Introduction

In the 2020s, ongoing rises in opioid-related overdose deaths have heightened awareness among health service professionals, community leaders, and politicians of the need for services to be provided to people who use drugs (PWUD) that will reduce the likelihood of overdose and other adverse drug-related outcomes. In the US, overdose deaths increased almost 15% from 2020 to 2021, rising to 107,622, of which more than 70,000 involved fentanyl [1], further extending the rises witnessed from 2019 to 2020 [2]. Deaths rose among all racial and ethnic groups [2], with relative increases observed in 2020 higher than any previous year since 1999 [3]. Mortality rate ratios from 2020 show overdose fatalities among older African Americans and American Indian/Alaska Native adults of all ages are even greater than Non-Hispanic White adults [4].Age-adjusted overdose rates from synthetic opioids other than methadone, which includes fentanyl and its analogs, continue to climb, as do overdose rates involving cocaine and psychostimulants [5]. These trends are mirrored by non-fatal overdoses, which rose more than 30% from January 2020 to January 2021, based on suspected drug overdoses in Emergency Departments [6]—likely a significant undercount of all non-fatal overdoses.

Fentanyl has been present in the US drug supply since at least 2013, becoming entrenched in the supply by initially adulterating and later supplanting heroin [7,8]. Polysubstance use, which can be both intentional and unintentional, is a significant driver of overdose [9,10] and awareness of this heightened risk has increased recently [10,11]. Specifically, fentanyl has co-adulterated stimulant-type drugs (e.g., methamphetamine & cocaine), and overdoses involving both fentanyl and stimulants have risen precipitously [12,13]. California (CA) has seen rising overdose deaths over this period. In 2018, one of the first years in which fentanyl-related deaths began to be observed in the state [14], CA had an overdose death rate of 12.8 per 100,000, slightly above the rates of the four years prior [15]. Each year since 2018 has seen notable increases, with 2021 being particularly deadly; 10,000 deaths and an overdose death rate of 26.6 per 100,000 [15]. Synthetic opioids accounted for nearly two thirds of deaths in 2021 [16]. San Francisco, like the rest of CA, saw the introduction of fentanyl in 2018 and overdoses began to rise, with 625 overdose deaths in 2021, 472 of which were attributed to fentanyl [17]. Since 2018, there have also been increases in cocaine- and methamphetamine-involved deaths, often in combination with opioids [17]. Deaths continue to rise, with 810 overdose deaths in 2023, higher than any previous year [18].

Drug checking services provide PWUD with voluntary information on the chemical composition of the drug compound in their possession. These services have been provided for several decades in Europe [19,20], but recently there has been a renewed focus on the role drug checking can play in public health surveillance and harm reduction services in North America. Particularly since the rise of fentanyl and other high-potency synthetic opioids in North American unregulated drug markets, drug checking has become a crucial technology for monitoring drugs at the point of use and improving risk communications to PWUD. Drug checking uses multiple modalities to test drugs with varying levels of sophistication. Each form of testing has related advantages and disadvantages with respect to analytical capabilities, cost of operation, time taken to test and provide feedback, and utility in respect to expectations [21–24]. Fieldwork for this study focused on two modes of drug checking, immunoassay strips and Fourier Transform Infrared (FTIR) spectrometry that were used in San Francisco, CA. Immunoassay strips, the earliest of which were fentanyl test strips (FTS), are the most widely used and recognized form of drug checking, and can rapidly indicate whether fentanyl (and a range of fentanyl analogs) is present in a sample with high sensitivity and specificity [25], typically offering detection in the ng/ml level [24]. FTS were initially designed for urine testing, and testing drugs put into solution directly is considered off-label use [24], potentially resulting in false positives at low cut-offs (below 40 ng/ml) or in the presence of other drugs [26]. Paradoxically, this is primarily how FTS are used currently. Additionally, FTS do not provide information on fentanyl concentrations or distinguish between analogs [24]. There has, however, been strong consensus that FTS are an accepted and useful modality among PWUD in many settings, and can promote harm reducing actions upon receipt of a positive test [27–29]. The success of FTS has led to further development of additional immunoassay strips as other drugs, notably benzodiazepines and the veterinary sedative xylazine, are increasingly found in the North American drug market [30,31].

