Skip to main content
NCBI home page
Drug and Alcohol Dependence Reports logoLink to Drug and Alcohol Dependence Reports
. 2024 Mar 13;11:100226. doi: 10.1016/j.dadr.2024.100226

Protocol for harmonization of randomized trials testing the addition of behavioral therapy to buprenorphine for opioid use disorder

R Kathryn McHugh a,b,⁎, Allen J Bailey a,b, Roger D Weiss a,b, Garrett M Fitzmaurice a,b
PMCID: PMC10966146  PMID: 38545409

Abstract

Background

Although buprenorphine is an effective treatment for opioid use disorder (OUD), much remains to be understood about treatment non-response and methods for improving treatment retention. The addition of behavioral therapies to buprenorphine has not yielded consistent benefits for opioid outcomes, on average. However, several studies suggest that certain subgroups may benefit from the combination of buprenorphine and behavioral therapy, highlighting the potential for personalized approaches to treatment. Furthermore, little is known about whether behavioral therapies improve buprenorphine retention or non-opioid (e.g., functional) outcomes.

Methods

The objective of this project is to harmonize four previously conducted clinical trials testing the addition of behavioral therapy to buprenorphine maintenance for OUD and to use this larger dataset to answer critical clinical questions about the role of behavioral therapy in this population. Study aims include identifying potential moderators of the effect of the addition of behavioral therapy and quantifying the effect of behavioral therapy on buprenorphine retention and functional outcomes.

Results

Analyses will consider outcomes of weeks of opioid use, weeks of retention in buprenorphine treatment, and functional outcomes as measured by the Addiction Severity Index. Analyses will include an indicator for each study to account for heterogeneity of samples and design.

Conclusion

Results will help to inform clinical and research efforts to optimize the use of behavioral therapies in the treatment of OUD.

Keywords: Opioid use disorder, Buprenorphine, Behavioral therapy, Treatment outcome, Treatment retention

Highlights

  • •

    The efficacy of behavioral therapy added to buprenorphine for non-opioid outcomes is unclear.

  • •

    This study will harmonize four clinical trials of treatment for opioid use disorder.

  • •

    Outcomes include treatment retention and functional outcomes.

  • •

    Moderators of treatment response will be examined.

1. Introduction

Medication for the treatment of opioid use disorder (MOUD) is highly effective for the reduction of opioid use, prevention of nonfatal and fatal overdose, and improvement in functioning (Committee on Medication-Assisted Treatment for Opioid Use Disorder, 2019, Pierce et al., 2016). Despite the effectiveness of this treatment for many people with OUD, approximately half of patients return to opioid use or discontinue MOUD prematurely (Lee et al., 2018, Weiss et al., 2011). The identification of strategies to improve treatment response is an urgent priority, particularly in the context of the continued escalation of opioid-related overdoses in the United States and globally (Fischer et al., 2019, Wilson et al., 2020).

Structured behavioral interventions, such as manualized drug counseling and cognitive-behavioral therapy, are efficacious for the treatment of substance use disorders, both when delivered alone and in combination with medication (Dutra et al., 2008, Magill et al., 2019, Ray et al., 2020). However, in contrast to clinical trials for other substance use disorders (Ray et al., 2020), studies suggest that the addition of behavioral therapy to buprenorphine plus structured medical management does not yield improvements—on average—on opioid use outcomes for OUD, compared to buprenorphine with structured medical management alone (Fiellin et al., 2006, Weiss et al., 2011). This unexpected finding has raised many questions about the role of behavioral therapy in the treatment of OUD (Carroll and Weiss, 2017).

One potential explanation for this finding is that not all people with OUD may need high-intensity behavioral therapy to reduce opioid use (Carroll and Weiss, 2017). Buprenorphine is an effective medication that may be sufficient for some people with OUD when combined with robust and structured medical management, whereas others may benefit from more intensive treatment that includes behavioral therapy. OUD is a highly heterogeneous disorder with respect to factors such as the type of opioid used, co-occurring medical and psychiatric disorders, and level of functional impairment; this heterogeneity may result in identifiable subgroups who respond adequately to MOUD with high-quality medical management. Indeed, preliminary data suggest that certain subgroups of people with OUD may benefit from adding behavioral therapy to buprenorphine maintenance (Moore et al., 2007, Weiss et al., 2014). However, such studies have yielded inconsistent results (Moore et al., 2007, Nielsen et al., 2015), which may be attributable to variability in inclusion criteria across studies and the limited power to detect moderational effects in single clinical trials. Further research in a study with greater statistical power and participant heterogeneity is needed to clarify these mixed results and ultimately to inform personalized decision-making in OUD treatment.

Furthermore, although behavioral therapy has not shown reliable additive benefits for opioid use outcomes when used in conjunction with MOUD and medical management, its efficacy for enhancing buprenorphine retention or for improving functional outcomes in this population has not been systematically studied in clinical trials. A recent report of the National Academies of Sciences, Engineering and Medicine on treatment for OUD highlighted this as a significant gap in need of further research (Committee on Medication-Assisted Treatment for Opioid Use Disorder, 2019). Indeed, analysis of claims data has suggested that receipt of psychosocial services was associated with greater buprenorphine retention compared to no psychosocial service receipt (Samples et al., 2022) and behavioral interventions targeting retention specifically have shown some promise (Brigham et al., 2014). Furthermore, expansion of endpoints for treatment outcome trials has been highlighted as a priority area for substance use disorder research, in general (Volkow, 2020). In other words, when evaluating treatments for substance use disorders, studies should consider the multitude of outcomes that are meaningful to patients, families, and communities, beyond substance use alone (e.g., social functioning, psychiatric symptoms).

