Abstract
Objectives
We evaluated the Veterans Aging Cohort Study (VACS) Index score, an index composed of age, CD4 count, viral load, hemoglobin, Hepatitis C (HCV) co-infection, Fibrosis Index-4 (FIB-4), and estimated glomerular filtration rate (EGFR), and psychosocial and clinical risk factors for all-cause hospitalization among HIV-infected women on HAART and HIV-uninfected women.
Methods
Data were collected from 2008 - 2014 from 1585 HAART-experienced HIV-infected and 692 uninfected women. Cox proportional hazards regression evaluated predictors of first hospitalization over 2 years.
Results
Among HIV-infected women, VACS Index score (per 5 points) (adjusted hazard ratio [aHR]1.08; 95% CI 1.06-1.11), Centers for Epidemiologic Studies-Depression (CES-D) scores ≥16 (aHR 1.61; 95% CI 1.30-1.99), smoking (aHR 1.26; 95% CI 1.02-1.55), abuse history (aHR 1.52; 95% CI 1.20-1.93), diabetes (aHR 1.63; 95% CI 1.31-2.04), and black race (aHR 1.28; 95% CI 1.03-1.59) increased risk of hospitalization. Among HIV-uninfected women, VACS Index score (aHR 1.08; 95% CI 1.03-1.13), CES-D ≥16 (aHR 1.38; 95% CI 1.02-1.86), diabetes (aHR 2.15; 95% CI 1.57-2.95), and black race (aHR 1.61; 95% CI 1.15-2.24) predicted subsequent hospitalization.
Conclusions
Psychosocial and clinical factors were associated with risk of hospitalization independently of the VACS Index score. Additional research on contextual and psychosocial influences on health outcomes among women is needed.
Keywords: HIV, hospitalizations, women
Background
Rates of hospitalization due to AIDS-related illnesses have declined substantially since the widespread availability of highly active antiretroviral therapy (HAART) 1,2. However, hospitalizations among HIV infected persons remain relatively common 3,4. The steep declines in hospitalizations observed soon after HAART became available have not declined as dramatically in recent years, likely due to aging of the HIV infected population, complications from co-morbidities, and side effects of HAART 5. Longer survival with HIV has increasingly been associated with aging-related conditions that may be accelerated among HIV infected persons, impacting morbidity and mortality even when HIV disease is controlled 6. HIV clinicians now treat a broad spectrum of chronic illnesses associated with aging and a comprehensive approach to care for HIV infected patients is needed 6,7.
Rates of hospitalization are higher and declines in hospitalizations have been less pronounced among women, African-Americans, and those with a history of injection drug use 5,8,9. Indeed, HIV-infected women disproportionately experience substance use, gender based violence, and mental health problems, which may lead to increased morbidity and mortality 10. Many of these issues also impact HIV uninfected women. Understanding factors that predict hospitalization in HIV infected and uninfected women may reduce morbidity, associated healthcare costs, and health disparities. During the earlier years of the HIV epidemic, HIV infected and uninfected women had very different clinical profiles and treatment needs. In contrast, HIV infected and uninfected women currently share many of the same aging-related comorbidities, particularly those from similar sociodemographic backgrounds. Understanding similarities and differences in patterns and predictors of hospitalization is important for effectively targeting interventions and improving care for women.
The Veterans Aging Cohort Study (VACS) Index, a score composed of age and HIV and non-HIV biomarkers, accurately predicts mortality among HIV infected men and women across varying stages of HIV disease progression and duration of HAART use. The VACS Index 11,12 has also been associated with a number of morbidities and markers of morbidity, including hospitalizations and readmissions, frailty indices, and cognitive function 13-15. Although the VACS Index also predicted mortality among treated HIV infected women in the Women's Interagency HIV Study (WIHS), depression and transactional sex history predicted mortality risk independently of the VACS Index score among WIHS women 16. Assessment of the impact of psychosocial factors on morbidity among women may thus be important for understanding and improving women's overall health. The goals of the current study were to 1.) compare trends over time in rates of hospitalization among HIV infected and uninfected WIHS women; 2.) examine associations between the VACS Index and risk for all-cause hospitalization among HIV infected and uninfected women; and 3.) examine associations between psychosocial and clinical factors and hospitalizations among WIHS women.
Methods
Sample and recruitment
The WIHS is an ongoing prospective cohort study of HIV infection among women in the United States. Data for this analysis were obtained from WIHS women enrolled during 1994-1995, 2000-2001, and 2011-2012 at six sites, including Bronx and Brooklyn, NY, Chicago, IL, Los Angeles and San Francisco, CA, and the Washington, D.C. metropolitan area. Detailed descriptions of the WIHS cohort, and recruitment, enrollment, retention, and data collection methods have been reported elsewhere 17. Of note, because the WIHS sought to enroll a comparison group of HIV uninfected women at risk of acquiring HIV based on sexual and substance use behaviors, with similar sociodemographic characteristics as HIV infected women, HIV uninfected WIHS women are not representative of HIV uninfected women in the general population. Women undergo clinical examination and structured face-to-face interviews and provide laboratory specimens at semi-annual visits. Analyses of predictors of hospitalization were restricted to the period beginning in 2008, the first year that reasons for hospitalizations were available (hereby defined as the index visit). To be included in the analysis women had to have at least one follow-up visit at which hospitalization data were ascertained. HIV infected women who had not initiated HAART were excluded because our goal was to examine predictors of hospitalization among women with relatively controlled HIV. Hospitalizations due to childbirth (<2% of hospitalizations during the analytic period) were also excluded. While the primary analysis focused on this more recent period, time trends in hospitalizations are reported descriptively for the entire WIHS cohort from 1994 through 2014. The study was approved by the Institutional Review Board at each participating site. All women provided written informed consent.
Variable definitions and classification
The VACS Index score was calculated using the formula developed by Justice and colleagues11,12 by adding points for age, CD4 count, viral load, hemoglobin, Hepatitis C (HCV) co-infection, Fibrosis Index-4 (FIB-4) 18, and estimated glomerular filtration rate (EGFR) 19. For analysis, the VACS Index score was categorized into 4 ordered groups with approximately equal number of events in each group, and it was analyzed as a continuous variable, using 5-point increments 11,12,20. VACS Index scores were calculated for HIV uninfected women by assigning the lowest possible number of points (0) for CD4 and viral load, a method consistent with prior research20.