Alternatively, spectrometry technologies can give a detailed breakdown of drug components, identifying emerging drug threats and adulterants, harmful byproducts of drug production and provide varying levels of quantification of the sample [21,22,24,32]. However, field-based drug checking has unique requirements in order to be effective, including remaining operable without excessive technical expertise, analyzing relatively small amounts of drugs and providing information in a time-sensitive manner [22]. One of the most commonly used approaches for field drug checking is Fourier Transform Infrared (FTIR) spectrometry, which produces spectra that can be analyzed and compared to spectral libraries to determine constituents contained within drug samples [32,33]. FTIR has some advantages for street- and mobile-based drug checking, particularly its accuracy in sample identification, small size and swift runtime (results provided within a few minutes) [24,33]. It is also preferred for its relatively easy operation, compared to more sophisticated spectrometry devices, and does not require specialized preparation of samples for analysis [33]. Particularly when optimized for street-based drug checking, FTIR is considered a qualitative drug checking modality, providing a breakdown of the main constituent ingredients. While it is possible to additionally provide semi-quantitative results, this is not believed to be particularly viable when not analyzed by expert technicians [32].

Although FTIR has notable strengths, it is not effective at detecting substances present in low concentrations (generally believed to be less than ~5% by weight) [33], a detection threshold that critically undermines the potential impact of FTIR, given the potential for overdose to occur with fentanyl samples under 5% by weight. It is also dependent on reference libraries, which require paid subscriptions and may not be particularly responsive to drug trends and detection of novel psychoactive substances [24,33]. Best practices in street-based drug checking services include the use of FTS alongside FTIR to offset the issue of low sensitivity [33]. In addition, there are significant disparities between states on the legality of drug checking technologies, and an overall lack of clarity about how existing laws may impact their use, complicating the rollout of these technologies [34].

Whilst the terminology and scope of drug checking services varies, the intended outcome frequently goes beyond solely providing chemical testing of drug samples. As such, Measham and Turnball argue “drug checking aims not simply to reduce drug-related deaths but also wider drug-related harm. Therefore its effectiveness should be measured not only by a reduction in drug-related deaths but by successful risk communications that enhance risk management and increase harm reduction practices.” [35] Amidst drastic increases in overdose-related mortality in CA, this study explores how currently deployed drug checking services in the state are being implemented to meet the needs of PWUD through qualitative interviews with a diverse set of PWUD based in San Francisco and drug checking experts in North America. In 2022, when this study was carried out, drug checking services in California remained in their nascent stage with few active providers. We deemed it worthwhile to supplement these few in-state provider interviews by drawing on the insights and experience of more established drug checking programs from locations across the US and Canada. Findings solely from the provider interviews have previously been published elsewhere [36]. This approach intends to understand how PWUD in the US think about and use available drug checking modalities and compare these findings with the experiences of drug checking experts working with these and other modalities across North America.

Methods

Research consisted of recruiting and interviewing two different samples of drug checking stakeholders, referred to in this paper as ‘providers’ and ‘clients’. After initiating interviews with California-based providers, additional key informants were recruited through a convenience sample of the PI’s known contacts. Using theoretical sampling methods, these participants were asked to identify other potential participants. Providers were recruited between June 21 and November 17 2022, with two authors (DC & LM) conducting n=11 in-depth semi-structured qualitative interviews over Zoom with providers, who were clinicians, technical experts in drug checking, and researchers on their experiences opening and operating drug checking services in the US and Canada. Inclusion criteria for providers consisted of being 18 years of age or older, employees or volunteers at drug checking service sites, and being willing and able to provide informed consent. Eight providers worked in the US, two worked in Canada, and one worked in both countries. Two were clinicians, four researchers, and five were harm reduction providers, with some currently and others formerly involved in drug checking projects. Some purposive sampling was used among this group to include perspectives on overdose prevention through drug checking services. Interviews lasted between 45–60 minutes and were recorded and transcribed in their entirety. Informants were offered a $100 honorarium in the form of a gift card at the end of the interview. Informed consent was sought by interviewers and consent was provided verbally. Participants have been pseudonymized for anonymity.

The second set of key informants interviewed were the ‘clients’, who were people who used drugs and lived in or commuted to the San Francisco Bay Area. Two authors (JO and NH) carried out n=13 semi-structured interviews over a week in November 2022. Clients had to be at least 18 years old and currently using fentanyl, heroin or illicit stimulants through any means of administration. Exclusion criteria covered any individuals who were intoxicated or could not otherwise provide verbal consent to participate. This was a non-random purposive sample. Individuals attending harm reduction programs at four locations in San Francisco were approached, asked about their current drug use patterns and subsequently if they would be interested in participating in the study before being formally verbally consented. Some participants were recruited directly by harm reduction site staff informed of the recruitment criteria while investigators carried out other interviews. The San Francisco-based research team was familiar with these locations from prior fieldwork, had good relationships with organizational staff and benefited from access to private rooms for interviewing. Additionally, the presence of a relatively new drug checking service in the city provided a unique opportunity to understand whether clients were aware of this service and ascertain their thoughts about its approach. Several participants (n=3) were recruited directly from the drug checking site while accessing these services. Interviews lasted approximately 30–60 minutes and were audio recorded and transcribed in their entirety. Informed consent was sought and participants provided consent verbally to interviewers. Clients were provided a nominal cash payment of $25 for their time and expertise.

