National findings from the Tracks survey of people who inject drugs in Canada, Phase 4, 2017–2019

Background: The Tracks survey of people who inject drugs (PWID) collected data in 14 sentinel sites across Canada (2017–2019). Objective: To describe the prevalence of human immunodeficiency virus (HIV) and hepatitis C and associated risk behaviours and to examine trends over time. Methods: Information regarding socio-demographics, social determinants of health, use of prevention services and testing, drug use, risk behaviours, and HIV and hepatitis C testing, care and treatment was collected through interviewer-administered questionnaires. Biological samples were tested for HIV, hepatitis C antibodies and hepatitis C ribonucleic acid (RNA). Descriptive statistics were calculated and trends over time were assessed. Results: Of the 2,383 participants, 65.6% were cisgender male, 42.2% were Indigenous, 48.0% completed some high school or less, 62.6% lived in unstable housing and 75.7% had ever been incarcerated. Average age was 40.1 years. The majority experienced stigma and discrimination (88.7%) and physical, sexual and/or emotional abuse in childhood (85.0%) or with a sexual partner (75.9%). The majority reported use of a needle/syringe distribution program (90.1%) and tested for HIV (90.5%) and hepatitis C (90.9%). Among participants who had ever had sex, the majority (59.2%) reported inconsistent condom use during vaginal and/or anal sex with a casual sex partner. Prevalence of HIV was 10.3% (82.9% were aware of infection status) and many (36.9%) were hepatitis C RNA-positive (50.1% were aware of infection status). Most surveillance indicators remained relatively stable from Phase 1 to Phase 4. Changes were found in substances used, and improvements were noted related to HIV and hepatitis C prevalence and care cascade indicators. Conclusion: Many PWID in Canada were living in unstable housing and experienced high levels of stigma and discrimination. Prevalence of HIV and hepatitis C was high in some areas. These findings contribute to the evidence base used to inform targeted prevention and control measures. Affiliations 1 Centre for Communicable Diseases and Infection Control, Public Health Agency of Canada, Ottawa, ON 2 National Mircrobiology Laboratory, Public Health Agency of Canada, Winnipeg, MB *Correspondence: jill.tarasuk@canada.ca Suggested citation: Tarasuk J, Zhang J, Lemyre A, Cholette F, Bryson M, Paquette D. National findings from the Tracks survey of people who inject drugs in Canada, Phase 4, 2017–2019. Can Commun Dis Rep 2020;46(5):138– 48. https://doi.org/10.14745/ccdr.v46i05a07


Introduction
People who inject drugs (PWID) represent an important group at risk for human immunodeficiency virus (HIV) and hepatitis C in Canada. It has been estimated that, of the 2,165 new HIV infections in 2016, the proportion attributed to injection drug use was 11.3%. This value had not decreased since the 2014 estimate of 11.2% (1). In North America, the burden of disease of hepatitis C attributable to injecting drug use is greater than for HIV, and was estimated to be 81% in 2013 (2). In Canada, HIV and hepatitis C antibody prevalence was high among PWID surveyed in 2010-2012 (11.2% and 68%, respectively) (3). These findings underscore the need for prevention and treatment efforts to mitigate HIV and hepatitis C morbidity and mortality in this population. Integrated bio-behavioural surveillance, an established World Health Organization (WHO)/Joint United Nations Programme on HIV/AIDS (UNAIDS) globally-endorsed approach (4), is critical to informing response and to guiding public health interventions. This surveillance provides information about risk practices and health-seeking behaviours among the populations most at risk for HIV and is necessary to better understand the factors driving transmission.
The Public Health Agency of Canada (PHAC), in conjunction with provinces and territories including regional and/or local public health partners, monitors trends in the prevalence of HIV and hepatitis C and associated risk factors in key populations, such as PWID, through the Tracks Surveillance Systems. The Tracks survey of PWID (formerly I-Track) involves repeated crosssectional surveys at selected sites across Canada. It was first implemented in 2003-2005 (Phase 1) in seven sentinel sites. This was followed by three subsequent data collection periods, including the most recent survey, Phase 4 (2017-2019), in 14 sentinel sites (Appendix 1).
The objective of this report is to present national surveillance findings from Phase 4 of the Tracks survey of PWID in Canada, conducted between January 1, 2017 and May 9, 2019, at participating sentinel sites in Canada. Findings include socio-demographic characteristics, social determinants of health, use of prevention services and testing, drug use and experiences with overdoses, sexual risk behaviours and HIV and hepatitis C care cascade, prevalence and awareness of infection status. Selected indicators from Phase 1 to Phase 4 of the Tracks survey of PWID are also presented to describe trends over time.

