Overview
We are conducting a randomized clinical trial in which individuals with a chronic pain condition who are currently taking prescription opioids are randomly assigned to one of two arms: a single-session skills-based pain management class or a single-session health education class (active control; “Control”) (Figs. 1 and 2). Participants will be followed for 12 months after treatment. Participants will be assessed via an online screening form, a phone screening, enrollment survey, pre-class survey, a two-week daily baseline period, 2-week daily follow-up period, and at 3, 6 and 12 months post-treatment. Team statisticians blinded to participant treatment assignment will assess outcomes three months after treatment, and after 6 and 12 months. The primary outcome is opioid use 3 months post-treatment. Secondary outcomes include reductions in pain-specific distress, pain interference and opioid misuse at 3 months.
The protocol for this trial has been approved by the Stanford Institutional Review Board (IRB). All participants will be required to give informed consent to a trained study team member prior to enrollment in the study.
Study sample and setting
Participants for this trial will be recruited through targeted emails and advertisements at Stanford’s pain management clinics, in addition to the Stanford Systems Neuroscience and Pain Lab (SNAPL) database. Recruitment efforts will extend to social media marketing, and local advertisements in clinics and in the community. All advertisements will direct interested individuals to an online screening form that assesses for initial eligibility. The study will enroll 136 adults (age 18-80 years) with a chronic non-cancer pain condition, currently using prescription opioids ≥20 morphine equivalent daily dose (MEDD) for ≥3 months and who meet study criteria (Table 1). The sample size accounts for expected attrition. Eligibility will be assessed by the research staff.
Inclusion and exclusion criteria
Tables 1 and 2 list the inclusion and exclusion criteria, respectively, as well as the rationale for each criterion and the sources where each criterion will be assessed. Additionally, we require that the participants be willing and available to participate in the full treatment session to which they are assigned and able to respond to the daily measures (at baseline and follow-up) and post-treatment (3, 6, and 12 months) follow-up questionnaires.
Recruitment procedures
Because the study intervention involves group treatment classes, we are recruiting participants in cohorts consisting of 7-12 participants per class cohort (minimum of 4 participants, maximum of 15 participants per cohort) for both treatment arms.
Interested individuals deemed initially eligible by the online screening will be further screened over the phone. Eligible individuals will then be invited to enroll in the study, and consented with a research staff over the phone, after which they provide an electronic signature to the consent form emailed to them. Participants are randomized following eligibility confirmation and informed consent procedures. Then participants complete the enrollment survey, which includes information related to their chronic pain, opioid use, non-opioid medication use, medical history, and psychosocial wellbeing. Additionally, measures of opioid misuse behaviors and severity, treatment expectancies and patient motivation factors will be administered.
Randomization
Enrolled participants will be randomized to one of two treatment arms: Empowered Relief (ER) and chronic pain health education class (HE). An automated program in REDCap will randomly assign a participant to a treatment arm when enrolled and will ensure blinded randomization, as well as equal numbers in both arms at the end of data collection.
Blinding
Participants cannot and will not be blinded to the intervention they are randomized to. A clinical psychologist who will be trained in the intervention, blind to participants and has no involvement in data analysis will deliver the intervention. The study coordinator will be responsible for handling the randomization process through REDCap, but will remain blinded to the randomization scheme. All data given to the statistician will be blinded, except as required when reporting adverse events. All research data will be kept separate from identifiers and linked using a participant number. An alternative research team member will have access to the data and will be responsible for the data monitoring. Only the principal investigators will have access to the file linking names and participant numbers and the file will be stored in their locked offices. The team will have access to the final unidentified dataset.
Study treatments
Both study arms (ER and HE) consist of a single-session 2-hour group class. Participants will leave the ER class with home-based resources that facilitate ongoing self-regulation and pain self-management.
Single-session skills-based pain class (Empowered Relief)
Our group developed the single-session skills-based pain management class in 2013 with a goal of rapidly equipping patients with skills to self-regulate pain-specific distress. Pilot data revealed significantly reduced pain-specific distress – as indexed by reductions in pain catastrophizing-- one month post-treatment regardless of comorbid depression and anxiety [4]. The single-session skills-based class (“Empowered Relief”) is also the subject of NIH-funded randomized controlled trial in chronic low back pain [5] with pain reduction as the primary outcome.
For this study, a clinical psychologist trained in delivering the 2-hour intervention will administer the class to groups of enrolled participants. Brief education on opioid reduction is included in the class, along with a one-page handout that summarizes key research findings. Materials are already in use in our existing projects [23]. The class is delivered by Powerpoint presentation and includes mind-body pain science, the importance of self-regulation in the context of pain and stress, and evidence-based skills that target pain-specific distress and enhance pain control. Participants will be guided in developing their own self-treatment plan and acquiring the skills necessary to decrease pain and stress related physiological hyperarousal, and to enhance regulation of cognition and emotion within the context of pain. At the end of the class, participants will leave with: 1) a self-tailored personalized plan to target pain-specific distress; 2) a 20-minute guided relaxation response electronic app; and 3) a printed copy didactic class content. Cohort effects are expected to be minimal due to the single session nature of the class, the highly structured and manualized nature of the class, and minimal participant interaction.
