Perceived effectiveness of antiretroviral therapy,self-rated health and treatment adherence among HIV-positive people who inject drugs in Estonia

Abstract

The HIV epidemic in Estonia affects the population of people who inject drugs (PWID) the most, but factors associated with adherence to antiretroviral therapy (ART) among PWID have not been thoroughly examined in Estonia, with particularly limited data regarding beliefs and attitudes of PWID. The objective of this study was to explore the association between ART adherence and individual beliefs, perceived effectiveness of ART, and self-rated health in particular, in this specific population. The study used baseline survey data from a longitudinal intervention study of HIV prevention among PWID in Estonia, in which 107 HIV-infected participants reported current use of ART. Current adherence was measured through the use of a visual analog scale. Approximately half (49%) of the participants reported optimal (≥95%) adherence. The vast majority (81%) believed in the effectiveness of ART. Less than a quarter of the participants (22%) rated their health as good or very good, and a half (52%) reported average health. Individual beliefs and self-reported health were not associated with ART adherence in both bivariate and multivariable analyses. Participants with problem drinking reported significant suboptimal adherence to ART (adjusted odds ratio [AOR] 0.42, 95% CI 0.19–0.97). Daily injection drug use was also associated with suboptimal adherence (AOR 0.34, 95% CI 0.13–0.91). Problem drinking has not been commonly reported as a factor of suboptimal ART adherence among PWID; further research would be useful to identify the pathways that might be involved.

Keywords

HIV AIDS treatment Europe ART

Introduction

HIV is an important public health concern in Europe. The World Health Organization (WHO) estimated that 2.2 million Europeans were HIV-positive in 2012, among whom 130,000 were newly infected.^1,2 The burden of HIV, however, is not the same between Eastern and Western Europe. The incidence of HIV is higher in Eastern Europe, and this is one of the few regions in the world with a continually rising number of new infections.³ While the incidence rate declined in Western Europe between 2006 and 2012, there was an approximate twofold increase in Eastern Europe, with people who inject drugs (PWID) the most affected by the epidemic.⁴

Estonia is one of the Eastern European countries with the highest HIV incidence, with 24.6 new cases per 100,000 people annually, and has a high HIV prevalence of 1.3% among persons aged 15–49.^5,6 PWID have been overrepresented in the HIV-positive population in Estonia since the beginning of the HIV epidemic in 2001.^5,7 In 2011, at least one half of the new HIV infections in the country were related to injecting drug use.⁸ Studies conducted among PWID in Estonia also reveal high HIV prevalence rates of 40–90%.^9–12 Unprotected sex between PWID and their non-PWID sexual partners may account for the increasing number of HIV infections acquired via heterosexual transmission in Estonia and may facilitate the expansion of the HIV epidemic into the general population.⁷

With advances in modern medicine, HIV is no longer immediately life threatening, but rather a chronic disease that can remain stable over time.¹³ Highly active antiretroviral therapy (ART) has extended the life expectancy of HIV patients and has decreased HIV-related mortality.¹⁴ Furthermore, studies show that patients on ART have reduced HIV transmission in comparison to those who are not on ART, possibly due to suppressed viral replication, suggesting the impact of ART on public health.^15,16 In Estonia, ART was available free to all patients in need regardless of insurance status, and over 40% of PWID who have HIV are currently receiving ART.¹⁷ The importance of treatment adherence has been widely emphasized in the HIV literature since nonadherence not only leads to failure of viral suppression, but also emergence of drug-resistant strains.^13,14 For instance, nonadherence among patients in a nonnucleoside reverse transcriptase inhibitor group was found to be associated with the development of class-specific resistance mutations, and in treatment-naïve patients, low adherence might be a motive to prescribe ART with a higher genetic barrier to resistance.¹⁸ In order to achieve long-term therapeutic effects and to prevent drug-resistant strains from emerging, high rates of adherence to ART are necessary.¹⁹

Physicians often assume that PWID are less likely to adhere to treatment due to psychosocial, co-occurring other drug use disorders (including alcohol) and lifestyle factors.²⁰ However, one meta-analysis of ART adherence among HIV-positive PWID suggests this population can achieve similar adherence levels as observed among the general HIV-positive population.²¹ Nonetheless, adherence to ART among PWID is commonly suboptimal and is an ongoing concern.^22,23

Given the significant burden of HIV among PWID in Estonia, it is important to ensure their access and optimal adherence to ART. However, factors associated with treatment adherence in this specific population have not been thoroughly studied. In general, studies regarding HIV medication adherence among PWID highlight the importance of mental health, opioid substitution therapy, active drug use, and social support.^21,24–28 Side effects have also been explored in a few studies.^20,24,27 Although patient beliefs and attitudes, such as perceived necessity of ART and intention to adhere, are consistently associated with medication adherence among HIV patients in general,^29–32 they are rarely investigated in studies targeting PWID. There are a few American studies focusing on beliefs/attitudes toward ART and medication-taking self-efficacy supporting an association with adherence to ART among PWID,^13,24,33 but no similar studies have been identified in Europe. In addition, though average or poor self-rated health among persons living with HIV has been found to increase the odds of patient-initiated changes to ART,³⁴ inadequate virological response to ART,³⁵ or nonperfect adherence to ART,³⁶ research on this topic area is lacking among PWID.

