A Feasibility Study of a Peer-Facilitated Physical Activity Intervention in Methadone Maintenance

Abstract

Purpose:

While methadone maintenance treatment (MMT) has been effective in improving opioid use outcomes, most patients continue to engage in unhealthy lifestyles that lead to significant mental and physical health consequences. Interventions targeting increases in physical activity (PA) in MMT patients could have a significant impact on reducing the overall morbidity in these individuals. The purpose of this study was to assess acceptability and feasibility of a 12-week peer-facilitated PA intervention for MMT patients called TREC (Transforming Recovery with Exercise and Community).

Method:

We developed and then pilot-tested TREC in 26 low-active MMT clients (73% female; mean age=41.2 years). TREC included: 1) an orientation session and intervention materials, 2) weekly PA discussion groups led by trained MMT clients, 3) peer-led walking groups and 4) a Fitbit activity tracker to facilitate self-monitoring of PA.

Results:

Participants attended 63% of eligible TREC sessions. Sixty-nine percent of the sample wore the Fitbit for at least 6 weeks (of the 12-week intervention). Participants reported that they enjoyed the group walks and that it was helpful to have a peer-facilitated PA group. There were small-to-moderate effect sizes for increases in PA, positive affect, and benefits of PA, and decreases in illicit opioid use and barriers to PA. No changes in depression, anxiety, and negative affect were observed from baseline to the end of the 12-week intervention.

Conclusion:

Indicators of feasibility and acceptability suggest that a peer-facilitated PA intervention can be incorporated in the context of MMT. Low active, opioid dependent clients showed increases in PA during the 12-week intervention. A future randomized clinical trial is necessary to determine the efficacy of TREC on long-term maintenance of PA and ancillary mental health and substance use outcomes.

INTRODUCTION

The most recent data from the 2018 National Survey on Drug Use and Health show that 2.1 million Americans met criteria for Opioid Use Disorder (OUD) in the last year (Substance Abuse and Mental Health Services Administration, 2019), reflecting more than a 50% increase in the last decade. The vast majority of those treated for OUD receive methadone maintenance treatment (MMT) (Substance Abuse and Mental Health Services Administration, 2019). Methadone is a long-acting synthetic opioid agonist that prevents opioid-related withdrawal and euphoria. While MMT is recognized as an effective treatment for individuals with OUD (Mattick et al., 2014), most of these patients continues to engage in unhealthy lifestyles (e.g., with low levels of physical activity) that lead to significant physical and mental health morbidities (Rosen et al., 2011). For example, patients in MMT have much higher rates of cardiovascular disease, diabetes, hypertension, obesity, depression, sleep difficulties, and cognitive impairments than age-matched controls, leading to premature death (Fareed et al., 2009; Maruyama et al., 2013; Rosen et al., 2008). Given the mental health, physical health, and drug use related benefits of physical activity (PA), interventions targeting increases in PA in patients receiving MMT could have a significant impact on reducing overall morbidity and mortality (Abrantes & Blevins, 2019).

In samples of patients receiving MMT, the majority (73%) are not meeting the public health recommendations for PA levels (Abrantes & Blevins, 2019; Beitel et al., 2016; Weinstock et al., 2012). For example, in a study of 713 MMT patients from 12 methadone clinics, low levels of PA were observed with most walking less than half a mile/day (Pieper et al., 2010). In our prior work, we found that among 305 MMT smokers, only 38% reported sufficient PA, while 25% were mostly sedentary reporting zero minutes of PA per week (Caviness et al., 2013). It is possible that patients in MMT experience important barriers to increasing PA (Caviness et al., 2013; Stein et al., 2013). For example, many are smokers, depressed, have chronic pain, and have financial struggles, all of which may affect adherence to PA and access to PA-related resources (Beitel et al., 2016). In addition, methadone can affect daily functioning as many patients experience daytime sleepiness (Wang et al., 2008). Therefore, PA interventions that can address these unique barriers will be necessary to effectively increase PA in this population.

Because most MMT patients go to a methadone clinic every day to receive a daily methadone dose, there is opportunity to intervene around PA in this context. Yet only one study has evaluated an exercise intervention with MMT patients (Cutter et al., 2014). In that study, 29 patients receiving MMT were randomized to daily 25-minute sessions of active Wii gaming or sedentary Wii gaming over the course of 8 weeks. Those in the active group reported greater minutes per week of moderate-to vigorous physical activity (MVPA) but no differences between groups on substance use and psychological outcomes were observed. An important limitation of this study was the enrollment of highly physically active participants (approx. 10 hours of MVPA per week at baseline). Thus, more research is necessary on approaches to engage physically inactive MMT patients who could benefit most from increases in PA.

Taking into consideration the socioecological model of behavior change (Sallis et al., 2008), multilevel interventions that target individual factors, interpersonal connections, and environmental context may be best suited to sustain changes in PA. Therefore, integrating a PA intervention into the MMT milieu could be a promising approach. However, relying on MMT staff and counselors to deliver PA interventions would be problematic; they have little time, large caseloads, and do not always see the same patients each month. As an alternative, an approach that is ideally suited for under-resourced settings is the use of peers to help deliver lifestyle interventions. Peer-facilitated interventions can be cost-effective and decrease the financial burden on a health care system (Colella & King, 2004). Peer facilitators have been described as individuals who share salient characteristics with a target population, allowing them to serve as empathic role models and provide support and encouragement (Hernandez et al., 2001). Peer-facilitated interventions are consistent with established theories of health behavior change (e.g., social-cognitive theory; Bandura, 2004) and self-determination theory (Ryan & Deci, 2000) that emphasize interpersonal relationships as a means for increasing self-efficacy and motivation for behavior change, often through vicarious experiences (e.g., modeling of the behavior) and verbal persuasion (e.g., encouragement). Also, peers can provide instrumental support, share information, help problem-solve barriers, and provide feedback.

