A Randomised Clinical Trial of a Theory-Based Fentanyl Overdose Education and Fentanyl Test Strip Distribution Intervention to Reduce Rates of Opioid Overdose: Study Protocol for A Randomized Controlled Trial

Background: Opioid overdose deaths involving synthetic opioids, particularly illicitly manufactured fentanyl, remain a substantial public health concern in North America. Responses to overdose events (e.g. administration of naloxone and rescue breathing) are effective at reducing mortality; however, more interventions are needed to prevent overdoses involving illicitly manufactured fentanyl. This study protocol aims to evaluate the effectiveness of a behaviour change intervention that incorporates individual counseling, practical training in fentanyl test strip use, and distribution of fentanyl test strips for take home use among people who use drugs. Methods: Residents of Rhode Island aged 18–65 years who report recent substance use (including prescription pills obtained from the street; heroin, powder cocaine, crack cocaine, methamphetamine; or any drug by injection) (n=500) will be recruited through advertisements and targeted street-based outreach into a two-arm randomised clinical trial with 12 months of post-randomisation follow-up. Eligible participants will be randomised (1:1) to receive either the RAPIDS intervention (i.e. fentanyl-specic overdose education, behaviour change motivational interviewing (MI) sessions focused on using fentanyl test strips to reduce overdose risk, fentanyl test strip training, and distribution of fentanyl test strips for personal use), or standard overdose education as control. Participants will attend MI booster sessions (intervention) or attention-matched control sessions at 1, 2, and 3 months post-randomisation. All participants will be offered naloxone at enrolment. The primary outcome is a composite measure of self-reported overdose in the previous month at six- and/or 12-month follow-up visit. Secondary outcome measures include administratively linked data regarding fatal (post-mortem investigation) and non-fatal (hospitalisation or emergency medical service utilisation) overdoses. Discussion: If the RAPIDS intervention is found to be effective, its brief MI and fentanyl test strip training components could be easily incorporated into existing community-based overdose prevention programming to help reduce the rates of fentanyl-related opioid overdose.


Background
Opioid overdose morbidity and mortality remain substantial public health threats in North America, particularly in the era of highly potent novel synthetic opioid agents (1). Mortality related to opioid overdose is also a global concern, with the greatest impact in North America, Europe, and Australasia (2).
Contamination of the street drug supply with illicitly manufactured fentanyl (IMF) and its analogs has precipitated dramatic increases in mortality in the United States and Canada, representing a period of substantially increased risk for people who use drugs (PWUD) (3). As IMF contamination expands to illicit stimulant and other non-opioid substances, the potential number of opioid-naïve individuals at risk of opioid overdose has also increased (4,5).
Overdose prevention efforts include medications for opioid use disorder, overdose education and naloxone distribution (OEND) programs, overdose prevention sites and supervised consumption services, and drug testing services (6)(7)(8). Studies show that PWUD modify or adapt their drug use behaviour in light of increased overdose risk associated with synthetic opioid contamination, including using smaller doses to test drug strength, maintaining consistent drug supply, consuming drugs with others present, reducing drug consumption, and having naloxone present (9,10). However, as contamination of the drug supply with IMF becomes more prevalent, novel tools to prevent and respond to opioid overdose events are urgently needed.
In recent years, analytic tests to detect fentanyl contamination have received substantial attention (11). Traditional forensic testing technologies such as mass spectrometry can identify fentanyl and other synthetic analogs in drug samples, but are expensive and require specialised training and reagents (12).
Rapid and simple point of care tests have been proposed as an alternative tool for increasing access to fentanyl testing services in highly marginalised communities, and in drug-using populations without access to supervised consumption facilities, as is currently the case in the United States. Paper-based immunoassays (e.g. Rapid Response™ fentanyl test strip, BTNX Corporation, Canada) provide similar sensitivity and speci city compared to portable mass spectrometers, with low technical requirements and results within ve minutes (13). People who use drugs report interest in knowing if fentanyl is in their drugs prior to use, with high willingness and acceptability to use fentanyl test strips in the future (14,15). Furthermore, early studies have shown that positive fentanyl test strip outcomes results in positive drug use behaviour change, and participants report a desire to perform testing prior to drug use and in private locations (16,17). Novel drug testing modalities that increase harm reduction self-e cacy will be paramount in assisting and supporting PWUD at risk of fentanyl overdose.
The study protocol described herein aims to test the e cacy of a novel behaviour change intervention that incorporates fentanyl testing strips with a theory-driven opioid overdose risk reduction component.
Guided by two health behaviour frameworks, we hypothesise that combining motivational interviewing counseling techniques to increase willingness and self-e cacy to use fentanyl test strips with fentanyl test strip training and teach back will result in enhanced overdose risk reduction practice, compared to participants receiving standard overdose education and naloxone training. We hypothesise that exposure to the Rhode Island Prescription and Illicit Drug Study (RAPIDS) intervention will reduce rates of fatal and non-fatal overdose. Mediation analysis will assess the differential uptake of behaviour changes in participants receiving the intervention.