This paper describes the protocol for a study that will leverage existing clinical trial data to answer pressing questions about the role of behavioral therapies in the treatment of OUD. In this project, we plan to harmonize data from four completed randomized controlled trials in which behavioral therapy and buprenorphine with standard medical management were compared to buprenorphine with standard medical management alone. With this harmonized dataset, we propose to answer important and timely questions with direct implications for clinical practice, including: (1) who benefits from the addition of behavioral therapy to buprenorphine maintenance, (2) does behavior therapy improve retention in buprenorphine treatment, and (3) is buprenorphine maintenance plus behavioral therapy more efficacious for functional outcomes compared to buprenorphine with medical management alone. Our overarching hypothesis is that the additive effects of behavioral therapy may be greater for people with more complex psychiatric and psychosocial needs, for whom the skills gained in behavioral therapy may offer broader benefits to both opioid use and other symptom and functional domains. Our approach represents an efficient way to answer these questions by leveraging existing data sources. Here, we describe the rationale for this project as well as the planned study procedures for harmonization and analysis of these data.

2. Material and methods

Based on a published review of trials of adding behavioral therapy to buprenorphine for the treatment of OUD (Carroll and Weiss, 2017) and a literature search by the study authors, we selected four NIDA-funded clinical trials for this harmonization study. These studies include three independent trials (R01 DA009803, R01 DA019511, R01 DA020210) and one NIDA Clinical Trials Network multi-site trial, representing over 860 participants recruited from sites across the United States (Fiellin et al., 2006, Fiellin et al., 2013, Ling et al., 2013, Weiss et al., 2011). These studies were selected based on several factors that would support rigorous harmonization, such as similarity in study design and the presence of common data elements. These trials all met the following criteria: (1) randomization to a structured behavioral therapy, (2) all participants received buprenorphine maintenance for the treatment of OUD, and (3) availability of common data elements to facilitate harmonization. This approach allows for use of existing data to answer urgent clinically meaningful questions, including those that individual clinical trials are not statistically powered to answer.

2.1. Summary of included studies

Below, we describe each of the studies included in our harmonization trial. Although studies were selected based on shared features, heterogeneity among studies remains in some design features. This heterogeneity will be addressed in the analysis through both inclusion of an indicator for study as a covariate in all models and in the use of sensitivity analyses to test the robustness of findings to specific elements of design heterogeneity (see Data Analysis section for more detail).

Fiellin et al. (2006) conducted a 24-week randomized controlled trial (N=166) that randomized patients with Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) opioid dependence (American Psychiatric Association, 1994) to one of three treatments: standard medical management (SMM) with medication dispensing once per week, SMM with medication dispensing three times per week, or enhanced medical management with medication dispensing three times per week. SMM consisted of one weekly 20-minute manualized and medically focused counseling sessions focused on medication adherence and support for abstinence efforts. Enhanced medical management consisted of similar, but extended counseling sessions (45-minute) that also focused on assessment and targeting of other functional impacts related to addiction (e.g., social functioning). Significant reductions in illicit opioid use were observed for all three conditions, with no significant differences between treatments in terms of reduction in opioid use or treatment discontinuation rates.

Fiellin et al. (2013) subsequently conducted a 24-week randomized controlled trial (N=141) that randomized patients with DSM-IV opioid dependence into either physician management or physician management plus cognitive behavioral therapy (CBT) for the first 12 weeks of treatment. Physician management, similar to SMM described above, consisted of brief, manualized, and medically focused counseling sessions. In the CBT condition, patients were offered weekly individual CBT sessions focused on behavioral activation, coping with drug craving, drug-refusal skills, management of high-risk situations, and functional analysis of behaviors (manual adapted from Carroll, 1998). The authors again found that both treatments were associated with significant decreases in illicit opioid use, but there were no significant differences between treatments on any of the outcomes tested (Fiellin et al., 2013).

Weiss et al. (2011) conducted a multisite randomized controlled trial with a 2-phase adaptive treatment design for treatment-seeking outpatients with DSM-IV opioid dependence, specific to prescription opioids. All patients (N=653) in phase 1 received brief buprenorphine treatment (i.e., 2-week stabilization, 2-week taper, and 8-week post-medication follow up) and either SMM or SMM plus opioid drug counseling (ODC; Pantalon et al., 1999). ODC provided longer counseling sessions and offered a wider variety of addiction skills and interactive exercises compared to SMM (manual adapted from Mercer and Woody, 1999 and Woody et al., 1977). Participants who were unsuccessful in phase 1, which was nearly 94% of the sample, were invited to participate in phase 2 of the study. In phase 2 (n=360), patients were again randomized into either SMM or SMM plus ODC and all patients received longer buprenorphine treatment (12 weeks of stabilization, 4-week taper, 8 weeks of follow up; outcomes were compared at the end of phase 1, at the end of 12 weeks of stabilization in phase 2, and at the end of follow-up in phase 2). Weiss et al. (2011) again found no significant treatment differences when comparing SMM + ODC vs. SMM in phase 1 or phase 2. The current study utilizes data from phase 2 of this study.