In addition, there is a strong history of PA interventions being effectively supported by activity monitors. For example, a recent meta-analysis of 38 studies (n=29 with pedometers and n=9 with other fitness trackers) found that wearable activity trackers were associated with significant increases in PA, particularly in the context of face-to-face facilitators (Hodkinson et al., 2021). Indeed, PA-related self-monitoring and goal setting are two theoretically informed strategies associated with increased adherence and maintenance of PA (Michie et al., 2009). With advances in technology, PA interventions can now be supported by Bluetooth and sensor-enabled activity monitors (e.g., Fitbits). A recent review of 67 studies found Fitbit devices are consistently accurate for determining steps/day, though less so for moderate-to-vigorous PA (MVPA) (Feehan et al., 2018). While commercially popular, the efficacy of Fitbit for increasing PA remain to be determined. Preliminary evidence from many patient populations suggest that people wear the device, find it appealing, and show moderate effect sizes in increased steps/day (approximately 2,000–3,000 steps/day over the course of 3 month interventions) (Gal et al., 2018; Kirk et al., 2018). Many studies include “health coaching” and supportive messages (e.g., Amorim et al., 2019; Gell et al., 2017; Phillips et al., 2018; J. B. Wang et al., 2015). On the whole, activity monitors hold promise for supporting PA interventions by facilitating self-monitoring and goal-setting, which are essential for initiating and maintaining PA in the long-term.

The purpose of this study was to develop a peer-facilitated PA intervention that could be delivered in the context of methadone maintenance treatment. The intervention is called TREC (Transforming Recovery with Exercise and Community); we describe components below. Results from a 12-week open pilot trial with low-active MMT patients are presented. In addition to the acceptability and feasibility of TREC, we also report on preliminary findings in the form of effect sizes representing the change from baseline to end-of-treatment (EOT; i.e., end of the 12-week peer-led TREC intervention) on the primary target of PA as well as substance use and mental health outcomes.

METHOD

Peer-facilitator recruitment and training

Recruitment occurred at Stanley Street Treatment and Resources (SSTAR) in collaboration with methadone clinic staff. SSTAR is a health care and social service agency providing a range of mental health and substance abuse treatment services to individuals throughout the communities of Southern New England in the United States. SSTAR offers methadone maintenance treatment daily to approximately 800 patients. All study procedures were approved by the Institutional Review Board. Peer-facilitators eligibility criteria included: (a) engaging in ≥150 min/week of MVPA for the last 6 months, (b) stable methadone treatment for ≥6 months (per MMT counselor), (c) planning to remain on MMT for the next year or more, (d) willing to undergo 8-hour training, and (e) willing to commit to 1.5 hours/week (1 hour peer-facilitated PA session and 30 minutes of supervision) to the project for the next 6 months.

Between January and May 2017, 23 potential peer-facilitators (either referred by their MMT counselors or who approached study staff directly) were screened by study staff. Twelve of the 23 individuals screened met eligibility criteria. Of these 12 individuals, 7 promising candidates then met with one of the study’s principal investigators for a face-to-face meeting where they discussed the candidate’s current availability, prior related work experience (not required), and expectations and duties of the study’s peer-facilitators. Four peer-facilitators were selected from this group.

Once enrolled, peer-facilitators completed 8 hours of training over the course of two days with study investigators. The training covered the following: a) rationale for the proposed program, b) public health recommendations for PA and the different approaches to achieving this goal, c) how to share personal experiences with exercise and serve as a role model, d) how the Fitbit works and its use in the PA program, e) managing group dynamics (e.g., making sure everyone has a chance to speak), f) how to handle instances when participants are not meeting weekly PA goals (e.g., providing encouraging remarks, sharing own examples of how to get back on track), g) a review of the 12-session TREC manual, and h) a mock group where peer-facilitators role played leading a session. Three of the 4 peer-facilitators completed the training and participated as peer leaders in the open pilot trial. During the trial, the first and second author listened to audio recordings of the previous week’s session and then provided feedback to the peers at a weekly scheduled supervision session. Peer-facilitators were paid $50/wk for conducting group discussion sessions, walks, and attending supervision.

Open Pilot Participants and Recruitment

Recruitment strategies included: accepting referrals from methadone counselors, posting flyers in clinic waiting areas, presenting the study at clinic group meetings, and hosting tabling events inside the clinic where peer-facilitators and research staff answered questions about the study and interested clients signed up to be contacted by a research staff member. Interested MMT clients were screened either over the phone or in-person, and those appearing to meet study criteria signed a release of information to contact the participant’s primary care provider requesting medical clearance for the intervention. Once medically cleared, eligibility was confirmed with a more comprehensive baseline assessment conducted at the clinic. In addition, participants were asked to wear an accelerometer for 7 days to obtain a measure of baseline PA levels. After these baseline procedures and prior to beginning the 12-week TREC intervention, participants were scheduled to participate in a brief orientation to TREC with study staff where they also received a Fitbit activity tracker (see below for details). At the end of the intervention (i.e., 3 months from randomization), participants completed a follow-up assessment that included another period of 7-day accelerometry. Participants received $25 gift certificates for the baseline and $50 for each of the follow-up assessments.

