Abstract
Objective
Our primary objective was to examine the syndemic effect of HIV/HCV co-infection and mental health disorders (MHD) on the acute care hospitalization rate among people living with HIV (PLW-HIV) in British Columbia, Canada. Secondarily, we aimed to characterize the longitudinal trends in the aforementioned rate, while controlling for the effect of several factors.
Methods
In this retrospective cohort study, individuals were antiretroviral therapy-naïve, ≥ 18 years old, initiated treatment between 1 January 2000 and 31 December 2014, and were followed for at least 6 months until 31 December 2015 or last contact. The outcome was acute care hospitalization rate (every 6-month interval) per individual. The exposure was the interaction between HIV/HCV co-infection and MHD. Generalized non-linear mixed-effects models were built.
Results
Of the 4046 individuals in the final analytical sample, 1597 (39%) were PLW-HIV without MHD, 606 (15%) were people living with HIV and HCV (PLW-HIV/HCV) without MHD, 988 (24%) were PLW-HIV with MHD, and 855 (21%) were PLW-HIV/HCV with MHD. The adjusted rate ratios for acute care hospitalizations were 1.31 (95% [confidence interval] 1.13–1.52), 2.01 (95% CI 1.71–2.36), and 2.53 (95% CI 2.20–2.92) for PLW-HIV with MHD, PLW-HIV/HCV without MHD, and PLW-HIV/HCV with MHD, respectively, relative to PLW-HIV without MHD.
Conclusion
The HIV/HCV co-infection and MHD interaction demonstrated a significant effect on the rate of acute care hospitalization, particularly for PLW-HIV/HCV with MHD. Implementing widely accessible integrative care model best practices may address this public health challenge.
Electronic supplementary material
The online version of this article (10.17269/s41997-019-00253-w) contains supplementary material, which is available to authorized users.
Keywords: HIV, Hepatitis C virus, Co-infection, Mental health, Inpatients, Public health
Résumé
Objectif
Notre principal objectif était d’examiner l’effet syndémique de la co-infection par le VIH et le VHC et des troubles mentaux (TM) sur le taux d’hospitalisation en soins de courte durée chez les personnes vivant avec le VIH (PVVIH) en Colombie-Britannique, au Canada. En second lieu, nous avons cherché à caractériser les tendances longitudinales du taux susmentionné, après avoir apporté des ajustements pour tenir compte des effets de plusieurs facteurs.
Méthode
Les sujets de cette étude de cohorte rétrospective étaient naïfs de traitement antirétroviral, avaient 18 ans ou plus, ont commencé un traitement entre le 1er janvier 2000 et le 31 décembre 2014 et ont été suivis pendant au moins 6 mois, jusqu’au 31 décembre 2015 ou au dernier contact. La résultante était le taux d’hospitalisation en soins de courte durée (à intervalles de 6 mois) par personne. L’exposition était l’interaction entre la co-infection par le VIH et le VHC et les TM. Nous avons construit des modèles non linéaires généralisés à effets mixtes.
Résultats
Sur les 4 046 sujets de l’échantillon d’analyse final, 1 597 (39 %) étaient des PVVIH sans TM, 606 (15 %) étaient des personnes vivant avec le VIH et le VHC (PVVIH-VHC) sans TM, 988 (24 %) étaient des PVVIH avec TM, et 855 (21 %) étaient des PVVIH-VHC avec TM. Les rapports de taux ajustés pour les hospitalisations en soins de courte durée étaient de 1,31 (IC de 95 % (1,13-1,52), de 2,01 (IC de 95 % 1,71-2,36) et de 2,53 (IC de 95 % 2,20-2,92) pour les PVVIH avec TM, les PVVIH-VHC sans TM et les PVVIH-VHC avec TM, respectivement, comparativement aux PVVIH sans TM.
Conclusion
L’interaction entre la co-infection par le VIH et le VHC et les TM exerce un effet significatif sur le taux d’hospitalisation en soins de courte durée, particulièrement chez les personnes vivant avec le VIH et le VHC et ayant des troubles mentaux. La mise en œuvre de modèles de soins intégrés exemplaires largement accessibles pourrait contribuer à résoudre ce problème de santé publique.
Mots-clés: VIH, Virus de l’hépatite C, Co-infection, Santé mentale, Patients hospitalisés, Santé publique
Introduction
In the modern combination antiretroviral therapy (ART) era, HIV infection has become a chronic condition (Deeks et al. 2013). In part due to the widespread use of ART, there has been a dramatic increase in the life expectancy of people living with HIV (PLW-HIV), particularly in high resource settings (Lima et al. 2015a). However, non-AIDS-related comorbidities are becoming increasingly prevalent in this population (Deeks et al. 2013).
In Canada, hepatitis C virus (HCV) is a leading infectious comorbidity among PLW-HIV (20%) due to shared routes of transmission (Buxton et al. 2010). The prevalence of HIV/HCV co-infection is highly contingent on the transmission route; percutaneous exposure to infected blood yields the highest rates of transmission (Taylor et al. 2012). As such, HIV/HCV co-infection rates are particularly elevated among people who have ever injected drugs (PWID) (50–90%) (Buxton et al. 2010).
HIV and HCV have fundamental similarities; they both produce subclinical, chronic conditions and are both capable of eluding immune function (Poles and Dieterich 2000). People living with HIV and HCV (PLW-HIV/HCV) may be at elevated risk of clinical progression of HIV and enhanced progression of HCV-related liver disease (Ingiliz and Rockstroh 2015).
PLW-HIV and PLW-HIV/HCV are also disproportionately affected by mental health disorders (MHD) (Blank et al. 2013). MHD has been shown to adversely impact ART adherence and virologic response, and has thus been linked to increased morbidity and mortality (Pence et al. 2007). A recent study has shown that a considerable amount of healthcare utilization among PLW-HIV/HCV can be attributed to mental health and substance use (Katrak et al. 2016). Taken together, this body of evidence brings into light the concept of syndemic theory.
Syndemic theory is based on a theoretical framework that takes into account multiple interrelated factors that lead to an excess in the burden of a disease, and represents an important tool that can be used to unmask the root causes of public health epidemics (Singer et al. 2017). Despite a number of studies examining the prevalence of blood-borne viruses (e.g., HIV, HCV) in individuals with MHD (Hughes et al. 2016), there remains a scarcity of literature investigating the impact of the HIV-HCV-MHD syndemic on disease burden, as proposed in the current study.
Our primary objective was to examine the syndemic effect of HIV/HCV co-infection and MHD on the acute care hospitalization rate among PLW-HIV in British Columbia (BC), Canada. Secondarily, we aimed to characterize the longitudinal trends in the aforementioned rate, while controlling for the effect of several demographic and clinical factors.
Methods
Study setting
In BC, the distribution of ART has been solely under the auspices of the BC Centre for Excellence in HIV/AIDS (BC-CfE) Drug Treatment Program (DTP) since 1992. HIV medical care and laboratory monitoring are also provided at no cost for all diagnosed PLW-HIV residing in BC, as per BC’s HIV therapeutic guidelines (British Columbia Centre for Excellence in HIV/AIDS 2018). HCV screening is recommended at HIV diagnosis, and annually for those with ongoing HCV risk behaviours or unexplained elevated liver enzymes (British Columbia Centre for Excellence in HIV/AIDS 2018).