Methods/design
Participants, interventions, and outcomes

Study setting
The RAPIDS Clinical Trial will be undertaken in the state of Rhode Island, a population of 1 million with sustained high rates of overdose mortality associated with any opioids (25.9 per 100,000 people in 2017) (1). Death involving synthetic opioids other than methadone (mostly fentanyl) increased substantially both nationally and within Rhode Island between 2013 and 2018, accounting for eight in 10 fatal overdose events in the state (1). Opioid overdose deaths in Rhode Island involving both fentanyl and other opioids appear geographically constrained, with hot spots of overdose deaths observed in a few urban centers in the state (18).

Eligibility criteria
Inclusion criteria for the study are: (a) 18 to 65 years of age; (b) reside in Rhode Island; (c) able to complete interviews in English; (d) able to provide informed consent; and (e) report past 30 day use of heroin, illicit stimulants (e.g., powder cocaine, crack cocaine, methamphetamine), counterfeit prescription pills, or any drug by injection, regardless of treatment status. Exclusion criteria are: (a) refusal of consent to participate, (b) unable to provide informed consent due to altered mental status; (c) unable to adequately hear and/or comprehend the consent process; (d) use of prescription medications exclusively obtained from a physician or diverted from someone else's prescription.

Patient and public involvement
No patient involved.

Interventions
All participants will be randomised to receive either the RAPIDS intervention or a control condition (usual care). An overview of the study design is shown in Figure 1; intervention components are described below.

Usual care
Participants in the control arm will receive standardised OEND training, as currently occurs in Rhode Island for people who are at risk for an overdose (19). Based on common components of many community-based overdose prevention interventions, the OEND training includes information on overdose recognition and response, including an educational video on recognising overdose symptoms, and basic overdose response (e.g., performing rescue breaths; administering naloxone; and calling for medical assistance). All participants will be offered a naloxone kit and education material at baseline, and referrals to community locations for no-cost or low-cost naloxone access will be provided at subsequent study visits.

