Preventing Opioid Use Disorder in Older Adolescents and Young Adults (ages 16-30): Expert Panel Planning Meeting

Overview

On September 26, 2018, the National Institute on Drug Abuse (NIDA) convened a group of scientific experts, state and federal partners, and other key stakeholders to provide input for informing research on preventing opioid misuse and opioid use disorder (OUD) in at-risk older adolescents and young adults, ages 16-30. Scientific experts shared ideas and provided input on research priorities focused on the development, testing, and implementation of interventions to prevent opioid misuse and OUD in this population, as well as strategies and considerations for identifying, accessing, engaging, and intervening with at-risk older adolescents and young adults.

Introduction

NIDA Director Nora D. Volkow, M.D., offered introductory remarks. She emphasized that the NIH has resources to support research on the prevention of opioid misuse and OUD. Improving pain management and prescribing practices is not enough to prevent these problems, as fentanyl and its derivatives are driving the opioid overdose epidemic.  

Meeting Goals and Structure

The meeting was organized into five sessions consisting of brief talks from speakers followed by a roundtable discussion among the participants, each guided by specific questions for consideration.  Below is a summary of key remarks and themes that emerged from the presentations and discussion.

Session 1: Epidemiology, Etiology, and Prevention of Opioid Misuse and Transition to Opioid Use Disorder in Older Adolescents and Young Adults

The prevalence of prescription opioid misuse among adolescents is higher among individuals who are in school with poor adjustment and among individuals who are not in school. Likewise, rates are higher among those ages 18–26 who are not in college, with the highest prevalence among individuals who did not complete high school. Prescription OUD is now much higher among individuals ages 26–34 than it was 10 years ago. A similar pattern is seen for heroin use. Macro level risk factors (e.g., passage of medical cannabis policies) also have an influence. For example, one presenter’s research shows that among U.S. 12^th grade youth living in states with a medical marijuana law, passage of the medical marijuana law predicted a significant increase in non-medical opioid use compared to pre-passage rates and overall rates in other states.

Data from young people who engage in nonmedical prescription opioid use indicate that they also use multiple other drugs. Reliable evidence is available to guide prevention efforts, as common risk and protective factors—at the individual, family, school, and community levels—have been identified for multiple drug and risk behaviors. In 2016, the U.S. Surgeon General identified more than 60 prevention programs and policies that have been found to prevent substance use problems. Prevention also is key to reducing the opioid epidemic. However, less is known about the unique proximal risk factors during young adulthood and about effective prevention programs and strategies for young adults specifically.

Session 2: At-Risk Populations and High-Priority Settings

Data presented during the meeting from the 3 Campus Alcohol and Marijuana study, conducted at three state universities, show relatively low past month rates of opioids use and minimal use of heroin among the college population at the campuses studied. Even among students who use other drugs, few use opioids. These data might suggest that at these campuses, students with opioid misuse are likely to drop out of college. The study also found that students who do misuse opioids also tend to engage in nonmedical use of prescription stimulant and sedative drugs. Various strategies and platforms could be used in prevention efforts with college students: (1) implementing Screening Brief Intervention and Referral to Treatment (SBIRT) in academic services for individuals who are struggling academically and (2) targeting programs for athletes (a potentially at-risk population) and staff who work with them. Brief motivational interventions are used for other types of substance use in this population (e.g., alcohol), but researchers have not determined whether this approach would work for opioids. The general risk factors for substance use among adolescents apply to college students’ opioid use.

For at-risk young adults, components of the Quit Using Drugs Intervention Trial (QUIT) model could be applicable for preventing opioid misuse and OUD, as this model was efficacious at reducing risky drug use among adult primary care patients of all ages and was feasible in routine practice. QUIT involves the use of the patient-completed Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) screener, and identifies action for the provider based on the screening result. This presenter recommended that primary care clinics and providers approach all patients for screening and that, ideally, screening be anonymous and data from adolescents not be visible to parents. A computerized version of the World Health Organization ASSIST has been validated for adolescents and adults. QUIT could be adapted for older adolescents and young adults and be used in student health clinics in high schools and colleges, U.S. Department of Veterans Affairs clinics, and other settings such as the workplace, criminal justice facilities, homeless shelters, and drop-in centers. Further support could come from mobile technology and social media.

It is clear that reducing opioid prescriptions will not solve the opioid crisis and that strategies for prevention are essential. The best settings for intervening with adolescents and young adults include schools, emergency departments (EDs), primary care, orthopedic specialty and dental clinics, Federally Qualified Health Centers, neonatal specialty facilities, mental health programs, and the Indian Health Service. Members of the military and veterans have unique considerations (problems with self-disclosure and access issues). Because the settings differ and the risk factors vary, prevention efforts must attend to the specific local context. SBIRT has potential, as does electronic delivery of prevention interventions. Visually engaging content, use of technology to enhance person-delivered content, smartphone applications, and social media all have potential for intervening with sustaining engagement, and measuring the impact of prevention in this population. Peer-based approaches also have potential.