Study data
Study data were obtained from the BC Seek and Treat for Optimal Prevention of HIV/AIDS (STOP HIV/AIDS) population-based cohort, which is comprised of longitudinal individual-level data on all diagnosed PLW-HIV in BC (Heath et al. 2014). This cohort is derived from a series of linkages between the DTP and several provincial databases; details and data steward references are presented in the Online Resource. The DTP captures nearly 85% of all CD4 counts and 100% of viral load tests. All viral load tests and most CD4 counts are performed at St. Paul’s Hospital laboratory, which is linked to the DTP. CD4 tests from other laboratories are only obtained via prescription forms, in which capture is incomplete.
Study design
A retrospective cohort study design was used. The eligibility criteria were as follows: (i) ART-naïve individuals aged ≥ 18 years; (ii) enrolled in the STOP HIV/AIDS cohort and initiated ART between January 1, 2000, and December 31, 2014; (iii) initiated ART on a regimen involving two nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) as backbone, plus either a non-nucleoside reverse transcriptase inhibitor (NNRTI) or a ritonavir-boosted protease inhibitor (bPI); (iv) had at least one CD4 cell count and a viral load test performed within 6 months of ART treatment; and (v) had been followed for at least 6 months. Note that those initiating treatment on other ART regimens were excluded due to insufficient numbers (Online Resource Table 1). Eligible individuals were followed until December 31, 2015, the last contact date (i.e., the last available laboratory test date, the last filled ART prescription refill date or physician visit date), or the date of death.
Outcome definition
The outcome was acute care hospitalization rate, calculated every 6-month interval, per individual. Rates were calculated by dividing the number of acute care hospitalizations by the number of person-years (PY) of follow-up, and were presented per 100PY. Thus, the events of acute care hospitalizations were combined from baseline (i.e., ART initiation) until the end of follow-up, as this was the time scale of this study.
Acute care hospitalizations were identified from the provincial Discharge Abstract Database (DAD) which captures all discharges, transfers, and deaths of inpatients from acute care hospitals. An acute care hospitalization is defined as an active, short-term admission requiring facility-based overnight stay and the presence of one of the following: (i) treatment for severe injury/illness, urgent medical or psychiatric conditions or recovery from surgery; (ii) around-the-clock care and monitoring; (iii) services required at a minimum level of frequency or intensity (e.g., clinical monitoring ≥ 3 times daily); (iv) access to diagnostic assessments required for care. Transfers of hospitalized individuals between acute care institutions were counted as a single hospitalization. Although visits to the emergency room are not recorded within the DAD database, individuals who were admitted to acute inpatient care via the emergency room were included in these analyses.
Analytical approach for the exposure variable definition
Eligible individuals were assigned an HIV/HCV co-infection status (yes/no) and an MHD status (yes/no) in accordance with their respective diagnoses’ status preceding/on the date of the first occurrence of an acute care hospitalization (i.e., the outcome event). For the purpose of this study, an interaction refers to the notion of a statistical interaction, which is defined as “a two-tailed hypothesis that two or more concepts “work together”/“have a combined effect” upon a third” (Hall and Sammons 2013). Thus, the syndemic of HIV/HCV co-infection and MHD among PLW-HIV was assessed using a four-level interaction defined by the product of HIV/HCV co-infection status and MHD status: (i) PLW-HIV without MHD, (ii) PLW-HIV/HCV without MHD, (iii) PLW-HIV with MHD, and (iv) PLW-HIV/HCV with MHD.
To ensure that the exposure reflects the status of the individual prior to/on the date of the first occurrence of the outcome event, the following strict set of rules were implemented: (i) diagnosis status must be known; (ii) diagnosis dates must be available; (iii) diagnosis dates must precede/be on the same date of the first acute care hospitalization admission. Thus, eligible individuals with an unknown diagnosis status, unknown diagnosis dates, or diagnosis dates succeeding the date of the first occurrence of an acute hospitalization admission were excluded from these analyses.
HIV/HCV co-infection diagnoses were determined by HCV antibody or RNA detection, or by physician report. MHD diagnoses were ascertained by applying a case-finding algorithm utilizing the International Classification of Diseases, Ninth and Tenth Revisions (ICD 9 and 10) coding system to the Medical Service Plan (MSP) (medical visits) and DAD (hospitalizations) databases (Online Resource Table 2). MHD were classified into four categories: (i) anxiety disorders, (ii) mood disorders, (iii) personality disorders, and (iv) schizophrenia-related disorders.
Study covariates and confounders
Time-fixed covariates measured at baseline comprised sex (male and female), HIV risk (gay, bisexual, and other men who have sex with men [gbMSM]; PWID; gbMSM/PWID; other; and unknown), and ART regimen (NNRTI and bPI). To account for differences in the efficacy and tolerability of antiretroviral medication over time, calendar year of ART initiation (continuous) or categorical (2000–2003, 2004–2007, 2008–2011, and 2012–2014) was considered.
Several time-varying (i.e., longitudinally measured) covariates, measured every 6-month interval, included age (< 30, 30–39, 40–49, and ≥ 50 years), the number of comorbidities excluding MHD, CD4 cell count (< 50, 50–199, 200–349, and ≥ 350 cells/mm3), viral load (log10 transformed), and adherence level (< 40%, 40–79%, 80–94%, and ≥ 95%). Covariates measured at the end of follow-up consisted of PY of follow-up time (continuous) and a continuous variable indicating which 6-month intervals the hospitalizations occurred (hereafter referred to as the 6-month interval index). This 6-month interval index was included to adjust for secular trends in acute care hospitalizations.
Adherence level was determined by calculating the cumulative amount of days of ART dispensed divided by the cumulative amount of days of follow-up (presented as a percentage). Adherence estimates were calculated based on varying regimen exposure per individual (Lima et al. 2010; Lima et al. 2015b). Comorbidities (i.e., 14 distinct conditions) were derived from a modified Charlson Comorbidity Index (Online Resource Table 3) (Quan et al. 2011). Modifications consisted of excluding HIV and liver disease as comorbidities due to the exposure variable comprising both HIV and HCV. Diabetes with and without complications were combined into a single comorbidity due to data capture limitations. CD4 counts were measured by flow cytometry, followed by fluorescent monoclonal antibody analysis (Beckman Coulter, Inc., Mississauga, Ontario, Canada). To account for evolving viral load testing technology over time, we re-coded our measurements to range from < 50 (coded as 49) to > 100,000 (coded as 100,010) copies/mL.
