Objective:
To compare the effectiveness of individual motivational interviewing plus group-based cognitive behavioral therapy (MI+CBT) versus patient-clinician shared decision making (SDM) on change in daily dosage of prescribed opioids, physical functioning, and pain interference for individuals with chronic non-cancer pain (CNCP).
Rationale:
About 24% of Americans suffer from CNCP and clinicians often prescribe opioids to treat it. Once on chronic opioid therapy (COT), individuals often continue with this class of medication for years. Evidence for the effectiveness of COT to treat CNCP is limited, exposing individuals to known risks. Modified or novel pharmacological and nonpharmacological strategies are needed to improve pain management and promote informed decision making regarding possible opioid dose reduction.
Study Design and Approach:
This was a large-scale, pragmatic randomized controlled trial implementing pharmacotherapy guidelines and behavioral interventions in real-world settings. A key eligibility criterion was individuals who were prescribed ≥ 20 milligrams of morphine equivalents for more than three months.
Interventions:
Researchers examined the comparative effectiveness of MI+CBT versus SDM for who are on COT. Neither approach is directive, and both support patient choice. Although MI+CBT and SDM are both behavioral intervention strategies, they differ in content covered and training of those delivering the intervention (behavioral vs. medical). Participants in both study arms received guideline-concordant pharmacotherapy treatment, based on clinical guidelines for opioid therapy for CNCP.
Outcomes:
- Primary: The change from baseline in average daily prescribed opioid dosage, measured in milligrams of morphine equivalents, at 12 months, with secondary time points at 6 months and 18 months
- Secondary: The change from baseline to 6 and 12 months (primary timepoint) in the PROMIS Short Form v1.0 Pain Interference and PROMIS Short Form v1.0 Physical Function scales.
Timeline:
The project commenced in February 2018. Participant recruitment occurred from June 2019 to March 2022. Delivery of the intervention occurred on a rolling basis through March 2023.
Recruitment, Screening, Enrollment, and Randomization:
The study randomized 525 participants from primary care and pain clinics at three medical centers in North Carolina and Tennessee. The researchers identified individuals who were potentially eligible through electronic health records and invited them to participate. A Research Coordinator contacted individuals to complete screening, enrollment, and randomization.
Data Collection:
The researchers collected validated patient-reported outcomes through Web-based and phone-based surveys and leveraged existing harmonized electronic health record (EHR) data for clinical outcomes.
Data Analysis:
Outcomes were analyzed for all randomized participants in an intent-to-treat fashion, irrespective of the amount of intervention received. Because some individuals did not attend an intervention session, sensitivity analyses were based on a modified intent-to-treat population limited to participants receiving at least 1 intervention session and a per-protocol population limited to participants receiving at least four CBT or SDM sessions.
Changes from baseline were analyzed with a repeated measures linear model accounting for correlation of measurements over time, with effects for intervention arm, time interval as a categorical predictor, intervention-by-time interaction, and adjustment for baseline value and study site. Differential treatment effects for two preplanned subgroups based on sex and prior mental health diagnosis, as defined by IDC-10 codes recorded in the EHR, were assessed at 12 months via a subgroup-by-intervention interaction within the repeated measures model.
Qualitative research methods were used to obtain participant input on their experiences.