Ling et al. (2013) conducted a randomized control trial in adults with DSM-IV opioid dependence receiving buprenorphine and medical management in addition to one of four behavioral treatments: CBT, contingency management (CM), CBT plus CM, or no additional behavioral treatment. Participants had a 2-week buprenorphine induction phase followed by a 16-week behavioral treatment plus medication phase, and then finally a medication-only phase that included 40- and 52-week follow up assessments. Consistent with the previous findings, they did not find any significant treatment differences in opioid use or other outcomes during the treatment phase or at 40- or 52-week follow ups (Ling et al., 2013).

The four included studies had similar inclusion/exclusion criteria. All studies required that patients meet for a current opioid dependence diagnosis based on DSM-IV criteria, did not have another significant current substance use disorder (e.g., alcohol dependence), were deemed in good overall physical health, were not experiencing other severe psychiatric or psychotic symptoms, were deemed not an acute suicide risk, and did not have current legal issues or other factors (e.g., planned relocation) that would likely lead to an inability to continue treatment. All women of childbearing age were required to be on some form of acceptable birth control, which was monitored throughout treatment. The POATS study (Weiss et al., 2011) had additional exclusion criteria related to the type and course of opioid use that are relevant to highlight. The POATS study focused on the treatment of prescription opioid dependence, excluding individuals whose opioid dependence diagnosis was accounted for by heroin use alone, had used heroin more than four days in the previous month before treatment, or had ever injected heroin in their lifetime. These additional exclusion criteria likely resulted in a lower severity sample than the other included studies; for example, participants who initiated heroin use during the POATS study had worse treatment response overall (Dreifuss et al., 2013).

2.2. Outcomes and endpoints in the trials

For analyses examining moderators of OUD treatment outcomes, the primary outcome will be the number of weeks of opioid use over 12 weeks of treatment. All studies assessed opioid use via urine-confirmed self-report, consistent with best practices in drug use disorder studies (Donovan et al., 2012). Urine drug screens included tests for both opioid analgesics and heroin (i.e., include oxycodone and opiate assays); however, because these studies were conducted prior to the proliferation of fentanyl and its analogs in the drug supply, the presence of fentanyl was not assessed. Self-reported days of opioid use was also consistently collected across all studies using the Timeline Followback Method (Sobell and Sobell, 1992), a calendar-based method for assessing substance use behaviors. Of note, three of the included studies collected urine drug screens weekly and one collected them twice per week. As with other dimensions of heterogeneity, this will be addressed in our data analysis (see 2.4.4).

The lack of consensus in the field regarding the definition of treatment response is a challenge in OUD trials. This is evident in the 4 trials we selected, with some studies defining the primary outcome as opioid use over the entire course of the trial (Fiellin et al., 2006) and others defining outcome as treatment response late in the treatment episode (e.g., last 4 weeks of treatment; Weiss et al., 2011). For this study, we will define our primary opioid-use outcome as the number of weeks of opioid use over 12 weeks of treatment. This will enhance statistical power relative to a binary outcome (e.g., response or non-response). Furthermore, early treatment response is a robust predictor of ultimate treatment response (McDermott et al., 2015) and thus we are confident that the use of the full course of treatment will enhance variability, while likely yielding a pattern of results similar to studies that focus on late (e.g., last- month) treatment response. Nonetheless, secondary analyses will be conducted to confirm the robustness of our findings to other commonly used definitions of treatment response, including those used in the primary studies included in our dataset. For example, we will consider the binary outcome of “treatment success” used in Weiss et al. (2011), which was defined as abstinence from opioids in both week 12 and at least 2 of the prior 3 weeks (weeks 9–11).

Retention is often conceptualized as a binary variable representing either a pre-determined cut-off of sessions attended or participation in the final session (or sessions) of treatment. However, to allow for understanding of the variability in participants’ retention in treatment, we will use a quantitative measure of retention, i.e., weeks of treatment received. We believe that this will offer a more precise characterization of retention rather than relying on binary indicators of retention that would not be sensitive to potentially meaningful differences in the receipt of treatment.

The analyses examining the effect of behavioral therapy on functioning will use the primary outcome of functional outcomes as assessed by the Addiction Severity Index (ASI) (McLellan et al., 1992). The ASI is a multidimensional semi-structured clinical interview meant to provide a comprehensive assessment to aid in the formulation of treatment plans for individuals with substance use disorders. The ASI provides severity estimates of functional impairment, during the last 30 days, within 7 domains that are commonly affected among people with substance use disorders, specifically, functional consequences related to alcohol, other drugs, employment/financial support, medical/physical health, psychiatric/mental health, legal issues, and family/social domains. The ASI provides composite scores to quantify functioning within each of these domains over the past 30 days, which can be used to assess changes in functional outcomes across the course of treatment compared to baseline functioning. All studies collected the ASI at baseline and at least one post-treatment follow-up time point.