Explanatory variables included the index visit VACS Index score; current smoking (yes vs. no); recent (since the last visit) hazardous alcohol use (defined according to NIAAA criteria 21 as >7 drinks per week or ≥4 drinks per day); a clinically relevant burden of depressive symptoms defined as scores ≥16 on the Center for Epidemiologic Studies-Depression (CES-D) scale 22; history of any physical, sexual, or emotional abuse; ever-use of crack, cocaine, or heroin; any history of transactional sex (defined as ever having had sex for drugs, money, or shelter); race/ethnicity (categorized as White non-Hispanic, Black non-Hispanic, Hispanic, Other non-Hispanic; and dichotomized as Black non-Hispanic vs. all other categories); current hypertension, defined as systolic blood pressure ≥140, diastolic blood pressure ≥90, or self-reported use of anti-hypertensives; and current or past history of diabetes, determined by self-report or use of anti-diabetic medications, fasting glucose ≥126 or HgbA1C ≥ 6.5%. Year of index visit and index visit hospitalization were examined to assess secular trends in hospitalization and because index visit hospitalization was expected to predict subsequent hospitalization. Sensitivity analyses yielded similar magnitude and direction of effects when index visit hospitalization was excluded from the models; models presented adjust for index visit hospitalization.
At each visit, women reported all hospitalizations since their last semiannual study visit. Reasons for hospitalization were assessed by self-report in pre-defined categories and reasons not listed on the questionnaire were reported via open-ended response. Open-ended responses were reviewed by the authors and classified into pre-existing categories when possible. The final categories included surgeries/procedures; pneumonia; infections other than pneumonia; cardiovascular disease; gastrointestinal problems; psychiatric or mental health problems; lung problems other than pneumonia, including asthma; injuries/accidents; central nervous system disorders; gynecologic procedures; kidney problems; blood disorders; liver problems (including Hepatitis related issues); cancer; endocrine disorders, and “other” reasons that did not fit into any of the above categories. Limitation in documentation precluded the classification of infections and cancers as AIDS-defining illnesses. Women could report multiple reasons for hospitalization and reasons were not classified as “primary” or “secondary”; thus, it was not possible to identify a primary admitting diagnosis. Reasons for hospitalization are reported descriptively, but predictors of cause-specific hospitalization were not analyzed due to the difficulty in interpreting such an analysis given data and sample size limitations.
Statistical Analysis
Sociodemographic, behavioral, and clinical characteristics were compared among HIV infected and uninfected women using Pearson chi-square tests for categorical variables and Wilcoxon rank-sum tests for continuous variables. Trends over time in hospitalizations among HIV infected and uninfected women were assessed using the Jonckheere-Terpstra nonparametric trend test. Time-adjusted analyses compared hospitalization (yes/no) at each visit by HIV status using generalized estimating equations with logit link to account for correlation among repeated measures. For multivariable analysis, the outcome was the time to first hospitalization within 2 years of the index visit. Cox survival analysis models were used to account for varying person-time contributions. Person-time was calculated as the time between the date of the index visit and the first visit at which hospitalization was reported, the last observed visit, or 2 years from the index visit. VACS Index scores and all additional variables were treated as fixed exposures at the index visit. The log-rank test was used to compare time to first hospitalization across categories of explanatory variables. Variables with p-values <0.2 in univariate analysis and those considered conceptually important based on published literature were entered in multivariable Cox proportional hazards regression models to evaluate associations with time to first hospitalization over 2-year follow-up. All variables were initially entered into the models and were then removed one by one in an iterative model selection process; those with p<0.05 were retained in the final multivariable models. The proportional hazards assumption was assessed by graphical examination of log-log survival curves for each explanatory variable and by statistical testing for non-zero slope of scaled Schoenfeld residuals on functions of time 23. The Efron method was used to approximate the exact probability of tied failures. Standard errors were calculated using robust variance estimation. Sensitivity analyses were conducted to examine the impact of predictors on 5-year hospitalization rates and use of time-varying vs. fixed covariates. Findings were similar using the extended follow-up and with fixed vs. time-varying exposures, so the analysis of fixed exposures is reported since we were interested in identifying index visit characteristics to predict risk of subsequent hospitalization. Data were analyzed using SAS software version 9.3 (SAS Institute, Cary, NC) and STATA/IC version 13.1 for Windows (Stata Corp, College Station, TX).
Results
Participant characteristics
Data were collected between April 2008 and March 2014 from 1622 HIV-infected women who had initiated HAART and 700 HIV uninfected women. The final analytic sample included 1585 HIV infected and 692 uninfected women who provided data on hospitalization at the index visit and at least one subsequent visit; follow-up occurred through March 2014. Over half of HIV infected and uninfected women reported history of substance use, nearly one third had history of transactional sex, and 60% had history of physical, sexual, and/or emotional abuse (table 1). At the index visit, 35% of HIV infected and 32% of uninfected women had CESD scores ≥16; over one third had hypertension, and 20% had a history of diabetes. HIV infected women were less likely than uninfected women to report current smoking (38% vs. 51%), hazardous alcohol use (18% vs. 31%), recent substance use (7% vs. 15%), recent abuse (6% vs. 9%) and recent transactional sex (1% vs. 4%) (p<0.05 for all comparisons). The median VACS Index score was 23 (IQR 16-39; range 0-106) among HIV infected women and 10 (IQR 10-22; range 0-86) among HIV uninfected women.
Table 1.