Between June 2017 and May 2018, a total of 80 individuals receiving MMT at SSTAR were screened for study eligibility. Eligibility criteria included: (a) between 18 and 65 years of age, (b) receiving MMT at SSTAR and planning to remain in treatment for the next 6 months, (c) inactive or low active, [i.e., <90 min/week of moderate-to-vigorous-intensity PA (MVPA) for the past 3 months], and (d) access to a smartphone compatible with the Fitbit application. Exclusion criteria included: (a) history of psychotic disorder or current psychotic symptoms, (b) current suicidality or homicidality, (c) current mania, (d) current eating disorder, (e) physical or medical problems that would not allow safe participation in a program of MVPA (i.e., not medically cleared by SSTAR primary care physician), or (f) current pregnancy or intent to become pregnant during the next 12 weeks.

Fifty-four individuals did not meet eligibility criteria for the following reasons: too physically active (n=13), did not receive medical clearance from primary care physician to participate (n=12), planned to discontinue MMT at SSTAR in the next 6 months (n=7), had a history of psychotic disorder or current psychotic symptoms (n=2), or were pregnant (n=2). Additionally, 18 individuals lost interest or we were not reachable after being screened and prior to study enrollment. Specifically, n=7 participants were lost with no explanation (e.g., repeatedly did not attend scheduled baseline appointments or abruptly stopped responding to attempts to schedule), n=5 were lost during the process of waiting for their primary care physician to medically clear them to initiate the trial, n=4 were asked by their primary care physician to schedule a medical visit prior to the provision of medical clearance and the participant was not responsive to this request, n=1 experienced legal issues and no longer wanted to enroll in the study, and n=1 got a job during the enrollment process and no longer wanted to participate. This left a final sample of 26 participants who were fully eligible and enrolled in the study.

Intervention

The TREC intervention consisted of the following components: 1) orientation session and intervention materials; 2) Fitbit physical activity tracker, 3) 12 weekly, peer-led group discussion of PA goals for individual PA; and 4) 12 weekly, peer-led group walks near the SSTAR MMT clinic.

Orientation session and intervention materials.

Prior to beginning the peer-led component of TREC, each participant met with a study research staff member to be oriented to the details of participation (where and when to meet, provided with PA manual, overview of PA goals, etc.) and to receive an introduction to using the Fitbit. Study staff provided public health guidelines for PA and discussed with participants how they would engage in PA on their own, in addition to the TREC walking group. Guided by the theoretical principals of self-determination theory (Ryan et al., 1997), participants were encouraged to select activities they enjoyed, as this would likely result in greater long-term adherence. Given that exercise preference research in alcohol and drug using populations has identified walking as the most preferred activity (Abrantes et al., 2011; Stoutenberg et al., 2015), participants in this study were given an example of a graduated schedule for building up to 10,000 steps/day (Tudor-Locke & Bassett, 2004) over the course of the 12 weeks. In addition, if they preferred to track and self-monitor minutes/week of exercise, participants were also provided a graduated schedule for goals focused on minutes of moderate-intensity exercise per week, building up to 150 minutes/week over the course of the 12 weeks. Participants were provided with examples of various PA options for achieving their goal that included both bouts of moderate-intensity aerobic exercise and integrating PA into their daily lives (e.g., through chores, indoor walking, taking stairs instead of elevators, getting off 1 bus stop earlier, walking the dog, etc.). Participants were encouraged to achieve PA goals through either short bouts (e.g., for 10 minutes, 3x/day) or longer bouts (e.g., 30 minutes, 1x/day) of moderate-intensity activity as these have been found to be equally effective in increasing fitness and physical health outcomes (Murphy et al., 2002). In addition, each participant was given the American Heart Association endorsed “Walk at Home” DVD to help facilitate the attainment of PA goals when the weather was a barrier to engaging in PA. Additional resources included lists of local bike paths, parks, trails, and gyms as well as tip sheets for getting extra steps throughout the day (e.g., stepping during tv commercials or while on the phone).

Fitbit activity tracker.

Both peer-facilitators and participants received a wrist-worn Fitbit Alta^™. A de-identified, study-generated Fitbit account (on Fitbit.com) and password were created for each participant. With the participant’s permission, the investigators had access to the participant’s activity data throughout the course of the intervention. Participants were instructed to self-monitor PA by looking at the data displayed on the Fitbit device itself and by checking their Bluetooth-enabled smartphone/tablet Fitbit application. Peer-facilitators encouraged participants to use their Fitbits as a way of setting and self-monitoring weekly PA goals. Both peer-facilitators and study staff were available to assist participants in troubleshooting any Fitbit-related questions or issues. Participants were instructed to wear the Fitbit during waking hours and throughout the entire 12-week intervention.

Weekly peer-led discussion.