Proposed moderators were selected based on prior research. These include lifetime history of heroin use (Moore et al., 2007, Weiss et al., 2014), cocaine use at baseline/treatment entry (Sullivan et al., 2010), number of prior treatment episodes (Dreifuss et al., 2013), age (Dreifuss et al., 2013, Marcovitz et al., 2016), presence of chronic pain (McDermott et al., 2019), and co-occurring psychiatric disorders (Dreifuss et al., 2013, McHugh et al., 2021). Sex will also be included in all analyses to consider both main and interaction effects on outcomes of interest (see 2.4.3). In addition, we hypothesize that people with greater levels of baseline functional impairment on each of the 7 domains of the ASI will benefit more from the addition of behavioral therapy to buprenorphine due to the benefits of behavioral treatment for psychosocial needs in these domains.

2.3. Data harmonization

Although studies were selected based on the presence of common data elements, these studies were not specifically designed to be combined. Thus, harmonization requires careful attention to accurately matching data elements and will entail some decision-points for outcomes and moderators of interest. Consistent with recommended best practices for harmonizing data sets, we will conduct an iterative process of defining common data elements (including data type and range of possible values), mapping the original datasets onto these data elements (e.g., conducting any recoding or other transformations), engaging study investigators’ expertise (e.g., harmonization methods, biostatistics, treatment of OUD), and dataset quality control (e.g., identifying out-of-range values and missing data; Rolland et al., 2015). To ensure the rigor and reproducibility of this process, each step of the process and each decision point (including all analysis code) will be carefully documented. During the harmonization process, we will also identify opportunities for secondary analyses; secondary analyses will be preregistered to ensure transparency of this process.

2.4. Data analysis

Preliminary analyses will include use of descriptive statistics to identify skewness or outliers in the data. Quantitative variables will be transformed, if indicated by the results of the descriptive analysis.

All study aims will use an intent-to-treat approach that includes all randomized participants and will adjust for both study condition (i.e., treatment group) and study effects through inclusion of an indicator for study (i.e., to represent the 4 included trials). The analysis for each aim is described below.

2.4.1. Moderators of behavioral therapy effects on opioid outcomes

Due to variability in treatment prior to enrollment in each study (e.g., participants may have completed inpatient detoxification or come directly from the community) that could confound baseline levels of opioid use, the analysis will not focus on change in use from baseline, but rather use during the 12 weeks of treatment. Specifically, due to the significant potential variability both within and between each study in the period of time immediately prior to enrollment (e.g., participants may have started the study after an inpatient detoxification where they could not have accessed opioids or could have entered the study at the outpatient level of care and thus had a more recent last opioid use), achieving a true “baseline” level of use that represents pre-treatment OUD severity is not possible. The hypothesized moderators of the effect of behavioral therapy on this opioid outcome will be tested using a linear count model for weeks of opioid use (urine-confirmed self-reported weeks of use) examining the interaction between treatment group (behavioral therapy vs. no behavioral therapy) and each hypothesized moderator on the outcome. Significance testing will be conducted at the 0.01 level to adjust for multiplicity. Standard errors will be based on the so-called “sandwich” or empirical variance estimator (i.e., standard errors, and test statistics, will be robust to any heterogeneity of the variance). We also acknowledge that the moderators of interest will likely be correlated. Accordingly, if two or more moderators are identified in these analyses, we will conduct supplemental analyses to consider whether each moderator has a unique effect or whether this is due to a shared effect.

2.4.2. Effect of behavioral therapy on buprenorphine retention

The hypothesis that behavioral therapy will be associated with better buprenorphine retention compared to buprenorphine alone will be tested using a linear count model for weeks of urine-confirmed buprenorphine retention examining the main effect of treatment group; to account for any heterogeneity of the variance, standard errors for the treatment group effect will be based on the empirical variance estimator. As a secondary/exploratory aim, we will repeat the moderator analyses from Aim 1 in this linear regression model to test for subgroup differences in weeks of retention. Although a longer period of retention would have benefits to understanding long-term outcomes and long-term retention is an important predictor of outcome, the majority of discontinuation from buprenorphine treatment happens early in treatment, most of which is within the first month (Hser et al., 2014, Marcovitz et al., 2016) thus we believe that this 3-month period provides an adequate timeframe for understanding the most clinically relevant retention outcomes.

2.4.3. Effect of behavioral therapy on functioning

The hypothesis that behavioral therapy will yield greater improvement in functioning than buprenorphine alone will be tested using an analysis of covariance (ANCOVA) of functioning (possibly transformed) in a model with both the treatment group and baseline functioning as covariates. Separate ANCOVAs will be conducted for each of the 7 Addiction Severity Index composite scores, testing significance at the 0.01 level to adjust for multiplicity.

2.4.4. Other analytic considerations

All planned analyses will include main and interaction effects of sex on all outcomes of interest. Our work has demonstrated that women with OUD have both greater functional impairment and higher psychiatric severity relative to men (Campbell 2018, McHugh 2013), highlighting the importance of evaluating sex differences in this study. Consistent with recommendations from NIH and the Office of Research on Women’s Health we will perform exploratory analyses of study aims 1) separately for males and females, and 2) by including sex, and its interactions with effects of interest, in all analyses. This combined data set will provide much higher statistical power for these analyses than prior studies. We will use this same approach to test main and interaction effects of race/ethnicity, recognizing that the issue of underrepresentation is even greater in this domain.