Index Visit Characteristics by HIV Status
| Characteristic at Index Visit | HIV Infected (n %) | HIV Uninfected n (%) | p-value |
|---|---|---|---|
| Total | 1585 (100.0) | 692 (100.0) | -- |
| VACS Index Score | |||
| 0-23 | 732 (46.2) | 547 (79.1) | <0.001 |
| 23-34 | 373 (23.5) | 85 (12.3) | |
| 35-50 | 272 (17.2) | 49 (7.1) | |
| ≥51 | 208 (13.1) | 11 (1.6) | |
| Median (IQR); Range | 23 (16-39); 0-106 | 10 (10-22); 0-86 | |
| CESD | |||
| ≥16 | 558 (35.3) | 218 (31.6) | 0.082 |
| <16 | 1022 (64.7) | 473 (68.5) | |
| Current smoking | |||
| Yes | 607 (38.4) | 353 (51.0) | <0.001 |
| No | 975 (61.6) | 339 (49.0) | |
| Hazardous alcohol use | |||
| Yes | 276 (17.5) | 211 (30.5) | <0.001 |
| No | 1306 (82.6) | 481 (69.5) | |
| Transactional sex ever | |||
| Yes | 505 (31.9) | 221 (31.9) | 0.994 |
| No | 1077 (68.1) | 471 (68.1) | |
| Transactional sex, recent | |||
| Yes | 16 (1.0) | 30 (4.3) | <0.001 |
| No | 1563 (99.0) | 661 (95.7) | |
| IDU ever | |||
| Yes | 354 (22.4) | 140 (20.2) | 0.260 |
| No | 1230 (77.6) | 552 (79.8) | |
| CCH ever | |||
| Yes | 841 (53.1) | 394 (56.9) | 0.091 |
| No | 743 (46.9) | 298 (43.1) | |
| Any IDU, recent | |||
| Yes | 19 (1.2) | 14 (2.0) | 0.132 |
| No | 1560 (98.8) | 677 (98.0) | |
| Any CCH, recent | |||
| Yes | 117 (7.4) | 104 (15.0) | <0.001 |
| No | 1465 (92.6) | 588 (85.0) | |
| Hypertension, current | |||
| Yes | 578 (36.5) | 245 (35.4) | 0.628 |
| No | 1007 (63.5) | 447 (64.6) | |
| Diabetes history | |||
| Yes | 306 (19.3) | 150 (21.7) | 0.194 |
| No | 1279 (80.7) | 542 (78.3) | |
| Race/Ethnicity | |||
| White | 208 (13.1) | 62 (9.0) | 0.012 |
| Black | 897 (56.6) | 430 (62.1) | |
| Hispanic | 418 (26.4) | 168 (24.3) | |
| Other race | 62 (3.9) | 32 (4.6) | |
| Abuse history | |||
| Yes | 762 (61.3) | 341 (58.9) | 0.327 |
| No | 481 (38.7) | 238 (41.1) | |
| Recent abuse | |||
| Yes | 67 (5.7) | 49 (9.1) | 0.010 |
| No | 1100 (94.3) | 488 (90.9) | |
| Hospitalized at index visit | |||
| Yes | 229 (14.5) | 83 (12.0) | 0.117 |
| No | 1356 (85.6) | 609 (88.0) | |
| WIHS recruitment wave | |||
| 1994-95 | 842 (53.1) | 296 (42.8) | <0.001 |
| 2001-2002 | 525 (33.1) | 307 (44.4) | |
| 2011-2012 | 218 (13.8) | 89 (12.9) |
Abbreviations: CESD, Centers for Epidemiologic Studies – Depression; CCH, crack, cocaine, or heroin use; IDU, injection drug use; VACS, Veterans Aging Cohort Study, WIHS, Women's Interagency HIV Study
Frequency of hospitalization by diagnostic category
Multiple hospitalizations were reported at 11% and 5% of visits among HIV infected and uninfected women respectively (Table 2). Across all visits at which reasons for hospitalization were reported, surgeries or procedures were reported at 22% and 23% of visits for HIV infected and uninfected women respectively. After surgeries the most common reasons for hospitalization among HIV infected women were infections (non-pneumonia at 17%, and pneumonia at 16% of visits), gastrointestinal disorders (13%), cardiovascular disease (11%), and lung problems other than pneumonia (9% of visits). For HIV uninfected women, the most common reasons for hospitalization after surgeries were cardiovascular disease (18%), gastrointestinal disorders (11%), followed by psychiatric/mental health problems, lung problems, and injuries (each reported at approximately 9% of visits) (Table 2).
Table 2.
Total Hospitalizations by HIV Status over 2 Years by Diagnostic Categories
| Visits among HIV infected women (N=5266) | Visits among HIV uninfected women (N=2263) | |||
|---|---|---|---|---|
| n | % | n | % | |
| Visits where hospitalization was reported | 673 | 12.8 | 235 | 10.4 |
| Total times hospitalized since last visit | n | % of 673 visits | n | % of 235 Visits |
| 1 | 480 | 71.3 | 184 | 78.3 |
| 2 | 121 | 18.0 | 39 | 16.6 |
| ≥ 3 | 72 | 10.7 | 12 | 5.1 |
| Reason for hospitalization | n | % of 673 visitsa | n | % of 235 Visitsa |
| Surgery/procedures | 148 | 22.0 | 55 | 23.4 |
| Infection other than pneumonia | 117 | 17.4 | 15 | 6.4 |
| Pneumonia | 111 | 16.5 | 9 | 3.8 |
| Gastrointestinal disordersb | 89 | 13.2 | 25 | 10.6 |
| Cardiovascular problems | 76 | 11.3 | 43 | 18.3 |
| Lung problems (non-pneumonia) | 58 | 8.6 | 21 | 8.9 |
| Psychiatric/mental health issues | 47 | 7.0 | 22 | 9.4 |
| Central nervous system disorders | 43 | 6.4 | 14 | 6.0 |
| Injury/accident | 39 | 5.8 | 21 | 8.9 |
| Gynecologic procedures/surgery | 24 | 3.6 | 12 | 5.1 |
| Kidney problems | 23 | 3.4 | 0 | 0.0 |
| Liver problemsc | 21 | 3.1 | 5 | 2.1 |
| Blood disorders | 13 | 1.9 | 8 | 3.4 |
| Endocrine disorders/diabetes | 8 | 1.2 | 9 | 3.8 |
| Cancer | 7 | 1.0 | 2 | 0.9 |
| Other illness | 15 | 2.2 | 10 | 4.3 |
Percentages sum to more than 100% because multiple reasons for hospitalization could be provided at each visit.
Includes stomach, intestinal, gallbladder, pancreatic issues
Includes Hepatitis related issues
Trends over time in hospitalization
Over two years of follow-up, 462 first hospitalizations occurred over 2308.6 person-years of observation among HIV infected women, resulting in an incidence rate of 20.01 (95% CI 18.27-22.92) per 100 person-years (Table 3). Among HIV uninfected women, there were 184 first hospitalizations in 1041.6 person-years, for an incidence rate of 17.66 (95% CI 15.29-20.41) per 100 person-years (p=0.095). In time-adjusted analysis, across all visits, HIV infected women were more likely than HIV uninfected women to be hospitalized (p<0.001). Among HIV infected women, hospitalizations declined over time (p for trend <0.001); the trend of declining hospitalizations over time among HIV uninfected women was not statistically significant (p=0.276) (Figure 1). Over time, the gap between hospitalization rates for women with and without HIV infection decreased.
Table 3.
