sociodemographic and psychological variables influencing adherence to antiretroviral therapy
6. Ickovics JR, Hamburger ME, Vlahov D, Schoenbaum EE, Schuman P, Boland RJ, et al. Mortality, CD4 cell count decline, and depressive symptoms among HIV-seropositive women: longitudinal analysis from the HIV epidemiology research study
7. Mohl PC, Huang L, Bowden C, Fischbach M, Vogtsberger K, Talal N. Natural killer cell activity in major depression
. Am J Psychiatry
8. Farinpour R, Miller EN, Satz P, Selenes OA, Cohen BA, Becker JT, et al. Psychosocial risk factors of HIV morbidity and mortality: findings from the Multicenter AIDS Cohort Study (MACS)
. J Clin Exp Neuropsychol
9. Lyketsos CG, Hoover DR, Guccione M, Senterfitt W, Dew MA, Wesch JE, et al. Depressive symptoms as predictors of medical outcomes in HIV Infection
10. Lyketsos CG, Hoover DR, Guccione M, Dew MA, Wesch JE, Bing EG, et al. Changes in depressive symptoms as AIDS develops. The Multicenter AIDS Cohort Study
. Am J Psychiatry
11. Lima VD, Geller J, Bangsberg DR, Patterson TL, Daniel M, Kerr T, et al. The effect of adherence on the association between depressive symptoms and mortality among HIV-infected individuals first initiating HAART
12. Asch SM, Kilbourne AM, Gifford AL, Burnam AM, Turner B, Shapiro MF, et al. Underdiagnosis of depression in HIV: who are we missing?
. J Gen Intern Med
13. Tempalski B, Cleland C, Pouget E, Chatterjee S, Friedman S. Persistence of low drug treatment coverage for injection drug users in large US metropolitan areas.Substance Abuse Treatment Prevent Policy
14. Conover C, Weaver M, Arno P, Ang A, Ettner S. Insurance coverage among people living with combined HIV/AIDS, chronic mental illness, and substance abuse disorders
. J Healthcare Poor Underserved
15. Department of Health and Human Services. Panel on antiretroviral guidelines for adults and adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. http://www.aidsinfo.nih.gov/
. [Accessed 6 June 2011]
16. Backus L, Mole L, Chang S, Deyton L. The immunology case registry
. J Clin Epidemiol
17. Rabeneck L, Menke T, Simberkoff MS, Hartigan PM, Dickinson GM, Jensen PC, et al. Using the national registry of HIV-infected veterans in research: lessons for the development of disease registries
. J Clin Epidemiol
18. Schneeweiss S, Wang PS, Avorn J, Glynn RJ. Improved comorbidity adjustment for predicting mortality in Medicare populations
. Health Serv Res
19. Schneeweiss S, Seeger JD, Maclure M, Wang PS, Avorn J, Glynn RJ. Performance of comorbidity scores to control for confounding in epidemiologic studies using claims data
. Am J Epidemiol
20. Solberg LI, Engebretson KI, Sperl-Hillen JM, Hroscikoski MC, O’Connor PJ. Are claims data accurate enough to identify patients for performance measures or quality improvement? The case of diabetes, heart disease, and depression
. Am J Med Qual
21. Peterson AM, Nau DP, Cramer JA, Benner J, Gwadry-Sridhar F, Nichol M. A checklist for medication compliance and persistence studies using retrospective databases
. Value Health
22. Centers for Disease Control and Prevention. Revised surveillance case definitions for HIV infection among adults, adolescents, and children aged <18 months and for HIV infection and AIDS among children aged 18 months to <13 years: United States, 2008
. Morb Mortal Wkly Rep
23. Bourdon KH, Rae DS, Locke BZ, Narrow WE, Regier DA. Estimating the prevalence of mental disorders in U.S. adults from the Epidemiologic Catchment Area Survey
. Public Health Rep
24. Bing EG, Burnam MA, Longshore D, Fleishman JA, Sherbourne CD, London AS, et al. Psychiatric disorders and drug use among human immunodeficiency virus-infected adults in the United States
. Arch Gen Psychiatry
25. Mkanta WN, Mejia MC, Duncan RP. Race, outpatient mental health service use, and survival after an AIDS diagnosis in the highly active antiretroviral therapy era
. AIDS Patient Care STDS
26. Tabbane K, Joober R, Spadone C, Poirier M, Olie‘ J. Mortality and cause of death in schizophrenia. Review of the literature [Abstract]
27. Chang CK, Hayes R, Broadbent M, Fernandes A, Lee W, Hotopf M, et al. All-cause mortality among people with serious mental illness (SMI), substance use disorders, and depressive disorders in southeast London: a cohort study
. BMC Psychiatry
28. Chwastiak LA, Rosenheck RA, Desai R, Kazis LE. Association of psychiatric illness and all-cause mortality in the National Department of Veterans Affairs Healthcare System
. Psychosom Med
29. Vinogradova Y, Coupland C, Hippisley-Cox J, Whyte S, Penny C. Effects of severe mental illness on survival of people with diabetes
. Br J Psychiatry
30. Cohen MH, French AL, Benning L, Kovacs A, Anastos K, Young M, et al. Causes of death among women with human immunodeficiency virus infection in the era of combination antiretroviral therapy
. Am J Med
31. Pence BW. The impact of mental health and traumatic life experiences on antiretroviral treatment outcomes for people living with HIV/AIDS
. J Antimicrob Chemother
32. Ironson G, O’Cleirigh C, Fletcher MA, Laurenceau JP, Balbin E, Klimas N, et al. Psychosocial factors predict CD4 and viral load change in men and women with human immunodeficiency virus in the era of highly active antiretroviral treatment
. Psychosom Med
33. Tsai AC, Weiser SD, Petersen ML, Ragland K, Kushel MB, Bangsberg DR. A marginal structural model to estimate the causal effect of antidepressant medication treatment on viral suppression among homeless and marginally housed persons with HIV
. Arch Gen Psychiatry
34. Peretti-Watel P, Spire B, Lert F, Obadia Y. Drug use patterns and adherence to treatment among HIV-positive patients: evidence from a large sample of French outpatients (ANRS-EN12-VESPA 2003)
. Drug Alcohol Depend
35. Kashner TM. Agreement between administrative files and written medical records: a case of the Department of Veterans Affairs
. Med Care
36. Richardson LK, Egede LE, Mueller M. Effect of race/ethnicity and persistent recognition of depression on mortality in elderly men with type 2 diabetes and depression
. Diabetes Care
37. Sajatovic M, Blow F, Ignacio RV. Psychiatric comorbidity in older adults with bipolar disorder
. Int J Geriatric Psychiatry
38. Lee TA, Bartle B, Weiss KB. Impact of NSAIDS on mortality and the effect of preexisting coronary artery disease in US veterans
. Am J Med
39. Al-Aly Z, Pan H, Zeringue A, Xian H, McDonald JR, El-Achkar TM, et al. Tumor necrosis factor-α blockade, cardiovascular outcomes, and survival in rheumatoid arthritis
. Transl Res
40. Al-Aly Z, Zeringue A, Fu J, Rauchman MI, McDonald JR, El-Achkar TM. Rate of kidney function decline associates with mortality
. J Am Soc Nephrol
41. The SMART/INSIGHT and the D:A:D Study Groups. Use of nucleoside reverse transcriptase inhibitors and risk of myocardial infarction in HIV-infected patients
42. Friis-Moller N, Weber R, Reiss P, Thibaut R, Kirk O, Monforte AA, et al. Cardiovascular disease risk factors in HIV patients: association with antiretroviral therapy. Results from the DAD study