Motivational interviewing-based brief intervention (RAPIDS intervention)
Participants in the RAPIDS intervention arm will receive: (1) a brief (15-minute) overdose education and behaviour change counseling session using motivational interviewing, (2) fentanyl test-strip training, roleplay, and teach back, and (3) training videos at the baseline study visit (Figure 1). The intervention is informed by the information-motivation-behavioural skills (IMB) model of behaviour change, which hypothesises that if a person possesses the information, motivation, and behavioural skills to act, there is an increased likelihood that they will ful ll and maintain the desired behaviours (20). The IMB model has been studied extensively and has empirical support as a framework for understanding both drug and sexual risk behaviours (21)(22)(23). Participants will receive ten Rapid Response™ fentanyl test strips for personal use at the end of the session (detection cut-off: 20 ng/ml) (13). Motivational interviewing "booster" sessions will occur at 1, 2, and 3 months post-randomisation, with additional test strips provided during the booster session on an as needed basis. The goals of this intervention are to provide participants with: (1) information about the risks of IMF exposure and how to reduce their risk of overdose, including naloxone training; (2) motivational interviewing to assess participants' willingness and enhance self-e cacy for engaging in overdose risk reduction behaviours; (3) an instructional video and hands-on training to increase behavioural skills in fentanyl rapid test strip use and interpretation, and opportunities to practice using the test strips; and (4) provision of a supply of test strips for personal use. A conceptual framework outlining intervention components (informed by the IMB model) is shown in Figure 2. Implementation of the intervention will be undertaken by research staff with a history of working in community settings to enable future scalability in resource constrained settings such as needle-syringe programs, emergency departments and health clinics (24)(25)(26).
During the brief intervention (approximately 10-15 minutes duration), interventionists will explore the participants' familiarity with basic overdose recognition and response. interventionists will discuss characteristics of fentanyl overdose and appropriate response with participants (e.g. rapidity of fentanyl overdose; naloxone as opioid overdose reversal) before checking with participants for understanding of participants and assessment of response. As part of fentanyl test strip training, participants will review the control arm educational video on overdose recognition and response, followed by an instructional video on the use and interpretation of fentanyl test strips. The fentanyl test strip training also includes practice test-strip use and role-play risk reduction behaviour given a positive or negative test-strip result.
Motivational interviewing processes that incorporate engaging, focusing, evoking and planning (27) with participants will guide the direction of subsequent conversations. Using the core skills of motivational interviewing (i.e., open-ended questions, offering a rmations, providing re ections, and summarising) (28), participants and interventionists will explore the participants' responses to the information component of the intervention. Current risk reduction behaviours endorsed by participants will be a rmed by interventionists, and possible ambivalence about reducing overdose risk will be re ected upon. Taken together, these elements will provide a foundation for the interventionist to elicit change talk from a participant as they increase motivation, strengthen self-e cacy, and plan overdose risk reduction behaviours.
In light of the high potency of fentanyl and di culty differentiating IMF from heroin and many illicit stimulants (11), the behavioural skills building component of the intervention will focus on overdose risk reduction behaviours considered most effective in the era of fentanyl (9,10). Interventionists will discuss with participants to build skills related to: 1) use of drugs with someone else present who could intervene and take action in the event of an overdose; 2) obtaining naloxone and having someone present to administer if necessary; 3) calling 911 immediately if an overdose occurs; and 4) using fentanyl test strips prior to drug consumption.

Outcomes
The primary outcome will be a composite measure of self-reported overdose (de ned as a "negative reaction from using too much drugs", as operationalised previously (29,30)) in the previous month at the six-and/or 12-month follow-up visit. This sensitive de nition is especially prudent in the era of IMF, where atypical presentations of opioid overdose may include chest wall rigidity, nausea, and vomiting (31). Selfreported overdose (compared to administrative data linkage) was selected as the primary outcome measure since a minority of participants in a pilot study reported attending a hospital at their last overdose (32). Furthermore, one of our target behaviours of the intervention includes calling 911 and seeking immediate medical attention if an overdose occurs, possibly leading to an increased rate of presentation at emergency departments for overdose events in the intervention arm. Nonetheless, exploratory analysis will utilise deterministic linkage of administrative health data systems in Rhode Island to identify fatal (through medical examiner reports) and non-fatal (through hospitalisation or emergency medical service utilisation) overdose events among participants.

Participant timeline
The trial consists of a 16-week behavioural intervention phase with a 36-week follow-up phase. The total trial period for participants will be 12 months. As shown in Table 1, measurements will be undertaken at six time points in each arm: at baseline, three times during the 12-week intervention program, and at six and 12-month follow-up visits.