Hospitalizations over 2-year follow-up by Patient Characteristics at Index Visit among HIV Infected and HIV Uninfected women
| HIV Infected (on HAART) | HIV Uninfected | |||||||
|---|---|---|---|---|---|---|---|---|
| Characteristic at Index Visit | Total n (%) | First Events/Person-years | Incidence Rate per 100 person-years (95% CI) | Log-rank p-value | Total n (%) | First Events/Person-years | Incidence Rate per 100 person-years (95% CI) | Log-rank p-value |
| Total | 1585 (100.0) | 462/2308.6 | 20.01 (18.27-22.92) | -- | 692 (100.0) | 184/1041.6 | 17.66 (15.29-20.41) | -- |
| VACS Index Score | ||||||||
| 0-23 | 732 (46.2) | 163/1110.5 | 14.68 (12.59-17.11) | <0.001 | 547 (79.1) | 130/840.7 | 15.46 (13.02-18.36) | <0.001 |
| 23-34 | 373 (23.5) | 99/554.5 | 17.85 (14.66-21.74) | 85 (12.3) | 32/117.9 | 27.14 (19.19-38.37) | ||
| 35-50 | 272 (17.2) | 90/395.4 | 22.76 (18.51-27.98) | 49 (7.1) | 15/67.8 | 22.11 (13.33-36.67) | ||
| ≥51 | 208 (13.1) | 110/248.1 | 44.33 (36.77-53.44) | 11 (1.6) | 7/15.1 | 46.28 (22.06-97.07) | ||
| On HAART since last visit | ||||||||
| Yes | 1378 (86.9) | 398/2008.7 | 19.81 (17.96-21.86) | 0.627 | -- | -- | -- | -- |
| No | 207 (13.1) | 64/300.0 | 21.34 (16.70-27.26) | -- | -- | -- | -- | |
| CESD | ||||||||
| ≥16 | 558 (35.3) | 231/744.2 | 31.04 (27.29-35.31) | <0.001 | 218 (31.6) | 74/314.7 | 23.51 (18.72-29.53) | 0.002 |
| <16 | 1022 (64.7) | 228/1557.9 | 14.64 (12.85-16.66) | 473 (68.5) | 109/725.9 | 15.02 (12.45-18.12) | ||
| Current smoking | ||||||||
| Yes | 607 (38.4) | 222/841.1 | 26.39 (23.14-30.10) | <0.001 | 353 (51.0) | 99/531.4 | 18.63 (15.30-22.69) | 0.395 |
| No | 975 (61.6) | 237/1465.5 | 16.17 (14.24-18.37) | 339 (49.0) | 85/510.3 | 16.66 (13.47-20.60) | ||
| Hazardous alcohol use | ||||||||
| Yes | 276 (17.5) | 88/395.6 | 22.25 (18.05-27.42) | 0.203 | 211 (30.5) | 55/322.5 | 17.05 (13.09-22.21) | 0.739 |
| No | 1306 (82.6) | 371/1911.0 | 19.41 (17.54-21.49) | 481 (69.5) | 129/719.1 | 17.94 (15.09-21.32) | ||
| Transactional sex ever | ||||||||
| Yes | 505 (31.9) | 180/711.1 | 25.31 (21.87-29.30) | <0.001 | 221 (31.9) | 75/321.1 | 23.36 (18.63-29.29) | 0.004 |
| No | 1077 (68.1) | 280/1595.0 | 17.55 (15.61-19.74) | 471 (68.1) | 109/720.5 | 15.13 (12.54-18.25) | ||
| Transactional sex, recent | ||||||||
| Yes | 16 (1.0) | 8/20.0 | 40.09 (20.05-80.16) | 0.041 | 30 (4.3) | 10/43.9 | 22.78 (12.26-42.35) | 0.395 |
| No | 1563 (99.0) | 451/2280.9 | 19.77 (18.03-21.69) | 661 (95.7) | 173/996.7 | 17.36 (14.95-20.15) | ||
| IDU ever | ||||||||
| Yes | 354 (22.4) | 137/484.4 | 28.28 (23.92-33.44) | <0.001 | 140 (20.2) | 47/206.3 | 22.78 (17.11-30.32) | 0.049 |
| No | 1230 (77.6) | 325/1823.1 | 18.09 (15.99-19.87) | 552 (79.8) | 137/835.3 | 16.40 (13.87-19.39) | ||
| CCH ever | ||||||||
| Yes | 841 (53.1) | 296/1184.2 | 25.00 (22.31-28.01) | <0.001 | 394 (56.9) | 122/579.7 | 21.05 (17.62-25.13) | 0.002 |
| No | 743 (46.9) | 166/1123.4 | 14.78 (12.69-17.20) | 298 (43.1) | 62/461.9 | 13.42 (10.46-17.21) | ||
| Any IDU, recent | ||||||||
| Yes | 19 (1.2) | 11/25.3 | 43.55 (24.12-78.64) | 0.014 | 14 (2.0) | 7/21.0 | 33.36 (15.91-69.99) | 0.090 |
| No | 1560 (98.8) | 448/2275.6 | 19.69 (17.95-21.60) | 677 (98.0) | 176/1019.6 | 17.26 (14.89-20.01) | ||
| Any CCH, recent | ||||||||
| Yes | 117 (7.4) | 47/161.1 | 29.18 (21.92-38.83) | 0.007 | 104 (15.0) | 37/150.5 | 24.59 (17.82-33.94) | 0.025 |
| No | 1465 (92.6) | 412/2145.5 | 19.20 (17.44-21.15) | 588 (85.0) | 147/891.2 | 16.50 (14.03-19.39) | ||
| Hypertension, current | ||||||||
| Yes | 578 (36.5) | 217/804.4 | 26.98 (23.62-30.82) | <0.001 | 245 (35.4) | 101/340.8 | 29.63 (24.38-36.01) | <0.001 |
| No | 1007 (63.5) | 245/1504.2 | 16.29 (14.37-18.46) | 447 (64.6) | 83/700.8 | 11.84 (9.55-14.69) | ||
| Diabetes history | ||||||||
| Yes | 306 (19.3) | 122/422.9 | 28.85 (24.16-34.45) | <0.001 | 150 (21.7) | 62/128.7 | 31.2 (24.3-40.0) | <0.001 |
| No | 1279 (80.7) | 340/1885.7 | 18.03 (16.21-20.05) | 542 (78.3) | 122/842.9 | 14.5 (12.1-17.3) | ||
| Race/Ethnicity | ||||||||
| White | 208 (13.1) | 54/302.5 | 17.85 (13.67-23.31) | 0.016 | 62 (9.0) | 15/96.4 | 15.56 (9.38-25.81) | 0.015 |
| Black | 897 (56.6) | 285/1277.4 | 22.31 (19.87-25.06) | 430 (62.1) | 131/630.6 | 20.77 (17.50-24.65) | ||
| Hispanic | 418 (26.4) | 111/633.0 | 17.53 (14.56-21.12) | 168 (24.3) | 31/264.5 | 11.72 (8.24-16.66) | ||
| Other race | 62 (3.9) | 12/95.7 | 12.54 (7.12-22.09) | 32 (4.6) | 7/50.1 | 13.97 (6.66-29.30) | ||
| Abuse historya (n=1,243) | ||||||||
| Yes | 762 (61.3) | 270/1082.3 | 24.95 (22.14-28.11) | <0.001 | 341 (58.9) | 100/505.1 | 19.80 (16.27-24.08) | 0.217 |
| No | 481 (38.7) | 104/726.2 | 14.32 (11.82-17.35) | 238 (41.1) | 59/359.0 | 16.43 (12.73-21.21) | ||
| Recent abuseb (n=1,167) | ||||||||
