Subthreshold opioid use disorder prevention (STOP) trial: a cluster randomized clinical trial: study design and methods

Responses to the opioid crisis have focused on decreasing opioid prescribing, reducing fatal overdoses, and increasing access to medications for opioid use disorder (OUD). Little attention has been paid to preventing the progression to moderate or severe OUD among people who exhibit risky opioid use behavior, which includes nonmedical use of prescribed opioids or any use of illicit opioids. Subthreshold OUD consists of risky opioid use (e.g. taking illicit opioid or nonmedical use of prescription opioid) that does not meet the threshold for treatment with opioid agonists or antagonists (e.g. moderate or severe OUD as defined by DSM-V criteria of 4 or more criteria) [1]. Patients with subthreshold OUD may experience few symptoms or consequences of their opioid use and rarely seek addiction treatment services [2, 3], but they are frequently encountered in primary care settings where they receive routine medical care and, in some cases, prescription opioids for pain management. Although prevalence in primary care settings varies based on the patient population, 4.7% of all adults are estimated to have risky opioid use, [4], and prevalence is much higher (21–29%) among those receiving prescribed opioids [5‐7]. Primary care clinics are optimally positioned to identify and provide early intervention for individuals with risky opioid use because they are the largest prescribers of opioid analgesics [8], and they are often the only point of health care contact for individuals with illicit opioid use that has not yet progressed to OUD. Yet there has been almost no research on interventions to reduce the potential harms of opioid use or halt the progression to OUD among primary care patients with subthreshold opioid use [9].

Most prior approaches to reducing subthreshold drug use in general medical settings have relied on single-contact brief interventions [10‐13], delivered using ‘screening, brief intervention, and referral to treatment (SBIRT)’ models. In randomized clinical trials, these single-contact brief interventions have had mixed results for reducing drug use, with most studies showing no difference compared to no intervention. Those that were positive showed small reductions and were limited by having short-term follow-up [13‐17]. A pilot study by Bohnert and colleagues in an Emergency Department (ED) setting utilized a 30-min therapist delivered motivational interview and found decreased self-reported opioid use and misuse in individuals at risk for overdose [15]. With a slightly more intensive but still limited intervention, Gelberg and colleagues conducted the Quit Using Drugs Intervention Trial (QUIT) which tested brief advice by the primary care provider (PCP) and video doctor, with 2 sessions of motivational counseling by a telephone health coach in primary care clinics for individuals with any moderate-risk illicit drug use. In two trials, the QUIT intervention reduced self-reported drug use by 33%-44% compared to usual care [16, 17].

Collaborative care is an established approach for improving mental health conditions in primary care patients, [18‐20] and could potentially be an effective approach for subthreshold opioid use. Formal collaborative care models include population management, patient-centered team-based care, evidence based treatment, measurement based care and accountability [18‐20]. Collaborative care is grounded in the Chronic Care Model (CCM) for chronic disease management, [21‐24] which seeks to improve patient outcomes through integrative systems changes, including organizational support, clinical information systems, delivery system redesign, decision support, patient self-management support, and community resources. Collaborative care teams typically consist of a PCP working with a care manager who maintains a clinical registry to proactively track and manage care delivery [24]. Prior interventions for substance use have included elements of formal collaborative care, such as interdisciplinary team-based care, as exemplified in the Massachusetts nurse-care model [25]. TOPCARE (Transforming Opioid Prescribing in Primary Care) and TEACH (Teaching Effective Analgesia in Clinics for HIV) were cluster randomized trials that employed a nurse care manager using population management tools, academic detailing, and web-based decision support. While these trials improved PCP delivery of guideline concordant opioid prescribing practices for patients receiving chronic opioids for pain in primary care and HIV care, respectively,[26‐28] they did not examine patient outcomes.

The Sub-Threshold Opioid Use Disorder Prevention (STOP) Trial is designed to study the efficacy of elements of a collaborative care intervention to reduce risky opioid use and to prevent progression to moderate or severe OUD in adult primary care patients with risky opioid use. The STOP intervention augments previously tested care models by combining PCP brief advice and telephone health coaching along with a nurse care manager working with the PCP. The STOP intervention thus combines patient-facing and provider-facing components. The purpose of this report is to describe the STOP trial protocol.

This paper describes the design and protocol of one of the first multi-site randomized controlled trials to examine interventions for adults presenting in primary care settings with subthreshold opioid use disorder [9]. It addresses an important gap in the literature: most trials addressing risky and potentially harmful opioid use have focused on either patients who already have moderate to severe opioid use disorder, or on prescribing opioids for pain. In contrast, our study focuses on individuals who have only risky opioid use behaviors and are currently experiencing few if any consequences to their health or functioning resulting from opioid use, but who nonetheless may be at risk for opioid-related harms, including to the development of moderate or severe OUD.

Little is currently known about the trajectories of individuals who have subthreshold OUD, including how likely, and within what timeframe, they may develop a more severe OUD or experience opioid-related overdose or other serious negative health impacts. OUD prevention efforts have focused on regulatory efforts and published guidelines about prescribing opioid analgesics, particularly by primary care physicians [54, 55]. There is increasing concern that curbing prescribing could be contributing to the opioid crisis when opioid tapers are not accompanied by interventions that actively engage patients and providers in preventing the development of OUD, treat underlying pain, and provide access to effective treatment and overdose prevention strategies for individuals who already have a moderate-severe OUD [56‐58].