Health education class
This is 2-hour class that will be delivered by a health educator using a PowerPoint presentation. The class will provide the participants with general health information related to exercise, nutrition and medication management. It includes information on managing flare-ups, working with healthcare professionals, evaluating treatments, and making informed decisions. This class is already in use in our NIH-funded research [5, 24]. This class serves as a control to the ER class, with matching factors, such as duration, structure, format, and location.
Class sites
All treatment sessions will occur at approved clinical or research sites within the Stanford University School of Medicine and Stanford HealthCare.
Instructors
For the ER treatment group, all instructors will be doctoral level clinical psychologists trained in the treatment of chronic pain. The HE class will be expert-led by experienced health educators or chronic pain professionals (e.g., chronic pain physician assistant).
Training and monitoring of instructors
ER instructors will be trained in the study protocol for their classes prior to administering treatment. Existing treatment manuals as well as highly structured and standardized class content will assure treatment fidelity. A research coordinator, serving as fidelity rater, will directly observe the first 3 classes of each treatment arm as well. Cohort effects are likely to be minimal, due to the single-session format and relatively minimal participant interaction.
Measures
Demographic data, chronic pain history, current and past opioid use, non-opioid medication, and current treatments will be collected at enrollment. The 17-item Current Opioid Misuse Measure [25] will be used to measure opioid misuse and change throughout the study. During screening, we will characterize the patient’s opioid misuse behaviors using DSM-5 Opioid (DSO) [26], a Clinical Trials network NIDA-supported instrument. PROMIS measures will be used to assess Pain Interference, Physical Function, Depression, Anxiety, Anger, Sleep Disturbance, Fatigue, Social Isolation, and Global Health using short forms [27]. Our group has applied the NIH PROMIS measures in multiple nationally funded clinical pain treatment trials and other studies [28-30]. Pain Catastrophizing will be assessed using the Pain Catastrophizing Scale (subscales: rumination, magnification, and feelings of helplessness) [31]. Motivational Factors including desire, confidence, readiness and motivation to reduce their opioid use, will be assessed using the questions developed by Goesling and colleagues [32]. Lastly, treatment expectancies will be assessed using the Stanford Expectations of Treatment Scale (SETS) [33].
Baseline period: During screening and enrollment, patients will complete an online baseline assessment with demographics, as well as measures inquiring about pain condition(s) and characteristics, opioid use and misuse, pain catastrophizing and the PROMIS measures. Within one month of starting treatment, participants will also complete two weeks of daily surveys assessing daily levels of pain catastrophizing and daily opioid use.
Pre-treatment assessment: Three days pre-treatment, patients will complete an online pre-class survey assessing pain condition and characteristics, opioid use and misuse, pain catastrophizing, and the PROMIS measures. Participants will not be asked to repeat demographic information again, but these measures will be identical to those assessed at baseline.
Post-treatment assessment: Immediately post-treatment, patients will complete a brief questionnaire assessing patient satisfaction with the intervention on an 11-point Likert scale.
One-month post-treatment, participants will complete daily surveys assessing daily levels of pain catastrophizing and daily opioid use. For patients in the ER class, they will also complete measures inquiring about daily app and skill use for two weeks. At 3, 6, and 12 months post-treatment, all participants will complete a set of questionnaires identical to those administered pre-class. The primary study endpoint is three months post-treatment. App use will also be tracked by REDCap throughout the duration of the trial.
Participants may receive up $160 for study completion.
Primary outcome measure. Our primary outcome measure is opioid use at three months post-treatment. We will convert opioid doses to morphine equivalent daily dose (MEDD). We will calculate a within-subject difference from baseline to determine if reduction in MEDD use occurred.
Secondary outcome measures: Our secondary outcome is pain-specific distress as indexed by pain catastrophizing scale scores at 3-months post treatment. Pain catastrophizing will be assessed using the Pain Catastrophizing Scale (PCS) and a total score will be generated using the sum of the 13-items of the PCS. The PCS consists of three subscales (helplessness, magnification, rumination) and has good psychometrics [31].
Tertiary outcome measures: NIH PROMIS measures [34] will be administered to assess Anxiety, Depression, Anger, Fatigue, Pain Interference, Physical Function, Global Health, and Sleep Disturbance. These measures have been successfully applied to pain research [13, 35-37].
Daily measures. All participants will complete daily measures of opioid medication use, pain intensity, and 3-item daily PCS during two 2-week time periods: at baseline (up to one month prior to the class), and at follow-up (one month post-class). Daily skills use will also be assessed for 2-weeks during the follow-up time-period only for the ER treatment arm. These include a 4-item questionnaire measuring frequency of use of cognitive, behavioral, or psychophysiological techniques over the past 24 hours from 0 times to 5+ times.
Data collection, quality control and confidentiality
The online assessments completed by participants will be gathered securely in a REDCap database. Though we do not expect any questionnaires to be collected on paper, if unforeseen circumstances require this to occur, they will be stored as source data and a member of the study team will manually enter the responses into the REDCap database. Additionally, members of the team will be trained to use and complete Case Report Forms (CRFs), how to review them for completeness, as well as how to maintain participant confidentiality. Patient flow will be reported according to the CONSORT (Consolidated Standards of Reporting Trials) guidelines [39].