| Yes | 67 (5.7) | 24/92.2 | 26.02 (17.44-38.83) | 0.326 | 49 (9.1) | 10/73.8 | 13.54 (7.29-25.17) | 0.228 |
| No | 1100 (94.3) | 336/1579.6 | 21.27 (19.11-23.67) | 488 (90.9) | 139/725.3 | 19.16 (16.23-22.63) | ||
| Hospitalized at index visit | ||||||||
| Yes | 229 (14.5) | 122/270.5 | 45.11 (37.77-53.87) | <0.001 | 83 (12.0) | 46/104.3 | 44.09 (33.03-58.87) | <0.001 |
| No | 1356 (85.6) | 340/2038.1 | 16.68 (15.00-18.55) | 609 (88.0) | 138/937.3 | 14.72 (12.46-17.40) | ||
| WIHS recruitment wave | ||||||||
| 1994-95 | 842 (53.1) | 278/1212.1 | 22.94 (20.39-25.80) | 0.006 | 296 (42.8) | 98/432.3 | 22.67 (18.60-27.63) | <0.001 |
| 2001-2002 | 525 (33.1) | 133/785.5 | 16.93 (14.28-20.07) | 307 (44.4) | 59/483.3 | 12.21 (9.46-15.76) | ||
| 2011-2012 | 218 (13.8) | 51/311.0 | 16.40 (12.46-21.58) | 89 (12.9) | 27/126.0 | 21.42 (14.69-31.24) | ||
Based on non-missing observations, N=1243 for HIV positive women and N=579 for HIV negative women
Based on non-missing observations, N=1167 for HIV positive women and N=537 for HIV negative women
Abbreviations: CESD, Centers for Epidemiologic Studies – Depression; CCH, crack, cocaine, or heroin use; HAART, highly active antiretroviral therapy; IDU, injection drug use; VACS, Veterans Aging Cohort Study, WIHS, Women's Interagency HIV Study
Figure 1.
Trends over Time in All-Cause Hospitalizations
Footnote: Across all visits, HIV positive women were more likely to be hospitalized (p<0.001) controlling for time and repeated measures correlation in a generalized estimating equations logistic regression model. Among HIV positive women, hospitalizations declined over time (p for trend <0.001); the trend among HIV negative women was not statistically significant (p>0.2).
Predictors of Hospitalization among HIV Infected Women
Among HIV infected women, in multivariable Cox regression, VACS Index score was associated with an 8% increased risk of hospitalization per 5-point increase (HR 1.08; 95% CI 1.06-1.11). Depressive symptoms (HR 1.61; 95% CI 1.30-1.99), current smoking (HR 1.26; 95% CI 1.02-1.55), history of abuse (HR 1.52; 95% CI 1.20-1.93), diabetes (HR 1.63; 95% CI 1.31-2.04), and Black race (HR 1.28; 95% CI 1.03-1.59) were statistically significant independent predictors of hospitalization in models including VACS Index score and index visit hospitalization (Table 4).
Table 4.
Univariable and Multivariable Cox Regression: Relative Hazard of First Hospitalization over 2-year follow-up
| HIV Infected (on HAART) | HIV Uninfected | |||||||
|---|---|---|---|---|---|---|---|---|
| Univariable HR (95% CI) | p-value | Multivariable HRa (95% CI) | p-value | Univariable HR (95% CI) | p-value | Multivariable HRa (95% CI) | p-value | |
| VACS at index visit (per 5 point increase) | 1.12 (1.10-1.15) | <0.001 | 1.08 (1.06-1.11) | <0.001 | 1.13 (1.08-1.19) | <0.001 | 1.08 (1.03-1.13) | 0.003 |
| CESD ≥16 | 2.15 (1.79-2.58) | <0.001 | 1.61 (1.30-1.99) | <0.001 | 1.63 (1.21-2.20) | 0.001 | 1.38 (1.02-1.86) | 0.038 |
| Smoking | 1.66 (1.38-1.99) | <0.001 | 1.26 (1.02-1.55) | 0.018 | 1.17 (0.87-1.57) | 0.295 | -- | -- |
| Hazardous alcohol use | 1.16 (0.92-1.47) | 0.203 | -- | -- | 0.97 (0.71-1.32) | 0.829 | -- | -- |
| Transactional sex ever | 1.45 (1.20-1.75) | <0.001 | -- | -- | 1.53 (1.14-2.05) | 0.005 | -- | -- |
| IDU history | 1.60 (1.31-1.95) | <0.001 | -- | -- | 1.42 (1.02-1.98) | 0.037 | -- | -- |
| CCH ever | 1.70 (1.41-2.06) | <0.001 | -- | -- | 1.67 (1.23-2.27) | 0.001 | -- | -- |
| CCH recent | 1.52 (1.13-2.03) | 0.006 | -- | -- | 1.54 (1.07-2.20) | 0.019 | -- | -- |
| Hypertension | 1.67 (1.39-2.00) | <0.001 | -- | -- | 2.74 (2.05-3.67) | <0.001 | -- | -- |
| Diabetes ever | 1.59 (1.30-1.96) | <0.001 | 1.63 (1.31-2.04) | <0.001 | 2.30 (1.69-3.14) | <0.001 | 2.15 (1.57-2.95) | <0.001 |
| Abuse history | 1.82 (1.44-2.29) | <0.001 | 1.52 (1.20-1.93) | 0.001 | 1.28 (0.92-1.77) | 0.138 | -- | -- |
| HAART use since last visit | 0.94 (0.72-1.22) | 0.630 | -- | -- | NA | NA | NA | NA |
| Time since HAART initiation | 1.01 (0.99-1.04) | 0.279 | -- | -- | NA | NA | NA | NA |
| WIHS recruitment wave | ||||||||
| 1994-95 | 1.0 (Ref) | -- | -- | 1.0 (Ref) | -- | -- | ||
| 2001-2002 | 0.74 (0.60-0.91) | 0.004 | -- | -- | 0.50 (0.36-0.70) | <0.001 | -- | -- |
| 2011-2012 | 0.74 (0.55-0.996) | 0.047 | -- | -- | 1.00 (0.66-1.53) | 0.986 | -- | -- |
| Year of index visit | 0.94 (0.87-1.02) | 0.158 | -- | -- | 1.11 (0.998-1.24) | 0.055 | -- | -- |
| Hospitalization at index visit | 2.84 (2.30-3.49) | <0.001 | 2.37 (1.87-2.99) | <0.001 | 3.43 (2.44-4.82) | <0.001 | 3.10 (2.19-4.38) | <0.001 |
| Black race | 1.34 (1.11-1.61) | 0.003 | 1.28 (1.03-1.59) | 0.028 | 1.62 (1.18-2.23) | 0.003 | 1.61 (1.15-2.24) | 0.005 |
Multivariable hazard ratios are adjusted for all variables for which estimates are presented.