Session 3: Study Designs and Outcomes

Various research designs can optimize interventions to prevent OUD. In the classical treatment-package approach, multiple components form a behavioral or biobehavioral intervention, which is evaluated via a randomized-controlled trial (RCT). RCTs can determine whether one intervention is more effective than another, but other research designs are needed to optimize interventions (e.g., to determine the effectiveness of individual components, interaction among components, and aspects of adaptive interventions). Optimization involves building the most effective intervention subject to realistic constraints so that the intervention is efficient, economical, cost-effective, and scalable (i.e., can be readily implemented as evaluated).  Additional benefits of optimization include development of a knowledge base and better positioning to respond to public health crises, less need for ad hoc modifications of interventions to make them scalable, and interventions become incrementally better over time.

The Multiphase Optimization Strategy can be used to optimize interventions. Key activities include optimization trials, and relevant research designs include factorial experiments, a fractional factorial experiment, a Sequential Multiple Assignment Randomized Trial, a micro-randomized trial, and system identification. These designs are highly efficient, and some can use cluster randomization. Resources (e.g., articles, trainings, technical assistance, software) are available on the optimization of behavioral and biobehavioral interventions.

The preferred study design to evaluate the resulting intervention will depend on the research question and on the expected intervention, setting, and population. Potential research designs include: 1) Group- or Cluster-Randomized Trial (GRT) design, 2) Stepped Wedge Design (SWD), 3) Individually Randomized Group-Treatment Trials (IRGT), 4) Regression Discontinuity Design (RDD), and 5) Time Series Designs (TSD). An intervention with a small effect size may not have a clinical or public health impact. Researchers can use modeling to estimate the effect of an integrated intervention if they have effect sizes of the components. Individual-level trials may be best for some questions, interventions, settings, and populations, whereas group-level designs may be best for other questions, interventions, settings and populations. The key is to settle on the research questions, intervention(s), target settings and population; those factors will determine which design is most appropriate.

Session 4: Adaptation and Integration of Existing Models and Approaches in Healthcare and Other Systems

Prevention professionals have implemented various screeners for youth (National Institute on Alcohol Abuse and Alcoholism Screener; Brief Screener for Tobacco, Alcohol, and other Drugs; and Screening to Brief Intervention). For adults, the Tobacco, Alcohol, Prescription medications, and other Substance (TAPS) Tool and ASSIST are available as screeners. A number of studies suggest potential prevention models—such as in-person or computerized brief interventions (BIs) in health care settings and in high schools or college. To adapt and integrate models, it is important to incorporate depression, anxiety, and sexual risk behaviors into screening. Models from alcohol or cannabis misuse could be adapted for opioids and incorporated into routine care in various settings. Although BIs may be effective for alcohol and marijuana misuse, it is important to determine whether implementing them to address the opioid epidemic is an effective strategy. Strategies for reaching and sustaining engagement for transition age youth and young adults include: strategies for handing off from pediatricians to internists (e.g., use of electronic health records), interventions for children of patients in OUD treatment, reaching and engaging new employees through employee assistance programs.

Existing models in other systems—such as juvenile justice and schools—include Treatment Foster Care Oregon and the Family Check-Up, which offer valuable lessons learned for reducing and preventing substance use. Systems gaps and transitions in and out of systems —when young people leave juvenile justice, child welfare, or educational systems—mean that they often do not receive services and it may be difficult to intervene with them. However, they often enter family medicine or pediatric care settings during pregnancy or as parents. Strategies are needed to reach young adults who are parenting, as the highest rates of opioid overdose fatalities were among adults ages 25–35, according to 2016 data. Overdose death rates among adults ages 15–24 and 25–34 showed the greatest increases from 2015 to 2016.

Session 5: Implementation and Uptake

Implementation science is the study of methods to integrate research findings and evidence into health care policy and practice. The focus is how to make prevention programs work in real-world settings. The Exploration, Preparation, Implementation, and Sustainment (EPIS) Framework is one approach that can be used. Suicide prevention offers a model for possible ingredients that can be incorporated into efforts to prevent opioid misuse and OUD. It is crucial to use hybrid effectiveness and implementation designs as well as scaling-out designs. Partnerships between implementation researchers, organizations and community stakeholders are crucial. Effectiveness research yields evidence-based programs that can be incorporated into the delivery system, and the quality of implementation determines the population benefit. EPIS provides a good model for sustainment, which must begin early in the research process.

The Prevention Economics Planning and Research Network works to advance economic analyses of substance misuse prevention efforts. It provides guidance on how to conduct high-quality, high-utility economic evaluations. Conducting prospective cost analyses during initial implementation is the best approach, as retrospective cost analyses can increase the uncertainty of estimates and can be more labor intensive. Prospective cost analyses lay the groundwork for future cost-effectiveness or benefit-cost analyses. They can also inform initial implementation questions, highlight issues of sustainability, and support program optimization. It is important to value the effects of programs as they are evaluated, as this informs the choices of decision-makers. Economic analyses should consider spillover across contexts—such as child welfare, criminal justice, education, and labor—and results in high-quality estimates that will inform budgets. These analyses should also map the impacts onto government administrative record systems.