All studies included in our proposal conducted repeated assessments over several months of treatment and thus inevitably include some missing data. We will use multiple imputation to address missing data. Observed and imputed data will then be combined in our statistical models. Then we will conduct all planned regression analyses within each of these imputed datasets; an overall summary of the analyses will be based on the average estimated regression effects over the imputed datasets, together with standard errors that appropriately account for between- and within-imputation sources of variability. We believe this strategy is superior to the approach to missing data used in many of our included studies, in which a missing urine screen was coded as positive; this approach is subject to biases, fails to account for uncertainty, and is not consistent with data suggesting that opioid outcomes vary following study dropout (Carroll and Weiss, 2017).

Although we selected studies with similar designs and assessments, there remains some heterogeneity across studies. We plan to account for study heterogeneity in two ways. In addition to the inclusion of an indicator for each study, we will account for potential differences across studies through conducting secondary analyses to test the robustness of our findings to study design features (e.g., behavioral therapy type, frequency of urine collection, maximum buprenorphine dose). Secondary analyses will be conceptualized as confirmatory, providing support of the robustness of our findings to methods modifications.

One important domain of heterogeneity is that the treatment lengths varied across studies. We selected week 12 as the primary endpoint as all studies had at least 12 weeks of weekly urine-confirmed self-report data. Nonetheless, we acknowledge that this time frame may be inadequate to see the full benefits of behavioral therapy, which can have an increasing effect over time (Kiluk and Carroll, 2013). Accordingly, we will conduct a secondary analysis of the main outcomes using the primary endpoint for each study; the opioid outcome will be defined as the proportion of opioid use weeks to account for variability in the number of weeks in this approach.

2.5. Statistical power

For this data harmonization study, our sample size was predetermined based on the samples of the included studies. By combining 4 trials, we seek to increase statistical power for the detection of moderational effects that may be underpowered in any one individual clinical trial. Our power analyses indicated that the inclusion of N=869 participants from 4 studies will provide more than adequate statistical power for testing hypotheses for our primary study aims. All analyses assumed an adjusted significance level of.01 to correct for multiplicity.

Based on summaries from the existing data, for Aim 1 (moderators of the effect of behavioral therapy on opioid use outcomes) we assumed the standard deviation (SD) of the outcome was approximately 4.2 and allowed the prevalence of the moderating factors to vary from 25% to 75% (e.g., in 3 of the studies the prevalence of heroin use varies from 26% to 69%). For Aim 1 hypotheses about moderating factors, with N=869 the study will have power of at least 80% to detect interactions with treatment effects of 1.95–2.25, depending on the prevalence of the particular moderating factor. These effect sizes correspond to the additional benefits (e.g., reductions) of approximately 2 weeks of opioid use during the first 12 weeks when one moderating subgroup is compared to the other. These effects correspond to standardized effect sizes ranging from 0.46 to 0.54 (moderate effects).

For the Aim 2 hypothesis about the effect of the addition of behavioral therapy to buprenorphine maintenance on retention, assuming SD=2.9 based on existing data summaries, with N=869 the study will have power of at least 80% to detect a treatment group difference of 0.55 weeks (out of 12 weeks) of retention, i.e., treatment group differences in retention of greater than ½ week. This corresponds to a standardized effect of 0.19 (small effect).

Finally, for Aim 3 hypotheses about the effect of the addition of behavioral therapy to buprenorphine maintenance on functional outcomes, the study will have power of at least 80% to detect standardized treatment effects sizes of 0.23; these are considered by many to be relatively small effect sizes and in our sample would translate, for example, to mean differences in the 7 Addiction Severity Index Composite Scores of approximately 0.02–0.07, depending on the variability of the specific composite score (e.g., SD = 0.32 for Medical, SD = 0.19 for Psychiatric, SD = 0.10 for Legal).

3. Discussion

The overarching objective of this trial is to leverage existing clinical trial data to advance understanding of the potential role of behavioral therapy in buprenorphine maintenance for OUD. In addition to considering the effects of behavioral therapy on non-opioid outcomes (buprenorphine retention and functional outcomes), by combining four RCTs, we have increased the statistical power to detect moderational effects. This will allow us to answer important questions about who benefits from the addition of high-intensity behavioral therapy.

Understanding of the role of behavioral therapy in the treatment of OUD has significant and direct implications for clinical practice. Access to psychosocial services and ease of referral systems to behavioral health providers are among the most commonly reported barriers to buprenorphine prescribing among waivered providers (Jones and McCance-Katz, 2019). The proposed study would fill important information gaps about the role of behavioral therapy in the treatment of OUD to help inform guidelines for the practice of buprenorphine prescribing. Furthermore, improving treatment outcomes for OUD will require an increase in person-centered approaches that consider individual differences in treatment needs and preferences (Volkow, 2020, Blanco 2020). Data on the heterogeneity of treatment response can help to guide shared decision-making and the personalization of treatment to individuals.

