Adherence to combination antiretroviral therapy (cART) is essential to the maintenance of health for people living with HIV (PLWH)  and to reduce individual and community viral load, keeping with the goals of treatment as prevention . Although the specific level of adherence required to decrease mortality, viral rebound, and drug resistance varies by medication and treatment regimen, higher levels of adherence have overwhelmingly been linked with better health outcomes, including the prevention of HIV transmission among serodiscordant sex partners [1,3].
The past decade has seen a global increase in the prevalence of HIV among women . In North America and in other resource-rich settings, the same sex and structural inequities that result in vulnerability to HIV infection among women may result in increased vulnerabilities to poor HIV-related health outcomes among women living with HIV . Even with favorable clinical results initially, women have shown worse health outcomes over time than men living with HIV  and disproportionately exhibit suboptimal cART adherence, regardless of the type of cART . In comparison with men living with HIV, studies have shown that women have lower use of primary care services and higher use of the emergency department , lower quality of initial HIV care , and an increased risk of death, even when controlling for cART use . These disparities may be exacerbated among certain subpopulations, including people who inject drugs (PWID) and ethnic minorities, including Indigenous people in Canada [11,12]. Canada's history of colonization, including the forced separation of families via the residential school system, the ‘60s scoop’, and the resultant multigenerational trauma, has had direct and indirect impacts on the myriad of health outcomes experienced by Indigenous people . Even after controlling for such potential confounding characteristics, poorer clinical outcomes among women living with HIV have been observed .
The current study was undertaken within a population-based cohort in British Columbia, Canada, where cART is universally available and offered free of charge to PLWH. Our objectives were to determine the effect of sex on cART adherence over time while adjusting for two known predictors of suboptimal adherence, IDU, and Indigenous ancestry [12,14]. The study is unique in that it examines sex differences in cART adherence longitudinally in a large population-based cohort without the limitation of financial barriers to treatment while accounting for IDU status, ethnicity, and elapsed time from initiation of therapy.
The current study is based upon HIV-positive individuals in British Columbia, Canada, enrolled in the HAART Observation Medical Evaluation and Research (HOMER) cohort. Established in 1996, HOMER is a population-based cohort of antiretroviral-naïve HIV-positive adults who have initiated a HAART regimen of at least three medications . On initiation of antiretroviral therapy (ART) in British Columbia, individuals are automatically enrolled in the provincial Drug Treatment Program. At initial prescription refill, informed consent for the use of medical and sociodemographic data is requested; with refusal in no way limiting access to ART or medical care . Prescription refill compliance was based on medication dispensation dates. Sociodemographic information, based on self-report, was collected at entry into HOMER. Clinical data are collected on an ongoing basis and is based on physician or laboratory record. In British Columbia, HAART is provided free of charge within a universal healthcare system and distributed to eligible individuals according to established guidelines that were updated regularly over the study period (2000–2004, 2005–2008, and 2009 onwards) . Ethical approval for the HOMER cohort and this study has been provided through University of British Columbia and Simon Fraser University Offices of Research Ethics.
Analyses were limited to individuals in HOMER who were at least 19 years of age and had accessed HAART for a minimum of 6 months between 1 January 2000 and 31 December 2014. Individuals were excluded from this study if they self-identified as transgender or had a prescribed treatment interruption. Transgender individuals were excluded due to small sample sizes (n = 39) and as they would experience unique barriers to optimal ART adherence. Individuals who had initiated HAART prior to 2000 were excluded because of changes in antiretroviral medications and guidelines, specifically the use of fixed-dose combination HAART after 2000 .
Optimal adherence to cART, defined as at least 95%, was the main outcome variable. Adherence was assessed for 6-month periods, with an individual considered optimally adherent for the period if the number of days for which they have been dispensed medications divided by the number of days between prescription refills was equal to or greater than 95%. For example, in a 6-month period or 183 days, an individual may be 7 days late to fill their prescription. In this case, adherence would be calculated by dividing 176 (the period for which the individual has medications) by 183, resulting in optimal adherence with medication coverage for 96.2% for the period. This study included adherence measurements from initiation of cART from January 2000 until the end of December 2014 or the date of last contact with the patient, which could include loss from care, movement from the province, or death. Prescription refill compliance has previously been validated as a reliable measure of cART adherence, using an adherence cutoff of at least 85% . Despite the ability to predict virologic failure with lower adherence cutoffs, the adherence cutoff of at least 95% was used in this study for the sake of comparability, as it is the most frequently used adherence cutoff .