Abbreviations: CESD, Centers for Epidemiologic Studies – Depression; CCH, crack, cocaine, or heroin use; IDU, injection drug use; VACS, Veterans Aging Cohort Study, WIHS, Women's Interagency HIV Study
Predictors of Hospitalization among HIV uninfected women
Among HIV uninfected women, in multivariable Cox regression, VACS Index score was associated with an 8% increased risk of hospitalization per 5-point increase (HR 1.08; 95% CI 1.03-1.13). Depressive symptoms (HR 1.38; 95% CI 1.02-1.86), diabetes history (HR 2.15; 95% CI 1.57-2.95), and Black race (HR 1.61; 95% CI 1.15-2.24) predicted subsequent hospitalization in models including VACS Index score and index visit hospitalization (Table 4).
We also assessed the association of index visit VACS Index score and other exposures on risk of hospitalization over 5 years of follow-up; these analyses yielded similar final models and equivalent conclusions in terms of the direction, magnitude, and statistical significance of effects, except that among HIV infected women, the effect of Black race was attenuated over 5 year follow-up (5-year HR 1.08; 95% CI 0.91-1.28). The effect of race remained statistically significant for HIV uninfected women over 5 year follow-up. Analyses excluding hospitalizations due to surgeries and procedures also yielded similar final models and effect estimates.
Discussion
Consistent with several recent studies 3,5, we observed declines over time in hospitalizations among HIV infected women in the WIHS, and while these declines appear to have been most pronounced shortly after HAART became widely available, they have persisted, though to a lesser extent, in recent years. However, a significant proportion of HIV infected and HIV uninfected women continue to be hospitalized at each visit (10-15%), and rates among HIV negative women have remained stable over time. While the convergence of hospitalization rates among HIV-infected and uninfected women are encouraging, rates of hospitalization are still higher among both groups of women than in the general population. Estimates of annual inpatient hospitalizations from recent population surveys range from approximately 11 – 14 stays per 100 women aged 18-64 24. While not directly comparable due to differences in populations and calculation of rates, in our study, incidences of first hospitalization were 20 and 17 per 100 person-years among HIV infected and uninfected women respectively. HIV infected and uninfected women in our sample reported high prevalence of smoking, history of substance and alcohol use, abuse, depression, and other comorbidities, all of which make women vulnerable to development and progression of comorbidities as they age. There is a need for interventions that address clinical and psychosocial factors that predispose women to comorbidities, increase risk of hospitalization, and potentially create added social and financial hardship.
The VACS index accurately predicted hospitalizations among HIV infected and uninfected women in our sample, consistent with other recent studies 13,25. The VACS index may thus be a useful tool for predicting risk in uninfected women despite normalized CD4 counts, since it includes markers of renal and liver function. However, additional psychosocial and clinical factors independently predicted hospitalization in models that included VACS for both HIV infected and uninfected women. Women who reported a clinically relevant depressive symptom burden were 1.4-1.6 times more likely to be hospitalized, suggesting that using VACS in combination with risk assessment tools such as depression screening may better quantify overall risk. Depression disproportionately impacts people with HIV 26, and is associated with inflammation 27 and negative clinical outcomes 28. Given the high prevalence of clinically significant depressive symptoms in our study, screening for and treating depression could have an important impact on women's health outcomes.
Approximately 60% of WIHS women reported history of sexual, physical, or emotional abuse, consistent with other literature showing high prevalence of physical and sexual abuse among HIV infected women 29. Among HIV infected women, history of abuse was associated with an increased risk of hospitalization independently of the VACS Index and depression. Studies have consistently demonstrated negative consequences on HIV disease progression and health in general among patients with history of abuse and trauma, with long-term effects on physical and cognitive functioning, health care utilization, including hospitalizations, emergency room and outpatient visits 29,30, and mortality 31 in HIV infected patients. It is unclear why history of abuse was not as strongly associated with hospitalization among HIV uninfected women in our study. This finding may reflect differences in sampling or self-selection factors among HIV infected and uninfected women, differences in cumulative exposures or specific types of abuse, or other unmeasured factors, such as differences in care seeking patterns or reporting of abuse that impacted the relative associations we observed. Alternatively, the intersection of HIV infection and abuse may uniquely impact additional predisposing conditions, such as social isolation, that in turn impact risk of hospitalization and development of comorbidities.
Among HIV infected women, smoking was associated with an increased risk of hospitalization independently of the VACS Index, though a statistically significant association was not observed for HIV uninfected women. The reason for the difference by HIV status is unclear, but may be due to unique interactions of smoking and HIV and/or HIV therapy, or differences between groups in the distribution of other unmeasured confounders, such as care-seeking behaviors or other comorbidities. Given the high prevalence of smoking among women in our study smoking cessation remains a priority and may have particularly important effects for HIV infected women if health effects of smoking are exacerbated in the presence of HIV disease 32.
Substance use (recent and former) was associated with increased rates of hospitalization in univariable analysis for HIV infected and uninfected women, though not statistically significant in multivariable analysis. This is likely due to correlation with VACS Index scores, depression, other co-morbidities, or a combination of these. In contrast to the recent study by Agkun and colleagues 13 among male veterans, alcohol use was not strongly associated with hospitalization in our study. This may reflect differences in measurement of alcohol use and cutoff for hazardous levels, reporting bias, other unmeasured confounding, or true gender differences. Despite the lack of a strong association with hospitalization in our analysis, the confidence intervals included fairly large increases in risk for alcohol abuse. Substance and alcohol use are associated with lower adherence to HAART and adverse clinical outcomes in HIV infected and uninfected women 33,34 and remain important targets for clinical interventions.
For HIV infected and uninfected women, black race was associated with an increased risk of hospitalization independently of the VACS index and other clinical and psychosocial risk factors, though the effect for HIV infected women was attenuated over 5-year follow-up. Reasons for these disparities are not entirely clear, but may reflect a combination of socioeconomic hardship, experience of discrimination, and other structural adversities that increase vulnerability to negative health consequences 35. Addressing factors contributing to racial disparities in access and care is a priority to reduce higher rates of hospitalization among these HIV infected and at-risk Black women.
