A Brief Negotiation Interview Chatbot to Improve Buprenorphine Engagement Among Justice-Involved Individuals (BNICHAT)

A Brief Negotiation Interview Chatbot to Improve Buprenorphine Engagement Among Justice-Involved Individuals

A Chatbot Utilizing Machine Learning and Natural Language Processing to Implement the Brief Negotiation Interview to Improve Engagement in Buprenorphine Treatment Among Justice-Involved Individuals

Primary Objective: Conduct pilot study to assess effects of brief negotiation interview (BNI) Chatbot among individuals involved in the Connecticut criminal justice system with opioid use disorder (OUD). Study Duration: Approximately 2 years (1 year for study activities, 1 year for data analysis) Study Design: This is a prospective, randomized study to evaluate the effectiveness of a BNI Chatbot on patients with OUD compared with Standard Care (SC). Number of Study Sites: The offices of the Center for Progressive Recovery, LLC and the New Haven Police Department Detention Center (NHPD). Study Population: The study population includes adult individuals with OUD who are involved in the Connecticut criminal justice system and not currently receiving medication-assisted treatment for their OUD. Number of Participants: Sixty participants Primary Outcome Variable: Attendance at participants first treatment appointment within four weeks of referral among participants in the BNI Chatbot vs. Standard Care (SC) groups. Secondary Outcome and Exploratory Outcome Variables: Secondary outcomes include readiness and intention to engage in buprenorphine (bup) treatment, and urine toxicology test-confirmed drug use at four weeks among participants in BNI Chatbot vs. SC groups. Exploratory outcomes include ratings of feasibility, acceptability and satisfaction between study groups, and a comparison of study findings to engagement data from previous in-person studies, including BNI+bup, and other digital programs, such as reSET-O, and DynamiCare.