The main explanatory variable of interest was sex (female vs. male). Additional explanatory variables included self-reported Indigenous ancestry (Indigenous vs. non-Indigenous), history of injection drug use (yes vs. no), and year of cART initiation grouped according to innovations in cART and treatment guideline revisions (2000–2004, 2005–2008, and 2009–2014) . Additional explanatory sociodemographic and psychosocial variables were not included in our analysis, as only limited sociodemographic information was collected with HOMER.
Chi-squared tests were used to assess differences in the proportion of women and men who met the criteria for optimal adherence for each 6-month period and to compare the mean proportion of individuals attaining optimal adherence over 90 months of follow-up. These analyses were conducted on the study sample as a whole and also among covariates, injection drug use status, and Indigenous ancestry, as women were more likely to belong to these subgroups; within the Canadian context, Indigenous women are overrepresented in the HIV epidemic, and a high proportion of women report IDU as a transmission category .
A generalized linear mixed model using a logistic link function was used to examine sex as an independent predictor of suboptimal adherence, adjusting for injection drug use, Indigenous ancestry, MSM, age, and the period in which cART was initiated. The logistic model included data up to 90 months of follow-up, within the 2000–2014 period. Statistical analyses were completed using SAS version 9.3 (SAS; Cary, North Carolina, USA) with a level of significance set at 0.05.
The sample comprised 4534 individuals, of whom 904 (19.9%) were women. A total of 273 (4.4%; 16.1% women) individuals were excluded from the study due to accessing cART for a period of less than 6 months. Of the 4534 individuals, 2502 (55.2%) had an unknown history of injection drug use and 2381 (52.5%) had unknown Indigenous ancestry.
A larger proportion of women than men identified as being of Indigenous ancestry (28.7 vs. 9.1%; P < 0.001) and having a history of injection drug use (55.3 vs. 30.4%; P < 0.001) (Table 1). The study population had a median time on therapy of 65.9 months (interquartile range: 37.0–103.2 months), with similar proportions of men and women who started therapy early enough to allow 90 months of observation for this study (75.4 vs. 72.3%; P = 0.205). Of note, a significantly larger proportion of women experienced mortality during the study period (15.9 vs. 11.6%; P < 0.001); however, in multivariate analysis (not shown), sex was not found to be independently associated with mortality [adjusted odds ratio (AOR): 0.99; 95% confidence interval (CI): 0.89–1.09]. Cause of death did not differ substantially between men and women (not shown).
Table 2 shows the mean proportion of men and women who attained optimal adherence over the full 90-month study period. Women were significantly less likely to be optimally adherent than men overall (57.0 vs. 77.1%; P < 0.001) and within stratified analyses. The largest difference between men and women was observed among non-Indigenous individuals who do not have a history of injection drug use (69.9 vs. 83.8%; P < 0.001). Of note, the proportion of women with lower vulnerability (no history of injection drug use or non-Indigenous ancestry) was found to be similar in adherence to the proportion of men with higher vulnerability (history of injection drug use or Indigenous ancestry) (Figs. 1 and 2).
Table 3 shows the multivariate confounder analysis of factors affecting adherence. Female sex was independently and significantly associated with suboptimal adherence when controlling for age, history of injection drug use, Indigenous ancestry, MSM, and year of therapy initiation (AOR: 0.55; 95% CI: 0.48–0.63).
Our findings indicate that women are significantly less likely to achieve optimal adherence to cART than men. These findings are consistent with other studies focusing on women in high and very high human development index countries . Although both Indigenous ancestry and history of injection drug use have previously been found to impact adherence [11,12,14], our study shows that sex maintains a significant and independent effect on adherence. The finding of significantly lower adherence among women relative to men continued across different cART guidelines and public health campaigns to improve access to care, early diagnosis, retention in care, and adherence (e.g. the Maximally Assisted Therapy and STOP programs) [22,23].