Protection of human participants and assessment of safety:
Protection of human participants
The Stanford University Institutional Review Board (IRB) approved this study.
Safety Monitoring
This trial will be monitored for safety by an independent Data and Safety Monitoring Board (DSMB) composed of a biostatistician and a clinical psychologists with knowledge in treatment of chronic pain conditions. A chairperson of the board will also be appointed who is an individual with expertise in treatment outcome research methodology and who has worked as a consultant on other clinical trial studies and DSMB boards. These members will have no other involvement in the study and will serve as independent reviewers of the DSMB. They will convene twice a year or as per needed basis. Members of the DSMB will make relevant safety decisions regarding reported participant cases. The DSMB report will be sent to NIH within 2 weeks of the meeting, twice a year. The report will be sent to the Stanford IRB after meetings have been held for the year, and prior to the continuing renewal.
Adverse experiences
The treatments in this study are not anticipated to pose any risks to participants. However, a study coordinator will review enrolled patient records weekly to monitor for adverse events. In the case of an adverse event, an Adverse Event Case Report Form will be completed. These will be discussed in monthly team meetings and reported to the DSMB and IRB annually. Serious adverse events will be reported to the Stanford IRB and NIH. The DSMB and PI will evaluate all serious adverse events within 24 hours after the study team becomes aware of the incident. All study-related adverse events will be included in the annual report to the NIH, and serious adverse events will be reported within 2 weeks.
Stopping rules
The trial will be stopped if (1) either treatment intervention is associated with adverse effects that calls into question the safety of the intervention, (2) difficulty in study recruitment or retention that will impact the ability to evaluable study endpoints, (3) new information becomes available during the trial that indicates the need to stop the trial, or (4) other unforeseen situations that would warrant stopping the study.
Statistical issues
Sample size and detectable differences
We chose our sample size to ensure adequate power to detect treatment effects on the primary outcome (i.e. opioid use) and to investigate pain catastrophizing reduction as a mediator between the two groups (ER and HE). The project will enroll 136 participants (ages 18-80 years) with diagnosis of chronic non-cancer pain (>3 months in duration) and currently using prescription opioids.
To compare the main effect of single-session skills-based pain management class on opioid use against the HE control condition, we will plan to enroll 136 participants and have 116 completers (58 per group). The proposed sample size accounts for 15% attrition in each treatment arm. This is lower than the current attrition rate seen in pain-CBT literature of 18-25% [40, 41], but we believe our less burdensome single-session groups will lead to lower rates. We hope to achieve 80% power to detect medium-large treatment effects on the primary outcome (i.e. opioid use).
To examine reduction of PC as a treatment mediator, we will use bias-corrected bootstapping. Previous data show a large effect of PC treatment on PCS scores (d=0.85~1.15) [4], and others show a medium-size effect of PCS scores on opioid use (d=0.4) [32]. When using bias-corrected bootstrap for mediation analysis, a total of 115 subjects are required for a large and medium effect associations with a mediator variable (PCS score), 80% power and α=0.05 [42].
Statistical analyses
Primary analyses
We will use an intent-to-treat approach in all analyses (i.e. the assessment of individuals will be analyzed by randomized group, regardless of participation in any classes). By doing so, we protect against any confounding that arises as a result of subject dropout.
The main effect of the single-session skills-based pain management class on opioid use will be compared against the HE control class using a 2-samples t-test. Clinically minimal reduction is defined as >15% reduction in opioid use [43] in MEDD, the recommended unit of measurement in studies of opioid use [38]. We will compare the rate of participants who reach clinically minimal reduction between the two groups. We will quantify absolute opioid reduction in addition to percent change reduction within subject and between classes. Finally, we will quantify percent achieving each group threshold for importance of change (15%, 30%, and 50% as minimally, moderately and substantially important change scores, respectively).
To test the hypothesis that the ER class will have greater reductions in pain-related distress, pain interference, and opioid misuse compared to the HE class, our endpoint is opioid misuse, pain catastrophizing, and pain interference at 3 months, and its within-subject difference from baseline is calculated. The mean difference in the ER group will be compared against the HE arm using the two-sample t-test. We will also compare the proportion of success rate, defined as ≥30% reduction in pain catastrophizing and pain interference for clinical significant treatment response [44] and a reduction to 13 as measured by the COMM (Current Opioid Misuse Measure). This number if the validated cut-point on COMM [45].
Secondary objectives
To test the hypothesis that Daily PC predicts same-day and next-day opioid use, a mixed effects model will be used to study this association. The subject specific random effects will be used to account for subject-level (level 2) effects, in particular the effect of the intervention as well as daily level variations in PC.
Mixed effects regression will be used to study the association between the 1-month change in daily PC mean and the post-treatment opioid use, the outcome of interest. The regression will adjust for baseline opioid use and other confounding factors. Same analyses will be conducted with opioid misuse, pain intensity and pain interference as outcomes.