Statistical analyses
The acute care hospitalization rate’s 95% confidence intervals (CI) were determined using Fisher’s exact test, while age-standardized rates for the 15-year study period were calculated using BC’s population estimates for 2010 as the reference (BC Stats 2017). Differences in the distribution among categorical variables were assessed using Fisher’s exact test or the χ2 test, while continuous variables were assessed using the Kruskal-Wallis test (McDonald 2009). Both confounder and explanatory generalized non-linear mixed-effects models were constructed, assuming a zero-inflated negative binomial distribution, the PY of follow-up time as the offset, a log link function, and a random intercept term (Brooks et al. 2017). These models were selected to accommodate the observed over-dispersion of the data through the negative binomial component, and to appropriately model the excess zero proportion in the longitudinal data resulting from the repeated intervals in which zero acute care hospitalizations occurred.
A backward-selection approach, published by our group based on the work by Maldonado and Greenland (1993), selected the potential confounders included in the confounder model built to assess the effect of HIV/HCV co-infection and MHD on the acute care hospitalization rate among PLW-HIV. This approach considers the magnitude of change in the coefficient of the exposure variable; potential confounders were dropped one at a time, using the relative change in the coefficient for the exposure variable as a criterion, until the maximum change from the full model surpassed 5% (Lima et al. 2007). Note that HIV risk (e.g., PWID) was not adjusted for in this confounder model due to high collinearity with the main exposure.
We built separate explanatory multivariable models, one for each exposure category to characterize longitudinal trends of acute care hospitalization rates, while controlling for the effect of several demographic and clinical factors. These models were used to estimate the predicted probability of acute care hospitalization (for every 6-month interval) from 2000 to 2015. Model selection was conducted using a backward elimination procedure based on the Akaike Information Criterion (AIC) and Type III P value (Lima et al. 2010). The significance level was set at 0.05. All analyses were performed using R© 3.2.2 glmmTMB package and SAS version 9.4 (SAS, Cary, NC).
Results
Study population characteristics
Initially, 4701 ART-naïve individuals met the eligibility criteria, with a total of 8921 acute care hospitalizations. Of these eligible individuals, 655 (14%) were subsequently excluded; 243 of which had an unknown HCV status, 248 had an HCV diagnosis date succeeding the first occurrence of an acute care hospitalization, 15 had an unknown HCV diagnosis date, and 149 had an MHD diagnosis following the first occurrence of an acute care hospitalization. These excluded individuals contributed 2770 acute care hospitalizations. Table 1 compares the characteristics between the analytical sample and the excluded individuals.
Table 1.
Study population characteristics comparison between the analytical sample and the excluded individuals
| Variables | Total eligible individuals N, % |
Included individuals N, % |
Excluded individuals N, % |
P value |
|---|---|---|---|---|
| 4701 (100) | 4046 (86) | 655 (14) | ||
| HIV/HCV co-infection and MHD status | < 0.0001 | |||
| PLW-HIV without MHD | 1597 (34) | 1597 (100) | 0 (0) | |
| PLW-HIV/HCV without MHD | 686 (15) | 606 (88) | 80 (12) | |
| PLW-HIV with MHD | 1059 (23) | 988 (93) | 71 (7) | |
| PLW-HIV/HCV without MHD | 1116 (24) | 855 (77) | 261 (23) | |
| Unknown HCV status | 243 (5) | 0 (0) | 243 (100) | |
| Anxiety disorder | 0.0003 | |||
| Yes | 927 (20) | 763 (82) | 164 (18) | |
| Mood disorder | < 0.0001 | |||
| Yes | 1780 (38) | 1477 (83) | 303 (17) | |
| Personality disorder | 0.0150 | |||
| Yes | 539 (11) | 445 (83) | 94 (17) | |
| Schizophrenia-related disorder | 0.0019 | |||
| Yes | 340 (7) | 273 (80) | 67 (20) | |
| Number of MHD at the end of follow-up, n (%) | < 0.0001 | |||
| 0 | 2419 (51) | 2203 (91) | 216 (9) | |
| 1 | 1125 (24) | 884 (79) | 241 (21) | |
| 2 | 679 (14) | 564 (83) | 115 (17) | |
| 3 | 303 (6) | 256 (84) | 47 (16) | |
| 4 | 175 (4) | 139 (79) | 36 (21) | |
| Anxiety disorder (at end of follow-up) | < 0.0001 | |||
| Yes | 1052 (22) | 855 (81) | 197 (19) | |
| Mood disorder (at end of follow-up) | < 0.0001 | |||
| Yes | 1892 (40) | 1561 (83) | 331 (17) | |
| Personality disorder (at end of follow-up) | < 0.0001 | |||
| Yes | 638 (14) | 524 (82) | 114 (18) | |
| Schizophrenia-related disorder (at end of follow-up) | < 0.0001 | |||
| Yes | 510 (11) | 396 (78) | 114 (22) | |
| Sex, n (%) | < 0.0001 | |||
| Male | 3789 (81) | 3312 (87) | 477 (13) | |
| Female | 912 (19) | 734 (80) | 178 (20) | |
| Age at ART initiation (years), n (%) | 0.5375 | |||
| < 30 | 629 (13) | 533 (85) | 96 (15) | |
| 30–39 | 1443 (31) | 1234 (86) | 209 (14) | |
| 40–49 | 1650 (35) | 1432 (87) | 218 (13) | |
| ≥ 50 | 979 (21) | 847 (87) | 132 (13) | |
| HIV risk, n (%) | < 0.0001 | |||
| gbMSM | 1041 (22) | 961 (92) | 80 (8) | |
| PWID | 1415 (30) | 1186 (84) | 229 (16) | |
| gbMSM/PWID | 254 (5) | 206 (81) | 48 (19) | |
| Other | 422 (9) | 371 (88) | 51 (12) | |
| Unknown | 1569 (33) | 1322 (84) | 247 (16) | |
| ART era, n (%) | < 0.0001 | |||
| 2000–2003 | 875 (19) | 664 (76) | 211 (24) | |
| 2004–2007 | 1223 (26) | 1032 (84) | 191 (16) | |
| 2008–2011 | 1750 (37) | 1569 (90) | 181 (10) | |