Data source and sampling methods
The data presented in this report are from the Tracks survey of PWID in Canada. The Tracks survey of PWID makes use of venue-based sampling, in which participants are recruited from settings in which they are likely to gather, most often, but not limited to, needle and syringe distribution programs. Individuals who had injected drugs six months prior to recruitment and who met the minimum age to provide consent, which was determined at each site according to local research ethics requirements, were eligible to participate in the survey. Eligible and consenting participants completed an interviewer-administered questionnaire and provided a biological sample in the form of a dried blood spot (DBS) specimen (or oral fluid exudate in the SurvUDI network sites).
The surveillance protocol and questionnaire were approved by the Health Canada/PHAC Research Ethics Board, and by local research ethics boards at each sentinel site where required. The same sampling and recruitment strategies and core questionnaire, with minor revisions, were used across all four phases to ensure comparability over time. Survey methods are described in more detail elsewhere (3).

Sentinel site selection
Sentinel sites were selected based on consultations with provincial/territorial representatives, who considered the epidemiology of HIV, hepatitis C and drug use and associated harms. Given this assessment, participating sentinel sites varied by phase of the Tracks Survey of PWID (Appendix 1). Data collection in Ottawa (Ontario) and in the province of Quebec was coordinated by the SurvUDI network (5). The SurvUDI network sites were divided into four geographical zones for the Phase 4-specific analyses (see Appendix 1).

Interviewer-administered questionnaire
The Tracks PWID questionnaire collects information about socio-demographic characteristics, social determinants of health, use of health and prevention services (including testing), drug use and injecting behaviours, sexual behaviours and care and treatment for HIV and hepatitis C. The questionnaire was first developed for a pilot phase by an expert working group to establish face validity. To ensure comparability, each subsequent phase retained most national-level questions to monitor change over time.
The Phase 4 questionnaire included a limited number of revisions, including new national-level questions that addressed gender identity, financial strain, mental health status, experiences of stigma and discrimination, physical, sexual and/or emotional abuse, borrowing used non-injection drug paraphernalia, overdose-related experiences, use of harm reduction services, condomless sex at last paid sex, substance use before or during sex, adherence to antiretroviral treatment and viral load status.

Biological sample
Dried blood spot samples were tested for HIV (antibody and antigen) and hepatitis C (antibody and RNA). Participants were not informed of their laboratory test results because no identifying information was collected to ensure participant anonymity. Sentinel sites were asked to provide on-site testing (e.g. point of care testing, full phlebotomy) during recruitment times so that participants who were not aware of their status could get tested, should they wish. Where on-site testing was not feasible, participants were referred to local testing sites and/or health care services. Updated laboratory testing algorithms for DBS were introduced in Phase 4 (see Appendix 2). Testing algorithms for SurvUDI samples are found in Appendix 2.

Analysis
Descriptive statistics for selected indicators were computed with SAS Enterprise Guide 7.1. Selected indicators from Phase 1 to Phase 4 were compared to examine trends over time. Small cell counts were assessed to determine the risk of identifying individual participants, and were left in when it was determined that there was no risk of reidentification, as per PHAC's Directive for the collection, use and dissemination of information relating to public health (PHAC, 2013, unpublished document). Participants who responded as "not stated", "don't know" or "refused" were excluded from each individual analysis.

