Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 14.
Published in final edited form as: Alcohol Clin Exp Res (Hoboken). 2024 Feb 27;48(3):556–566. doi: 10.1111/acer.15272

Understanding effects of the group process on drinking outcomes for emerging adults experiencing homelessness

Elizabeth J D’Amico 1, Jon M Houck 2, Eric R Pedersen 3, David J Klein 1, Anthony Rodriguez 4, Joan S Tucker 1
PMCID: PMC10939735  NIHMSID: NIHMS1967383  PMID: 38411540

Abstract

Background.

There is little research on group process for motivational interviewing-based group interventions with young people. We examine how participants’ change talk, group climate and cohesion, and facilitator empathy affect drinking outcomes among emerging adults experiencing homelessness.

Methods.

Data come from a clinical trial at three drop-in centers serving emerging adults experiencing homelessness in Los Angeles County and focus on those who received the intervention (n = 132). Participants completed baseline, 3, 6-, and 12-month follow up surveys. They were predominantly male and non-white. Group sessions were digitally recorded and coded for percentage change talk (PCT), group climate and cohesion, and facilitator empathy.

Results.

Baseline alcohol use was significantly higher at site 1, compared to sites 2 and 3, and thus we examined associations separately by site. At six months, higher PCT was associated with fewer drinks per drinking day for sites 2 and 3, whereas higher PCT was associated with more drinks per drinking day for site 1. There were no effects of PCT at 12 months. Higher group cohesion scores were associated with fewer drinking days at six months; higher facilitator empathy was associated with fewer maximum drinks in one day at both 6- and 12-months. Group climate was not associated with drinking outcomes.

Conclusions.

Findings highlight the importance of measuring multiple factors in the group process to understand outcomes. Both what is ‘uttered’ during group and what is observed provide different methods of evaluation of the group process and allow us to better bridge the gap between research and practice.

Keywords: emerging adults, homelessness, motivational interviewing, change talk, alcohol use

Introduction:

Emerging adults experiencing homelessness have higher rates of heavy drinking and alcohol use disorder than do emerging adults in the general population (Burke et al., 2022) and also are more likely to report engaging in risky sexual behavior, such as condomless sex and sex with multiple casual partners (Harris et al., 2017; Santa Maria et al., 2018). To date, however, very few programs have attempted to address both alcohol use and risky sexual behavior among this population. Programs that have addressed both behaviors often require numerous sessions over a long period of time (e.g., Rotheram-Borus et al., 2003), which can be difficult to maintain, especially among this transient and often difficult-to-engage population. Although several briefer interventions of 1–3 sessions with young people experiencing homelessness have been tested, any intervention effects are usually only found in the short-term. For example, a two-session program that included a smart phone app showed reductions in drinking and condomless sex at a two week-follow up among 60 participants (Thompson et al., 2020), whereas another two-session program with 61 participants found no significant effects on either substance use or sexual risk behavior after one month (Thompson et al., 2017).

Effective programming could be facilitated through drop-in centers, which offer services to address the basic needs of their clients such as food and hygiene, case management, and other programs to meet health and social service needs (Pedersen et al., 2016). For example, they conduct intake and crisis care, classes on parenting, safety and survival, health and wellness, and provide opportunities and case management for education, employment, and housing. Yet drop-in centers have limited resources and it may be difficult to retain participants over time. Thus, brief interventions need to be intensive enough to reduce risk behaviors in this population, but they also need to be feasible enough to be implemented in the often busy and resource-constrained settings such as drop-in centers where these young people routinely seek services (Slesnick et al., 2009). With this in mind, AWARE was developed to address alcohol use, other drug use, and risky sexual behavior in a brief four-session group format (Tucker et al., 2017). Two evaluations of AWARE conducted in drop-in centers found that this group motivational interviewing (MI) intervention was efficacious: a pilot trial that assessed outcomes over three months (Tucker et al., 2017) and a larger randomized controlled trial (RCT) that followed participants for one year (Tucker et al., 2023). In both studies, emerging adults ages 18–25 experiencing homelessness were randomized to either receive AWARE or usual care at the drop-in centers. Participants in the usual care control condition had access to the full range of programs and services already available at the drop-in center (these programs and services were available to intervention participants as well).