Given the lack of knowledge in this area, our objective was to identify factors associated with less than optimal ART adherence, particularly perceived effectiveness of ART and self-rated health, among PWID in Estonia.

Methods

Overview

We used baseline data from a longitudinal study aimed at evaluating the feasibility and acceptability of multiple interventions for HIV prevention among PWID in Estonia.¹⁷ Adult residents of Tallinn, Estonia aged 18 years or older reporting injecting illegal drugs in the past two months were eligible for the study, while people enrolled in another HIV prevention or drug use intervention study, or had a significant psychological or cognitive disorder that would affect their ability to provide informed consent, were excluded. Respondent-driven sampling, which has been recognized as an effective tool to sample hard-to-reach populations (such as PWID), was employed.³⁷ Nonrandom selection of seeds (n = 6) representative of the PWID population by age, gender, ethnicity, main type of drug used, and HIV status was employed, and each seed referred up to three other PWID from their existing social networks.

Each participant was asked to complete a 40 min long interviewer-administered survey in either Estonian or Russian. The survey was adapted from the WHO Drug Injecting Study Phase II Survey and included information such as sociodemographic characteristics, drug use behaviors, drug treatments, sexual behaviors, HIV diagnosis and care, and AIDS knowledge.³⁸ Venous blood was collected from participants and tested for the presence of HIV antibodies using commercially available test kits (ADVIA Centaur CHIV Ag/Ab Combo [SIEMENS]). Incentive payment was provided for participation (10 euros) as well as referral of peers (5 euros for each new referral). Details on the study protocol were approved by the Ethics Review Board at the University of Tartu in Estonia and the Institutional Review Board at Beth Israel Medical Center in New York, USA. A total of 328 PWID completed the baseline survey between 2 July and 15 August 2013; our final analytical sample consisted of 107 of them who reported a diagnosis of HIV and were currently on an ART regimen.

Measures

A visual analog scale (VAS) was used to measure ART adherence.^39,40 Participants were presented with a horizontal line anchored by 0 and 100%, where the endpoints were labeled ‘You do not take ART medication’ and ‘You take ART medication exactly as suggested by doctor,’ respectively. Patients were asked to mark a point on the line representing the percentage of prescribed medications taken in general. While this question does not specify a time period for the reported adherence, the use of present tense in the wording of the question implies that the question concerns current adherence. Previous research supports the association between VAS and undetectable viral load.^39,41–43 Optimal adherence was defined as reporting at least 95% adherence in this study.¹⁹ Among the 107 participants who reported current use of ART, 48.6% were categorized as optimally adherent. Perceived effectiveness of ART was assessed with the question ‘Do you believe that treatment for HIV is effective?,’ where ‘Yes,’ ‘No,’ and ‘Don’t know’ were the possible responses. Self-rated health was measured on a 5-point Likert scale.

In addition to sociodemographic factors, other variables included health insurance status, history of incarceration, number of years since HIV diagnosis, diagnosis of tuberculosis, hepatitis B, or hepatitis C in the past, number of current HIV symptoms (tiredness, fever, loss of weight, night sweats, loss of appetite, cough that has lasted over two weeks, blood in sputum, and chest pain), number of years injecting drugs, and frequency of injection in the past month. Since problem alcohol use frequently coincides with drug use, we were also interested in examining the independent effect of alcohol use on ART adherence. The Cut Down, Annoyed, Guilty, and Eye-Opener (CAGE), a four-item questionnaire, was utilized to screen for problem drinking in the past six months. Participants giving affirmative responses to two or more items were classified as problem drinkers.⁴⁴

Psychosocial factors were also examined. Satisfaction with life was assessed by asking ‘How satisfied have you been with your life over the past four weeks?,’ and used a 5-point Likert scale. Mental health was evaluated using the five-item Mental Health Inventory (MHI-5), with scores ranging from 0 to 100, with higher scores indicating a better state of mental health in the past month.⁴⁵ A cutoff point of 52 has been suggested to perform the best in detecting major depression.⁴⁶ Additionally, participants were asked whether they had received mental health services in the past six months, were currently receiving medication-assisted treatment for drug addiction (methadone maintenance treatment), and knew anyone who died from AIDS.