Material covered in the provider interview guide included experience opening and operating drug checking services, navigating logistical and political barriers for these services, scaling up programming, knowledge about drug checking technologies and experiences integrating drug checking program into existing services and settings. For the client interview guide, questions explored drug use trajectory, current use practices, experience accessing harm reduction services, and awareness and experience of using drug checking services in the city, focusing particularly on the use of FTIR and FTS. Clients were also read a description of spectrometry drug checking occurring in San Francisco, and asked questions specifically about this service, what barriers and facilitators to its use they saw and how the service might be combined with other services and agencies. Both providers and clients were asked whether they thought drug checking could reduce opioid-related overdoses. The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the University of California, San Francisco (IRB numbers 22–36262 on 22 April 2022 and 22–36640 on 12 September 2022).

Following publication of a paper focused solely on providers [36], analysis of the client material followed. Interviews were transcribed and checked for accuracy before being uploaded to the Dedoose qualitative software package. SO created an initial coding structure using the client interview guide as its basis; this was refined in subsequent discussion with the other authors. SO and JO coded all the client interviews separately and JO checked both sets of coded interviews for consistency. It is estimated that consistency between the two reviewers was approximately 85–90% (consistency was high as the coding scheme was based largely on interview questions). Four authors (EAA, LM, SO and JO) read both datasets, identifying key themes across both datasets that would provide a structure for a fuller analysis. Using the tenets of thematic analysis, authors reduced and organized the provider data into analytic memos exploring patterns, contrasts and disagreements across this set of interviews [37,38]. The team met weekly during this process, discussing the data in-depth to further refine these thematic areas. Key thematic areas identified from the provider interviews including topics such as embedding drug checking programs into existing harm reduction services and information deficits in different drug checking technologies were also integrated into the coding for the client interviews. Coded material was read and discussed by all authors while outlining this paper, and alongside the provider analytic memos, forms the basis of these findings.

Results

We present our findings within the following five thematic areas:

1. (1). Complex dangers of fentanyl, polysubstance use, and adaptations to fentanyl in the drug supply;
2. (2). Social and policy contexts
3. (3). Assessing risk profiles;
4. (4). Fentanyl test strips and the challenge of focusing solely on fentanyl;
5. (5). Consumer knowledge and community overdose prevention;

Discussion

The introduction of fentanyl and other contaminants to the North American drug supply are contributing to increased polysubstance use, whether intentional or unintentional, driving overdose mortality ever higher. This crisis has contributed to greater interest in developing a more nuanced understanding of the drug supply among both service providers and PWUD. Various drug checking modalities are being employed in North America to fill this need and provide feedback on drug constituents directly to users of these services. However, while fentanyl identification has been the raison d’être for the uptake of drug checking, providers and clients complicated this rationale by identifying several weaknesses that limit this approach, while simultaneously recognizing groups who would benefit from expanded drug checking. Spectrometry was thought to be particularly impactful and yet still underutilized as a means to understand both the drug supply and overdose patterns in a swift-moving drug market. Providers and clients were also acutely aware of the delicate social, legal and political context framing the debate around drug checking and how these programs and harm reduction services more generally are facing significant opposition locally and federally in North America.

Drug checking services as currently constituted are attempting to strike a difficult balance: provide detailed and relevant information on the drug supply that is both timely and actionable. But given the urgency of the North American overdose crisis, this may be the wrong lens for scrutinizing the success of drug checking services, particularly when they are already so underleveraged. Giving people more information about the drugs they are using is extraordinarily important and returns agency to individuals who are often marginalized—providers and clients were emphatic on this point. However, it also places all the responsibility for a tainted drug supply back on the individual using drugs in an illicit, unregulated market where their means of control and self-protection are limited. Indeed, research among structurally vulnerable populations in Canada has found strong sentiment against using drug checking technologies, due to factors such as having to give up drugs for testing, time constraints, ambivalence towards overdose and limited recourse in the event of an unwanted result [40]. While other research has shown people with the means to do so will dispose of drugs they find out are contaminated [35], or change how they use them [41], structurally vulnerable populations including street-based individuals and people with opioid use disorder may not be as able to take these actions. A better drug checking service means providing clients with more reliable information, including communicating knowledge about concentrations, contaminants and other harmful components of a drug sample. But it also means moving beyond the fentanyl binary as expressed by the participants in this study.