Sample size
Sample size/power calculations are based on ndings from our pilot intervention study (14). In this study, 37% of participants reported a lifetime history of overdose, and 10% reported a suspected fentanyl-related overdose in the past six months. Therefore, for the present study, we assumed that 20% of participants in the control arm will experience an overdose during the 12-month follow-up period. We conservatively assume an 80% retention rate in the trial, below the 90% retention rate observed in the pilot study. Given these assumptions, with a sample size of 500 participants we have >80% power to detect a 50% reduction in overdoses (i.e., the primary endpoint described above).

Recruitment and retention
As in our pilot intervention study, we will employ a combination of eld-based recruitment techniques, Internet-based advertising, and state-wide public transport advertisement to recruit potential study participants. We will employ participant retention strategies that are effective at engaging PWUD, including dedicated study phones; comprehensive participant contact information tracking database; regular outreach in areas that participants are known to frequent; study visit honoraria (Brown University IRB-approved: $35 for baseline visit, $25 per follow-up visit); providing free parking to participants; and routinely searching the public Rhode Island Department of Corrections database to determine participant incarceration (considered lost to follow-up if incarcerated for greater than 90 days). For each visit, contact will be limited to a maximum of 10 attempts, after which participants will be considered lost to follow up.
If participants initiate contact after this, follow-up will continue according to the study schedule and contact attempts count reset.

Assignment of interventions Allocation and Blinding
Eligible and consenting participants will be randomly assigned to either control or intervention arm using a simple 1:1 allocation using a permuted block randomisation schedule, operationalised using the Randomization Module in REDCap™. Randomisation will be assigned during the baseline visit following informed consent and completion of a baseline assessment (see below). Intervention allocations will be accessible in REDCap™ by authorised users only, such as the biostatistician, and revealed to interventionists only through REDCap™ by prompts during the baseline and follow-up visits.

Data collection, management, and analysis
Data collection methods Training for data collection staff All staff will receive training for human subjects research. Additionally, staff will receive eight hours of relevant education (e.g. harm reduction, naloxone administration, overdose recognition and response, and stigma related to drug use) and 16 hours of didactic and hands-on training in motivational interviewing processes and skills. The motivational interviewing component will be audio recorded and a subset reviewed for delity of implementation as an ongoing practice (33). Prior to provision of the intervention, all interventionists will be trained in MI by a certi ed trainer, with random delity audits throughout the study period. Interventionists will undertake additional training where initial or audit assessments are below the 100% and 85% benchmarks, respectively.

Data management Data management guidelines and procedures
The primary source of data will be survey responses, captured using REDCap™ software on laptops and tablet devices for direct, secure, and remote data entry. Probabilistic linkage of administrative datasets will provide an additional source of information regarding opioid overdose events for exploratory analyses. Ongoing surveillance efforts by the Rhode Island Department of Health captures data on all fatal overdose events occurring in the state through the Rhode Island O ce of State Medical Examiners, and hospitalisation and emergency medical service utilisation related to suspected nonfatal opioid overdose, as described above (34). Data will be transferred from the Rhode Island Department of Health to Brown University School of Public Health using Health Insurance Portability and Accountability Actcompliant secure le transfer protocols.

Primary outcome
Using the intention-to-treat procedure, we will determine the e cacy of the intervention by comparing selfreported rates of overdose among those assigned to the intervention arm (RAPIDS intervention) versus the control arm (usual care). A chi-square test of a two-by-two table (statistically signi cant at p<0.05, two sided) will compare exposure to the intervention and the composite measure of self-reported overdose (de ned as a "negative reaction from using too much drugs") in the previous month at the sixand/or 12-month follow-up visit.