Statistical analysis

Bivariate analyses were performed to explore differences in individual characteristics between optimally adherent and nonadherent PWID. Independent samples t-tests were used to compare continuous variables; with regard to categorical variables, Chi square or Fisher’s exact tests (as appropriate) were performed. Factors statistically significant at the 0.10 alpha level in bivariate analyses were included into the regression model.⁴⁷ Multivariable logistic regression was used to determine whether our two primary variables of interest – perceived ART effectiveness and self-rated health – were each significantly associated with a higher odds of optimal adherence upon adjustment for age, gender, and other variables potentially associated with adherence. Additionally, mental health was retained in our final model regardless of its statistical significance in bivariate analysis since previous PWID studies highlight its association with ART adherence.^13,21,25 All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Sample characteristics

There were 107 participants (75.7% men) ranging in age between 23 and 54 years (mean 33.1, SD 5.3). The vast majority were either ethnic Russians (83.0%) or Estonians (11.3%). Less than half (47.7%) completed secondary or higher education (≥ 9 years). Most participants were single (61.7%), and over a half (57.0%) were unemployed. A large proportion (85.8%) reported having health insurance.

Length of HIV diagnosis ranged from 0 to 23 years (mean 8.7, SD 4.0). Duration of injecting drug use also varied, ranging from 1 to 32 years (mean 14.5, SD 4.8). Approximately a quarter (27.1%) reported injecting drugs daily in the past month. Less than half (42.9%) were receiving methadone maintenance treatment. Nearly half had either problem drinking (CAGE scores ≥2) or poor mental health (MHI-5≤52) (48.6 and 49.5%, respectively). The vast majority (81.3%) believed in the effectiveness of ART. Less than a quarter (22.4%) rated their health as good or very good, and approximately half (52.3%) reported average health.

Bivariate analysis

Perceived effectiveness of ART (p=0.90) and self-rated health (p=0.34) each were not statistically significantly associated with treatment adherence (Table 1). Additionally, none of the sociodemographic factors, nor was mental health statistically significantly related to adherence. However, frequency of injecting drug use and problem drinking were each found to be statistically significant. Daily drug users had a 63% reduced odds of optimal adherence versus nondaily users (OR 0.37, 95% CI 0.15–0.91). Problem drinkers had a 55% reduced odds of optimal adherence versus nonproblem drinkers (OR 0.45, 95% CI 0.21–0.97).

Table 1.

Characteristics of optimally adherent (≥95%) PWID versus suboptimally adherent (<95%) PWID in Estonia (n=107), 2013.

	Optimal adherence (n=52)	Suboptimal adherence (n=55)
Characteristics	n (%) or mean±SD	n (%) or mean±SD	P
Age (years)	33.4±5.5	32.9±5.1	0.61
Gender
Male	40 (76.9)	41 (74.6)
Female	12 (23.1)	14 (25.4)	0.77
Ethnicity
Estonian	5 (9.8)	7 (12.7)
Russian	44 (86.3)	44 (80.0)
Other	2 (3.9)	4 (7.3)	0.73^a
Education
Basic (≤9 years)	26 (50.0)	25 (54.5)
Secondary or higher (>9 years)	26 (50.0)	30 (45.5)	0.64
Marital status
Married/cohabitation	15 (28.9)	14 (25.5)
Divorced/separated/widowed	4 (48.6)	8 (14.5)
Single	33 (63.5)	33 (60.0)	0.53
Employment
Yes	24 (45.1)	22 (40.0)
No	28 (53.9)	33 (60.0)	0.52
Health Insurance
Yes	45 (86.5)	46 (85.2)
No	7 (13.5)	8 (14.8)	0.84
History of incarceration in lifetime
Yes	38 (73.1)	47 (85.5)
No	14 (26.9)	8 (14.5)	0.11
Duration since diagnosis of HIV (years)	8.4±3.7	9.0±4.3	0.40
Number of current HIV symptoms^b	2.6±1.5	2.7±1.6	0.76
Ever diagnosis of tuberculosis
Yes	3 (5.8)	5 (9.3)
No	49 (94.2)	49 (90.7)	0.72^a
Ever diagnosis of hepatitis B
Yes	30 (58.8)	34 (64.2)
No	21 (41.2)	19 (35.8)	0.58
Ever diagnosis of hepatitis C
Yes	38 (74.5)	46 (86.8)
No	13 (25.5)	7 (13.2)	0.11
Know anyone who died from AIDS
Yes	30 (57.7)	27 (50.0)
No	22 (42.3)	27 (50.0)	0.43
Perceived effectiveness of ART
Effective	42 (80.8)	45 (81.8)
Ineffective/don’t know	10 (19.2)	10 (18.2)	0.90
Self-rated health
Very good/good	10 (19.2)	14 (25.4)
Average	31 (59.6)	25 (45.5)
Very poor/poor	11 (21.2)	16 (29.1)	0.34
Satisfaction with life in the past month
Very satisfied/satisfied	14 (26.9)	20 (36.4)
Average	27 (51.9)	20 (36.4)
Very dissatisfied/dissatisfied	11 (21.2)	15 (27.2)	0.27
Mental health in the past month (MHI-5 score)^c
Good (>52)	29 (55.8)	25 (45.5)
Poor (≤52)	23 (44.2)	30 (54.5)	0.29
Problem drinking in the past six months (CAGE score)^d
Yes (≥2)	20 (38.5)	32 (58.2)
No (<2)	32 (61.5)	23 (41.8)	0.04*
Duration of injecting drug use (years)	14.7±5.1	14.3±4.7	0.68
Frequency of injecting drug use in the past month
Daily (≥28 days)	9 (17.3)	20 (36.4)
Nondaily (<28 days)	43 (82.7)	35 (63.6)	0.03*
Current methadone maintenance treatment
Yes	26 (50.0)	20 (36.4)
No	26 (50.0)	35 (63.4)	0.15
Received mental health services in the past six months
Yes	14 (26.9)	11 (20.0)
No	38 (73.1)	44 (80.0)	0.40