Of note, our selection of potential moderators was informed by the extant literature on OUD outcomes. Given the vast variability in what baseline characteristics are assessed, how they are assessed and how many are tested in individual clinical trials, we understand that this is certainly not a comprehensive list of possible treatment moderators. Accordingly, secondary analyses will use exploratory approaches to understanding the potential contribution of a wider array of baseline variables (e.g., baseline other drug use) on outcomes. Furthermore, many of our hypothesized moderators reflect greater overall clinical severity (e.g., cocaine use, high functional impairment). Although we anticipate that this group may show greater additive benefit of behavioral therapy, it is possible that this group may in fact have a worse overall treatment response across both conditions due to their clinical acuity and complexity.

Although data harmonization offers many benefits, such as increasing statistical power, and answering questions in a time- and cost-efficient manner, there are also limitations to this approach. Any harmonization trial is limited by (1) any weaknesses in the clinical trials included (e.g., limitations in the sample selection strategy), and (2) the inevitability of some design and method variation across studies (e.g., different measurement strategies or treatment durations). In this study, we will seek to minimize these weaknesses, where possible, through inclusion of sensitivity analyses to test the robustness of findings to variation across the studies. Another limitation is that the opioid drug supply is continually evolving. These trials were all completed prior to the drastic increase in fentanyl and fentanyl analogs in the opioid supply in the US beginning in 2013. Accordingly, these studies did not systematically test for fentanyl. We do not anticipate that this will be a limitation due to its rare use at that time; nonetheless, any observed effects will require replication in more recently collected samples. Likewise, the available formulations of medication for OUD continue to evolve, with the introduction of long-acting injectable buprenorphine formulations. Another limitation is the lack of a well-validated measure of functional outcomes in OUD research. We will use the ASI in this study as it is a validated and widely used tool, nonetheless, it is important to note that some of these functional domains (e.g., legal status) may be unlikely to change over a relatively short time period (12 weeks), and thus any findings in this domain will be interpreted with caution and contextualized with respect to overall change (i.e., if substantive reductions are observed within subjects as well as between conditions).

The use of data harmonization to answer critical questions about individual treatment response allows for continuous updating as new data sources are available. Furthermore, as the opioid epidemic and the tools at the disposal of the health care system evolve, the continued use of existing data sources can provide essential preliminary data to inform future study designs to further the goal of improving treatment outcomes for people with OUD.

Funding

Effort for this project was supported by NIH grant R01 DA054113.

Author contributions

Drs. McHugh, Fitzmaurice and Weiss designed the trial. Drs. McHugh, Bailey and Fitzmaurice wrote the first draft of the manuscript. All authors edited the manuscript and contributed to the interpretation of findings. All authors approved the final manuscript.

CRediT authorship contribution statement

Garrett M. Fitzmaurice: Writing – original draft, Conceptualization. Allen J. Bailey: Writing – original draft. Roger D. Weiss: Writing – review & editing, Conceptualization. R. Kathryn McHugh: Writing – original draft, Funding acquisition, Conceptualization.

Declaration of Competing Interest

In the past 12 months, Dr. Weiss has consulted to Alkermes. All other authors have no disclosures or potential conflicts of interest to report.