To that end, spectrometry and other drug checking programs that employ quantitative (e.g., drug concentration) as well as qualitative reporting (i.e., the psychoactive compounds and other adulterants present in a sample) can provide individuals with a strong basis for informed decision-making regarding use of unregulated or unknown drugs. Indeed, providers and clients in our study believed that there are significant populations of PWUDs who are at risk of overdose for whom knowing the components and concentrations in a drug sample would be meaningful, whereas simply knowing if fentanyl is present in a sample would not be. Recent research in Chicago supports our findings, where among people who use opioids, most expected fentanyl contamination and believed more specific information, including drug concentrations, specific drugs found in the sample and the presence of potentially harmful diluents, would be more useful [42]. Areas where fentanyl has not yet reached saturation, e.g., Europe, may markedly benefit from integrating drug checking programs into existing services if fentanyl begins to enter these markets.

While the technologies currently being deployed are possibly adequate for reducing overdose mortality, their individual weaknesses require multiple, overlapping methods to be used in combination to provide the most comprehensive and timely information to PWUD, clinicians and policymakers. Drug checking technicians should therefore use not one, but many kinds of testing strips (fentanyl, xylazine and benzodiazepine, etc.) to investigate the tainted supply and provide useful information. However, because test strips can only provide targeted qualitative testing of specific compounds and cannot indicate the range of analogs that might be present, they are inherently limited. Simply adding further test strips to a drug checking regime may not be terribly useful in the long-term as novel psychoactive substances enter the marketplace and contribute to an unstable drug supply. Current evidence also indicates they do not accurately identify the presence of carfentanil [25], a potent opioid that has periodically emerged in the drug supply to devastating effect [43]. As one provider noted, “There is just no perfect model or perfect technology at this point,” requiring providers and clients to make do with available tools. However, existing evidence [33] and the experience of providers in this study finds that using FTS alongside FTIR provides more information than either modality alone. Early indications from drug supply surveillance, such as crime lab data surveillance [44,45], have been promising and require greater study and implementation. The potential for aggregating drug checking data within and across communities about local trends and emerging substances of concern also requires further exploration.

There is an urgent need to devote financial and technical resources to providing drug checking services, much of which work is currently being advanced by harm reduction service providers themselves [46] without largescale federal or state support. California and other US states are beginning to see large amounts of money roll into state coffers from several opioid litigation settlements with pharmaceutical companies and drug distributors. The first of these, settled in 2021 with Janssen Pharmaceuticals and three drug distributors (McKesson, AmerisourceBergen and Cardinal Health) will see California receive $2.05 billion dollars through 2038 [47]. The settlement funds will be used for current and future ‘opioid remediation activities,’ defined as “care, treatment, and other programs and expenditures designed to (1) address the misuse and abuse of opioid products, (2) treat or mitigate opioid use or related disorders, or (3) mitigate other alleged effects of, including on those injured as a result of, the opioid epidemic including naloxone distribution, support and expansion of existing medication assisted treatments (MAT), research and myriad other uses” [48]. However, outside of a call to support ‘fentanyl checking’ and research into FTS and other novel methods, these funds are lacking a crucial support for drug checking services in the state and broader monitoring of the drug supply outside of law enforcement channels. However, a bill passed the California legislature and was signed by Governor Gavin Newsom to clarify the legality of drug checking services in the state and increase access to these services [49,50].

Potential limitations of this study include its focus on drug checking solely in the North American context and limiting the client portion to one week of data collection in San Francisco. Providers from across North America were included to understand how drug checking has been implemented in diverse settings, some of which data was not germane to this manuscript. Provider experience using FTS and FTIR was prioritized for this manuscript. Some participants were unaware of the availability of drug checking services and had not utilized these resources in the city. It would be useful for a follow-up study to determine how drug checking implementation has impacted PWUD in California now that these services have been available for several years. The recent introduction of these services meant most PWUD interviewed had not heard of or used them before, with the exception of point-of-use technologies. The inclusion of multiple sites in San Francisco for client recruitment, including an active site for street-based drug checking, was intended to capture perspectives of PWUD from across the city and the utility of drug checking to individuals using different drugs, including non-opioid drugs. Additionally, the familiarity of the study team with San Francisco-based harm reduction services and the substance use landscape helped reach thematic saturation with a relatively small sample size.

Conclusions

Drug checking is a promising intervention for overdose prevention, albeit one that remains severely underleveraged, particularly with respect to the most vulnerable drug-using populations in North America. Providers and clients identified a need for these services to become far more mainstream in the US and Canada, democratizing access and pushing beyond the fentanyl binary, recognized as a limitation of the potential of these technologies and one that is challenged by the realities of substance use in North America. Law enforcement channels for detecting emerging drug threats and communicating these findings to the public have been inadequate in their response to the rapidly changing drug supply—leaving grassroots programs to confront the challenges of the unstable North American drug supply.