Enrollment and Randomization: Individuals meeting inclusion/exclusion criteria with the research associate (RA), who will be following a "Recruitment Script" (see below) will be offered participation and written informed consent will be obtained. Participants will be randomly assigned (via http://www.randomization.com) to receive either the BNI Chatbot or Standard Care (SC). SC: The RA will provide participants with an electronic link, which provides OUD and bup education, and lists OUD treatment options and descriptions and locations, including bup prescribers. After reviewing, participants will be asked to choose from among the list of and a referral will be made based on the receiving treatment providers' procedures. These participants will not have access to the BNI chatbot. However, the referral facilitation, the electronic link will direct participants to a short Feasibility, Acceptability and Satisfaction (FAS) Assessment to obtain feedback on the process. BNI Chatbot: Participants will receive a link to register for the chatbot. Following registration, they will complete the initial chat, which includes instructions for the entire study. The chatbot will then guide the participant through multiple BNIs focused on the goal of bup engagement, with the following steps: 1) Raise the Subject/Establish Rapport; 2) Enhance Motivation; 3) Provide & Elicit Feedback; and 4) Negotiate a Plan. The chatbot then reminds the participant of their reasons why they might engage in OUD treatment based on their responses to Steps 2 & 3, and their negotiated plan. The chatbot will electronically connect participants to a treatment provider of their choice, based on available options. Based on their choice, individuals in the BNI Chatbot group choosing to engage in bup treatment will be assessed by the Clinical Opiate Withdrawal Scale (COWS) and those in active withdrawal may be offered a telehealth appointment immediately (to the extent possible), which would occur during the enrollment and help facilitate a prescription for bup. The telehealth appointment may be with a different treatment provider than the one that they choose to meet with post-enrollment. If the participant is prescribed bup during this telehealth appointment, the RA will pick up the prescription for participant and deliver it to the participant immediately upon release from the police, courthouse or probation office (to the extent possible). At the end of the encounter, participants will be prompted to complete the feasibility, acceptability and satisfaction (FAS) assessment electronically to obtain feedback on the process. Note: While participants are financially responsible for their own healthcare, we will only refer them to treatment options that are listed as being covered by their health insurance. Baseline Assessment: Demographics and status of the participant's case in the criminal justice system will be collected by the RA via a mobile tablet. One Month Assessment: At four weeks post-randomization, we will collect the following data: Patient-facing assessment: Participants will complete a brief follow-up assessment to collect information on readiness and intention to engage in bup treatment and ratings of satisfaction regarding their assigned intervention (BNI chatbot or Web-based OUD & Treatment Referral Resources). Additionally, they will be contacted by phone 1-month post-randomization to assess self-reported treatment attendance and drug use. Treatment providers: The RA will call treatment providers to collect information regarding participants' attendance and record urine toxicology results during the one month post-randomization, as well as any treatment/prescriptions provided for their opioid use disorder. Information will be provided to the RA over the phone and recorded in the electronic study database. Participant Payments: Participants will receive a $25 gift card after the completing the initial enrollment, including their first use of whichever web application the are assigned (BNI Chatbot or SC). Participants will receive an additional $25 gift card following completion of the one-month follow-up assessment. Study Schedule: Participants will participate in two visits, including 1) enrollment, randomization, and baseline assessment, web application completion (BNI Chatbot or SC, both of which are delivered via the web) and 2) one-month follow-up. The initial visit will take up to 45 minutes and the follow-up assessment will take up to 15 minutes. Informed Consent: An in-person recruitment/written consent has been submitted with this protocol, which describes in detail the study intervention, study procedures, and risks. The consent will be given to the participant via a mobile tablet and paper (if they request), and participants will be given as much time as needed to review the information and ask any questions they may have. will be answered by the RA in a way that does not reveal, or ask the individual to reveal, any PHI. If the individual's questions have been satisfactorily answered or if they have no questions, then they are given the option of signing the consent form on the tablet. If they sign the consent form, a copy of it will be emailed to them. They are then asked to sign a release of information allowing us to contact the substance use disorder treatment center that has appointments available at that time to confirm attendance and treatment one month later. If the individual signs the release of information, then we will email them a copy. Randomization and study interventions are then completed as outlined in the protocol. Written documentation of informed consent is required prior to enrollment and randomization. Screening & Recruitment: Below is a detailed outline of how recruitment, enrollment/informed consent and data collection will occur throughout the study. Overall, the NHPD's only involvement in this study is to provide space and access for recruitment and enrollment, and they will not provide or receive any study data. The NHPD detention center (i.e., "lockup") at 1 Union Ave, New Haven, CT, provides space for the study's RA to sit during set hours. NHPD shares with the RA which cells have nonviolent offenders (i.e., offenses in the past 5 years will not have been violent in nature) and can be approached. All nonviolent offenders will be given a Study Flyer by RA and then asked if they are willing to hear more about a research study. If yes, the prospective individual will be handed a study mobile tablet and a pair of disposable headphones so that they may hear the RA read the "Recruitment Script" (see attached), which will guide the RA through describing the study and answering any questions from the prospective participant in such a way as not to reveal any protected health information (PHI). Following the reading of the script, the prospective participant is handed a mobile tablet (e.g., iPad) where they can note their interest by clicking one of several options, including "I want to answer eligibility questions," which then provides the questions for the patient to read and answer on the tablet. The individual also has the option of having the RA read the questions but still provide the answers on the tablet. If the individual is eligible, then they can note their interest in participating by clicking one of several options, including "I want to read the Consent Form" or "I want the RA to review the Consent Form without revealing any private information". The RA will enroll subjects in the study after they have been referred, screened for eligibility criteria, and provided informed consent. If screening/enrollment is completed in person, written informed consent will be collected. If screening/enrollment is completed by phone or video chat, a copy of the consent form will be emailed to the participant and the time/date of verbal consent will be documented and retained for study records. On Study Visits: Baseline visit (30-45 minutes, depending on group) Consent Randomization SC: The RA will provide participants with an electronic link, which provides OUD and bup education, and lists OUD treatment options and descriptions and locations, including bup prescribers. After reviewing, The RA will ask participants to choose a treatment option and, if the participant so chooses, a referral will be made based on the receiving treatment providers' procedures. These participants will not have access to the BNI chatbot but will be prompted to complete a brief FAS assessment electronically. BNI Chatbot: Participants will receive a link to register for the chatbot. Following registration, they will complete the initial chat, which includes instructions for the entire study. The chatbot will then guide the participant through multiple BNIs focused on the goal of bup engagement and electronically connect participants to the treatment provider of their choice. They may be offered and connected with a telehealth appointment immediately, as outlined above, and may receive the physical prescriptions immediately once released from police, probation or the courthouse. After the encounter, participants will be prompted to complete a brief FAS assessment electronically. Baseline Assessment - Demographics and status of case in criminal justice system, provided by the participant. Participant payment ($25 gift card) One-month assessment (5-15 minutes, depending on group) At 4 weeks post-randomization, participants will complete a brief follow-up assessment to collect information on satisfaction of the treatment referral process. Participants in the BNI Chatbot group will also provide feedback on feasibility, acceptability and satisfaction of the chatbot and participants in the SC group will provide feedback on feasibility, acceptability and satisfaction of the electronic education and referral process. Assessments will be conducted via the chatbot (BNI Chatbot group only) or in-person or by phone. To facilitate scheduling, the RA will have access to all participants' scheduled probation appointments during the study. Participant payment ($25 gift card) End of Study and Follow-up: Following completion of study procedures, the research team will review data for completeness and attempt to complete any missing data, as possible. If patients withdraw from the study early, investigators will retain data collected up until discontinuation of the study.