In its 2020 recommendation to screen adult primary care patients for drug use, the US Preventive Services Task Force (USPSTF) noted critical gaps in the evidence for interventions to reduce drug use, risk behaviors, and negative health consequences. In particular, evidence was noted to be lacking for treatment interventions among non-treatment seeking patients who are identified via screening.[59‐61] Although it is limited specifically to patients with subthreshold OUD, our study will contribute to filling this evidence gap, and may inform future studies of primary care interventions for other types of drug use.

The STOP collaborative care model brings together intervention components that individually improved outcomes in opioid prescribing and substance use but have not previously been tested in combination or with a focus on subthreshold OUD. Collaborative care has been shown to improve patient outcomes and increase quality of care for a variety of conditions, including alcohol use disorder and depression [21, 27, 62‐65]. The small number of studies that have previously examined collaborative care interventions for drug use did not show significant changes in opioid use [21, 66]. However, these trials enrolled a different treatment population, comprising individuals with moderate-severe SUD rather than subthreshold OUD. Patients with subthreshold OUD, who are the focus of our trial, have less severe symptoms and different treatment needs, requiring less emphasis on medications for OUD treatment and more focus on motivating behavior change, medical care coordination, and self-management. A feature of the STOP collaborative care intervention, which was designed to address these needs, is the addition of an off-site telephone health coach providing support and counseling using techniques from motivational interviewing and cognitive behavioral therapy. The Health Coach has the potential to engage patients in setting goals for behavior change, while the ongoing support of the nurse care manager and PCP can enable them to access treatment and other resources that are needed to execute and sustain reductions in risky opioid use.

The intervention is designed to be integrated into regular primary care practice in clinical settings with collaborative care teams, which are increasingly common in primary care practices, but generally focus on mental health rather than substance use. The telephone health coach is also scalable in that a single health coach can service multiple sites. A strength of the STOP intervention model is that it avoids over-burdening the PCP and makes use of existing resources. Apart from the brief advice, the case management and health coaching do not require direct PCP involvement and can offload some of the work that generally falls to the PCP. Many primary care practices have allied health resources to assist with chronic care management, (for example, diabetes educators) that could be trained in providing case management and support for these high-risk patients. Telephone health coaching is commonly available through health insurance plans or large health systems. Health insurance companies could contract with national experts who do short term coaching telephonically who also collaborate with the affiliated clinical nursing resources. Should our trial have positive findings, there is thus potential for adapting and integrating the STOP intervention into existing team-based primary care models.

A notable aspect of the study design is the use of monthly self-administered assessments of opioid misuse, which can avoid recall bias and offers multiple opportunities for data collection. The strategy of inviting patient participants by text message or email to complete brief self-administered assessments that can be completed on their phone or computer has similarities to ecological momentary assessment (EMA) approaches, which often achieve high rates of completion. One of the longest duration EMA studies in the literature was an alcohol study that used weekly assessments collected by text message for 25 weeks and found that 82% of participants completed all 25 of the assessments [67]. In prior systematic reviews, allowing surveys to be completed using e-mail and text messages typically results in higher response rates [68, 69]. By easing barriers to their completion we thus anticipate having a high participation rate. While there is biological verification with a urine drug screen, this verification captures a short window of substance use, and is used as an exploratory measure.

Limitations include the challenges in identifying individuals meeting the eligibility criteria and in quantifying low level risky opioid use. Risky opioid use and OUD are not well captured in electronic health records ,[70] and so we rely on screening delivered as part of the research to identify eligible participants. Changes in patient receipt of prescription opioids may be limited as the intervention takes place in primary care settings, but the controlled substance may be prescribed elsewhere. The use of prescribed opioids, including those from outside providers, is only measured if patients report using them nonmedically.

The primary outcome of days of risky opioid use for this trial comes from validated tools but itself has not been validated. Existing opioid-specific screening tools can predict likelihood of opioid misuse but do not quantify or identify risky opioid use. ¹⁶ To overcome this limitation, the study team used items from the validated COMM assessment that relate to specific opioid consumption behaviors (taking more than prescribed or taking medication that belonged to/was borrowed from someone else). These measures were then quantified for the primary outcome. However, these measures still require the participant to distinguish nonmedical from medically indicated use to accurately respond to these questions. For example, if a patient has an opioid prescribed to treat severe pain but then takes it to help with their sleep or for mood elevation, they may not perceive it as taking a medication not as prescribed (since they have not exceeded the intended daily dose), even though this would be considered nonmedical use of their medication. Furthermore, the stigma of nonmedical use and fear of having their opioid medications discontinued may reduce reporting of misuse behaviors [71‐73]. A key feature of addiction is an element of impaired self-awareness, including cognitive distortions around potential misuse of substances.[74, 75] Therefore, the study may not have the sensitivity and specific for detecting and confirming risky opioid use based on the underlying need for self-identification.

The study design relies on frequent self-reported assessments of substance use to assess the primary outcome. Assessment itself has been observed to motivate change in trials of early intervention approaches for risky alcohol use, [52, 53] and this could occur in our study population. Our RCT design minimizes the possibility of this as the control condition (EUC) receives identical assessments as the intervention group, the participants are blinded to the focus on opioids by including questions on other health behaviors, and planned statistical analyses account for assessment reactivity. Furthermore, we rely on confidential self-administered assessments to minimize social desirability bias in reporting.

The Subthreshold Opioid Use Disorder Prevention Trial (STOP) is a cluster randomized trial that draws on successful interventions combining nurse care management and telephone health coaching with primary care to intervene with patients experiencing subthreshold OUD. Given the growing interest in interventions for managing patients with risky opioid use, and the need for primary care-based interventions, this study could offer a blueprint for a feasible and effective approach in this population.