ART: antiretroviral therapy; MHI-5: Mental Health Inventory; PWID: people who inject drugs.

Note. Chi square or Fisher’s exact tests were used to compare categorical variables, while independent samples t-tests were used to compare continuous variables. Percentages are calculated based on the number of responses available for each variable.

P value from Fisher’s exact test was reported.

Including tiredness, fever, loss of weight, night sweats, loss of appetite, cough that has lasted over two weeks, blood in sputum, and chest pain.

Possible MHI-5 scores: 0–100. Higher scores indicate better mental health.

Possible CAGE scores: 0–4.

P<0.10.

Multivariable analysis

Perceived ART effectiveness (yes versus no/don’t know) (p=0.95) and self-rated health (very good/good versus poor/very poor: p=0.69; average versus poor/very poor: p=0.36) remained nonsignificant after adjustment, whereas frequency of drug use and problem drinking stayed significant (Table 2). Injection drug users had a 66% lower odds of optimal adherence (adjusted odds ratio [AOR] 0.34, 95% CI 0.13–0.91). Problem drinkers had a 58% lower odds of being adherent (AOR 0.42, 95% CI 0.19–0.97). An additional model excluding the nonsignificant variables in our bivariate analysis (i.e. age, gender, and mental health) showed no meaningful change on the coefficients of the two belief variables (results not shown).

Table 2.

Crude and adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) for factors associated with optimal adherence (≥95%) among PWID in Estonia, 2013.

Characteristics	Crude ORs (95% CIs)	Adjusted ORs (95% CIs)^a	P ^b
Age (years)	1.02 (0.95–1.10)	1.02 (0.94–1.10)	0.72
Gender
Male	1.00	1.00
Female	0.88 (0.36–2.13)	0.92 (0.34–2.54)	0.88
Perceived effectiveness of ART
Effective	0.93 (0.35–2.47)	0.96 (0.32–2.86)	0.95
Ineffective/don’t know	1.00	1.00
Self-rated health
Very good/good	0.70 (0.28–1.75)	0.77 (0.21–2.87)	0.69
Average	1.77 (0.82–3.82)	1.66 (0.56–4.92)	0.36
Very poor/poor	1.00	1.00
Problem drinking in the past six months (CAGE score)^c
Yes (≥2)	0.45 (0.21–0.97)	0.42 (0.19–0.97)	0.04*
No (<2)	1.00	1.00
Frequency of drug use in the past month
Daily (≥28 days)	0.37 (0.15–0.91)	0.34 (0.13–0.91)	0.03*
Nondaily (<28 days)	1.00	1.00
Mental health in the past month (MHI-5 score)^d
Good (>52)	1.51 (0.71–3.24)	1.20 (0.50–2.86)	0.68
Poor (≤52)	1.00	1.00

ART: antiretroviral therapy; MHI-5: Mental Health Inventory; PWID: people who inject drugs.

Multivariable model included all factors listed in the table (n=104).

P values for adjusted ORs.

Possible CAGE scores: 0–4.

Possible MHI-5 scores: 0–100. Higher scores indicate better mental health.

Significant p value (<0.05).