In the pilot, AWARE participants reported greater reductions over the 3 month period in drinking frequency in the past 3 months and likelihood of being a frequent heavy drinker in the past 30 days (Tucker et al., 2017). In the larger RCT, AWARE participants reported significant declines over the 12 month period in past month frequency of drinking, heavy drinking, number of drinks on drinking days, and alcohol consequences, as well as a significant increase in their use of drinking protective strategies (Tucker et al., 2023).

Given the efficacy of the AWARE intervention on drinking behavior among this group, it is important to better understand the group processes that may have contributed to these observed effects over time. More generally, an important question for clinicians who conduct group interventions with this population is understanding the types of communication dynamics (e.g., supportive or non-supportive language of others in the group) that occur in the group and how these dynamics may contribute to better outcomes. There is little research on process, structure, and clinician behavior for motivational interviewing-based group interventions with young people, particularly among those interventions addressing alcohol use (D’Amico et al., 2017). Though the literature base on individual speech coding in interventions is quite extensive (see meta-analysis by Magill et al., 2018), to date, there are only six published studies (five from our group) addressing the group process by coding the speech that occurs during the group and examining how this speech relates to behavior change. Briefly, all of these studies highlight the important role of change talk in the group, with greater change talk language (e.g., “I’m quitting for summer”) during the group session being associated with significant decreases in alcohol use (D’Amico et al., 2015; D’Amico et al., 2017; Engle et al., 2010; Houck et al., 2015; Ladd et al., 2016; Osilla et al., 2015). One study was part of the pilot trial of AWARE and showed that group change talk was associated with a lower likelihood of being a heavy drinker three months later, whereas group sustain talk was associated with decreased readiness and confidence to change alcohol use (D’Amico et al., 2017). Given the success of AWARE in reducing drinking behavior (Tucker et al., 2023; Tucker et al., 2017), the current study adds to this rapidly developing literature by examining not only the association of actual change talk during the group session with alcohol-related outcomes, but also by assessing the effects of group climate, group cohesion, and facilitator empathy on group member alcohol outcomes.

Cohesion is the degree to which group members evince bonding and solidarity. It is one of the most studied relationship constructs in group therapy. A 2018 meta-analysis of 55 studies from 2009 to 2016 showed that the correlation between cohesion and treatment outcome was statistically significant, and had an overall moderate effect size. Interestingly, the average group lasted 22 sessions, comprised participants with an average age of 36, and was guided by therapists using cognitive behavioral, psychodynamic, or humanistic approaches (Burlingame et al., 2018). Thus, there is a need to examine cohesion for younger participants and for briefer interventions that are based on MI. Group climate is typically defined in terms of engagement, conflict, and avoidance and indicates the extent to which group members are supportive and affirming of each other. It has also been associated with treatment outcomes. For example, several studies examining group climate with group cognitive behavioral therapy found that only higher levels of engagement over time (and not conflict or avoidance) were associated with reduced psychological distress at the end of the intervention (Arrow et al., 2021; Bonsaksen et al., 2011; Ryum et al., 2009). More work is needed in this area with larger samples and younger populations, and to date, neither group climate nor group cohesion have been examined in relation to outcomes for participants receiving group MI.

Building off the success of established coding systems for individual MI such as the MITI and MISC (Houck et al., 2010; Moyers et al., 2009), Wagner and Ingersoll developed the Assessment of Motivational Interviewing Groups-Observer Scales (AMIGOS) to assess group cohesion and climate in MI groups by identifying key constructs in the group therapy, person-centered, and MI theoretical and research literatures that fit the model of MI groups (Wagner & Ingersoll, 2012). They validated this measure by having three coders rate 18 groups utilizing various therapeutic approaches (Wagner & Ingersoll, 2018). Briefly, coders used the AMIGOS and two other measures, including the well-known Group Climate Questionnaire (MacKenzie, 1981), to code all sessions. Overall, the AMIGOS showed good scale reliability (.93–.95) and intra-class correlation coefficients indicated excellent inter-rater agreement in this sample (.82–.88). To date, however, there has been no assessment of how this measure of group cohesion and climate is associated with subsequent treatment outcomes. Thus, the current study uses this MI-developed group process measure to better understand how observed group cohesion and climate are associated with drinking outcomes at six and 12 months after the intervention among emerging adults experiencing homelessness.