The peer-facilitated PA groups were co-facilitated by 2 MMT peers. The 12-week intervention involved participants attending a 1x/week peer-facilitated PA group session at SSTAR. Group sessions were scheduled to start at 12:30 PM; this time was selected because it immediately followed dosing hours. Each peer-facilitated group session lasted approximately 20–30 minutes and consisted of the following: 1) Introductions and group rules, 2) review of the SMART goal framework (i.e., making goals Specific, Measurable, Attainable, Relevant, and Time-bound), 3) discussion of the previous week’s PA including whether goals were met, 4) problem-solving any barriers to achieving PA goals, 5) setting PA goals for the upcoming week, and 6) troubleshooting any Fitbit related issues. Recruitment consisted of rolling admission; therefore, some participants could be starting as others were ending their participation. Participants were paid $10 in the form of a gift certificate to a local store for each of these sessions attended, with a $50 bonus at the end of the intervention for attending at least 9 out of 12 sessions. The selection of 12 weeks as the duration of the TREC intervention was consistent with other PA studies conducted in the context of addiction treatment (Thompson et al., 2020). The rationale for once weekly group discussion and walking sessions was to provide support and opportunity to be active while still allowing for individual participants to learn how to integrate PA as part of their daily lives the rest of the week. In doing so, PA may be more sustainable long-term. For a period of 3 months, the study offered 2 additional, optional group walks where peer facilitators were available. No participant attended these additional scheduled walks, and this offering was discontinued.

Peer-led walking group.

Immediately following the peer-led group discussion, the peer led group participants on a 30-minute walk around a park near to the MMT clinic. During times of inclement weather, a group room at the clinic was cleared and participants completed the Walk at Home DVD together, led by the peers who provided encouragement.

Measures

Substance Use:

The Timeline Follow Back (TLFB) (Sobell & Sobell, 1992) was administered at baseline and end of treatment to assess for frequency of alcohol, cannabis, cocaine, and illicit opioid use over the previous 90 days. The TLFB uses anchor dates to prompt participant recall (e.g., holidays, birthdates). With data from the TLFB, variables for the total number of days of use over the last 3 months were created for each of the 4 substances.

Self-Report Physical Activity:

Consistent with the literature describing exercise as a “vital sign” (Coleman et al., 2012), self-reported levels of PA were assessed by asking participants: 1) “Over the last 3 months, on average how many days/week did you exercise?” and 2) “On those days, on average how many minutes/day did you exercise?” Time spent exercising was calculated based on the multiplication of these 2 responses.

Objectively Measured Physical Activity:

At baseline and EOT, participants were instructed to wear an Actigraph GT3x accelerometer for 7 days. Steps/day were derived using the algorithm provided by the Actilife software. MVPA was estimated using the Freedson cutpoint of ≥1,952 counts/min (Freedson et al., 1998). Only days of 8 or more hours of Actigraph use, using Choi’s algorithm (Choi et al., 2011), were considered valid days and included in analyses. For the purposes of this study, step counts and MVPA were calculated for participants with at least 3 days of validated wear time data. In addition, average daily step counts were collected during the 12-week intervention objectively via the Fitbit. Only days which participants wore their Fitbit at least 8 hours were used for step count totals. An average daily step count was calculated for participants who had at least 6 weeks of Fitbit data (n =17).

Depressive Symptoms:

Depressive symptoms were assessed using the Center for Epidemiologic Studies Depression Scale Revised (CES-D; Radloff, 1977). The CES-D is a 20-item measure that assesses depressive symptoms on a 4-point scale (0–3). After reverse coding appropriate items, the depressive symptom score is obtained by summing the 20 items, with higher scores indicating higher levels of depressive symptoms.

Anxiety Symptoms:

The GAD-7 (Spitzer, Kroenke,Williams, & Lo, 2017)was utilized as a measure of generalized anxiety disorder symptomatology. The GAD-7 asks participants to rate how often over the last two weeks that they have experienced seven symptoms of anxiety on a scale of 0 (not at all) to 3 (nearly every day).

Positive and Negative Affect:

The Positive and Negative Affect Scale (PANAS; Watson, Clark, & Tellegen, 1988) was administered at baseline and EOT. The PANAS is comprised of two subscales, which correspond to positive affect and negative affect. Ten items comprise the negative affect scale and ten items are in the positive affect scale.

Usability and Acceptability:

Participants completed the 8-item Client Satisfaction Questionnaire (CSQ; Attkisson and Zwick, 1982) at EOT. This measure assessed level of satisfaction with the overall TREC intervention on a scale of 1–4, with higher numbers indicating greater satisfaction. Participants were also asked to rate their experience specifically with the Fitbit tracker using the 19-item Participant Experience Questionnaire of Wearable Activity Trackers (PEQ; Mercer et al., 2016), with items rated on a 5-point Likert scale from 1=strongly disagree to 5=strongly agree.

Beliefs about Exercise:

The Exercise Benefits and Barriers Scale (Sechrist et al., 1987) is a 43-item scale designed to assess an individual’s perceptions of exercise. They answer each item on a scale of 1 (strongly disagree) to 4 (strongly agree). The scale can be divided into a Benefits subscale and a Barriers subscale. The Benefits scale can range between 29 and 116, while the Barriers scale score can range between 14 and 56.

Physical Activity Enjoyment:

The Physical Activity Enjoyment Scale (PACES; Kendzierski and DeCarlo, 1991) is a self-report measure used to assess enjoyment of PA. The PACES includes 18 items in which respondents are asked to rate “how you feel at the moment about the physical activity you have been doing” using a 7-point rating scale (e.g., “It’s not at all stimulating/It’s very stimulating” or “I feel bored/I feel interested”); higher scores reflect greater levels of exercise enjoyment. The PACES has strong psychometric properties across various populations, and is the most widely used measure of exercise enjoyment (Kendzierski & DeCarlo, 1991; Mullen et al., 2011).