As an example, the Promoting School-community-university Partnerships to Enhance Resilience (PROSPER) Prevention System is an evidence-based way to deliver interventions. PROSPER dissemination costs of more than $3 million resulted in a project public benefit of about $6 million—including labor, health, and criminal justice savings. Universal school and family programs can be cost-effective for reducing opioid misuse.

Summary

The following major themes emerged during the discussions:

Populations at risk and settings for accessing and intervening

• Populations of young people at risk for opioid misuse and OUD are heterogeneous, as are the potential settings for prevention interventions.
• One recommendation is to include a focus on young people who are not engaged in the education system. Opportunities for intervention exist in community-based settings, child welfare settings, and various workforce pathways and settings (e.g., including vocational-technical schools and training programs).
• Health care settings have great potential for delivering prevention interventions in general; at the time of pregnancy or parenting an infant is an opportune time. It is recommended that primary care providers routinely ask about substance use as part of overall health behaviors. Interventions implemented in urgent care, emergency departments, and prenatal care clinics to reach individuals not engaged in primary care may yield positive outcomes. For example, fewer boys and young men seek treatment in primary care; they may be more likely to seek care in urgent care settings.
• Youth transitioning out of systems (e.g., foster care, juvenile justice, or K–12 education) experience gaps in services. Strategies to consider for engaging with these populations include: a) a case-finding approach that is developmentally sensitive, b) a coordinated specialty-care model in which a team follows patients (e.g., the National Institute of Mental Health’s Recovery After an Initial Schizophrenia Episode initiative model), c) a case management wrap-around approach for young parents with OUD, d) one-stop (drop-in) center models for young people disconnected from parents and caregivers.
• In American Indian and Alaskan Native communities, lack of opportunity and trauma are contributing to opioid misuse and OUD and devastating communities. There is a need for strategies for treating adults and reducing OUD and overdose deaths, along with implementing community-based interventions. Technology as a way to deliver education and training and boost economic opportunity are important to explore.

Prevention strategies and delivery formats 

• It is recommended that theory-based prevention interventions be developed to address the multiple pathways to opioid misuse and OUD.  
• It is recommended that prevention programs and strategies account for the fact that nonmedical opioid misuse is distinct from prevention of the escalation of opioid misuse to OUD.
• Research is needed to determine which technologies and strategies are most useful for specific populations. Technology-delivered programs that are complemented by face-to-face human connections may yield better results than those without some face-to-face connection.

Public health messaging and communication strategies

• Public health messaging and communication can be conducted relatively quickly, reach a large number of people, and can be relatively straightforward to evaluate.
• Key messages for the general public that focus on fentanyl and other synthetics (addictiveness and danger of overdose), address harm to others, and present the human face of the opioid epidemic (stories are more effective than data) may be the most effective.
• Key messages for young people need to be developmentally sensitive and reach parents as well as youth who are not engaged with caregivers (perhaps with peer messages). In addition, communication efforts could be designed and targeted to medical practitioners and policymakers to change the dialogue on prevention.

Screening and intervention

• Screening can take place in a wide variety of settings (e.g., health care, schools, and mental health centers); the field needs a systematic way to screen and intervene with young people.
• The effectiveness of SBIRT is clear for alcohol (although the active/essential ingredients are unknown). The effectiveness of SBIRT for opioids needs to be determined.
• An important consideration is that screening information is captured in medical records; some patients have concerns about this, which may influence the data they provide.

Scale-up and sustainability of prevention programs

• Scale-up and sustainability of prevention programs can be enhanced by taking into account the context and obtaining local knowledge early in the research process.
• Stigma must be addressed at all levels, particularly misunderstandings about medication-assisted treatment among service providers and communities.
• Training providers on prevention interventions takes time and requires a stable prevention infrastructure and funding across the states.
• Determining the effectiveness of intervention components, optimizing interventions, and evaluating cost-effectiveness should take place early in the process. Interventions that are optimized based on theory and research based criteria (e.g., risk and protective factors) may be most effective. Hybrid effectiveness and implementation designs are useful tools for optimization trials.
• Costs analyses that include benefits of prevention on labor and justice costs and the economic impact of overdose deaths are needed.
• Prior to a large trial, it is important to do work to determine: (a) the fidelity, dosage, and quality of the intervention; (b) how to engage people in the intervention.

Leveraging existing research

• Rigorous meta-analyses can be valuable, and Bayesian approaches could be useful.
• Universal prevention interventions have demonstrated effects on long-term outcomes. A synthesis of long-term effects and impacts of universal prevention programs on opioid use behaviors would advance understanding of the benefits of universal interventions implemented during earlier developmental stages on opioid outcomes later in life, into later adolescence and young adulthood.