Vulnerabilities that women face may be amplified by multiple intersecting identities, including ethnic minority status, histories of trauma, and injection drug use [24,25]. Within this study, we observed marked disparity between Indigenous and non-Indigenous individuals, compounded by sex. Among Indigenous people in Canada, there is a disproportionately high prevalence of HIV , lower quality of HIV-related care , as well as poorer health outcomes among Indigenous PLWH despite more favorable clinical characteristics at baseline . Poorer health outcomes among Indigenous people, such as HIV-related morbidity and mortality, cannot be discussed without connecting these conditions with Canada's history of colonization, marginalization, and criminalization of culture . The negative impact of Canada's colonial legacy can certainly still be observed in the disproportional rates of injection drug use, alcoholism, and suicide, in addition to the high HIV incidence [12,13,28]. In the context of colonialism, there is a need for culturally competent HIV-related care to support cART adherence and to build trust between Indigenous people and non-Indigenous HIV-care providers .
Although injection drug use does increase the likelihood of suboptimal adherence among PLWH, this study shows that sex maintains a central role in predicting adherence. This is consistent with existing literature, though the effect of sex may be modified by drug use characteristics (e.g. frequent heroin injection or frequent crack use may further lower cART adherence) . Injection drug use can yield additional barriers to cART adherence, including unstable housing, food insecurity, social instability, competing priorities, and low perception of adherence self-efficacy [11,31]. Women who inject drugs may be more likely to be involved in street-based survival sex work and experience greater marginalization than men , undermining self-efficacy and their relationship with healthcare services.
Reasons for differing levels of adherence were not directly explored in this study. Clinically, women have been observed to experience greater risk for adverse drug-related events relative to men, including gastrointestinal, hematological, metabolic, and toxicity-related symptoms , which could affect adherence and retention to therapy. Psychosocial and structural factors may also result in the sex differences observed in this study. Psychosocial differences, including self-efficacy and social support, between men and women may be related to differences in coping with HIV and sustaining cART adherence [33,34]. Anticipated or experienced HIV-related stigma may isolate the woman from previously supportive relationships and cause additional emotional stress . Many women also need to contend with competing life demands, particularly the care of children and her partner , in addition to the basic needs of housing and food security [22,37]. These demands can be further exacerbated by experiences of abuse and an absence of local women-centered care [38,39].
Among high-risk groups targeted for HIV care interventions, such as PWID and ethnic minorities, it is simply not enough to develop group-specific interventions without considering sex. Acknowledging women's care needs, specific barriers to entry and retention into care and adherence to therapy require more attention, both within research and in program implementation. However, as observed within our study and elsewhere , women living with HIV cannot be considered a homogenous group. Though women are less likely than men to attain adherence goals, a unique hierarchy emerged indicating increasing vulnerability to suboptimal cART adherence among PWID, followed by Indigenous peoples. More research is necessary to identify women-centered, culturally well tolerated interventions if we are to reduce outcome disparities between women and men, as well as between the various intersections of women's identities .
Although this study highlights the overall vulnerability of women to suboptimal cART adherence, it cannot speak to other aspects of vulnerability, such as sexual orientation, transgender identity, experiences of pregnancy, HIV-related stigma, mental health, or health service usage. In addition, this study did not address discontinuation of cART, which has previously been found to be associated with female sex . Another limitation of this study is that pharmacy refill compliance does not indicate whether PLWH are taking medications as prescribed, or picking them up at the time of dispensation, as may be the case with automated refills. Underreporting of injection drug use due to social desirability bias may also be present, appearing as either unknown history or no history of injection drug use. This may have the effect of artificially reducing the proportion of optimally adherent individuals within the subpopulations that did not identify as PWID. Lastly, the results of this study may not be generalizable outside of Canada, due to the accessibility of the Canadian healthcare system and the specific impacts of colonialism on the Indigenous peoples of Canada.
In our study, we have demonstrated sex disparities in adherence to cART, compounded by Indigenous ancestry and injection drug use history. To reach the goals of 90-90-90 and Treatment as Prevention, there is a need to identify where women are being lost along the Cascade of Care, under what circumstances, and how they can best be supported in their care at the varying levels of the Cascade. Existing interventions to support cART adherence may not identify and care for the specific needs of women and, thus, may not be sufficient for the women who access them. The care needs of women and barriers to women's care along the Cascade are particularly important to address if we are to meet the goals of UNAIDS’ 90-90-90 campaign to end AIDS by 2030.
The study was conceived and designed by C.M.P., C.L.M., R.S.G., and A.K.; B.Y. assisted in the development of the longitudinal cART adherence data design; W.Z. developed the statistical design and conducted the analysis. C.M.P. wrote the manuscript, and all authors edited and shaped the final manuscript.