As might be expected, hospitalization at the index visit was a strong predictor of subsequent hospitalizations, and a small but important proportion of women were hospitalized multiple times during the study. Repeat hospitalizations may reflect greater disease severity and may have uniquely deleterious effects on morbidity and mortality, quality of life, and associated healthcare costs. Understanding factors associated with repeated hospitalizations warrants further study.
Limitations
The WIHS has a broad geographic representation and demographic characteristics of the WIHS women reflect those of women living with HIV in the United States, but those retained in the study may be generally healthier than other women living with HIV due to repeated study assessments and ongoing clinical care; additionally, women who died or left the study prior to 2008 were not included. Our findings are not generalizable to HIV infected women not on HAART. Importantly, HIV uninfected WIHS participants are not representative of HIV uninfected women in the general population, due to eligibility criteria for enrollment in WIHS for HIV uninfected women 17, although HIV infected and uninfected WIHS women represent vulnerable and often understudied populations.
Hospitalizations were self-reported and subject to recall bias and reasons for hospitalization were crudely categorized. We were unable to distinguish AIDS defining from non-AIDS defining hospitalizations or to identify the primary cause of hospitalization. Many of the surgeries and other procedures may be secondary to co-morbid conditions. There were also insufficient numbers of events in many of the hospitalization categories to examine multivariable predictors of cause-specific hospitalizations. Nonetheless, the rate of self-reported hospitalizations and the relative associations between VACS and all-cause hospitalizations in our study are similar in magnitude and direction to those reported in other studies 13,25.
Conclusions
Despite decreases in hospitalizations among HIV infected women since the widespread availability of HAART, we found persistently elevated rates of hospitalization among both HIV infected and uninfected WIHS women relative to rates in the general population. Our data suggest that higher rates of infection are the primary source of the disparity between HIV infected and uninfected women. We observed a 3-fold higher rate of infections resulting in hospitalization among HIV infected relative to HIV uninfected women, despite availability of HAART and improved CD4 counts. Further research is needed to understand differences in cause-specific predictors of hospitalization among HIV infected and uninfected women, and to quantify the contribution of clinical and psychosocial influences on women's health outcomes. In addition to improving clinical markers such as those captured in the VACS Index, addressing racial inequity in access and quality of care and proritizing treatment for depression and smoking cessation may reduce morbidity among HIV infected and uninfected women.
Acknowledgements
Data were collected by the Women's Interagency HIV Study (WIHS). The contents of this publication are solely the responsibility of the authors and do not represent the official views of the National Institutes of Health (NIH). WIHS (Principal Investigators): Bronx WIHS (Kathryn Anastos), U01-AI-035004; Brooklyn WIHS (Howard Minkoff and Deborah Gustafson), U01-AI-031834; Chicago WIHS (Mardge Cohen and Audrey French), U01-AI-034993; Metropolitan Washington WIHS (Mary Young), U01-AI-034994; Connie Wofsy Women's HIV Study, Northern California (Ruth Greenblatt, Bradley Aouizerat, and Phyllis Tien), U01-AI-034989; WIHS Data Management and Analysis Center (Stephen Gange and Elizabeth Golub), U01-AI-042590; Southern California WIHS (Joel Milam), U01-HD-032632 (WIHS I – WIHS IV).
Footnotes
Ethical Standards: Procedures for the protection of human subjects were in accordance with the ethical standards of the institutional review board of the Cook County Health and Hospitals System and with the Helsinki Declaration of 1975, as revised in 2000
References
- 1.Casalino E, Wolff M, Ravaud P, Choquet C, Bruneel F, Regnier B. Impact of HAART advent on admission patterns and survival in HIV-infected patients admitted to an intensive care unit. AIDS (London, England) 2004 Jul 2;18(10):1429–1433. doi: 10.1097/01.aids.0000131301.55204.a7. [DOI] [PubMed] [Google Scholar]
- 2.Gebo KA, Fleishman JA, Moore RD. Hospitalizations for metabolic conditions, opportunistic infections, and injection drug use among HIV patients: trends between 1996 and 2000 in 12 states. Journal of acquired immune deficiency syndromes (1999) 2005 Dec 15;40(5):609–616. doi: 10.1097/01.qai.0000171727.55553.78. [DOI] [PubMed] [Google Scholar]
- 3.Berry SA, Fleishman JA, Moore RD, Gebo KA. Trends in reasons for hospitalization in a multisite United States cohort of persons living with HIV, 2001-2008. Journal of acquired immune deficiency syndromes (1999) 2012 Apr 1;59(4):368–375. doi: 10.1097/QAI.0b013e318246b862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Crum-Cianflone NF, Grandits G, Echols S, et al. Trends and causes of hospitalizations among HIV-infected persons during the late HAART era: what is the impact of CD4 counts and HAART use? Journal of acquired immune deficiency syndromes (1999) 2010 Jul;54(3):248–257. doi: 10.1097/qai.0b013e3181c8ef22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yehia BR, Fleishman JA, Hicks PL, Ridore M, Moore RD, Gebo KA. Inpatient health services utilization among HIV-infected adult patients in care 2002-2007. Journal of acquired immune deficiency syndromes (1999) 2010 Mar;53(3):397–404. doi: 10.1097/QAI.0b013e3181bcdc16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Justice AC. HIV and aging: time for a new paradigm. Current HIV/AIDS reports. 2010 May;7(2):69–76. doi: 10.1007/s11904-010-0041-9. [DOI] [PubMed] [Google Scholar]
- 7.Palella FJ, Jr., Baker RK, Moorman AC, et al. Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV outpatient study. Journal of acquired immune deficiency syndromes (1999) 2006 Sep;43(1):27–34. doi: 10.1097/01.qai.0000233310.90484.16. [DOI] [PubMed] [Google Scholar]
- 8.Floris-Moore M, Lo Y, Klein RS, et al. Gender and hospitalization patterns among HIV-infected drug users before and after the availability of highly active antiretroviral therapy. Journal of acquired immune deficiency syndromes (1999) 2003 Nov 1;34(3):331–337. doi: 10.1097/00126334-200311010-00012. [DOI] [PubMed] [Google Scholar]
- 9.Schoenbaum EE, Lo Y, Floris-Moore M. Predictors of hospitalization for HIV-positive women and men drug users, 1996-2000. Public health reports (Washington, D.C. : 1974) 2002;117(Suppl 1):S60–66. [PMC free article] [PubMed] [Google Scholar]
