Towards cross-Canada monitoring of the unregulated street drug supply

Trends in drug use and related harms can be monitored by multiple methods of public health surveillance, including surveys, law enforcement drug seizures, coroners’ data, and drug checking services. This information is critical to detecting and responding to drug use trends with potential for harm such as intentional polysubstance use [5, 6] or unknown use of fentanyl [7]. However, these methods are often not timely, not comparable across jurisdictions, or do not reach key populations.

For example, national population-based surveys (e.g., Canadian Alcohol and Drugs Survey, Canadian Alcohol and Drug Use Monitoring Survey, Canadian Community Health Survey) provide data on self-reported use of illegal drugs but likely underestimate trends due to response and sampling bias affecting marginalized populations [8, 9]. Nationally coordinated cross-sectional surveys such as I-Track, M-Track and Y-Track (formerly Enhanced Surveillance of Canadian Street Youth or E-SYS) have improved reach to marginalized populations [10, 11], but survey cycles are infrequent. Regional cohort or cross-sectional studies of underserved populations [12‐21] have improved upon the limitations described above, but methodological differences between studies (e.g., variation in reporting period of substance use) preclude comparing trends across regions. Regional studies also variably capture important subgroups of PWUD, such as those who live outside of urban centers or use drugs by non-injection means.

Other means of monitoring unregulated drug contents include analyzing samples seized by law enforcement (e.g., Health Canada’s Drug Analysis Service) [22], substances identified in urine drug screens collected in health care settings [23, 24], and substances identified by toxicology of decedents [25]. However, these methods are limited in their ability to confirm what individuals intended or expected to consume, and, in the case of toxicology data showing multiple substances, whether these were consumed at the same time or separately.

Drug checking has emerged as a quick and inexpensive method to generate data on drug contents and consumer expectations [26‐28]. However, the range of detectable substances and levels of specificity and sensitivity vary among different techniques. Surrendering and handling illegal substances can also pose challenges [29].

The methods above contribute vital information to our understanding of substance use trends and related harms, but there remain gaps in timeliness, representativeness, and sustainability of these existing systems [30].

In 2015, the British Columbia Centre for Disease Control (BCCDC) began analyzing expected and actual drug use by combining an existing harm reduction survey with a urine norfentanyl test. This project — known as the Fentanyl Urine Screen Study (FUSS) — provided timely access to information on unintentional fentanyl use and associated behaviours (e.g., concurrent methamphetamine use) [31]. The FUSS survey tool and methodology were based on an annual survey implemented at harm reduction sites (health or social services aiming to prevent a wide range of drug-related harms) across BC from 2012 to 2015. This annual survey was identified as a feasible, low-barrier means to obtain locally-relevant drug use data from clients while also illustrating important regional differences [32]. The development, implementation and adaptation of the original survey tool has been described in detail previously [32]. In 2017, the Direction régionale de santé publique (Regional Public Health Department) of the Centre intégré universitaire de santé et de services sociaux (CIUSSS) du Centre-Sud-de-l’Île-de-Montréal (DRSP-CCSMTL) conducted a similar study that built on the FUSS methodology. Broad spectrum urine toxicology tests were employed to examine local prevalence of fentanyl and compare expected and actual substances used by participants. This study demonstrated the feasibility of using broad spectrum urine toxicology tests to examine contents of urine samples submitted by participants. Additionally, both studies identified a discrepancy between self-reported and detected use of drugs; presence of fentanyl among participants who did not report its use ranged from 73% in BC in 2015 [31] to 18% in Montreal in 2017 [33].

To scale up this pilot monitoring process, partners from the BCCDC, DRSP-CCSMTL, University of Alberta School of Public Health and Streetworks Edmonton obtained a Health Canada, Substance Use and Addiction Program grant that aimed to develop a cross-Canada surveillance system of illegal drug content using standardized tools that can be implemented across the country. The aim was to develop a system that could:

This paper describes the feasibility of implementing this system and preliminary indications that these goals were met.

The pilot system was developed collaboratively with representatives from the BCCDC, DRSP-CCSMTL, University of Alberta and Streetworks Edmonton, and the Canadian Centre on Substance Use and Addiction (CCSA). These partners built upon the design used in the FUSS and Montreal drug testing projects described above. For this pilot system, the survey focused on past three-day drug use, history of overdose and injection drug use and basic demographics (age, gender). The three-day reporting period was chosen since it is consistent with the time frame of detection of most drugs in urine [34] and improves recall [35]. The survey consisted of “core” questions common to all sites (e.g., drugs used, experience of opioid overdose, injection drug use), and optional additional items based on local and emerging needs. All survey content was developed with input from PWUD.

In parallel with survey development, protocols were also developed for analysis of urine samples by different laboratories (LifeLabs, Ontario for BC and Edmonton sites, and Centre de toxicologie du Québec for Montreal sites). Briefly, samples were to be screened with liquid chromatography and mass spectrometry for approximately 150 drugs, including 40 opioids and several fentanyl analogues. The broad spectrum urine toxicology screen method used by LifeLabs has been described in detail previously [36]. The cost of urine toxicology ranged between 75 and 90 Canadian dollars per sample.