Results
Sample sizes for Phase 1, Phase 2 and Phase 3 were 2,986, 2,982 and 2,687, respectively. A total of 2,383 individuals were eligible and consented to participate in the Phase 4 survey, among whom 2,379 (99.8%) completed a questionnaire and 2,162 (90.7%) provided a biological sample. Findings for selected indicators by socio-demographic characteristic and social determinants of health of participants are provided in Supplemental tables A (prevention and testing indicators), B (injecting behaviours and drug use), C (sexual risk behaviours) and D (selected indicators by Phase).

Socio-demographic characteristics
In Phase 4, 65.6% identified their gender as cisgender male, 32.7% as cisgender female and 1.0% as transfeminine (i.e. those assigned male at birth who identified with either female or a non-binary gender) and 0.7% transmasculine (i.e. those assigned female at birth who identified with either male or a non-binary gender) ( Table 1

Social determinants of health
Among Phase 4 participants, just under half (48.0%) had completed some high school or less and a large proportion

SURVEY REPORT
(86.0%) experienced financial strain (i.e. difficulty making ends meet) in the 12 months prior to the interview ( Table 2). Overall, 62.6% of participants lived in unstable housing in the six months prior to the interview and 75.7% reported having ever been incarcerated. A large proportion (84.0%) reported their mental health as "fair to excellent" and 16.0% reported poor mental health status. Among Indigenous participants, 83.1% had attended a residential school themselves or had a family member who had attended a residential school.
Experiences of stigma and discrimination (related to racial or cultural background, hepatitis C status, HIV status, sexual orientation, use of drugs or alcohol or sex work) were reported by the majority of participants (88.7%). Large proportions of participants had experienced physical, sexual and/or emotional abuse in childhood (85.0%) or with a sexual partner (75.9%).
Over the past four phases, the social determinant indicators stayed relatively stable with the exception of an increase in the proportion reporting living in unstable housing in the six months prior to the interview (51.1%-62.6%) (Supplemental table D).

Use of prevention services and testing
In Phase 4, the majority of participants (90.1%) reported using a needle and syringe distribution program in the 12 months prior to the interview, with lower proportions reporting use of methadone, suboxone or other opioid substitution therapy (47.3%) and use of a supervised injection or consumption site (13.5%). The majority of participants reported ever testing for HIV (90.5%) and hepatitis C (90.9%) ( Table 3). Some (14.3%) of the participants had heard about preexposure prophylaxis (PrEP). Among participants who did not report an HIV diagnosis, 0.4% had used PrEP in the 12 months prior to the interview to reduce the risk of contracting HIV. The proportion of participants who had ever tested for HIV (90.0%-92.9%) and hepatitis C (87.5%-91.3%) was high and varied slightly across all phases (Supplemental table D).

Injecting behaviours
In Phase 4, over one-third of participants (38.1%) reported injecting daily in the month prior to the interview and over half (52.7%) reported injecting in a public space in the six months prior to the interview. Overall, 11.6% of participants injected with used needles and/or syringes in the six months prior to the interview, of whom the majority (85.2%) borrowed needles and/ or syringes from people who they knew well (i.e. family, friends or sex partners). Over one-third (38.0%) injected with other used injection equipment such as water, filters, cookers, tourniquets, swabs or acidifiers in the six months prior to the interview. Among those who borrowed used equipment, the majority (82.9%) reported borrowing from people they knew well. More than half of participants (56.0%) borrowed used non-injection a Total represents total counts for the corresponding indicator excluding "don't know", "refused" and "not stated" values b Defined as ever having difficulty making ends meet in the year prior to the interview c Unstable housing included living in a hotel or motel room, rooming or boarding house, shelter or hostel, transition or halfway house, psychiatric institution or drug treatment facility, public place or correctional facility d Only partial data available at the SurvUDI network sites e Defined as ever experienced any stigma or discrimination (e.g. avoidance, pity, blame, shame, rejection, verbal abuse or bullying) based on racial or cultural background, hepatitis C status, HIV status, sexual orientation, use of drugs or alcohol or sex work  drug paraphernalia such as straws, dollar bills, or pipes in the six months prior to the interview ( Table 4).
The proportion of participants who reported borrowing used needles and/or syringes decreased by almost half from 20.2% in Phase 1 and 21.8% in Phase 2 to 11.6% in Phase 4. In contrast, the proportion who reported borrowing other used injection equipment (such as water, filter, cooker, spoons, tourniquets, ties, swabs and acidifiers) increased by almost a third from Phase 1 (29.8%) to Phase 4 (38.0%) (Supplemental table D).
Among Phase 4 participants, the majority had heard of overdose kits (87.5%) and reported that kits were available in their community (96.4%); a lower proportion had ever used one on someone else (32.0%). Nearly one-quarter (22.6%) had overdosed in the six months prior to the interview and the drugs most commonly reported at last overdose were fentanyl (43.0%), heroin (38.3%), methamphetamine (28.4%), cocaine (23.1%) and alcohol (15.9%) ( Table 5).