References

  1. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association, Washington, DC.
  2. Blanco C., Wiley T. R.A., Lloyd J.J., Lopez M.F., Volkow N.D. America’s opioid crisis: the need for an integrated public health approach. Transl. Psychiatry. 2020;10 doi: 10.1038/s41398-020-0847-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Brigham G.S., Slesnick N., Winhusen T.M., Lewis D.F., Guo X., Somoza E. A randomized pilot clinical trial to evaluate the efficacy of Community Reinforcement and Family Training for Treatment Retention (CRAFT-T) for improving outcomes for patients completing opioid detoxification. Drug Alcohol Depend. 2014;138:240–243. doi: 10.1016/j.drugalcdep.2014.02.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Campbell A.N.C., Barbosa-Leiker C., Hatch-Maillette M., Mennenga S.E., Pavlicova M., Scodes J., Saraiya T., Mitchell S.G., Rotrosen J., Novo P., Nunes E.V., Greenfield S.F. Gender differences in demographic and clinical characteristics of patients with opioid use disorder entering a comparative effectiveness medication trial. Am. J. Addict. 2018;27:465–470. doi: 10.1111/ajad.12784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Carroll, K.M., 1998. A Cognitive Behavioral Approach: Treating Cocaine Addiction. Rockville, MD.
  6. Carroll K.M., Weiss R.D. The role of behavioral interventions in buprenorphine maintenance treatment: a review. Am. J. Psychiatry. 2017;174:738–747. doi: 10.1176/appi.ajp.2016.16070792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Committee on Medication-Assisted Treatment for Opioid Use Disorder, 2019. Medications for Opioid Use Disorder Save Lives. The National Academics of Sciences, Engineering and Medicine, Washington, DC.
  8. Donovan D.M., Bigelow G.E., Brigham G.S., Carroll K.M., Cohen A.J., Gardin J.G., Hamilton J.A., Huestis M.A., Hughes J.R., Lindblad R., Marlatt G.A., Preston K.L., Selzer J.A., Somoza E.C., Wakim P.G., Wells E.A. Primary outcome indices in illicit drug dependence treatment research: systematic approach to selection and measurement of drug use end-points in clinical trials. Addiction. 2012;107:694–708. doi: 10.1111/j.1360-0443.2011.03473.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dreifuss J.A., Griffin M.L., Frost K., Fitzmaurice G.M., Potter J.S., Fiellin D.A., Selzer J., Hatch-Maillette M., Sonne S.C., Weiss R.D. Patient characteristics associated with buprenorphine/naloxone treatment outcome for prescription opioid dependence: results from a multisite study. Drug Alcohol Depend. 2013;131:112–118. doi: 10.1016/j.drugalcdep.2012.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dutra L., Stathopoulou G., Basden S.L., Leyro T.M., Powers M.B., Otto M.W. A meta-analytic review of psychosocial interventions for substance use disorders. Am. J. Psychiatry. 2008;165:179–187. doi: 10.1176/appi.ajp.2007.06111851. [DOI] [PubMed] [Google Scholar]
  11. Fiellin D.A., Pantalon M.V., Chawarski M.C., Moore B.A., Sullivan L.E., O'Connor P.G., Schottenfeld R.S. Counseling plus buprenorphine-naloxone maintenance therapy for opioid dependence. N. Engl. J. Med. 2006;355:365–374. doi: 10.1056/NEJMoa055255. [DOI] [PubMed] [Google Scholar]
  12. Fiellin D.A., Barry D.T., Sullivan L.E., Cutter C.J., Moore B.A., O'Connor P.G., Schottenfeld R.S. A randomized trial of cognitive behavioral therapy in primary care-based buprenorphine. Am. J. Med. 2013;126 doi: 10.1016/j.amjmed.2012.07.005. 74.e11-77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fischer B., Pang M., Tyndall M. The opioid death crisis in Canada: crucial lessons for public health. Lancet Public Health. 2019;4:e81–e82. doi: 10.1016/S2468-2667(18)30232-9. [DOI] [PubMed] [Google Scholar]
  14. Hser Y.I., Saxon A.J., Huang D., Hasson A., Thomas C., Hillhouse M., Jacobs P., Teruya C., McLaughlin P., Wiest K., Cohen A., Ling W. Treatment retention among patients randomized to buprenorphine/naloxone compared to methadone in a multi-site trial. Addiction. 2014;109:79–87. doi: 10.1111/add.12333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jones C.M., McCance-Katz E.F. Co-occurring substance use and mental disorders among adults with opioid use disorder. Drug Alcohol Depend. 2019;197:78–82. doi: 10.1016/j.drugalcdep.2018.12.030. [DOI] [PubMed] [Google Scholar]
  16. Kiluk B.D., Carroll K.M. New developments in behavioral treatments for substance use disorders. Curr. Psychiatry Rep. 2013;15:420. doi: 10.1007/s11920-013-0420-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lee J.D., Nunes E.V., Jr, Novo P., Bachrach K., Bailey G.L., Bhatt S., Farkas S., Fishman M., Gauthier P., Hodgkins C.C., King J., Lindblad R., Liu D., Matthews A.G., May J., Peavy K.M., Ross S., Salazar D., Schkolnik P., Shmueli-Blumberg D., Stablein D., Subramaniam G., Rotrosen J. Comparative effectiveness of extended-release naltrexone versus buprenorphine-naloxone for opioid relapse prevention (X:BOT): a multicentre, open-label, randomised controlled trial. Lancet. 2018;391:309–318. doi: 10.1016/S0140-6736(17)32812-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ling W., Hillhouse M., Ang A., Jenkins J., Fahey J. Comparison of behavioral treatment conditions in buprenorphine maintenance. Addiction. 2013;108:1788–1798. doi: 10.1111/add.12266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Magill M., Ray L., Kiluk B., Hoadley A., Bernstein M., Tonigan J.S., Carroll K. A meta-analysis of cognitive-behavioral therapy for alcohol or other drug use disorders: treatment efficacy by contrast condition. J. Consult. Clin. Psychol. 2019;87:1093–1105. doi: 10.1037/ccp0000447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Marcovitz D.E., McHugh R.K., Volpe J., Votaw V., Connery H.S. Predictors of early dropout in outpatient buprenorphine/naloxone treatment. Am. J. Addict. 2016;25:472–477. doi: 10.1111/ajad.12414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. McDermott K.A., Griffin M.L., Connery H.S., Hilario E.Y., Fiellin D.A., Fitzmaurice G.M., Weiss R.D. Initial response as a predictor of 12-week buprenorphine-naloxone treatment response in a prescription opioid-dependent population. J. Clin. Psychiatry. 2015;76:189–194. doi: 10.4088/JCP.14m09096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. McDermott K.A., Griffin M.L., McHugh R.K., Fitzmaurice G.M., Jamison R.N., Provost S.E., Weiss R.D. Long-term naturalistic follow-up of chronic pain in adults with prescription opioid use disorder. Drug Alcohol Depend. 2019;205 doi: 10.1016/j.drugalcdep.2019.107675. [DOI] [PubMed] [Google Scholar]
