We will also examine intervention efficacy on other opioid outcomes , other substance misuse and consequences, and impaired driving. Further, we will also measure intervention costs and identify moderators/mediators of efficacy. In this paper, we describe the RCT protocol with a focus on evaluation of primary and secondary outcomes.We are testing the efficacy of behavioral interventions as secondary prevention for opioid misuse and opioid use disorder among AYAs in a 2 x 2 factorial RCT design with two intake and two post-intake conditions. Conditions delivered at intake are: 1) a single video delivered health coach session, and 2) an EUC control condition of a community resource brochure. Post-intake conditions are: 1) portal messaging that lasts for 30 days, or 2) EUC control brochure delivered at 30 days. This combination effectively results in 4 groups as shown in Figure 1 . The health coach session occurs during intake and the portal begins at intake and occurs over the following 30 days. Depending on randomization, the EUC resource brochure is either delivered at intake or 30 days later . EUC essentially functions as a control condition, with the provision of minimal resources that exceed the current standard of care in the ED pertaining to opioid screening and prevention as done in a number of prior ED-based studies.In total, we seek to enroll 1170 ED patients ages 16 to 30 . Recruitment occurs either in-person or remotely, a secondary recruitment approach that was added because of restrictions on in-person research activities during the COVID-19 pandemic. All procedures are IRB approved.As part of our funding mechanism and with IRB approval, we piloted the interventions and EUC described in detail below in the Spring of 2020, with N = 40 AYAs enrolled with informed consent/assent and the same trial eligibility criteria described below.
Each participant received the EUC at enrollment,cut flower transport bucket the health coach session, 30 days of portal messaging, and a 1-month follow-up invitation e-mail . After enrolling the first 10 participants using ED-based recruitment methods, we paused recruitment to refine interventions based on participants’ feedback. During this pause, in-person recruitment efforts were suspended due to the COVID-19 pandemic, and the remaining 30 participants were recruited remotely after their ED visit and all interventions were delivered remotely in private from staff homes to participants’ homes. Among the participants who participated in pilot testing the health coach session, 38 provided acceptability ratings. Their intervention satisfaction was M = 9.3 on a 10-point scale with 10 being the highest possible rating. Similarly, on a 10-point scale, ratings for recommending the intervention were M = 9.0 . Fidelity coding of selected sessions was based on the Motivational Interviewing Treatment Integrity Code33 with fidelity exceeding “fair” thresholds, and nearly all markers exceeding “good” thresholds. These included the following means: M = 4.3 relational scale, M = 3.9 technical score, M = 71.5% complex reflections, and M = 1.7:1 reflection/question ratio. Next with regard to the portal, among our first 10 participants, portal engagement was lower than expected with only 60% replying at all. Thus, we used participant feedback to make several functional and design changes prior to recruiting the next 30 participants. These changes included adding a “remember me” function to avoid needing to remember a password, updating the look and feel of the portal, adding use of an emoji avatar chosen by participants, branded token items valuing ~$1 mailed weekly . We also evaluated use of an incentive structure for portal messaging with the final 30 participants.
Thus, we enrolled 10 participants each into cohorts: a) no incentive for message replies, b) $1 incentive for each reply with ability to earn up to $20, and c) $5 incentive for each reply with ability to earn up to $20. Engagement, as measured by number of replies from participants, was higher in these latter cohorts. The no incentive cohort had a M = 11.8 messages received with 100% replying, the $1 incentive group had M = 10.0 messages received with 80% replying, and the $5 incentive group had M = 15.6 messages received with 100% replying. Given that functional and design changes in the portal appeared to result in increased engagement without the use of additional incentives, we elected to not incentivize participation in the planned RCT, especially since there would be additional challenges associated with later implementation of incentives within healthcare systems. Finally, among 37 pilot participants who provided feedback ratings for the portal, ratings of satisfaction were M = 8.5 and recommendation were M =8.9 , on a 10-point scale with 10 being the highest rating. Modified MITI fidelity coding for complex reflections and reflection-to-question ratio exceeded benchmarks such that were on average 59.5% complex reflections and the average reflection-to-question ratio was 2.9:1. Overall, the feasibility of our pilot testing supported the inclusion of the intervention conditions in the planned RCT. The acceptability data provided by participants at follow-up was promising, particularly given refinements to the portal condition, and staff demonstrated acceptable fidelity to the intervention model. Participants are recruited from the Michigan Medicine pediatric and adult EDs in Ann Arbor, Michigan. Combined, these EDs have about 100,000 visits per year and are located in adjacent, but separate hospitals and maintain independently functioning systems of triage, medical staffing, and physical space allocation.