Sexual risk behaviours
In Phase 4, in the six months prior to the interview, among participants who had ever had sex, 35.2% had two or more sexual partners, 59.2% had inconsistent condom use during vaginal and/or anal sex with a casual sex partner, 84.9% had inconsistent condom use during vaginal and/or anal sex with a regular sex partner and 15.7% had engaged in transactional sex at least once ( Table 6). Among those that engaged in transactional sex, 30.7% did not use condoms at last transactional sex. The majority of participants (84.2%) reported substance use before or during sex (Table 6).
Across all phases, of participants who had ever had sex in the six months prior to the interview, the proportion who had two or more sex partners and who had engaged in transactional sex stayed relatively stable (Supplemental table D).

HIV and hepatitis C prevalence and awareness
Based on the laboratory testing, HIV prevalence was 10.3% and of those who were HIV-positive, 82.9% were aware of their HIV-positive status ( Table 7). The proportion who tested positive for hepatitis C antibodies was 64.2%, which is a measure of lifetime exposure to hepatitis C infection. Many

HIV and hepatitis C care cascade
Indicators measuring the HIV care cascade were examined among participants aware of their HIV-positive status ( Table 7). The majority were under the care of a doctor or health care provider for HIV-related services at the time of the interview (95.0%). The majority had also taken antiretroviral therapy (ART) (97.2%) and were currently taking ART at the time of the interview (87.7%). Adherence to ART, measured as no missed doses in the month prior to the interview, was 42.5%. Among participants currently taking ART at the time of the interview, 62.8% reported an undetectable HIV viral load. Nearly half of all  participants who were aware of their HIV-positive status reported avoiding HIV services because of stigma and discrimination in the 12 months prior to the interview (45.3%).
Indicators measuring the hepatitis C care cascade were examined among participants who were aware of their current hepatitis C infection (Table 7). Nearly half (48.5%) reported being linked to care for hepatitis C; a smaller proportion (10.6%) had ever taken hepatitis C treatment; and an even smaller proportion (3.8%) were currently taking hepatitis C treatment.
From Phase 1 to Phase 4, among participants aware of their HIVpositive status, linkage to care for HIV-related services increased (88.1%-95.0%) as did the proportion of those currently taking ART treatment (52.0%-87.7%). Across all phases, only about half of the participants who were aware of their hepatitis C infection status were under the care of a doctor for their hepatitis C infection and the proportion currently taking hepatitis C treatment was very low (Supplemental table D).