Adverse Events.

At EOT, adverse events and safety of the intervention was measured using the Systematic Assessment of Treatment-Emergent Events (SAFTEE; Levine & Schooler, 1984).

Data Analytic Plan

Descriptive statistics (mean and standard deviation for continuous variables and percentages for categorical variables) are presented for participant characteristics at baseline, including demographic information, PA levels, depression, anxiety, affect and substance use. Descriptive statistics are also used to report on participant adherence to treatment components (i.e., in-person sessions and Fitbit use) as well as participants’ perceptions of acceptability and feasibility. Changes in outcome variables of the intervention (PA outcomes, depression, anxiety, affect, and substance use) are then evaluated using paired samples t-tests. Given the preliminary nature of the study and small sample size, effect sizes (Leon et al., 2011) in the form of Cohen’s d (Cohen, 1988) along with 95% Confidence Intervals are presented.

RESULTS

Participant characteristics

Participants (n=26) were 73% female, mean age of 41.2 years (SD=11.0 years). The majority of participants identified as Caucasian (88%) and 8% as Black, 8% as Hispanic, and almost half of participants (42%) identified culturally as Portuguese, reflecting the demographic of the local area. See Table 1 for more details of baseline demographic characteristics. End-of-treatment (EOT) follow-up assessments were completed with 92.3% of participants.

Table 1.

Demographics

Demographic Characteristic	Mean (SD) or % of Sample
Gender (Female)	73%
Mean Age (Years)	41.2 (11.0)
Race
White	88%
Black	8%
Other	4%
Ethnicity
Hispanic	8%
Education
Completed College	0%
Some College	30.8%
High School	53.8%
8 years or less of school	15.4%
Employment Status
Full-time	3.8%
Part-time	7.7%
Unemployed	42.3%
On Disability	38.5%
Student	7.7%
Marital Status
Married	0%
Single, Never Married	42.3%
Divorced	15.4%
Separated	15.4%
Living Together	26.9%

Intervention Adherence

Of the enrolled participants, during the intervention, n=1 left the methadone clinic and n=4 experienced scheduling conflicts (e.g., new job) that resulted in the participant not being able to attend the TREC intervention sessions (i.e., peer-led group discussion and walks). All of these participants were encouraged to continue to wear the Fitbit and attend the EOT follow-up assessment. Of the remaining participants (n=21), the average number of intervention sessions attended was 7.5 (SD=3.6) out of a possible 12 sessions (equating to 63% of all sessions). The average number of valid days (8 hours or more) participants wore the Fitbit was 52.5 (SD=27.6) out of 84 possible days. Sixty-nine percent (69%) of the sample wore the Fitbit for at least 6 weeks, 57.7% wore it for at least 9 weeks, and 26.9% wore it the entire 12 weeks of the intervention.

A few factors contributed to participants’ level of adherence to intervention sessions. The most common reason for non-adherence was health-related issues (not associated with participating in PA) such as oral surgery, flu, sinus infections, etc., experienced by n=5 participants. The second most common reason (n=4) was substance use relapse. Other reasons reported by participants that impacted adherence included becoming homeless, significant personal relationship issues, court-mandated appearances, loss of transportation, and getting a new job. Further, several participants experienced disruptions in their smartphone cellular data plans, ranging from a few days to a few weeks, due to not being able to pay their phone bills. Though the Fitbit Alta was able to store daily-level data for several weeks when worn without syncing, there is the possibility some of this disruption in cellular data could have contributed toward missing data.

Baseline demographic (gender and age) and clinical (depression, anxiety, positive/negative affect, substance use) predictors of session attendance and Fitbit adherence were examined. None of the demographic and clinical characteristics predicted session attendance. Correlations representing medium effects were observed for higher levels of depressive symptoms (r=−.57), greater number of days of cocaine use (r=−.51), and greater number of days of cannabis use (r=−.41) at baseline were predictive of fewer days of Fitbit wear time during the 12-week intervention. Session attendance was also correlated with Fitbit wear duration (r=.54).

Satisfaction and acceptability

The mean score on the Client Satisfaction Questionnaire was 28 (SD=2.9), indicating high levels of satisfaction with the intervention. See Table 1 for responses to CSQ-8 items. We also inquired about specific aspects of the TREC intervention. Only one participant (4.3%) indicated that they did not find it helpful to have the TREC leaders be peers from the methadone clinic, whereas 17% found it moderately helpful and 73.9% found it very helpful. All participants reported satisfaction with the peer-led PA group discussion (8.3% somewhat satisfied, 20.8% moderately satisfied, and 70.8% very satisfied). Also, all participants indicated liking the weekly group walks at the nearby park (4.2% somewhat liked it, 20.8% moderately liked it, and 75% very much liked it). Further, participants had very positive experiences utilizing the Fitbit physical activity tracker; see Table 2. Lastly, while several adverse events were reported during the 12-week intervention, none was related to PA or participating in the TREC intervention. Overall, participants reported 5 adverse events: psychiatric inpatient hospitalization, opioid detoxification admission following a relapse, upper respiratory illness, bruised torso, and minor traumatic knee injury.