Recruiting from both EDs enhances generalizability to pediatric and adult ED settings. Historically, the average ED length of stay as been 3-4 hours which facilitates completion of in-person research protocols during the visit, although the COVID-19 pandemic has altered ED patient flow. Currently, there are no opioid-focused prevention programs in the study setting and opioid-related clinical care focuses primarily on treatment referrals and prescription monitoring. We attempt to recruit English-speaking patients ages 16-30 presenting to the ED for any reason in-person or remotely to complete an eligibility screening. Staff approach participants in the ED, and remote recruitment is used when staff are unable to recruit in the ED or to supplement in-person recruitment because of physical distancing limitations on approaching patients with suspected COVID and the number of staff who can be present at any one time due to COVID-related restrictions. When approaching potential participants, they must be medically and cognitively able to provide consent/assent; thus patients presenting with acute substance intoxication are excluded until able to consent. Individuals presenting to the ED with a chief complaint of acute sexual assault or acute suicidality will be excluded from screening. Those presenting with a current cancer diagnosis or currently receiving cancer treatment and pregnant women will also be excluded . AYAs who participated in our pilot study described below or who may be taking part in other current behavioral intervention trials at this study site are also excluded . Screened participants are eligible for the trial based on past 12-month prescription opioid use plus at least 1 other risk factor , or 12-month opioid misuse as described in the measures below . Screened individuals reporting injecting drugs or screening as high risk for current opioid use disorder based on a NIDA-Modified ASSIST V229–32 score of 27+ are excluded due to the study focus on prevention of development of opioid use disorder. These individuals are instead referred for treatment; staff direct them to options listed in study resource brochure if recruited remotely and if in recruited in-person staff offer referral to ED social worker. All procedures herein have been piloted in the study setting and among AYAs,procona flower transport containers including consenting and enrolling minors. Research assistants identify patients ages 16-30 via the electronic health record and tracking system, with a waiver of HIPAA authorization. Patients who meet screening exclusion criteria are not approached. Staff approach screening eligible patients based on triage time/status or discharge date/time using a standard script. Two-stage consent for ages 18+ and parental consent/child assent rare obtained for screening and, subsequently, the RCT. Staff review limits to confidentiality and study procedures during consenting/assenting. Specifically, parents are consented concurrently for screen and baseline . Subsequently, youth assent is obtained for screen and baseline. Study data is not shared with parents or guardians and confidentiality would only be broken in the case of acute risk, such as acute suicidality, homicidality, or child abuse, to preserve adolescents’ or others’ safety. For in-person recruitment, research staff approach ED patients ages 16-30 for consent/assent to self-administer a brief web-based screening survey on iPads. Surveys are paused during medical procedures and consultations such as x-rays, assessments, and blood draws.
For remote recruitment, research assistants contact ED patients ages 16-30 after discharge for consent/assent to complete the screening survey on their personal device or by phone. Participants screened in person receive a gift valuing $1.00 for survey completion. AYAs meeting eligibility criteria above are invited to participate in the RCT. After consenting/assenting, RCT participants complete a baseline survey and are randomized to conditions. Randomization is stratified by sex, age , and opioid risk severity : none or higher and occurs in blocks of 8 within strata to equalize over time. Randomization is computer generated in Qualtrics and is not be known by recruitment staff until after completion of the baseline survey when staff run the randomization program to reveal condition. Staff then orient participants to their assigned condition. Because this is a behavioral intervention where participants are receiving counseling interventions, participants and coaches are not blind to their condition. Participants then receive their assigned intake condition at intake and are then oriented to their post-intake condition . If needed, those unable to complete all intake activities in the ED may complete them in person, by telephone, or video call; these activities are scheduled prior to leaving the ED. Participants who join the study remotely, or who do not complete enrollment during the ED visit, may have up to 30 days to complete intake activities. Enrolled participants receive a token gift in-person or by mail. Participants are asked to complete follow-up assessments lasting approximately 25-30 minutes that mirror baseline measures at 3-, 6-, and 12-months. Participants are remunerated in cash or gift card . Follow-ups are primarily online, without staff interaction, however, in the event a participant elects in-person follow-up staff administering assessments are blinded to condition assignment.Consistent with a telemedicine hub model, research assistants connect participants to the remote health coach using a telehealth platform such as Facetime or Zoom. Health coaches are bachelor’s-level or higher staff hired who have relevant backgrounds and experience in fields such as social work, psychology, counseling. Health coaches are further trained in MI and CBT strategies by the study team and supervised weekly by experienced master’s-level coordinators and doctoral-level investigators. The ~45 minute session blends elements of efficacious brief motivational interventions from our prior work, including an opioid-focused brief intervention and developmentally appropriate content for ages 16-30 based on our prior alcohol-focused brief interventions.The intervention guide is housed in an internal website that is only accessible to study staff and includes decision-support screens to guide intervention delivery and enhance fidelity , allowing health coaches to collect within session para-data , consistent with prior work that found para-data are associated with important MI mechanisms and clinical outcomes.Our session, outlined in Table 1 is rooted in the Why Change and How Change model of MI.The remote health coach session uses MI strategies to engage and explore individuals’ situations, allowing for in-the-moment tailoring to address needs relevant for AYAs using MI concepts . During Why Change, health coaches review and affirm participants’ goals and strengths, and invite discussion of opioid misuse as well as other substance use. MI skills , seeking permission, Elicit-Provide-Elicit, elaboration, rulers, and autonomy support are used. Elicit-Provide-Elicit is a tool for collaborative psychoeducation used to explore risk perceptions of opioid medications and/or overdose in a highly tailored manner .Personallyspecific benefits of change are elicited and reinforced; as many young people may be pre-contemplative about change, health coaches learn to elicit hypothetical benefits of future change or scenarios for choosing to avoid use . After summarizing change talk, health coaches explore How Change, eliciting personally tailored tools to address misuse of substances and risk factors. Tools include brainstorming alternatives to address motives and risks . If applicable, participants are encouraged to consult their primary care or other medical providers with questions.