Discussion
People who inject drugs represent an important risk group in Canada's HIV and hepatitis C epidemics (1). Information gathered from the Tracks survey of PWID in Canada help contextualize the epidemiology of HIV, hepatitis C and associated risk behaviours among this population, providing comparisons over time and new baseline data on key emerging indicators, such as experiences of stigma and discrimination, overdoses and the use of PrEP. Factors associated with increased vulnerability to HIV and hepatitis C in previous studies were also identified among this survey sample of PWID. Markers of poverty and marginalization, including high numbers living in unstable housing and/or ever incarcerated, were common. Lived experience of stigma and discrimination, as well as physical, sexual and/or emotional abuse (in childhood or with a sexual partner), were also identified by the majority of participants.
High rates of testing for HIV and hepatitis C and the use of needle and syringe distribution programs were noted. However, access to other key harm reduction services was lower, with less than half of the participants reporting the use of opioid-substitution therapy or the use of a supervised injection or consumption site in the previous year. Drug use and injecting behaviours reported in Phase 4 signalled important proportions of participants who borrow needles and/or syringes and other used injecting equipment. The majority of participants (59.2%) reported inconsistent condom use with a casual sex partner and 84.2% reported substance use before or during sex, both of which are associated with the transmission of STBBI including syphilis. Preexposure prophylaxis awareness was low among participants (14.3%), and the use of PrEP was only 0.4% among those who did not report an HIV-positive diagnosis. Abbreviations: ART, anti-retroviral therapy; HCV, hepatitis C virus; HIV, human immunodeficiency virus; RNA, ribonucleic acid a Total represents total counts for the corresponding indicator, excluding "don't know", "refused", and "not stated" values b Among participants who provided a biological sample of sufficient quantity for HIV testing c HIV and hepatitis C testing algorithms are provided in Appendix 2 d Among participants who tested positive for HIV antibodies and who reported their HIV diagnosis. Participants who reported that their last HIV test result was positive and who were found to be HIV positive based on testing of the biological specimen provided at the time of interview were classified as being aware of their HIV positive status e Defined as under the care of a doctor or health care provider for HIV-related services at the time of the interview (in the six months prior to the interview in the SurvUDI network and London sites) f This question was not asked at the SurvUDI network (n=65) and London sites (n=17). The denominator also excludes participants with missing data g Among participants currently on ART treatment at the time of the interview. This question was not asked at the SurvUDI network sites (n=62). The denominator also excludes participants with missing data h Among participants who provided a biological sample of sufficient quantity for HCV antibody testing i Among participants who provided a biological sample of sufficient quantity for HCV antibody and RNA testing. HCV RNA testing was not conducted at the SurvUDI network sites j Among participants who tested HCV RNA positive and who reported their current hepatitis C status. Participants who reported being currently infected with hepatitis C and who were hepatitis C RNA positive based on testing of the biological specimen provided at the time of interview were classified as being aware of their hepatitis C RNA positive status k Defined as under the care of a health care provider for hepatitis C-related services at the time of the interview. The denominator excludes participants with missing data l The denominator excludes participants with missing data m Among participants who provided a biological sample of sufficient quantity for testing for both HIV antibodies and HCV RNA testing. HCV RNA testing was not conducted at the SurvUDI network sites Currently taking hepatitis C treatment l 9 236 3.8

SURVEY REPORT
Many of the national surveillance findings are consistent with the findings from other integrated bio-behavioural surveillance systems with comparable HIV and hepatitis C epidemics. Specifically, among PWID surveyed in the United States, Australia and the United Kingdom, similar levels of prevention and testing indicators (i.e. testing for HIV and hepatitis C, use of opioid substitution therapy), injecting behaviours (i.e. borrowing used needles and/or syringes, borrowing used other injection equipment) and sexual practices (i.e. transactional sex, condomless sex) were found (6-8). Previous regional studies among PWID in Canada have also found similar levels of unstable housing (9,10), and high proportions who have experienced violence (10), and abuse (9,10).
The ongoing opioid crisis and other drug-related overdose deaths have greatly affected the population of PWID in Canada. Increased use of methamphetamine, fentanyl and opioid analgesics found among Phase 4 participants echo this alarming trend. Phase 4 surveillance findings provided information regarding new overdose-related indicators. While awareness and access of overdose kits was high, 22.6% had overdosed in the six months prior to the interview with fentanyl and heroin the most commonly reported drugs used at last overdose.
While HIV prevalence among Phase 4 participants (10.3%) had decreased since Phase 1 (conducted in 2003-2005) it was nevertheless high-nearly 10-fold higher compared with rates among PWID in Australia and the United Kingdom (7,8). A slightly higher proportion of participants were aware of their HIV-positive status in Phase 4 (82.9%) compared with the previous phases. For the first time, hepatitis C RNA prevalence was measured in the Tracks survey of PWID and found to be high (36.9%). In addition, only 50.1% of participants were aware of their hepatitis C RNA-positive status (i.e. current hepatitis C infection).
Nearly all participants who were aware of their HIV-positive status were linked to care for HIV-related services and were currently taking ART; however, less than two-thirds (62.8%) reported an undetectable viral load and 45.3% reported avoiding HIV services because of experienced stigma and discrimination. Much lower rates for linkage to care (48.5%) and current treatment use (3.8%) were found among participants who self-reported current infection with hepatitis C. Low numbers of PWID who are linked to hepatitis care and treatment have been observed in other regional studies in Canada (11).
The results from the Phase 4 Tracks survey of PWID can inform evidence-based strategies to address gaps in prevention, testing and linkage to care approaches. This can include better linkage to opioid substitution therapy and supervised injection or consumption sites, and improve access to health and social services for mental health and addictions (12). The confluence of high rates of hepatitis C combined poor awareness, continued but reduced needle sharing and inconsistent condom use despite increased rates of program uptake highlights the need harm reduction programs to continue to evolve to meet these challenges.