- 10.Meyer JP, Springer SA, Altice FL. Substance abuse, violence, and HIV in women: a literature review of the syndemic. Journal of women's health (2002) 2011 Jul;20(7):991–1006. doi: 10.1089/jwh.2010.2328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Justice AC, McGinnis KA, Skanderson M, et al. Towards a combined prognostic index for survival in HIV infection: the role of 'non-HIV' biomarkers. HIV medicine. 2010 Feb;11(2):143–151. doi: 10.1111/j.1468-1293.2009.00757.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Justice AC, Modur SP, Tate JP, et al. Predictive accuracy of the Veterans Aging Cohort Study index for mortality with HIV infection: a North American cross cohort analysis. Journal of acquired immune deficiency syndromes (1999) 2013 Feb 1;62(2):149–163. doi: 10.1097/QAI.0b013e31827df36c. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Akgun KM, Gordon K, Pisani M, et al. Risk factors for hospitalization and medical intensive care unit (MICU) admission among HIV-infected Veterans. Journal of acquired immune deficiency syndromes (1999) 2013 Jan 1;62(1):52–59. doi: 10.1097/QAI.0b013e318278f3fa. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Marquine MJ, Umlauf A, Rooney AS, et al. The veterans aging cohort study index is associated with concurrent risk for neurocognitive impairment. Journal of acquired immune deficiency syndromes (1999) 2014 Feb 1;65(2):190–197. doi: 10.1097/QAI.0000000000000008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Womack JA, Goulet JL, Gibert C, et al. Physiologic frailty and fragility fracture in HIV-infected male veterans. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2013 May;56(10):1498–1504. doi: 10.1093/cid/cit056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cohen MH, Hotton AL, Hershow RC, et al. Gender-related risk factors improve mortality predictive ability of VACS Index among HIV-infected women. J Acquir Immune Defic Syndr. 2015 Dec 15;70(5):538–44. doi: 10.1097/QAI.0000000000000795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Barkan SE, Melnick SL, Preston-Martin S, et al. The Women's Interagency HIV Study. WIHS Collaborative Study Group. Epidemiology. 1998 Mar;9(2):117–125. [PubMed] [Google Scholar]
- 18.Sterling RK, Lissen E, Clumeck N, et al. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology. 2006 Jun;43(6):1317–1325. doi: 10.1002/hep.21178. [DOI] [PubMed] [Google Scholar]
- 19.Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function--measured and estimated glomerular filtration rate. N Engl J Med. 2006 Jun 8;354(23):2473–2483. doi: 10.1056/NEJMra054415. [DOI] [PubMed] [Google Scholar]
- 20.Tate JP, Justice AC, Hughes MD, et al. An internationally generalizable risk index for mortality after one year of antiretroviral therapy. AIDS. 2013 Feb 20;27(4):563–572. doi: 10.1097/QAD.0b013e32835b8c7f. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.U.S. Department of Health and Human Services . Helping Patients Who Drink Too Much: A Clinician's Guide. National Institute on Alcohol Abuse and Alcoholism; 2005. NIH Publication No. 07–3769. [Google Scholar]
- 22.Radloff L. The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1:385–401. [Google Scholar]
- 23.Schoenfeld D. Partial Residuals for the Proportional Hazards Regression Model. Biometrika. 1982;69:239–241. [Google Scholar]
- 24.Wier LM, Pfuntner A, Maeda J, et al. HCUP Facts and Figures: Statistics on Hospital-based Care in the United States, 2009. Agency for Healthcare Research and Quality; Rockville, MD: 2011. [PubMed] [Google Scholar]
- 25.Akgun KM, Tate JP, Pisani M, et al. Medical ICU admission diagnoses and outcomes in human immunodeficiency virus-infected and virus-uninfected veterans in the combination antiretroviral era. Critical care medicine. 2013 Jun;41(6):1458–1467. doi: 10.1097/CCM.0b013e31827caa46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ciesla JA, Roberts JE. Meta-analysis of the relationship between HIV infection and risk for depressive disorders. The American journal of psychiatry. 2001 May;158(5):725–730. doi: 10.1176/appi.ajp.158.5.725. [DOI] [PubMed] [Google Scholar]
- 27.Stewart JC, Rand KL, Muldoon MF, Kamarck TW. A prospective evaluation of the directionality of the depression-inflammation relationship. Brain Behav Immun. 2009 Oct;23(7):936–944. doi: 10.1016/j.bbi.2009.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hartzell JD, Janke IE, Weintrob AC. Impact of depression on HIV outcomes in the HAART era. The Journal of antimicrobial chemotherapy. 2008 Aug;62(2):246–255. doi: 10.1093/jac/dkn193. [DOI] [PubMed] [Google Scholar]
- 29.Brezing C, Ferrara M, Freudenreich O. The syndemic illness of HIV and trauma: implications for a trauma-informed model of care. Psychosomatics. 2015 Mar-Apr;56(2):107–118. doi: 10.1016/j.psym.2014.10.006. [DOI] [PubMed] [Google Scholar]
- 30.Leserman J, Whetten K, Lowe K, Stangl D, Swartz MS, Thielman NM. How trauma, recent stressful events, and PTSD affect functional health status and health utilization in HIV-infected patients in the south. Psychosomatic medicine. 2005 May-Jun;67(3):500–507. doi: 10.1097/01.psy.0000160459.78182.d9. [DOI] [PubMed] [Google Scholar]
- 31.Leserman J, Pence BW, Whetten K, et al. Relation of lifetime trauma and depressive symptoms to mortality in HIV. The American journal of psychiatry. 2007 Nov;164(11):1707–1713. doi: 10.1176/appi.ajp.2007.06111775. [DOI] [PubMed] [Google Scholar]
- 32.Feldman JG, Minkoff H, Schneider MF, et al. Association of cigarette smoking with HIV prognosis among women in the HAART era: a report from the women's interagency HIV study. Am J Public Health. 2006 Jun;96(6):1060–1065. doi: 10.2105/AJPH.2005.062745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Chander G, Lau B, Moore RD. Hazardous alcohol use: a risk factor for non-adherence and lack of suppression in HIV infection. Journal of acquired immune deficiency syndromes (1999) 2006 Dec 1;43(4):411–417. doi: 10.1097/01.qai.0000243121.44659.a4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hinkin CH, Barclay TR, Castellon SA, et al. Drug use and medication adherence among HIV-1 infected individuals. AIDS and behavior. 2007 Mar;11(2):185–194. doi: 10.1007/s10461-006-9152-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Williams DR, Mohammed SA, Leavell J, Collins C. Race, socioeconomic status, and health: complexities, ongoing challenges, and research opportunities. Annals of the New York Academy of Sciences. 2010 Feb;1186:69–101. doi: 10.1111/j.1749-6632.2009.05339.x. [DOI] [PMC free article] [PubMed] [Google Scholar]