Conflicts of interest
There are no conflicts of interest.
1. Maggiolo F, Airoldi M, Kleinloog HD, Callegaro A, Ravasio V, Arici C, et al. Effect of adherence to HAART on virologic outcome and on the selection of resistance-conferring mutations in NNRTI- or PI-treated patients
. HIV Clin Trials
2. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Antiretroviral therapy for the prevention of HIV-1 transmission
. N Engl J Med
3. Lima VD, Bangsberg DR, Harrigan PR, Deeks SG, Yip B, Hogg RS, Montaner JS. Risk of viral failure declines with duration of suppression on highly active antiretroviral therapy irrespective of adherence level
. J Acquir Immune Defic Syndr
5. Gahagan J, Ricci C, Jackson R, Prentice T, Mill J, Adam B. Gahagan J. Advancing our knowledge: findings of a meta-ethnographic synthesis
. Canadian Scholars’ Press Inc, Women and HIV prevention in Canada, in Women and HIV prevention in Canada: implications for research, policy, and practice
. Toronto: 2013.
6. Cescon A, Patterson S, Chan K, Palmer AK, Margolese S, Burchell AN, et al. Gender differences in clinical outcomes among HIV-positive individuals on antiretroviral therapy in Canada: a multisite cohort study
. PLoS One
7. Puskas CM, Forrest JI, Parashar S, Salters KA, Cescon AM, Kaida A, et al. Women and vulnerability to HAART nonadherence: a literature review of treatment adherence by gender from 2000 to 2011
. Curr HIV/AIDS Rep
8. Sohler NL, Li X, Cunningham CO. Gender disparities in HIV healthcare utilization among the severely disadvantaged: can we determine the reasons?
. AIDS Patient Care STDS
9. Carter A, Min JE, Chau W, Lima VD, Kestler M, Pick N, et al. Gender inequities in quality of care among HIV-positive individuals initiating antiretroviral treatment in British Columbia, Canada (2000–2010)
. PLoS One
10. Lemly DC, Shepherd BE, Hulgan T, Rebeiro P, Stinnette S, Blackwell RB, et al. Race and sex differences in antiretroviral therapy use and mortality among HIV-infected persons in care
. J Infect Dis
11. Knowlton AR, Arnsten JH, Eldred LJ, Wilkinson JD, Shade SB, Bohnert AS, et al. Antiretroviral use among active injection-drug users: the role of patient-provider engagement and structural factors
. AIDS Patient Care STDS
12. Lima VD, Kretz P, Palepu A, Bonner S, Kerr T, Moore D, et al. Aboriginal status is a prognostic factor for mortality among antiretroviral naive HIV-positive individuals first initiating HAART
. AIDS Res Ther
13. Bombay A, Matheson K, Anisman H. The intergenerational effects of Indian Residential Schools: Implications for the concept of historical trauma
. Transcult Psychiatry
14. Lima VD, Kerr T, Wood E, Kozai T, Salters KA, Hogg RS, Montaner JS. The effect of history of injection drug use and alcoholism on HIV disease progression
. AIDS Care
15. Patterson S, Cescon A, Samji H, Cui Z, Yip B, Lepik KJ, et al. Cohort profile: HAART observational medical evaluation and research (HOMER) cohort
. Int J Epidemiol
16. Hogg RS, Yip B, Chan KJ, Wood E, Craib KJ, O'Shaughnessy MV, Montaner JS. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy
17. Thompson MA, Mugavero MJ, Amico KR, Cargill VA, Chang LW, Gross R, et al. Guidelines for improving entry into and retention in care and antiretroviral adherence for persons with HIV: evidence-based recommendations from an International Association of Physicians in AIDS Care panel
. Ann Intern Med
2012; 156:817–833. W-284–294.
18. Vella S, Schwartländer B, Sow SP, Eholie SP, Murphy RL. The history of antiretroviral therapy and of its implementation in resource-limited areas of the world
19. Grossberg R, Zhang Y, Gross R. A time-to-prescription-refill measure of antiretroviral adherence predicted changes in viral load in HIV
. J Clin Epidemiol
20. Bezabhe WM, Chalmers L, Bereznicki LR, Peterson GM. Adherence to antiretroviral therapy and virologic failure: a meta-analysis
. Medicine (Baltimore)
21. PHAC. Population-specific HIV/AIDS report: women
. Ottawa: Public Health Agency of Canada; 2012.