  23. McHugh R.K, DeVito E.E., Dodd D., Carroll K.M., Potter J.S., Greenfield S.F., Connery H.S., Weiss R.D. Gender differences in a clinical trial for prescription opioid dependence. J. Subst. Abuse Treat. 2013;45:38–43. doi: 10.1016/j.jsat.2012.12.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. McHugh R.K., Hilton B.T., Chase A.M., Griffin M.L., Weiss R.D. Do people with opioid use disorder and posttraumatic stress disorder benefit from dding Individual opioid Drug Counseling to buprenorphine? Drug Alcohol Depend. 2021;228 doi: 10.1016/j.drugalcdep.2021.109084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. McLellan A.T., Kushner H., Metzger D., Peters R. The fifth edition of the Addiction Severity Index. J. Subst. Abus. Treat. 1992;9:199–213. doi: 10.1016/0740-5472(92)90062-s. [DOI] [PubMed] [Google Scholar]
  26. Mercer, D.E., Woody, G.E., 1999. Individual Drug Counseling, NIH Pub. No 99-4380. University of Pennsylvania, Veterans Affairs Medical Center.
  27. Moore B.A., Fiellin D.A., Barry D.T., Sullivan L.E., Chawarski M.C., O'Connor P.G., Schottenfeld R.S. Primary care office-based buprenorphine treatment: comparison of heroin and prescription opioid dependent patients. J. Gen. Intern. Med. 2007;22:527–530. doi: 10.1007/s11606-007-0129-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Nielsen S., Hillhouse M., Mooney L., Ang A., Ling W. Buprenorphine pharmacotherapy and behavioral treatment: comparison of outcomes among prescription opioid users, heroin users and combination users. J. Subst. Abus. Treat. 2015;48:70–76. doi: 10.1016/j.jsat.2014.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Pantalon, M.V., Fiellin, D.A., Schottenfeld, R.S., Gordon, L., O'Connor, P.G., 1999. Manual for Enhanced Medical Management of Opioid Dependence with Buprenorphine. Yale University.
  30. Pierce M., Bird S.M., Hickman M., Marsden J., Dunn G., Jones A., Millar T. Impact of treatment for opioid dependence on fatal drug-related poisoning: a national cohort study in England. Addiction. 2016;111:298–308. doi: 10.1111/add.13193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ray L.A., Meredith L.R., Kiluk B.D., Walthers J., Carroll K.M., Magill M. Combined pharmacotherapy and cognitive behavioral therapy for adults with alcohol or substance use disorders: a systematic review and meta-analysis. JAMA Netw. Open. 2020;3 doi: 10.1001/jamanetworkopen.2020.8279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rolland B., Reid S., Stelling D., Warnick G., Thornquist M., Feng Z., Potter J.D. Toward rigorous data harmonization in cancer epidemiology research: one approach. Am. J. Epidemiol. 2015;182:1033–1038. doi: 10.1093/aje/kwv133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Samples H., Williams A.R., Crystal S., Olfson M. Psychosocial and behavioral therapy in conjunction with medication for opioid use disorder: patterns, predictors, and association with buprenorphine treatment outcomes. J. Subst. Abus. Treat. 2022;139 doi: 10.1016/j.jsat.2022.108774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sobell L.C., Sobell M.B. In: Measuring Alcohol Consumption: Psychosocial and Biochemical Methods. Litten R.Z., Allen J.P., editors. Humana Press; Totowa, NJ: 1992. Timeline follow-back: a technique for assessing self-reported alcohol consumption; pp. 41–72. [Google Scholar]
  35. Sullivan L.E., Moore B.A., O'Connor P.G., Barry D.T., Chawarski M.C., Schottenfeld R.S., Fiellin D.A. The association between cocaine use and treatment outcomes in patients receiving office-based buprenorphine/naloxone for the treatment of opioid dependence. Am. J. Addict. 2010;19:53–58. doi: 10.1111/j.1521-0391.2009.00003.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Volkow N.D. Personalizing the treatment of substance use disorders. Am. J. Psychiatry. 2020;177:113–116. doi: 10.1176/appi.ajp.2019.19121284. [DOI] [PubMed] [Google Scholar]
  37. Weiss R.D., Potter J.S., Fiellin D.A., Byrne M., Connery H.S., Dickinson W., Gardin J., Griffin M.L., Gourevitch M.N., Haller D.L., Hasson A.L., Huang Z., Jacobs P., Kosinski A.S., Lindblad R., McCance-Katz E.F., Provost S.E., Selzer J., Somoza E.C., Sonne S.C., Ling W. Adjunctive counseling during brief and extended buprenorphine-naloxone treatment for prescription opioid dependence: a 2-phase randomized controlled trial. Arch. Gen. Psychiatry. 2011;68:1238–1246. doi: 10.1001/archgenpsychiatry.2011.121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Weiss R.D., Griffin M.L., Potter J.S., Dodd D.R., Dreifuss J.A., Connery H.S., Carroll K.M. Who benefits from additional drug counseling among prescription opioid-dependent patients receiving buprenorphine-naloxone and standard medical management? Drug Alcohol Depend. 2014;140:118–122. doi: 10.1016/j.drugalcdep.2014.04.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wilson N., Kariisa M., Seth P., Smith Ht, Davis N.L. Drug and opioid-involved overdose deaths - United States, 2017-2018. MMWR Morb. Mortal. Wkly. Rep. 2020;69:290–297. doi: 10.15585/mmwr.mm6911a4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Woody, G., Stockdale, D., Hargrove, E., 1977. A Manual of Drug Counseling with Opiate Addicts. University of Pennsylvania.

Articles from Drug and Alcohol Dependence Reports are provided here courtesy of Elsevier

RESOURCES