The survey and urine sample collection were piloted between May 2018 and March 2019 at harm reduction sites in BC (May to August, 2018), Edmonton (March, 2019) and Montreal (August to September, 2018). Sites included harm reduction supply distribution services that stood alone or were integrated with supervised consumption sites or other health and social services for PWUD. Sites were recruited based primarily on site capacity (e.g., availability of staff to support data collection, washroom facilities for urine sample collection, hours of operation), as well as willingness to participate and feedback from regional harm reduction coordinators.

Participant recruitment and data collection ranged from two to four weeks. Direct service providers or PWUD involved in providing site services approached clients and invited those eligible to participate. Participants provided verbal consent, a low-barrier approach that does not require personal identifying information and is consistent with low-barrier requirements for receiving harm reduction services [32]. Participants were included if they were over the provincial age of majority and reported use of an illegal drug other than cannabis in the past 6 months (BC and Edmonton) or 3 days (Montreal). In Edmonton and Montreal, all participants completed the survey and urine sample and were offered ten dollars in compensation. In BC, the urine sample was optional, and participants were compensated with five dollars for each of the survey completion and providing a urine sample. Participating sites in BC and Edmonton were also provided five dollars per participant enrolled to remunerate the site for staff assistance with recruitment and data collection. Staff were encouraged to assist with administration of the paper-based survey at all sites, though participants wishing to self-administer were permitted to do so.

The completed survey and urine sample obtained from each participant were assigned matching anonymous ID codes to allow for data linkage while maintaining participant anonymity. Survey and toxicology data were received, entered, linked and analyzed by BCCDC for BC and Edmonton sites and by DRSP-CCSMTL for Montreal sites. Results were summarized by site and region, and presented to participating sites, local harm reduction coordinators and other stakeholders (e.g., harm reduction and overdose response teams, provincial ministries of health).

The quantitative results presented here can help understand local trends and needs, and can justify future work to understand and address the emerging needs of PWUD. For example, stimulant use was prevalent in all regions, raising the need to better understand associated demands (e.g., for pipes and other harm reduction supplies [37]) and supports for those who use stimulants only or in combination with opioids. On a similar note, polysubstance use was another important trend. Most participants reported using two (BC, Montreal) or three (Edmonton) substances. Reported polysubstance use was consistent with previous research in similar populations [38‐40]. There is a need to better understand why PWUD use multiple substances in specific patterns given accumulating evidence on risk of harms related to poor physical and mental health [41] and overdose [1, 42, 43]. Harms may be compounded by adulterants such as levamisole, a non-psychoactive substance found in cocaine or crack that poses risk of multiple serious adverse effects [2, 3, 44, 45].

Results may also be used to track trends over time. Rates of unintentional fentanyl use observed in the present study have decreased compared to earlier prototype studies. In BC, among those with fentanyl detected, only 36.1% used unintentionally in 2018 versus 72.8% in 2015 [31]. Among the entire sample in Montreal, fentanyl was detected but not reported used among 8.9% of respondents in 2018 versus 18.0% in 2017 [33]. It will be important to continue to monitor trends in intentional and unintentional fentanyl use to ensure overdose prevention efforts remain relevant and evidence-informed.

This system also provides an opportunity to examine use of prescription opioids (POs) (e.g., methadone, morphine, hydromorphone) by individuals accessing harm reduction services. Participants reporting use of POs in this study may have obtained them with a prescription or from the diverted supply in the unregulated market. The latter is preferred by some PWUD who face barriers to obtaining a prescription but wish to use opioids of predictable type and quantity [46]. To further complicate matters, apparent diverted POs in the street supply may also be illegally manufactured. Counterfeit POs often contain unexpected substances of varying potency, including novel synthetic opioids such as fentanyl and fentanyl analogues [47]. Given the interplay between PO use and harms experienced by PWUD in Canada [48], a nuanced understanding of use by clientele of harm reduction services is needed to inform interventions such as safer supply.

In addition to assessing expectations of drug contents, the survey also provides an opportunity to gather detailed contextual information about drug use. Results presented here are a small snapshot of information available from the survey, which includes, for example, reasons for using drugs alone [49], unintentional fentanyl use [36], naloxone kit access [50], acceptability of overdose monitoring applications [51], and changes in methadone formulation [52]. This information has been used to provide feedback for improvement and quality assurance of harm reduction services, and findings have helped create targeted knowledge translation approaches. For instance, naloxone kit ownership improved after identifying lower ownership among those who smoke substances [50]. Other trends currently under investigation include knowledge and understanding of the Good Samaritan Drug Overdose Act, concurrent methamphetamine and opioid use, and shifting modes of opioid use (i.e., to inhalation). This information will likely inform future harm reduction interventions that are responsive to the realities of PWUD. In the future, optional content can be further adapted to allow regions flexibility to ask questions that are of local interest, reflect the local drug use and harm reduction landscape and its emergent issues, or evaluate local interventions.