Table 2.

Client Satisfaction Questionnaire-8 item

CSQ-8 Question	Percent
1. How would you rate the quality of our program?
Excellent	50%
Good	45.8%
Fair	4.2%
Poor	0%
2. Did you get the program you wanted?
Yes, definitively	56.5%
Yes, generally	43.5%
No, not really	0%
No, definitively not	0%
3. To what extent has our program met your needs?
Almost all of my needs have been met	37.5%
Most of my needs have been met	58.3%
Only a few of my needs have been met	4.2%
None of my needs have been met	0%
4. If a friend were in need of similar help in increasing physical activity, would
Yes, definitively	70.8%
Yes, generally	25.0%
No, not really	4.2%
No, definitively not	0%
5. How satisfied are you with the amount of help you have received?
Very satisfied	66.7%
Mostly satisfied	25.0%
Indifferent or mildly dissatisfied	0%
Quite dissatisfied	8.3%
6. Has this program helped you increase your physical activity?
Yes, it helped a great deal	47.8%
Yes, it helped somewhat	47.8%
No, it really didn’t help	4.3%
No, it seemed to make things worse	0%
7. In an overall, general sense, how satisfied are you with this program?
Very satisfied	75%
Mostly satisfied	20.8%
Indifferent or mildly dissatisfied	0%
Quite dissatisfied	4.2%
8. If you were to seek help in increasing physical activity again, would you
Yes, definitely	91.%
Yes, generally	8.7%
No, not really	0%
No, definitively not	0%

Physical activity outcomes

Physical activity assessments included self-reported average minutes/week of exercise (assessed with the Exercise as a Vital Sign questions) over the last 3 months, 7-day accelerometry-derived steps/day and minutes of MVPA, and daily steps/day from the Fitbit during the 12 weeks of the intervention. See Table 3 for correlations between these assessment methods. While almost all the participants (n=25; 96%) wore the accelerometer at baseline, only 69% (n=18) provided valid accelerometry data (i.e., at least 3 days of 8 hours or more of wear time). At end of treatment, 62% of participants wore the accelerometer for some period of time and 46% (n=12) provided valid data.

Table 3.

Participant Experience Questionnaire (PEQ) of Wearable Activity Trackers (Range: 1=strongly disagree to 5=strongly agree)

PEQ Item	Mean (SD)
1. Overall, I was satisfied with the activity tracker.	4.38 (.71)
2. Using the activity tracker helped me set activity goals.	4.25 (.79)
3. Using the activity tracker helped me reach my activity goals more	4.08 (.83)
4. Using the activity tracker helped me to be more active.	4.04 (.96)
5. Using the activity tracker make it easier to be more active.	4.08 (.93)
6. Using the activity tracker supported my recoverya.	3.71 (.99)
7. I found it easy to operate the activity tracker.	4.38 (.77)
8. I found the activity tracker to be clear and understandable to use.	4.29 (.81)
9. I found the activity tracker to be flexible to work with.	4.29 (.75)
10. Overall, the activity tracker was easy to use.	4.42 (.72)
11. People who are important to me think I should use the activity tracker.	3.54 (1.02)
12. I have the technology necessary to use the activity tracker.	4.13 (.80)
13. I have the knowledge necessary to use the activity tracker.	4.25 (.79)
14. I am very knowledgeable about my physical activity needs.	4.25 (.68)
15. I understand how to use physical activity to manage my moodb.	4.13 (.68)
16. I understand how to use physical activity to manage my	4.13 (.74)
17. The activity tracker was comfortable to wear.	4.13 (.68)
18. The activity tracker accurately tracked my physical activity.	4.42 (.88)

^aThe original item was worded “Using the activity tracker supported me in managing my disease”

^bThe original item referred to managing “my health problems”.

A moderate effect (Cohen’s d=.60) was observed for self-reported minutes per week of exercise over the last 3-months from baseline to EOT. Among the small number of participants who provided valid 7-day accelerometry data for both baseline and EOT (n=11), effect sizes were relatively small for change in steps/day (Cohen’s d=.18) or MVPA (Cohen’s d=.23). See Table 5 for baseline and EOT means as well as 95% Confidence Intervals of these effects. Among participants who wore the Fitbit for any length of time (n=25), the average number of steps per day was 10,572 (SD=4409). Participants who wore the Fitbit for at least 6 weeks of the intervention period averaged higher steps/day (11,540, SD=5095) during the intervention relative to their baseline 7-day accelerometry steps/day (6222, SD=2926), Cohen’s d=1.2; a large effect with 95% CI=.64 to 1.85.

Table 5.