Chatbot, Conversational Agent, Artificial Intelligence, Natural Language Processing, Machine Learning, Medication-Assisted Treatment, Buprenorphine, Treatment Engagement, Criminal Justice, Brief Intervention, Brief Negotiation Interview

Participants will receive a link to register for the chatbot. Following registration, they will complete the initial chat, which includes instructions for the entire study. The chatbot will then guide the participant through multiple BNIs focused on the goal of bup engagement, with the following steps: 1) Raise the Subject/Establish Rapport; 2) Enhance Motivation; 3) Provide & Elicit Feedback; and 4) Negotiate a Plan. The chatbot then reminds the participant of their reasons why they might engage in OUD treatment based on their responses to Steps 2 & 3, and their negotiated plan. The chatbot will electronically connect participants to a treatment provider of their choice, based on available options.

• SC: The RA will provide participants with an electronic link, which provides OUD and bup education, and lists OUD treatment options and descriptions and locations, including bup prescribers. After reviewing, participants will be asked to choose from among the list of and a referral will be made based on the receiving treatment providers' procedures. These participants will not have access to the BNI chatbot. However, the referral facilitation, the electronic link will direct participants to a short Feasibility, Acceptability and Satisfaction (FAS) Assessment to obtain feedback on the process.

The BNI Chatbot is a web application that guides participants through a brief motivational intervention that motivates them to engage in treatment for opioid use disorder.

The OUD Education & Treatment Resources intervention is a web application that provides OUD education and treatment resources.

Inclusion Criteria: Can speak, read and write in English Provision of signed and dated informed consent form Adult aged >18 years Screened positive for opioid use disorder (OUD), moderate or severe, based on the Diagnostic and Statistical Manual-5th Edition Have health insurance Have an existing electronic medical record in Epic Have a working device that can access a web browser and receive texts (i.e., smartphone, tablet or computer) Can provide a working email address Can provide a working cell phone number Exclusion Criteria: Current use of buprenorphine (bup), methadone, or naltrexone for a substance use disorder Known allergy to bup Pregnancy or lactation Known current suicide risk based on interview Mandatory transport to a Department of Correction Facility, or strong likelihood of immediate transport, resulting from current legal case