22. Parashar S, Palmer AK, O’Brien N, Chan K, Shen A, Coulter S, et al. Sticking to it: the effect of maximally assisted therapy on antiretroviral treatment adherence among individuals living with HIV who are unstably housed
. AIDS Behav
23. Heath K, Samji H, Nosyk B, Colley G, Gilbert M, Hogg RS, Montaner JS. STOP HIV/AIDS Study Group. Cohort profile: seek and treat for the optimal prevention of HIV/AIDS in British Columbia (STOP HIV/AIDS BC)
. Int J Epidemiol
24. Logie CH, James L, Tharao W, Loutfy MR. HIV, gender, race, sexual orientation, and sex work: a qualitative study of intersectional stigma experienced by HIV-positive women in Ontario, Canada
. PLoS Med
25. Myers HF, Sumner LA, Ullman JB, Loeb TB, Carmona JV, Wyatt GE. Trauma and psychosocial predictors of substance abuse in women impacted by HIV/AIDS
. J Behav Health Serv Res
26. Public Health Agency of Canada. HIV/AIDS among aboriginal people in Canada
. C.f.C.D.a.I. Control. editor. HIV/AIDS Epi update
. Public Health Agency of Canada (author/government body) Centre for Communicable Diseases and Infection Control (Department). 2010.
27. Thira D. Aboriginal youth suicide prevention: a postcolonial community-based approach
. Int J Child Youth Fam Stud
28. Elias B, Mignone J, Hall M, Hong SP, Hart L, Sareen J. Trauma and suicide behaviour histories among a Canadian indigenous population: an empirical exploration of the potential role of Canada's residential school system
. Soc Sci Med
29. Reimer RJG. Canadian aboriginal people living with HIV/AIDS: care. Treatment and support issues
. Ottawa: Canadian Aboriginal AIDS Network; 2005.
30. Tapp C, Milloy MJ, Kerr T, Zhang R, Guillemi S, Hogg RS, et al. Female gender predicts lower access and adherence to antiretroviral therapy in a setting of free healthcare
. BMC Infect Dis
31. Krusi A, Wood E, Montaner J, Kerr T. Social and structural determinants of HAART access and adherence among injection drug users
. Int J Drug Policy
32. Loutfy MR, Sherr L, Sonnenberg-Schwan U, Walmsley SL, Johnson M. d’Arminio Monforte A: Women
for Positive Action. Caring for women living with HIV: gaps in the evidence
. J Int AIDS Soc
33. Ubbiali A, Donati D, Chiorri C, Bregani V, Cattaneo E, Maffei C, Visintini R. Prediction of adherence to antiretroviral therapy: can patients’ gender play some role? An Italian pilot study
. AIDS Care
34. Vosvick M, Martin LA, Smith NG, Jenkins SR. Gender differences in HIV-related coping and depression
. AIDS Behav
35. Chaudoir SR, Earnshaw VA, Andel S. ‘Discredited’ versus ‘Discreditable’: understanding how shared and unique stigma mechanisms affect psychological and physical health disparities
. Basic Appl Soc Psych
36. Knowlton AR, Yang C, Bohnert A, Wissow L, Chander G, Arnsten JA. Main partner factors associated with worse adherence to HAART among women in Baltimore, Maryland: a preliminary study
. AIDS Care
37. Kalichman SC, Hernandez D, Cherry C, Kalichman MO, Washington C, Grebler T. Food insecurity and other poverty indicators among people living with HIV/AIDS: effects on treatment and health outcomes
. J Community Health
38. Trimble DD, Nava A, McFarlane J. Intimate partner violence and antiretroviral adherence among women receiving care in an urban Southeastern Texas HIV clinic
. J Assoc Nurses AIDS Care
39. Carter AJ, Bourgeois S, O’Brien N, Abelsohn K, Tharao W, Greene S, et al. Women-specific HIV/AIDS services: identifying and defining the components of holistic service delivery for women living with HIV/AIDS
. J Int AIDS Soc
40. Gonzalez-Serna A, Chan K, Yip B, Chau W, McGovern R, Samji H, et al. Temporal trends in the discontinuation of first-line antiretroviral therapy
. J Antimicrob Chemother