| 2012–2014 | 853 (18) | 781 (92) | 72 (8) | |
| Baseline CD4 cell count (cells/mm3), n (%) | < 0.0001 | |||
| ≥ 350 | 1278 (27) | 1154 (90) | 124 (10) | |
| 200–349 | 1433 (30) | 1244 (87) | 189 (13) | |
| 50–199 | 1455 (31) | 1220 (84) | 235 (16) | |
| < 50 | 535 (11) | 428 (80) | 107 (20) | |
| Adherence (first 6 months), n (%) | < 0.0001 | |||
| ≥ 95% | 3683 (78) | 3237 (88) | 446 (12) | |
| 80–94% | 239 (5) | 197 (82) | 42 (18) | |
| 40–79% | 527 (11) | 423 (80) | 104 (20) | |
| < 40% | 252 (5) | 189 (75) | 63 (25) | |
| Number of comorbidities at baseline, n (%) | 0.0001 | |||
| 0 | 2454 (52) | 2164 (88) | 290 (12) | |
| 1 | 1410 (30) | 1181 (84) | 229 (16) | |
| 2 | 537 (11) | 455 (85) | 82 (15) | |
| ≥ 3 | 300 (6) | 246 (82) | 54 (18) | |
| Initial ART regimen, n (%) | 0.7404 | |||
| NNRTI | 2229 (47) | 1914 (86) | 315 (14) | |
| bPI | 2472 (53) | 2132 (86) | 340 (14) | |
| Number of acute care hospitalizations, n (%) | < 0.0001 | |||
| 0 | 2459 (52) | 2298 (93) | 161 (7) | |
| 1 | 815 (17) | 693 (85) | 122 (15) | |
| 2 | 451 (10) | 364 (81) | 87 (19) | |
| 3 | 256 (5) | 196 (77) | 60 (23) | |
| ≥ 4 | 720 (15) | 495 (69) | 225 (31) | |
| Baseline viral load (log10 copies/mL), median (Q1–Q3) | 4.88 (4.37–5.00) | 4.87 (4.35–5) | 4.98 (4.47–5.00) | 0.0230 |
| Follow-up time (years), median (Q1–Q3) | 5.98 (3.49–8.97) | 5.98 (3.49–8.97) | 6.48 (3.49–9.96) | 0.0204 |
| 6-month interval index, median (Q1–Q3) | 12 (7–18) | 12 (7–18) | 13 (7–20) | 0.0202 |
Q1–Q3, 25th–75th percentiles; gbMSM, gay, bisexual, and other men who have sex with men; PWID, people who have ever injected drugs; HCV, hepatitis C; MHD, mental health disorders; PLW-HIV, people living with HIV; PLW-HIV/HCV, people living with HIV and HCV; ART, antiretroviral therapy; NNRTI, non-nucleoside reverse transcriptase inhibitor; bPI, ritonavir-boosted protease inhibitor
Among the 4046 individuals included in the analytical sample, 6151 acute care hospitalizations occurred (Online Resource Table 4). Overall, 1597 (39%) were PLW-HIV without MHD, 606 (15%) were PLW-HIV/HCV without MHD, 988 (24%) were PLW-HIV with MHD, and 855 (21%) were PLW-HIV/HCV with MHD. Table 2 presents the demographic and clinical characterization of the study population by exposure category. Among the 1843 individuals with MHD preceding/at the time of the first occurrence of the outcome event, 763 (41%) had anxiety disorders, 1477 (80%) had mood disorders, 445 (24%) had personality disorders, and 273 (15%) had schizophrenia-related disorders. A total of 764 (41%) had co-occurring MHD (Online Resource Table 5).
Table 2.
Study population characteristics stratified by PLW-HIV without MHD, PLW-HIV/HCV without MHD, PLW-HIV with MHD, and PLW-HIV/HCV with MHD
| Variables | Whole sample N, % |
PLW-HIV without MHD N, % |
PLW-HIV/HCV without MHD N, % |
PLW-HIV with MHD N, % |
PLW-HIV/HCV with MHD N, % |
P value |
|---|---|---|---|---|---|---|
| 4046 (100) | 1597 (39) | 606 (15) | 988 (24) | 855 (21) | ||
| HIV/HCV co-infection, n (%) | < 0.0001 | |||||
| Yes | 1461 (36) | 0 (0) | 606 (41) | 0 (0) | 855 (59) | |
| MHD, n (%) | < 0.0001 | |||||
| Yes | 1843 (46) | 0 (0) | 0 (0) | 988 (54) | 855 (46) | |
| Anxiety disorder | < 0.0001 | |||||
| Yes | 763 (19) | 0 (0) | 0 (0) | 347 (45) | 416 (55) | |
| Mood disorder count | < 0.0001 | |||||
| Yes | 1477 (37) | 0 (0) | 0 (0) | 815 (55) | 662 (45) | |
| Personality disorder | < 0.0001 | |||||
| Yes | 445 (11) | 0 (0) | 0 (0) | 146 (33) | 299 (67) | |
| Schizophrenia-related disorder | < 0.0001 | |||||
| Yes | 273 (7) | 0 (0) | 0 (0) | 88 (32) | 185 (68) | |
| Number of MHD at end of follow-up, n (%) | < 0.0001 | |||||
| 0 | 2203 (54) | 1597 (72) | 606 (28) | 0 (0) | 0 (0) | |
| 1 | 884 (22) | 0 (0) | 0 (0) | 575 (65) | 309 (35) | |
| 2 | 564 (14) | 0 (0) | 0 (0) | 279 (49) | 285 (51) | |
| 3 | 256 (6) | 0 (0) | 0 (0) | 94 (37) | 162 (63) | |
| 4 | 139 (3) | 0 (0) | 0 (0) | 40 (29) | 99 (71) | |
| Anxiety disorder (at end of follow-up) | < 0.0001 | |||||
| Yes | 855 (21) | 0 (0) | 0 (0) | 396 (46) | 459 (54) | |
| Mood disorder (at end of follow-up) | < 0.0001 | |||||
| Yes | 1561 (39) | 0 (0) | 0 (0) | 850 (54) | 711 (46) | |
| Personality disorder (at end of follow-up) | < 0.0001 | |||||
| Yes | 524 (13) | 0 (0) | 0 (0) | 181 (35) | 343 (65) | |
| Schizophrenia-related disorder (at end of follow-up) | < 0.0001 | |||||
| Yes | 396 (10) | 0 (0) | 0 (0) | 148 (37) | 248 (63) | |
| Sex, n (%) | < 0.0001 | |||||
| Male | 3312 (82) | 1428 (43) | 484 (15) | 849 (26) | 551 (17) | |
| Female | 734 (18) | 169 (23) | 122 (17) | 139 (19) | 304 (41) | |
| Age at baseline (years), n (%) | < 0.0001 | |||||
| < 30 | 533 (13) | 255 (48) | 58 (11) | 128 (24) | 92 (17) | |
| 30–39 | 1234 (30) | 507 (41) | 188 (15) | 288 (23) | 251 (20) | |
| 40–49 | 1432 (35) | 493 (34) | 257 (18) | 343 (24) | 339 (24) | |
| ≥ 50 | 847 (21) | 342 (40) | 103 (12) | 229 (27) | 173 (20) | |
| HIV risk, n (%) | < 0.0001 | |||||
| gbMSM | 961 (24) | 526 (55) | 57 (6) | 342 (36) | 36 (4) | |
| PWID | 1186 (29) | 52 (4) | 405 (34) | 68 (6) | 661 (56) | |
| gbMSM/PWID | 206 (5) | 37 (18) | 43 (21) | 53 (26) | 73 (35) | |
| Other | 371 (9) | 213 (57) | 19 (5) | 121 (33) | 18 (5) | |
| Unknown | 1322 (33) | 769 (58) | 82 (6) | 404 (31) | 67 (5) | |
| ART era, n (%) | < 0.0001 | |||||
| 2000–2003 | 664 (16) | 234 (35) | 135 (20) | 144 (22) | 151 (23) | |
| 2004–2007 | 1032 (26) | 352 (34) | 167 (16) | 255 (25) | 258 (25) | |
| 2008–2011 | 1569 (39) | 609 (39) | 237 (15) | 398 (25) | 325 (21) | |