Strengths and limitations
The Tracks survey of PWID is a rich source of information on HIV and hepatitis C among PWID from sites across the country, and provides trends on key indicators since 2003. Notably, it is the only national source of such information, and has been used at the local, provincial and federal levels to inform and guide public health interventions in this population. However, it is important to note that the Tracks survey uses non-probability-based sampling; therefore, findings may not be representative of all PWID at any given site or in Canada. With the exception of the laboratory results, these findings were based on interviewer-administered questionnaires and self-reported data and it is possible that certain risk behaviours were over-or underrepresented.

Conclusion
High levels of unstable housing, experienced stigma and discrimination, borrowing of used injection equipment and inconsistent condom use were found. Both HIV prevalence and hepatitis C RNA-positive prevalence is high among PWID in some areas of Canada. Important gaps related to linkage to care and treatment for hepatitis C were identified. These findings highlight the need for: continued access to testing and prevention services, targeted strategies to address barriers to accessing HIV and hepatitis C treatment and care and improvements in ongoing supports for housing, mental health and addictions. provincial, regional and local health authorities, researchers and community-based organizations. The authors gratefully acknowledge the contribution of the participants, the sentinel site teams, including survey coordinators and interviewers, and the site principal investigators: M Alary; K Chokani; J DeMille; B Enns; M Gully; B Hanley; M Hennink; P Leclerc; C Mackie; S Marshall; C Morissette; É Roy; S Shaw; C Smith; A Vanderlaan; and D Warren. The authors would also like to thank C Archibald for his advice prior to and throughout data collection and for his review of the manuscript and C Daniuk for her support with the dried blood spot testing.

Funding
The Tracks survey of people who inject drugs in Canada is supported by the Public Health Agency of Canada, with in-kind contributions provided by regional and local health authorities and community-based organizations.

Hepatitis C testing algorithms
For all non-SurvUDI network sites: hepatitis C screening testing was performed using the Ortho ® HCV version 3.0 EIA (Ortho). A non-reactive result indicated never having been infected with hepatitis C. A reactive result indicated lifetime exposure to hepatitis C. Confirmatory testing was performed on screened reactive and indeterminate results (i.e. absorbance results that were near, but did not overlap, the cut-off value for a reactive/non-reactive result) using the Roche COBAS AmpliPrep/COBAS Taqman HCV Quantitative test v2.0 (Roche). A detected result indicated a current hepatitis C infection and a not detected result indicated a lifetime exposure to hepatitis C. For those that screened indeterminate on the Ortho, a detected result on the Roche indicated a current hepatitis C infection and a not detected result on the Roche was interpreted as an indeterminate result.
SurvUDI network sites: hepatitis C antibody testing for oral fluid specimens was performed using the Ortho ® hepatitis C version 3.0 EIA at the Institut national de santé publique du Québec laboratories. Confirmatory testing was not performed for samples that tested positive. A positive result indicated past or present hepatitis C infection and did not discriminate acute from chronic or resolved infections. Validation of this test for use with oral fluid was performed in the SurvUDI study.