Sensitivity analysis of primary outcome
We will repeat our analysis of the primary outcome using: (1) pooled logistic regression to determine the independent effect of the intervention on self-reported overdose at either or both follow-up visits, adjusting for known or anticipated prognostic variables measured at baseline (e.g., injection drug use, history of overdose); (2) generalised linear mixed effects models (GLMM) with a logit link function to examine differences in self-reported overdose rates at six and 12-month visits; (3) a per-protocol subanalysis to determine the effect of the intervention on self-reported overdoses among those who used the fentanyl test strips compared to those who did not.

Secondary outcome
Using administrative-linked data of fatal and nonfatal overdose events, Kaplan-Meier estimators and Cox proportional hazards models will be used to determine the hazard of overdose events in the intervention and control groups. Right censoring for time-to-event analyses will occur at time of rst identi ed overdose, estimated date of loss to follow-up, non-opioid overdose related mortality, or end of study.

Missing data
At the completion of the study, patterns of outcome and covariate data missingness will be explored to determine the appropriate assumptions required for each affected variable. Using chained equations, the model will specify the conditional models for the missing outcome and covariates as missing completely at random (MCAR), missing at random (MAR), or missing not at random (MNAR). In the case of MNAR, we will specify a subset of the observations to derive the imputation model and may adjust imputed values by using speci ed shift and scale parameters for a set of selected observations in sensitivity analysis with a tipping-point approach. Finally, models with complete case, available data, and fully imputed data will be compared based on their coe cients, standard errors (SE), and p-value for the covariates included in the model.

Data monitoring
This study includes an independent Data and Safety Monitoring Board (DSMB) and a Data and Safety Monitoring Plan (DSMP). Further details on the DSMB charter are available on request. The DSMB will meet on a monthly basis for the rst six months with the PI, Project Director, Biostatistician, and Study Coordinator to review the DSMP and protocol adherence. Thereafter, the DSMB will determine the schedule for future review meetings. The DSMB will also review any adverse or severe adverse events as outlined in the DSMP. Adverse events (including severe adverse events) will be captured using REDCap™ software during routine participant assessment and spontaneously as required, and reported to relevant parties as necessary.
Interim analyses will be performed when approximately one-third of participants are recruited to assess potential harms, and when approximately two-thirds of participants are recruited to assess e cacy. Using the alpha spending rule (35), p-values will be constructed to maintain the overall study power of 0.05, two-sided, assessing the proportion of expected events (i.e., self-reported non-fatal overdose) in each arm. If the test statistic exceeds the boundary, then the study could be considered for early termination due to emerging differences between the two arms. The study may be stopped due to: 1) evidence of harm based on adverse and severe adverse event reporting and other safety data; 2) early evidence of e cacy from the second planned interim analysis; and 3) if any investigator judges it necessary for medical, safety, regulatory, or other reasons consistent with applicable laws, regulations, or good clinical practice. We do not propose that the trial be stopped for futility, as additional information collected at the 12month follow-up may be useful for public health program planning and practice.

Auditing
The Principal Investigator is ultimately responsible for data safety and monitoring of the study. This process will be monitored on a weekly basis by his team, including the Project Director, Biostatistician, Study Coordinator, and Data Manager, with quarterly updates to the entire team of Co-Investigators. The project team will be responsible for ensuring that all policies and processes outlined in the Data Use Agreement (with Rhode Island Department of Health) are followed accordingly, and that data are transferred and shared on the agreed-upon timeline.

Ethics and dissemination
Research ethics approval Ethical approval for this study has been obtained from the Brown University Institutional Review Board (ref: 1904002388) and other relevant Institutional Review Boards. Any modi cations to the trial protocol (e.g. eligibility, potential bene t or harms, study objectives or design, study procedures) will be subject to approval by Brown University Institutional Review Board and reported to relevant parties as necessary. Administrative amendments resulting in minor changes will be subject to approval by the Brown University Human Research Protection Program.