Discussion

To our knowledge, this is one of the first studies exploring individual beliefs and other factors related to ART adherence among PWID in Eastern Europe, where PWID continue to bear a significant burden of HIV infections.⁴ The self-reported optimal (95%) adherence rate in the present study was 48.6% and differed from existing studies of PWID, which generally report higher rates, ranging from 70 to 84%.^{13,24,25,48,49} However, adherence estimates are strongly sensitive to the measures and cutoff values chosen.⁵⁰ Our study used a 95% cutpoint, while others have used 80 or 90%.^13,24,25 Furthermore, our study measured adherence based on VAS, whereas recall on the number of pills taken or missed is frequently used in previous studies.^13,24,25 VAS has been reported to have better agreement with viral load, which is the biological (‘gold standard’) measure of adherence, as compared to self-reported recall measures.³⁹ Our study used a respondent-driven sampling method, resulting in a community-based sample of PWID in Tallinn, while others rely on institutional-based samples (participants from clinics or drug treatment programs),^13,24,25 which may introduce selection bias and overestimate true population adherence. Additionally, our study recruited only people who were currently (within the previous two months) injecting drugs, while others include drug users who were not currently injecting.^13,25

Patient beliefs and/or attitudes have been identified as correlates of treatment adherence among the HIV-positive population in several European studies.^29,30,32,51 The present study found no significant independent association between individual beliefs and adherence to ART among PWID, however. Our analysis focused on perceived effectiveness of ART and self-rated health, but we are not aware of research targeting the same areas. Studies suggest attitudes toward ART, self-efficacy for taking medication, and concerns about ART side effects are also associated with adherence, though these studies were conducted in the United States.^13,24,33

Our null findings on self-rated health can be explained by the two-sided relationship of ‘feeling healthy’ to ART adherence. Adherence may be reinforced if an individual attributes his/her healthiness to adhering to medication, but it may also be reduced if one believes he/she is healthy enough to skip medications (a person who is less than optimally adherent may still be symptom-free). Both thought patterns might have occurred in our study population, thus producing the observed null association. With respect to perceived effectiveness of ART, Arnsten et al.¹³ reported that positive attitudes toward ART correlate with better adherence, which differs from our findings. Such distinction, however, may be reflective of the instruments used to measure beliefs related to ART (a continuous 11-item scale in Arnsten et al.’s study versus a dichotomized single-item measure in our study). We also had very little variance in the belief in the effectiveness of ART.

To the best of our knowledge, the present study was the first to report a significant association between alcohol use (problem drinking) and ART adherence among PWID in the region. Though alcohol abuse among PWID is commonly reported in studies from St. Petersburg, Russia (66%) and Estonia (58%),⁵² drinking patterns are frequently ignored by treatment and prevention interventions. This is somewhat surprising, since alcohol use disorders among PWID have been found to have potential detrimental effects on health.^53–55 In studies from USA, patients with co-occurring alcohol and injecting drug use disorders had a significantly lesser increase in CD4 cell count after initiation of ART when compared to patients who were neither alcohol users nor injecting drug users⁵³; hazardous alcohol use alone was associated with decreased ART use, adherence, and viral suppression and with injection drug use coaddiction further exacerbating these negative effects.⁵⁴ It has been suggested that heavy alcohol drinking has significant negative effects on a variety of clinical outcomes (i.e. HIV viral load, CD4+ T-cell count, and estimated liver fibrosis), some which are mediated through reduced ART adherence.⁵⁵

Our findings are compatible with a few studies supporting the association of hazardous drinking with ART adherence and HIV-1 RNA suppression.^56,57 However, one French study failed to find any association with alcohol use; though the number of units of alcohol consumption and not problem drinking was explored.²⁵ A few studies also report problem drinking as a significant factor of adherence among HIV patients in general.^39,57 However, in a recent study using a HIV clinic-based sample in Estonia, no association was found between harmful patterns of alcohol use and ART adherence.³² The difference between the present and the previous Estonian studies may indicate that problem drinking correlates to adherence exclusively among PWID, but further studies are required to confirm our results. It would also be important to understand the pathways through which alcohol use is associated with poor adherence and whether alcohol use is a predictor of adherence.

Our study also highlighted the potential influence of drug use frequency (nondaily injection versus daily injection) on ART adherence. Existing literature supports the association of active injecting drug use with virological failure and poor adherence to ART.^25,58,59 Our findings provide additional evidence on this topic. The mechanism of how active drug use is related to ART adherence is not exactly known. While lifestyle factors and lack of social support may play a role, one study suggests that active drug use predicts lapses in HIV care, which may affect treatment adherence and outcomes.⁵⁹

Our study had several limitations. First, the measurements of individual beliefs were limited. The current study was not designed to solely evaluate beliefs related to ART and health and employed single-item measures for these variables. Our study also explored limited aspects of individual beliefs, and a definitive conclusion regarding the association between individual beliefs and ART adherence cannot be reached. The Health Belief Model suggests that beliefs can be broken down into areas such as perceived susceptibility and self-efficacy.⁶⁰ Thus, future research exploring topics such as self-efficacy for taking medications and perceived risk of disease progression, in addition to perceived ART effectiveness and self-rated health, would allow enhanced understanding on how beliefs are related to treatment adherence among PWID.