| 2012–2014 | 781 (19) | 402 (51) | 67 (9) | 191 (24) | 121 (15) | |
| Baseline CD4 cell count (cells/mm3), n (%) | < 0.0001 | |||||
| ≥ 350 | 1154 (29) | 502 (44) | 118 (10) | 341 (30) | 193 (17) | |
| 200–349 | 1244 (31) | 477 (38) | 180 (14) | 314 (25) | 273 (22) | |
| 50–199 | 1220 (30) | 439 (36) | 227 (19) | 234 (19) | 320 (26) | |
| < 50 | 428 (11) | 179 (42) | 81 (19) | 99 (23) | 69 (16) | |
| Last CD4 cell count (cells/mm3), n (%) | < 0.0001 | |||||
| ≥ 350 | 2840 (70) | 1212 (43) | 344 (12) | 771 (27) | 513 (18) | |
| 200–349 | 464 (11) | 167 (36) | 104 (22) | 84 (18) | 109 (23) | |
| 50–199 | 266 (7) | 68 (26) | 80 (30) | 29 (11) | 89 (33) | |
| < 50 | 87 (2) | 24 (28) | 25 (29) | 9 (10) | 29 (33) | |
| Unknown | 389 (10) | 126 (32) | 53 (14) | 95 (24) | 115 (30) | |
| Adherence (first 6 months), n (%) | < 0.0001 | |||||
| ≥ 95% | 3237 (80) | 1397 (43) | 429 (13) | 826 (26) | 585 (18) | |
| 80–94% | 197 (5) | 54 (27) | 37 (19) | 53 (27) | 53 (27) | |
| 40–79% | 423 (10) | 113 (27) | 93 (22) | 78 (18) | 139 (33) | |
| < 40% | 189 (5) | 33 (17) | 47 (25) | 31 (16) | 78 (41) | |
| Adherence (last 6 months), n (%) | < 0.0001 | |||||
| ≥ 95% | 2845 (70) | 1211 (43) | 355 (12) | 757 (27) | 522 (18) | |
| 80–94% | 288 (7) | 95 (33) | 63 (22) | 63 (22) | 67 (23) | |
| 40–79% | 493 (12) | 152 (31) | 95 (19) | 97 (20) | 149 (30) | |
| < 40% | 420 (10) | 139 (33) | 93 (22) | 71 (17) | 117 (28) | |
| Number of comorbidities at baseline, n (%) | < 0.0001 | |||||
| 0 | 2164 (53) | 994 (46) | 376 (17) | 480 (22) | 314 (15) | |
| 1 | 1181 (29) | 421 (36) | 159 (13) | 319 (27) | 282 (24) | |
| 2 | 455 (11) | 129 (28) | 50 (11) | 129 (28) | 147 (32) | |
| ≥ 3 | 246 (6) | 53 (22) | 21 (9) | 60 (24) | 112 (46) | |
| Number of comorbidities at end of follow-up, n (%) | < 0.0001 | |||||
| 0 | 1445 (36) | 721 (50) | 255 (18) | 300 (21) | 169 (12) | |
| 1 | 1265 (31) | 497 (39) | 188 (15) | 325 (26) | 255 (20) | |
| 2 | 688 (17) | 220 (32) | 98 (14) | 189 (27) | 181 (26) | |
| ≥ 3 | 648 (16) | 159 (25) | 65 (10) | 174 (27) | 250 (39) | |
| Initial ART regimen, n (%) | < 0.0001 | |||||
| NNRTI | 1914 (47) | 788 (41) | 337 (18) | 414 (22) | 375 (20) | |
| bPI | 2132 (53) | 809 (38) | 269 (13) | 574 (27) | 480 (23) | |
| Total number of acute care hospitalizations, n (%) | < 0.0001 | |||||
| 0 | 2298 (57) | 1127 (49) | 301 (13) | 584 (25) | 286 (12) | |
| 1 | 693 (17) | 239 (34) | 109 (16) | 181 (26) | 164 (24) | |
| 2 | 364 (9) | 99 (27) | 66 (18) | 91 (25) | 108 (30) | |
| 3 | 196 (5) | 50 (26) | 33 (17) | 42 (21) | 71 (36) | |
| ≥ 4 | 495 (12) | 82 (17) | 97 (20) | 90 (18) | 226 (46) | |
| Baseline viral load (log10 copies/mL), median (Q1–Q3) | 4.87 (4.35–5.00) | 4.89 (4.38–5.00) | 4.81 (4.34–5.00) | 4.9 (4.37–5.00) | 4.81 (4.31–5.00) | 0.0112 |
| Last viral load (log10 copies/mL), median (Q1–Q3) | 1.54 (1.54–1.54) | 1.54 (1.54–1.54) | 1.54 (1.54–1.75) | 1.54 (1.54–1.54) | 1.54 (1.54–1.65) | < 0.0001 |
| Follow-up time (years), median (Q1–Q3) | 5.98 (3.49–8.97) | 5.48 (2.99–8.47) | 5.98 (3.98–9.46) | 5.98 (3.98–9.46) | 5.98 (3.98–8.97) | 0.00013 |
| 6-month interval index, median (Q1–Q3) | 12 (7–18) | 11 (6–17) | 12 (8–19) | 12 (8–19) | 12 (8–18) | 0.00012 |
Q1–Q3, 25th–75th percentiles; gbMSM, gay, bisexual, and other men who have sex with men; HCV, hepatitis C; MHD, mental health disorders; PLW-HIV, people living with HIV; PLW-HIV/HCV, people living with HIV and hepatitis C; PWID, people who have ever injected drugs; ART, antiretroviral therapy; NNRTI, non-nucleoside reverse transcriptase inhibitor; bPI, ritonavir-boosted protease inhibitor
Reasons for acute care hospitalizations
The prominent causes of hospitalizations were diseases and disorders of blood, blood forming organs, and immunological disorders (25%); lymphoma, leukemia, or unspecified site neoplasms (18%); diseases and disorders of the digestive system (8%); disease and disorders of the respiratory system (8%); and mental disease and disorders (6%). Online Resource Table 6 presents the five leading causes of acute care hospitalizations for each exposure category.
Age-standardized acute care hospitalization rates
The exacerbation of the age-standardized acute care hospitalization rate was dependent on the complexity of the syndemic (Fig. 1). This is particularly exemplified by the fact that the highest rate was observed among PLW-HIV/HCV with MHD who were PWID (56/100PY [95% CI 54–59]). Further exploratory analyses demonstrated that individuals with schizophrenia-related disorders had remarkably higher rates relative to those with other MHD (data not shown).
Fig. 1.
Venn diagram demonstrating the syndemic of HIV/HCV co-infection and MHD, including the influence of PWID on the 15-year age-standardized acute care hospitalization rate (per 100PY) among PLW-HIV in British Columbia from 2000 to 2015. PLW-HIV, people living with HIV, PWID, people who have ever injected drugs, HCV, hepatitis C virus, MHD, mental health disorders, PY, person-years, CI, confidence interval
Multivariable models
Our adjusted confounder model demonstrated that the adjusted rate ratios for acute care hospitalizations were 1.31 (95% CI 1.13–1.52), 2.01 (95% CI 1.71–2.36), and 2.53 (95% CI 2.20–2.92) for PLW-HIV with MHD, PLW-HIV/HCV without MHD, and PLW-HIV/HCV with MHD, respectively, relative to PLW-HIV without MHD. Confounders selected in the adjusted model were sex, time-varying CD4 cell count, time-varying ART adherence, time-varying number of comorbidities, and the 6-month interval index.