Consent
Consent will be obtained at the start of pre-screening (verbal or written) for eligibility, and e-signed informed consent at the start of the baseline visit for eligible participants. Participants will be offered a copy of the e-signed informed consent form via email. All participants will be given detailed explanations of their rights as human subjects, including the purpose of the study, length of time for the interview process, study requirements, and risks and bene ts of the interventions. At each in-person visit, research personnel will complete a brief verbal assessment using the "Alert & Oriented X4" procedure, where the participants' awareness of person, time, place and situation is assessed. Furthermore, all research personnel are trained to recognise and respond to an overdose, including calling for medical assistance, administering naloxone, and performing rescue breaths. On-call clinicians and local behavioural health services are available in the event of an emergency.

Con dentiality
All trial-related information will be securely stored at the study site, either in locked ling cabinets in secured areas (paper documents) or in secure Windows-based servers using two-factor authentication (electronic records).

Dissemination policy
This study has been registered on www.ClinicalTrials.gov and all summaries of the results will be published on ClinicalTrials.gov by the Principal Investigator as soon as they become available. Trial results will be shared in peer-reviewed journals and presented at relevant local, regional, and national scienti c meetings, conferences and invited lectureships in a timely fashion. Publication authorship will follow the International Committee of Medical Journal Editors guidelines. All nal peer-reviewed manuscripts from the trial will be submitted to PubMed Central digital archive.

Discussion
To date, no randomised clinical trials have examined the e cacy of rapid fentanyl test strips in combination with behaviour change intervention for the prevention of opioid overdose events. Our pilot study of young people at risk of fentanyl exposure demonstrated high willingness to use fentanyl test strips as a harm reduction tool, and the feasibility of a brief training module for non-expert use (14,16). Furthermore, positive test strip results were associated with drug risk behaviour changes (16). The current randomised clinical trial aims to demonstrate the application of theory-driven motivational interviewing sessions that improve participant self-e cacy in drug risk behaviour reduction alongside fentanyl test strip distribution and brief training in their use.
The results from this randomised clinical trial will need to be considered in light of some limitations. Firstly, participants are being drawn from an unknown sample frame and may not be representative of the greater population of PWUD in Rhode Island or other geographic locations. However, our participant recruitment and retention plan is informed by pilot studies of PWUD in Rhode Island, and will consist of numerous and diverse strategies for participant engagement (36). Secondly, evaluating the effectiveness of a behaviour change and fentanyl test strip intervention versus standard overdose education and naloxone distribution is complicated by the subjective nature of overdose events. Our measure of selfreported overdose captures participants' perceived negative reactions of using drugs, in line with the literature (29,30). In addition, we will collect symptomatology of overdose events to better describe the severity of nonfatal overdose events experienced by participants. To address issues with subjective measures of exposure to fentanyl, we have incorporated urine drug testing-for 12 drugs including fentanyl-in the protocol and a three-day recall period for exposure to substances. Finally, loss to followup is a key issue for studies of PWUD, compounded by the illicit nature of substance use. We will implement comprehensive participant retention strategies, including text message and email reminders, participant honoraria and subsidised parking, targeted canvassing in spaces where PWUD are known to frequent. To address potential bias related to loss to follow up, administrative data linkage-capturing fatal overdose events reported by the medical examiner, and nonfatal overdose events attended to by emergency medical services or presented at emergency departments-will be utilised in exploratory analyses.
This randomised clinical trial serves as the rst such work to evaluate the e cacy of novel drug testing technology combined with a behavioural change component. If found to be effective, the RAPIDS intervention will provide the basis for an affordable program that can be feasibly implemented by harm reduction services for the prevention of opioid overdoses around the world. In light of IMF contamination of non-opioid substances, novel interventions are necessary to protect the health of people who use drugs.

Availability of data and materials
The principal investigator (PI) will oversee management of the nal trial dataset with the Biostatistician and Data Manager. The PI will determine access to the nal trial dataset in instances of formal requests for substudy analyses from co-investigators or collaborators.