Changes in Other Physical Activity and Clinical Outcomes

Outcome	Baseline Mean (SD)	End of Treatment (EOT) Mean (SD)	Cohen’s d	95% Confidence Interval for Cohen’s d
Mood and Affect
Depressive Symptoms	22.5(11.8)	22.1(13.6)	−.06	−.46 to .34
Anxiety Symptoms	7.8(5.2)	8.9(5.5)	.18	−.22 to .58
Positive Affect	24.9(7.0)	30.0(10.0)	.63	.18 to 1.06
Negative Affect	20.4(8.9)	21.3(8.6)	.05	−.35 to .45
Substance Use (total days over last 3
Illicit Opioids	10.9(18.5)	3.7(9.7)	−.41	−.82 to .01
Cocaine	10.1(17.4)	4.7(9.7)	−.37	−.78 to .05
Cannabis	14.6(30.6)	9.6(24.9)	−.18	−.58 to .23
Alcohol	6.2(13.5)	3.5(9.2)	−.18	−.58 to .22
Physical Activity
Self-reported mins/week of exercise in the last 3 months	67(140)	243(387)	.60	.11 to 1.07
Steps/day (7-day Accelerometry)	7291(3352)	8230(6314)	.18	−.42 to .77
MVPA (7-day Accelerometry)	35.5(27.8)	47.7(60.5)	.23	−.37 to .83
Physical Activity Enjoyment	84.5(15.2)	96.8(23.3)	.46	.03 to .88
Perceived Barriers to Physical	14.5(5.1)	11.9(5.9)	−.50	−.92 to −.07
Perceived Benefits to Physical	58.4(8.6)	62.0(11.6)	.32	−.09 to .73

Changes in other physical activity and clinical outcomes

Table 4 presents effect sizes for changes from baseline to EOT on mood/affect, substance use, and other PA variables. While there were no changes in depressive symptoms, anxiety symptoms, and negative affect, participants increased in positive affect from baseline to EOT. Small-to-moderate effects were observed for decreases in illicit opioid use and cocaine use. Small-to-moderate effects were also observed for increases in PA enjoyment and perceived benefits of PA, as well as decreases in perceived barrier to engaging in PA.

Table 4.

Correlations between Physical Activity Assessment Methods

	Baseline 3-Month Min/Week (Self-Report)	Baseline 7-day Steps/Min (Acc)	Baseline 7-day MVPA (Acc)	EOT 3-Month Min/Week (Self-Report)	EOT 7-day Steps/Day (Acc)	EOT 7-day MVPA (Acc)
Baseline 7-day Steps/Day (Acc)	.77**
Baseline 7-day MVPA (Acc)	.59*	.89**
EOT 3-Month Min/Week (Self-Report)	.75***	.64**	.43
EOT 7-day Steps/Day (Acc)	.48	.56	.54	.19
EOT 7-day MVPA (Acc)	.35	.46	51	.08	.96***
Fitbit Mean Steps/Day During 12-Week Intervention	.39	.55*	.31	.49*	.40	.31

Note. Acc = Accelerometer; EOT = End-of-Treatment;

^*p<.05;

^**p<.01;

^***p<.001

DISCUSSION

This study reports on the acceptability and feasibility of a peer-facilitated PA program, TREC, in the context of methadone maintenance treatment for individuals with opioid use disorder. Despite less-than-optimal levels of compliance with Fitbit use and session attendance, participants who were enrolled in this study reported high levels of satisfaction with the 12-week TREC intervention, including peer-led group discussion, walking sessions, and the use of the Fitbit activity tracker. Preliminary findings point toward small-to-moderate effect sizes with large 95% confidence intervals for increases in PA (95% CI: .11 to 1.07), positive affect (95% CI: .18 to 1.06), benefits of PA (95% CI: −.09 to .73) and decreases in illicit opioid (95% CI: −.82 to .01) and cocaine use (95% CI: −.78 to .05) and barriers to PA (95% CI: −.92 to −.07). No changes in depression, anxiety, and negative affect were observed from baseline to the end of the 12-week intervention.

While adherence to the intervention protocol (63% of sessions) was lower than other exercise interventions (Stubbs et al., 2016), it is consistent with rates observed in other substance-using populations engaged in PA trials (Hallgren et al., 2017). Also, in the only other PA trial conducted with patients on methadone, participants similarly attended an average of 65% of exercise sessions (Cutter et al., 2014). In another example, Barry and colleagues (Barry et al., 2014) found less than 50% attendance at various behavioral treatments for pain management, one of which included a walking meditation intervention. As such, level of adherence may be less about the specific intervention but rather a reflection of the varied health and life concerns, at times chaotic and unpredictable, that may that disrupt clinical study participation (Stein et al., 2015).

Similarly, compliance with Fitbit use during the intervention was lower than general population samples (Cadmus-Bertram et al., 2015; Hartman et al., 2018; Lyons et al., 2017). We also observed that higher levels of depression and use of other substances (cocaine and cannabis) at baseline predicted lower durations of Fitbit use during the intervention period. Mean levels of CES-D depression scores in this sample were above the clinical cut-off for depression (i.e., a score of 16). Given the effect of depression on lower PA adherence (Batra et al., 2016; Jefferis et al., 2014), it may not be surprising that the Fitbit may have been underutilized in this vulnerable population. It is also important to note that participants largely had very positive feedback on the use of the Fitbit, suggesting it is worthwhile to continue to incorporate it as a self-monitoring and goal-setting tool in PA interventions with MMT patients. Future efforts to increase PA in the context of MMT may need to ensure that depression and other drug use are being concurrently addressed to optimize effectiveness of any intervention.

Most participants in this study found the peer-led discussions very helpful, including having the leader be a peer from the methadone program. While peer-facilitated interventions have been widely used to effectively address many self-management behaviors (e.g., medication adherence; nutrition), the involvement of peers to target increasing PA has been less studied and considered to be an “overlooked opportunity” by some researchers – particularly given the impressive within-subject increases in PA in existing peer-led intervention studies (Ginis et al., 2013). Further, when compared to professionally delivered PA interventions, peer-facilitated PA interventions were found to be just as effective (Ginis et al., 2013) and lead to greater long-term maintenance of PA (Buman et al., 2011). Peer-facilitated interventions have also been found to be cost-effective (Colella & King, 2004) and therefore, could play an important role in decreasing the financial burden on the under-resourced setting of MMT. Lastly, by a peer-facilitated model allows for sustainability of the intervention, as participants in the intervention could, in turn, transition to the role of peer upon successful completion of the program.