For each exposure category, a fully adjusted explanatory multivariable model was built to estimate the predicted probability of acute care hospitalization (for every 6-month interval) from 2000 to 2015 (Table 3; Fig. 2; Online Resource Table 7). PLW-HIV/HCV with MHD had the greatest median predicted probability over the study period, by a considerable margin; however, a downward trend was observed (0.140 to 0.107; 24% decrease). Comparatively, the remaining exposure categories exhibited greater reductions in the aforementioned predicted probability: PLW-HIV without MHD (0.030 to 0.014; 55% decrease), PLW-HIV/HCV without MHD (0.093 to 0.046; 51% decrease), and PLW-HIV with MHD (0.061 to 0.031; 55% decrease).
Table 3.
Multivariable explanatory models built to estimate the predicted probability for the acute care hospitalization among PLW-HIV without MHD, PLW-HIV/HCV without MHD, PLW-HIV with MHD, and PLW-HIV/HCV with MHD
| Variables | PLW-HIV without MHD Final model aRR (95% CI) |
PLW-HIV/HCV without MHD Final model aRR (95% CI) |
PLW-HIV with MHD Final model aRR (95% CI) |
PLW-HIV/HCV with MHD Final model aRR (95% CI) |
|---|---|---|---|---|
| Sex | ||||
| Male | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) |
| Female | 2.58 (1.87–3.55) | 1.76 (1.32–2.36) | 1.41 (0.99–2.00) | 1.25 (1.02–1.53) |
| Age (time-varying) | ||||
| < 30 | NS | NS | 1.43 (0.98–2.10) | 1.76 (1.32–2.34) |
| 30–39 | 1.42 (1.11–1.82) | 1.30 (1.09–1.54) | ||
| 40–49 | 1.00 (REF) | 1.00 (REF) | ||
| ≥ 50 | 1.02 (0.81–1.27) | 0.86 (0.73–1.02) | ||
| CD4 (time-varying) | ||||
| ≥ 350 | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) |
| 200–349 | 1.39 (1.12–1.72) | 1.39 (1.12–1.72) | 1.13 (0.91–1.40) | 1.26 (1.09–1.45) |
| 50–199 | 2.06 (1.59–2.66) | 1.97 (1.56–2.49) | 1.87 (1.42–2.46) | 1.59 (1.35–1.88) |
| < 50 | 4.65 (3.29–6.56) | 3.46 (2.45–4.89) | 4.29 (2.93–6.26) | 2.89 (2.26–3.69) |
| Unknown | 0.83 (0.55–1.25) | 1.36 (0.88–2.08) | 0.83 (0.59–1.18) | 0.95 (0.77–1.18) |
| Adherence (time-varying) | ||||
| ≥ 95% | 1.00 (REF) | NS | 1.00 (REF) | 1.00 (REF) |
| 80–94% | 1.17 (0.85–1.60) | 1.65 (1.29–2.11) | 1.26 (1.07–1.49) | |
| 40–79% | 1.64 (1.32–2.04) | 1.42 (1.15–1.75) | 1.36 (1.20–1.54) | |
| < 40% | 1.33 (0.98–1.81) | 0.97 (0.72–1.32) | 1.20 (1.03–1.41) | |
| Number of comorbidities (time-varying) | ||||
| 0 | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) |
| 1 | 2.18 (1.73–2.76) | 2.07 (1.63–2.62) | 1.62 (1.25–2.10) | 1.33 (1.08–1.63) |
| 2 | 5.44 (4.12–7.19) | 3.96 (2.99–5.24) | 2.62 (1.98–3.48) | 2.56 (2.06–3.18) |
| ≥ 3 | 13.61 (10.00–18.52) | 6.40 (4.60–8.90) | 5.34 (3.93–7.26) | 4.32 (3.45–5.42) |
| HIV risk | ||||
| gbMSM | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) | 1.00 (REF) |
| Other | 0.71 (0.49–1.00) | 1.19 (0.42–3.32) | 1.06 (0.72–1.57) | 1.72 (0.64–4.62) |
| PWID | 1.48 (0.89–2.47) | 3.90 (2.05–7.40) | 1.55 (0.99–2.42) | 4.06 (2.15–7.66) |
| MSM/PWID | 1.92 (1.08–3.43) | 3.27 (1.50–7.14) | 2.20 (1.39–3.50) | 4.49 (2.25–8.97) |
| Unknown | 0.94 (0.74–1.21) | 3.10 (1.51–6.40) | 0.73 (0.55–0.97) | 2.31 (1.10–4.87) |
| Calendar year of ART initiation | NS | NS | NS | NS |
| 6-month interval index | 0.99 (0.97–1.00) | NS | NS | NS |
| Initial regimen | ||||
| NNRTI | NS | NS | NS | NS |
| bPI | ||||
| Viral load (log10 copies/mL) (time-varying) | 1.29 (1.21–1.38) | 1.23 (1.15–1.30) | 1.18 (1.10–1.26) | 1.08 (1.04–1.13) |
aRR, adjusted rate ratio; CI, confidence interval; PLW-HIV, people living with HIV; PLW-HIV/HCV, people living with HIV and hepatitis C; MHD, mental health disorders; ART, antiretroviral therapy; gbMSM, gay, bisexual, and other men who have sex with men; PWID, people who inject drugs; NNRTI, non-nucleoside reverse transcriptase inhibitor; bPI, ritonavir-boosted protease inhibitor; NS, not selected
Fig. 2.
Predicted probability of acute care hospitalization (within a 6-month interval) among PLW-HIV in British Columbia from 2000 to 2015 by exposure category. The predicted probability is based on the fully adjusted explanatory models. MHD, mental health disorders, PLW-HIV, people living with HIV, PLW-HIV/HCV, people living with HIV and hepatitis C
Discussion
To our knowledge, this large population-based cohort study represents one of the first to quantify the inpatient healthcare utilization attributable to the syndemic of HIV/HCV co-infection and MHD among PLW-HIV by characterizing acute care hospital discharge rates. We demonstrated that PLW-HIV with MHD, PLW-HIV/HCV without MHD, and PLW-HIV/HCV with MHD had an increased rate of acute care hospitalization by 31%, 101%, and 153%, respectively, compared with PLW-HIV without MHD, after adjusting for several confounders.
Additional key predisposing and modifiable factors were identified as important targets for interventions aimed at alleviating the excess of acute care hospitalizations aside from HIV/HCV co-infection and MHD. These included (i) history of injection drug use, (ii) lower ART adherence level, (iii) initiating treatment late in the course of HIV infection, and (iv) high multimorbidity.