The pilot system leveraged existing harm reduction supply distribution sites and used a low-barrier design suitable to these environments (e.g., anonymous participation). This allows the system to be administered regularly in many types of sites accessed by PWUD with respect to service delivery model (standalone or integrated harm reduction services), geography (urban, rural or remote), and capacity for data collection. In addition, unlike with methods such as drug checking, participants are not faced with the choice of surrendering drugs for testing, and staff are not required to handle illegal samples or interpret complex results. Toxicological screening also allows systematic detection of a wide range of substances across participating sites, which is not always possible with other drug checking methods (e.g., fentanyl immunoassay strips) [29].

The original survey tool implemented in BC led to additional positive benefits in harm reduction sites [32]. Its anonymous nature encouraged client participation, and involvement of site staff (including PWUD) in data collection provided an opportunity to engage and build positive rapport with clients. The current system preserves this reciprocal process: participants are able to describe their experiences with the unregulated street drug supply, and site staff are able to learn about these experiences. Aggregate results are then shared with clients to ensure they feel heard and that their knowledge is accurately represented. As discussed by Wallace et al. [53] in the context of drug checking services, trust is the keystone of harm reduction wherein any intervention is “implemented in a context of stigma, trauma and other risks”.

The study incorporated principles of participatory research, which greatly improved the relevance, reach and uptake of the pilot through relationships with networks of PWUD and harm reduction service providers. For instance, PWUD were encouraged and compensated for leading participant recruitment and data collection. In some sites, these partners also guided development of the survey tool and disseminated findings through their networks. The latter is crucial as knowledge shared within communities of PWUD can be highly influential [54]. Such participatory processes form the backbone of an effective drug content monitoring system [30] and improve the rigor, relevance and reach of results for PWUD and their communities – the “3 R’s” of participatory research [55].

There are several limitations to the urine toxicology screening method. Drug detection timeframes and the reporting period (past three-day use) align for most but not all substances [34]. For instance, prevalence of use may be underestimated for drugs such as fentanyl and heroin which clear rapidly from urine. It can also be difficult to determine the origin of certain detected substances (e.g., morphine and metabolized heroin-related compounds are indistinguishable). The screening method only indicates whether a substance is present but not how much. Additionally, one can examine associations between expected and unexpected substances (e.g., cocaine/crack and levamisole) but cannot conclude precisely where contamination occurred if multiple substances are reported used. For these reasons, urine toxicology screening is complementary to drug checking services that can provide detailed qualitative and quantitative information about drug contents [29]. Various monitoring methodologies may be combined to further clarify contents of the street drug supply year to year; the BC Drug Overdose and Alert Partnership is one example of this synthesis of monitoring data [23].

Availability of staff time and washrooms to facilitate urine collection were among the primary factors in site selection. Risk of selection bias is low as a systematic relationship between these factors and drug use is unlikely. While non-response rates were not available, evidence of low refusal and non-completion rates in an earlier prototype [32] and the low-barrier context of data collection suggest non-response bias would be minimal. However, findings from this pilot project may not be generalizable to all harm reduction sites and PWUD in each region.

Additional limitations relate to the reliance on anonymous self-report survey data. A small number of individuals may have participated multiple times. Any effect on results is likely minimal, especially as drug use data would be novel if contributed three or more days apart. Recall of used substances may be biased, though the chosen reporting period (past three-day use) improves recall accuracy compared to longer reporting periods [35] and is consistent with most urine drug detection timeframes [34]. Underreporting of certain self-reported behaviours due to social desirability bias may also have occurred. However, evidence that the survey tool provides an opportunity for rapport-building suggests that respondents have positive and trusting interactions with the staff or PWUD collecting data [32].

Finally, there are limits to the comparability of the three sites in this study. In this pilot phase, inclusion criteria differed slightly between regions and may have led to samples with slightly different substance use characteristics. This criterion differed due to historical eligibility for the studies upon which this system is based. In addition, some results from the pilot phase are not strictly comparable across sites due to differences in survey question reporting periods. Future phases of the project will standardize participant eligibility criteria and reporting periods to improve comparison of results across all regions.

In light of the results presented here, trends in unintentional or unknown substance use will be monitored closely in future iterations of this project. This need is further justified by the record-high numbers of accidental drug poisoning deaths coinciding with the COVID-19 pandemic [56, 57]. Validity of self-reported substance use may also be measured, as done recently for methamphetamine use in BC [58]. Future work may also examine how substances such as POs are obtained (e.g., with or without a prescription) to provide further context on the state of the unregulated street supply.

In the study’s next phase, partners will collaborate to standardize materials and methods and facilitate scale-up to additional sites across Canada. A “project toolkit” will be developed, including a standardized survey containing core and optional content. Adaptation of optional content according to local relevance, need, and consultation with PWUD and harm reduction stakeholders will be encouraged; question wording will be kept as similar as possible to maintain comparability. The “project toolkit” will also contain guidelines based on best practices in implementation, data analysis and result dissemination to expedite these processes for new sites, particularly those with few resources available for research. This form of knowledge translation prioritizes capacity-building and system sustainability. Following scale-up to additional regions, results from all participating sites will be analyzed to form a national picture on key drug use trends.