Second, our study used self-reported adherence, which might be subject to overestimation of the true adherence level. However, VAS was reported to have good correlation with objective measures such as medication event monitoring system.⁴² Research has also supported the validity of VAS when measuring medication adherence.^39–41 Lastly, causal inferences could not be drawn from the present study due to its cross-sectional design.

Nonetheless, this very first study of ART adherence among PWID in Estonia provides insight into the factors related to adherence in this specific population. Our study shows that perceived effectiveness of ART and self-rated health are not independently associated with ART adherence, but in order to fully understand the relationship of individual beliefs to adherence among PWID, future research with comprehensive assessments on beliefs and attitudes would be necessary. In addition, we found that problem drinking and daily injecting drug use are significantly associated with less than optimal adherence to ART among the sampled PWID population. Alcohol use had been rarely investigated in adherence studies targeting PWID; thus, further research would be useful to identify the potential pathways.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institutes of Health (USA) under grant #R01AI083035 and the Estonian Ministry of Education and Research under grant #TARTH15017I.

References

European Centre for Disease Prevention and Control/World Health Organization. HIV/AIDS surveillance in Europe 2012, http://www.euro.who.int/__data/assets/pdf_file/0018/235440/e96953.pdf (2013, accessed 1 July 2015).

World Health Organization. HIV/AIDS, http://www.euro.who.int/en/health-topics/communicable-diseases/hivaids/hivaids (2015, accessed 1 July 2015).

Joint United Nations Programme on HIV/AIDS. AIDS by the numbers, http://www.unaids.org/sites/default/files/media_asset/JC2571_AIDS_by_the_numbers_en_1.pdf (2013, accessed 1 July 2015).

World Health Organization. Figures talk: HIV/AIDS in the WHO European Region, 2012. Halting and reversing the spread of HIV by 2015 are at stake, http://www.euro.who.int/__data/assets/pdf_file/0006/235599/WAD-Fact-sheet-Eng.pdf (2013, accessed 1 July 2015).

European Centre for Disease Prevention and Control / World Health Organization. HIV/AIDS surveillance in Europe 2013, http://ecdc.europa.eu/en/publications/Publications/hiv-aids-surveillance-report-Europe-2013.pdf (2014, accessed 1 July 2015).

Joint United Nations Programme on HIV/AIDS. The gap report, http://www.unaids.org/sites/default/files/media_asset/UNAIDS_Gap_report_en.pdf (2014, accessed 1 July 2015).

Laisaar

Avi

DeHovitz

et al.

Estonia at the threshold of the fourth decade of the AIDS era in Europe. AIDS Res Hum Retroviruses 2011; 27: 841–851.

World Health Organization. Key Facts on HIV Epidemic in Estonia and Progress in 2011, http://www.euro.who.int/__data/assets/pdf_file/0004/188752/Estonia-HIVAIDS-Country-Profile-2011-revision-2012-final.pdf (2013, accessed 1 July 2015).

Abel-Ollo

Rahu

Rajaleid

et al.

Knowledge of HIV serostatus and risk behaviour among injecting drug users in Estonia. AIDS Care 2009; 21: 851–857.

10.

Platt

Bobrova

Rhodes

et al.

High HIV prevalence among injecting drug users in Estonia: implications for understanding the risk environment. AIDS 2006; 20: 2120–2123.

11.

Uusküla

Des Jarlais

Kals

et al.

Expanded syringe exchange programs and reduced HIV infection among new injection drug users in Tallinn, Estonia. BMC Public Health 2011; 11: 517–517.

12.

Vorobjov

Uusküla

Des Jarlais

et al.

Multiple routes of drug administration and HIV risk among injecting drug users. J Subst Abuse Treat 2012; 42: 413–420.

13.

Arnsten

Mizuno

et al.

Factors associated with antiretroviral therapy adherence and medication errors among HIV-infected injection drug users. J Acquir Immune Defic Syndr 2007; 46: S64–71.

14.

Degroote

Vogelaers

Vermeir

et al.

Determinants of adherence in a cohort of Belgian HIV patients: a pilot study. Acta Clin Belg 2014; 69: 111–115.

15.

Cohen

Chen

McCauley

et al.

Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med 2011; 365: 493–505.

16.

Donnell

Baeten

Kiarie

et al.

Heterosexual HIV-1 transmission after initiation of antiretroviral therapy: a prospective cohort analysis. Lancet 2010; 375: 2092–2098.

17.

Uusküla

Des Jarlais

Raag

et al.

Combined prevention for persons who inject drugs in the HIV epidemic in a transitional country: the case of Tallinn, Estonia. AIDS Care 2015; 27: 105–111.

18.

von Wyl V, Klimkait T, Yerly S, et al. Adherence as a predictor of the development of class-specific resistance mutations: the Swiss HIV cohort study. PLoS One 8: e77691.

19.

Maggiolo

Ravasio

Ripamonti

et al.

Similar adherence rates favor different virologic outcomes for patients treated with nonnucleoside analogues or protease inhibitors. Clin Infect Dis 2005; 40: 158–163.

20.

Vlahov

Celentano

. Access to highly active antiretroviral therapy for injection drug users: adherence, resistance, and death. Cad Saude Publica 2006; 22: 705–718.

21.

Malta

Magnanini

Strathdee

et al.

Adherence to antiretroviral therapy among HIV-infected drug users: a meta-analysis. AIDS Behav 2010; 14: 731–747.

22.

Kamarulzaman

Altice

. Challenges in managing HIV in people who use drugs. Curr Opin Infect Dis 2015; 28: 10–16.

23.

Nolan

Milloy

Zhang

et al.

Adherence and plasma HIV RNA response to antiretroviral therapy among HIV-seropositive injection drug users in a Canadian setting. AIDS Care 2011; 23: 980–987.

24.

Altice

Mostashari

Friedland

. Trust and the acceptance of and adherence to antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 28: 47–58.

25.

Carrieri

Chesney

Spire

et al.

Failure to maintain adherence to HAART in a cohort of French HIV-positive injecting drug users. Int J Behav Med 2003; 10: 1–14.

26.

Clarke

Delamere

McCullough

et al.

Assessing limiting factors to the acceptance of antiretroviral therapy in a large cohort of injecting drug users. HIV Med 2003; 4: 33–37.

27.

Mimiaga

Safren

Dvoryak

et al.

“We fear the police, and the police fear us”: structural and individual barriers and facilitators to HIV medication adherence among injection drug users in Kiev, Ukraine. AIDS Care 2010; 22: 1305–1313.

28.

Wolfe

. Paradoxes in antiretroviral treatment for injecting drug users: access, adherence and structural barriers in Asia and the former Soviet Union. Int J Drug Policy 2007; 18: 246–254.

29.

Fumaz

Munoz-Moreno

Molto

et al.

Sustained antiretroviral treatment adherence in survivors of the pre-HAART era: attitudes and beliefs. AIDS Care 2008; 20: 796–805.

30.

Horne

Buick

Fisher

et al.

Doubts about necessity and concerns about adverse effects: identifying the types of beliefs that are associated with non-adherence to HAART. Int J STD AIDS 2004; 15: 38–44.

31.

Johnson

Catz

Remien

et al.

Theory-guided, empirically supported avenues for intervention on HIV medication nonadherence: findings from the Healthy Living Project. AIDS Patient Care STDS 2003; 17: 645–656.

32.

Uusküla

Laisaar

Raag

et al.

Antiretroviral therapy (ART) adherence and correlates to nonadherence among people on ART in Estonia. AIDS Care 2012; 24: 1470–1479.

33.

Waldrop-Valverde

Dong

Ownby

. Medication-taking self-efficacy and medication adherence among HIV-infected cocaine users. J Assoc Nurses AIDS Care 2013; 24: 198–206.

34.

Zúñiga

Muñoz

Kozo

et al.

A binational study of patient-initiated changes to antiretroviral therapy regimen among HIV-positive Latinos living in the Mexico–U.S. border region. AIDS Behav 2012; 16: 1622–1629.

35.

Boileau

Nguyen

-K Sylla

et al.

Low prevalence of detectable HIV plasma viremia in patients treated with ART in Burkina Faso and Mali. J Acquir Immune Defic Syndr 2008; 48: 476–484.

36.

Uusküla

Laisaar

K-T

Raag

et al.

ART adherence and correlates to nonadherence among people on ART in Estonia. AIDS Care 2012; 24: 1470–1479.

37.

Heckathorn

. Respondent-driven sampling: a new approach to the study of hidden populations. Soc Probl 1997; 49: 11–34.

38.

Des Jarlais

Perlis

Stimson

et al.

Using standardized methods for research on HIV and injecting drug use in developing/transitional countries: case study from the WHO Drug Injection Study Phase II. BMC Public Health 2006; 6: 54–54.