The steady decrease observed in the secular trends of the predicted probability of acute care hospitalization may be reflective of BC’s multi-phase response to HIV/AIDS: (i) the harm reduction and health service scale-up phase (2000–2005), (ii) the early Treatment as Prevention phase (2006–2009), and (iii) the STOP HIV/AIDS phase (2010–present) (Olding et al. 2017). The last phase was particularly instrumental as BC’s HIV therapeutic guidelines recommended ART treatment, regardless of CD4 count (Olding et al. 2017).
However, disparities in the acute care hospitalizations among the distinct categories of the exposure variable continue to persist, particularly when comparing PLW-HIV/HCV with MHD with the other categories. This may be in part related to the larger proportions of individuals with more severe and debilitating MHD, namely personality disorders and schizophrenia-related disorders relative to the PLW-HIV with MHD category. Further, the PLW-HIV/HCV with MHD exposure category had a greater proportion of individuals diagnosed with multiple co-occurring MHD. Although only 7% of admissions were primarily due to MHD, it is entirely plausible that these disorders may have played a secondary role in admissions related to other causes, and clinical liaison with psychiatric multidisciplinary teams may have been utilized.
To fully address the unique health needs of individuals affected by the syndemic of HIV/HCV co-infection and MHD, expanded accessibility, availability, and coordination of treatment should be prioritized across the province. The integration of care model successfully employed by the Dr. Peter Centre in Vancouver, BC, highlights the feasibility of a cohesive approach designed for PLW-HIV who face multiple barriers to optimal health, including homelessness, mental health, and addiction (Jeal 2015). Dissemination of integrative care model best practices in various settings has been published and may be applicable to BC.
Readers may be reflecting on whether integrating HIV, HCV, and MHD care remains relevant at this current time, given that our study period concluded in 2015. However, as recently as 2018, the BC College of Family Physicians asserted that a need remains for integrating quality mental health and addictions care into primary and community care accessible to all BC residents, particularly in rural and remote settings (British Columbia College of Family Physicians 2018), thus rendering the current study’s findings timely and pertinent.
It can be argued that since the individuals in this study have been linked to care through HIV treatment, the severity of this syndemic is likely underestimated. Our findings can shed light on the increased magnitude of this syndemic in the developing world, particularly where the epidemics of injection drug use and HIV are interlaced, and a simultaneous lack of resources for mental health and harm reduction to properly address this crisis persist (Vlahov et al. 2010).
Note that our study period does not extend beyond 2015 due to the unavailability of data; as a result, it does not encompass the dramatic increase in opioid overdoses subsequent to 2015. Up to 2014, HCV treatment consisted of interferon-based therapy, which had limited efficacy and uptake due to tolerability issues and a restrictive treatment coverage criteria (Janjua et al. 2016). The emergence of direct-acting antivirals in 2014 dramatically changed the landscape of HCV treatment with cure rates reaching 95% (Janjua et al. 2016). Effective March 2018, coverage of direct-acting antivirals expanded to all BC residents living with HCV, regardless of disease stage, setting the stage for the implementation of a Treatment as Prevention (TasP) strategy (British Columbia Ministry of Health 2018).
Biological sex considerations
It has been recognized that biological sex may modify the response to ART, the natural history of HCV, and the development of MHD via a multifaceted interaction between physiological factors (e.g., genetic, hormonal, immunological factors) (Marcus et al. 2015; Green et al. 2018). In light of these reflections, biological sex was considered and adjusted for in our models assessing the effect of our main exposure on the rate of acute care hospitalizations.
Limitations
Our study should be interpreted in light of a number of limitations. First, HIV/HCV co-infection status cannot ascertain that individuals had active HCV infection throughout the entirety of the follow-up period, given that we were unable to determine and adjust for the rates of spontaneous viral clearance, HCV treatment, and re-infection due to data unavailability. Second, though individuals often present to care with comorbid MHD, only the primary diagnosis within a specified physician visit is recorded in the MSP database. Thus, the number of MHD diagnoses may have been underestimated. Third, the STOP HIV/AIDS cohort did not contain data on emergency room visits; therefore, these analyses were confined to acute care hospitalizations. Fourth, while there is no gold standard for measuring ART adherence, our refill compliance measurement has been previously shown to be associated with different outcomes (Lima et al. 2010; Lima et al. 2015b). Finally, healthcare administrative data are susceptible to coding error.
Conclusion
The HIV/HCV co-infection and MHD interaction term demonstrated a significant effect on the rate of acute care hospitalization at each level of the interaction, particularly for PLW-HIV/HCV with MDH. Additional key modifiable factors were identified as important targets for interventions. This excess in disease burden indicates a need to shift away from a fragmented model of service delivery to an integrated system, a viewpoint shared by BC College of Family Physicians. Implementing widely accessible integrative care model best practices may be meaningful in addressing this public health challenge.
Electronic supplementary material
(DOCX 75 kb)
Funding
This work was supported by the following sources of funding: JSGM’s Treatment as Prevention (TasP) research, paid to his institution, has received support from the Public Health Agency of Canada, the British Columbia Ministry of Health, and the US National Institutes of Health (R01DA036307 and CTN 248). VDL is funded by a grant from the Canadian Institutes of Health Research (PJT-148595), by a Scholar Award from the Michael Smith Foundation for Health Research and a New Investigator award from the Canadian Institutes of Health Research. The sponsors had no role in the design, data collection, data analysis, data interpretation, or writing of the report.
Compliance with ethical standards
Conflict of interest
JSGM has received institutional grants from Gilead Sciences, J&J, Merck, ViiV Healthcare, and a Knowledge Translation Award from the Canadian Institutes of Health Research. JSGM has also served as an advisor to the Government of Canada and the Government of British Columbia in the last year. All other authors declare no competing interests.
Ethics approval
The BC-CfE received approval for this study from the University of British Columbia Ethics Review Committee at the St. Paul’s Hospital, Providence Health Care site (H18-02208; H16-02036).
Disclaimer
All inferences, opinions, and conclusions drawn in this manuscript are those of the authors and do not reflect the opinions or policies of the data stewards.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- BC Stats. (2017). Population estimates. http://www.bcstats.gov.bc.ca/StatisticsBySubject/Demography/PopulationEstimates.aspx. Accessed 27 February 2017.
- Blank MB, Himelhoch S, Walkup J, Eisenberg MM. Treatment considerations for HIV-infected individuals with severe mental illness. Current HIV/AIDS Reports. 2013;10(4):371–379. doi: 10.1007/s11904-013-0179-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- British Columbia Centre for Excellence in HIV/AIDS. (2018). Therapeutic guidelines: antiretroviral treatment of adult HIV infection. http://cfenet.ubc.ca/sites/default/files/uploads/Guidelines/Therapeutic-Guidelines-for-Antiretroviral-ARV-Treatment-of-Adult-HIV-Infection-[29-MAY-2018].pdf. Accessed 05 June 2018.
- British Columbia College of Family Physicians. (2018). Mental health and addictions strategy development https://bccfp.bc.ca/wp-content/uploads/2019/01/BCCFP-Mental-Health-and-Addictions-Strategy-Development.pdf. Accessed 23 April 2019.