Moderate effect sizes were observed for self-reported PA increases from baseline to the end of the 12-week intervention. However, due to the compliance issues with the Fitbit and accelerometer, objective changes in PA were more challenging to determine. Less than half of participants provided useable accelerometry data at the EOT assessment. Though not ideal, given the different methodological sources, a large effect for increased steps/day was observed when comparing baseline accelerometry and average Fitbit steps counts during the intervention. Therefore, overall, it is likely participants made increases in their PA levels during the intervention. Importantly, increases in PA enjoyment and perceived benefits of exercise and decreases in perceived barriers to exercise were observed. In addition, given the high rates of time spent sitting in this population (Stein et al., 2013) and the unique and independent contribution of sedentary behaviors on poor physical health (Lavie et al., 2019), future studies should employ rigorous methods of objectively measuring time spent sedentary.

Surprisingly, changes in depression, anxiety, and negative affect were not observed in this study. The effect of increased PA on these outcomes has been well documented in the literature(Wegner et al., 2014). It is possible that participants did not sufficiently increase PA to produce an effect on these mental health outcomes. It is also possible that the high rate of psychosocial stressors in MMT patients, including low income, homelessness, transportation problems, accidents, physical assault or fights, and incarceration (Hayaki et al., 2005) could have outweighed the effect of PA on mood and negative affect. Encouragingly and consistent with prior exercise research, increases in positive affect were observed here as well. Lastly, small to moderate effects were observed for decreases in cocaine and illicit opioid use. Based on prior research (e.g., Dongshi Wang, Zhou, Zhao, Wu, & Chang, 2016), it is possible bouts of PA may decrease drug cravings and, in turn, other drug use.

Based on the experience of conducting this feasibility study, the emergence of the role of social determinants of health (i.e., social, behavioral, and environmental factors known to impact health disparities and outcomes (Northwood et al., 2018)) on the ability to engage with the various intervention features became clear. In order for future iterations of this intervention to be effective, supporting participants by addressing these factors will be critical. For example, in addition to financially incentivizing wear of Fitbits and accelerometers, participants may also need to receive help paying their cellular data plans. Orientation materials could incorporate information for accessing publicly available Wi-Fi hotspots as well as community resources for necessary psychosocial help (e.g., family therapy, free legal services). Technology-supported approaches such as videoconferencing to conduct peer-led discussions may make the intervention more accessible to participants. It may also be important to lengthen the interventions to provide participants more time to re-engage if they experience life disruptions (e.g., relapse; relationship issues). Despite these challenges, this is a population that is indeed interested in increasing PA and, due to the prevalence of mental and physical health comorbidities, could greatly benefit from effectively doing so.

There are several limitations that merit discussion. First, given the challenges associated with compliance and adherence to intervention components, it is important to interpret the Cohen’s d effective sizes with some caution. Missing data can contribute to large ranges of effect, as demonstrated by the 95% confidence intervals listed in Table 4. Second, this is a small sample of patients in one methadone clinic in a geographical location that consisted of primarily White individuals. Aspects of the intervention may have to be modified if implemented in other MMT clinics located in areas with different demographics and built environments. In addition, the majority of the sample (73%) was female, which is much higher than the broader MMT population (approximately 30% (Guerrero et al., 2021)), suggesting this intervention may be less appealing to men receiving MMT. Third, as mentioned above, objective measures of PA were challenging to obtain. Future research should consider appropriate incentives (e.g., gift cards to local supermarkets) and population-specific strategies for increasing compliance to device use and wear time. For example, sending reminder text messages or having clinic staff provide encouragement may be helpful. Fourth, this was intended to be a study of feasibility and acceptability, and there was no control condition. As such, it is not possible to infer any causality of the intervention on clinical and substance use outcomes. Similarly, we did not examine mechanisms of intervention effects. Future studies with larger samples may consider assessing and testing the role of social support and increased self-efficacy as mediators of intervention effects on PA outcomes.

Despite these limitations, this study introduces a novel approach for increasing PA in the context of MMT for patients with opioid use disorder. This multilevel intervention was found to be acceptable and feasible with some preliminary evidence of impacting changes in PA, positive affect, and drug use. As such, an important next step will be to conduct a randomized controlled trial to determine the efficacy of this intervention on PA and relevant outcomes. If this future study could be designed to isolate the effects of various intervention components (e.g., Fitbit alone vs with peer-facilitation), it would help determine the most effective intervention features. A successful PA intervention that is delivered by peers in methadone clinics could improve the overall quality of life, mental health symptoms, substance use, and health outcomes of MMT patients.

Highlights.

• Patients with opioid use disorder engaged in methadone maintenance treatment (MMT) may benefit from increases in physical activity.

• Peer-facilitation of physical activity may be sustainable strategy in the context of MMT.

• Participants were very satisfied with their participation in the peer-facilitated TREC intervention.

• Increases in physical activity and indicators of acceptability and feasibility were observed.