39.

Chkhartishvili

Rukhadze

Svanidze

et al.

Evaluation of multiple measures of antiretroviral adherence in the Eastern European country of Georgia. J Int AIDS Soc 2014; 17: 18885–18885.

40.

Meyer JC, Summers B, Lentsoane PP, et al. Is a simple self-rating or visual analogue scale (VAS) more reliable than prescription refill (PR) data as an indicator of non-adherence in a resource-limited setting in South Africa? 7th International Conference on HIV Treatment and Prevention Adherence, Miami, USA, 3–5 June 2012. Available at: http://www.iapac.org/AdherenceConference/assets/Adherence2012_Program.pdf (accessed 6 June 2017).

41.

Giordano

Guzman

Clark

et al.

Measuring adherence to antiretroviral therapy in a diverse population using a visual analogue scale. HIV Clin Trials 2004; 5: 74–79.

42.

Kalichman

Amaral

Swetzes

et al.

A simple single-item rating scale to measure medication adherence: further evidence for convergent validity. J Int Assoc Physicians AIDS Care 2009; 8: 367–374.

43.

Walsh

Mandalia

Gazzard

. Responses to a 1 month self-report on adherence to antiretroviral therapy are consistent with electronic data and virological treatment outcome. AIDS 2002; 16: 269–277.

44.

Ewing

. Detecting alcoholism. The CAGE questionnaire. JAMA 1984; 252: 1905–1907.

45.

Berwick

Murphy

Goldman

et al.

Performance of a five-item mental health screening test. Med Care 1991; 29: 169–176.

46.

Holmes

. A short, psychiatric, case-finding measure for HIV seropositive outpatients: performance characteristics of the 5-item mental health subscale of the SF-20 in a male, seropositive sample. Med Care 1998; 36: 237–243.

47.

Vittinghoff

Glidden

Shiboski

et al.

Regression methods in biostatistics: linear, logistic, survival, and repeated measures models, New York: Springer, 2015.

48.

Feelemyer

Des Jarlais

Arasteh

et al.

Adherence to antiretroviral medications among persons who inject drugs in transitional, low and middle income countries: an international systematic review. AIDS Behav 2015; 19: 575–583.

49.

Mann

Milloy

Kerr

et al.

Improved adherence to modern antiretroviral therapy among HIV-infected injecting drug users. HIV Med 2012; 13: 596–601.

50.

Turner

. Adherence to antiretroviral therapy by human immunodeficiency virus-infected patients. J Infect Dis 2002; 185: S143–151.

51.

Langebeek

Gisolf

Reiss

et al.

Predictors and correlates of adherence to combination antiretroviral therapy (ART) for chronic HIV infection: a meta-analysis. BMC Med 2014; 12: 142–142.

52.

Uusküla

Raag

Vorobjov

et al.

Non-fatal overdoses and related risk factors among people who inject drugs in St. Petersburg, Russia and Kohtla-Järve, Estonia. BMC Public Health 2015; 15: 1255–1255.

53.

Henrich

Lauder

Desai

et al.

Association of alcohol abuse and injection drug use with immunologic and virologic responses to HAART in HIV-positive patients from urban community health clinics. J Community Health 2008; 33: 69–77.

54.

Chander

Lau

Moore

. Hazardous alcohol use: a risk factor for non-adherence and lack of suppression in HIV infection. J Acquir Immune Defic Syndr 2006; 43: 411–417.

55.

Kahler CW, Liu T, Cioe PA, et al. Direct and indirect effects of heavy alcohol use on clinical outcomes in a longitudinal study of HIV patients on ART. AIDS Behav. Epub ahead of print 8 July 2016. DOI: 10.1007/s10461-016-1474-y.

56.

Palepu

Tyndall

et al.

Alcohol use and incarceration adversely affect HIV-1 RNA suppression among injection drug users starting antiretroviral therapy. J Urban Health 2003; 80: 667–675.

57.

Tran

Nguyen

et al.

Associations between alcohol use disorders and adherence to antiretroviral treatment and quality of life amongst people living with HIV/AIDS. BMC Public Health 2014; 14: 27–27.

58.

Weber

Huber

Rickenbach

et al.

Uptake of and virological response to antiretroviral therapy among HIV-infected former and current injecting drug users and persons in an opiate substitution treatment programme: the Swiss HIV Cohort Study. HIV Med 2009; 10: 407–416.

59.

Westergaard

Hess

Astemborski

et al.

Longitudinal changes in engagement in care and viral suppression for HIV-infected injection drug users. AIDS 2013; 27: 2559–2566.

60.

Becker

. The health belief model and personal health behavior. Health Educ Monogr 1974; 2: 354–386.