- British Columbia Ministry of Health. (2018). Adults with chronic hepatitis C: PharmaCare expanded coverage for adults with chronic hepatitis C infection. https://www2.gov.bc.ca/assets/gov/health/health-drug-coverage/pharmacare/chc_expanded_coverage.pdf. Accessed 25 July 2018.
- Brooks ME, Kristensen K, van Benthem KJ, Magnusson A, Berg CW, Nielsen A, et al. glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. The R Journal. 2017;9(2):378–400. doi: 10.32614/RJ-2017-066. [DOI] [Google Scholar]
- Buxton JA, Yu A, Kim PH, Spinelli JJ, Kuo M, Alvarez M, et al. HCV co-infection in HIV positive population in British Columbia, Canada. BMC Public Health. 2010;10:225. doi: 10.1186/1471-2458-10-225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Deeks SG, Lewin SR, Havlir DV. The end of AIDS: HIV infection as a chronic disease. Lancet. 2013;382(9903):1525–1533. doi: 10.1016/S0140-6736(13)61809-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Green Tamar, Flash Shira, Reiss Allan L. Sex differences in psychiatric disorders: what we can learn from sex chromosome aneuploidies. Neuropsychopharmacology. 2018;44(1):9–21. doi: 10.1038/s41386-018-0153-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hall J, Sammons P. Mediation, moderation & interaction. In: Teo T, editor. Handbook of quantitative methods for educational research. Rotterdam: SensePublishers; 2013. pp. 267–286. [Google Scholar]
- Heath K, Samji H, Nosyk B, Colley G, Gilbert M, Hogg RS, et al. Cohort profile: seek and treat for the optimal prevention of HIV/AIDS in British Columbia (STOP HIV/AIDS BC) International Journal of Epidemiology. 2014;43(4):1073–1081. doi: 10.1093/ije/dyu070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hughes E, Bassi S, Gilbody S, Bland M, Martin F. Prevalence of HIV, hepatitis B, and hepatitis C in people with severe mental illness: a systematic review and meta-analysis. Lancet Psychiatry. 2016;3(1):40–48. doi: 10.1016/S2215-0366(15)00357-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ingiliz P, Rockstroh JK. Natural history of liver disease and effect of hepatitis C virus on HIV disease progression. Current Opinion in HIV and AIDS. 2015;10(5):303–308. doi: 10.1097/COH.0000000000000187. [DOI] [PubMed] [Google Scholar]
- Janjua NZ, Kuo M, Yu A, Alvarez M, Wong S, Cook D, et al. The population level cascade of care for hepatitis C in British Columbia, Canada: the BC Hepatitis Testers Cohort (BC-HTC) EBioMedicine. 2016;12:189–195. doi: 10.1016/j.ebiom.2016.08.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jeal, B. (2015). Welcome home: impact and effectiveness of the Dr. Peter Centre’s harm reduction model for those living with HIV/AIDS and who use illicit drugs: part of the mixed method study titled: a mixed method evaluation of the impact of the Dr. Peter Centre on health care access and outcomes for persons living with HIV/AIDS.
- Katrak S, Park LP, Woods C, Muir A, Hicks C, Naggie S. Patterns of healthcare utilization among veterans infected with hepatitis C virus (HCV) and human immunodeficiency virus (HIV) and coinfected with HIV/HCV: unique burdens of disease. Open Forum Infectious Diseases. 2016;3(3):ofw173. doi: 10.1093/ofid/ofw173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 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. AIDS. 2007;21(9):1175–1183. doi: 10.1097/QAD.0b013e32811ebf57. [DOI] [PubMed] [Google Scholar]
- Lima VD, Bangsberg DR, Harrigan PR, Deeks SG, Yip B, Hogg RS, et al. Risk of viral failure declines with duration of suppression on highly active antiretroviral therapy irrespective of adherence level. Journal of Acquired Immune Deficiency Syndromes. 2010;55(4):460–465. doi: 10.1097/QAI.0b013e3181f2ac87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lima VD, Eyawo O, Ma H, Lourenco L, Chau W, Hogg RS, et al. The impact of scaling-up combination antiretroviral therapy on patterns of mortality among HIV-positive persons in British Columbia, Canada. Journal of the International AIDS Society. 2015;18:20261. doi: 10.7448/IAS.18.1.20261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lima VD, Reuter A, Harrigan PR, Lourenco L, Chau W, Hull M, et al. Initiation of antiretroviral therapy at high CD4+ cell counts is associated with positive treatment outcomes. AIDS. 2015;29(14):1871–1882. doi: 10.1097/QAD.0000000000000790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Maldonado G, Greenland S. Simulation study of confounder-selection strategies. American Journal of Epidemiology. 1993;138(11):923–936. doi: 10.1093/oxfordjournals.aje.a116813. [DOI] [PubMed] [Google Scholar]
- Marcus JL, Leyden WA, Chao CR, Xu L, Quesenberry CP, Jr, Tien PC, et al. Differences in response to antiretroviral therapy by sex and hepatitis C infection status. AIDS Patient Care and STDs. 2015;29(7):370–378. doi: 10.1089/apc.2015.0040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McDonald JH. In: Handbook of biological statistics. 2nd Edition ed. Baltimore SHP, editor. 2009. [Google Scholar]
- Olding M, Enns B, Panagiotoglou D, Shoveller J, Harrigan PR, Barrios R, et al. A historical review of HIV prevention and care initiatives in British Columbia, Canada: 1996-2015. Journal of the International AIDS Society. 2017;20(1):21941. doi: 10.7448/IAS.20.1.21941. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pence BW, Miller WC, Gaynes BN, Eron JJ., Jr Psychiatric illness and virologic response in patients initiating highly active antiretroviral therapy. Journal of Acquired Immune Deficiency Syndromes. 2007;44(2):159–166. doi: 10.1097/QAI.0b013e31802c2f51. [DOI] [PubMed] [Google Scholar]
- Poles MA, Dieterich DT. Hepatitis C virus/human immunodeficiency virus coinfection: clinical management issues. Clinical Infectious Diseases. 2000;31(1):154–161. doi: 10.1086/313892. [DOI] [PubMed] [Google Scholar]
- Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. American Journal of Epidemiology. 2011;173(6):676–682. doi: 10.1093/aje/kwq433. [DOI] [PubMed] [Google Scholar]
- Singer M, Bulled N, Ostrach B, Mendenhall E. Syndemics and the biosocial conception of health. Lancet. 2017;389(10072):941–950. doi: 10.1016/S0140-6736(17)30003-X. [DOI] [PubMed] [Google Scholar]
- Taylor LE, Swan T, Mayer KH. HIV coinfection with hepatitis C virus: evolving epidemiology and treatment paradigms. Clinical Infectious Diseases. 2012;55(Suppl 1):S33–S42. doi: 10.1093/cid/cis367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vlahov D, Robertson AM, Strathdee SA. Prevention of HIV infection among injection drug users in resource-limited settings. Clinical Infectious Diseases. 2010;50(Suppl 3):S114–S121. doi: 10.1086/651482. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 75 kb)