Another posited mechanism of the group process in many different psychotherapeutic approaches, including MI, is empathy – the extent to which a clinician is interested in and attempts to understand the client’s perspective. Across therapeutic methods, clinician empathy is also one of the more robust predictors of intervention outcomes (Elliott et al., 2018; Flückiger et al., 2018; Norcross & Lambert, 2018). Within MI, higher levels of clinician empathy have consistently been related to better outcomes in individual therapy settings (Magill et al., 2018; Moyers, Houck, et al., 2016; Moyers & Miller, 2013). Although studies of group interventions have tested the effects of the therapeutic relationship (Johnson et al., 2005), empathy has only rarely been examined explicitly in group psychotherapy. In fact, a recent narrative review by Hogue and colleagues (Hogue et al., 2021) highlights that there is limited empirical evidence for the specific types of behaviors that may affect participant outcomes in the group setting. Thus, work is needed to better understand the factors that may contribute to iatrogenic versus positive outcomes among those who participate in group interventions, particularly over the long-term.

The current study therefore significantly adds to this literature by examining how the group process affects drinking outcomes among emerging adults experiencing homelessness. We focus on several different observer ratings: participants’ change talk, group climate and cohesion, and facilitator empathy. As noted earlier, we targeted drinking outcomes specifically in the current study because we found long-term effects of AWARE on drinking behavior (Tucker et al., 2023). Based on our previous group MI work (D’Amico et al., 2015; D’Amico et al., 2017) we hypothesized that a greater percentage of change talk in the group would be associated with better drinking outcomes at six and 12 months. The AMIGOS has not been used previously to examine the association of climate and cohesion ratings with group participant outcomes; however, we hypothesized based on previous literature in this area (e.g., Arrow et al., 2021; Burlingame et al., 2018) that higher ratings of group cohesion and climate would be associated with better drinking outcomes. Finally, given that higher levels of clinician empathy have consistently been related to better outcomes in individual therapy (Magill et al., 2018; Moyers, Houck, et al., 2016; Moyers & Miller, 2013), we hypothesized that higher ratings of empathy would be associated with better drinking outcomes among the group members.

Materials and Methods

Participants and setting

The study occurred at three drop-in centers serving emerging adults experiencing homelessness in Los Angeles County. All materials and procedures were approved by the RAND Internal Review Board. This study was preregistered with Clinical Trials, registration NCT03735784, and the study protocol has been published (Tucker et al., 2020). The drop-in centers are diverse in terms of geographic area (e.g., Hollywood, Venice/Santa Monica) and population served (e.g., one drop-in offers services specifically for sexual and gender minority (SGM) clients). Eligibility criteria for study participation included: (a) being between the ages of 18 to 25; (b) seeking services at one of the drop-in centers; (c) planning to be in the study area for the next month; (d) willing to provide name and information that would allow us to contact them again about completing follow-up surveys; (e) could be reached by e-mail or phone for follow-up; and (f) no evidence of cognitive impairment during the screening and consent process. AWARE was designed to be used as both prevention and intervention; as such there were no eligibility criteria based on substance use severity. Although participants were not initially screened for homelessness status, most (89%) reported at baseline that in the past 3 months they had spent at least one night at a shelter, transitional housing program, outdoors/street/park, car/other private vehicle, abandoned building, hotel/motel, or someone else’s apartment or house that was a temporary place for them to stay. The sample for the current study focuses on participants who were randomized to AWARE (n=132); they were predominantly male, heterosexual, and non-White (see Table 1), similar to the demographic profile of the population of young people experiencing homelessness in Los Angeles County (Los Angeles Homeless Services Authority, 2020). A detailed description of the AWARE intervention, including the recruitment plan and a summary of all content in the four sessions can be found elsewhere (Tucker et al., 2020).

Table 1.

Baseline descriptive information

Measure Site 1 (n=67) Site 2 (n=35) Site 3 (n=30) All (n=132)
Mean/n (SD/%) Mean/n (SD/%) Mean/n (SD/%) Mean/n (SD/%)
Age 22.25 (1.94) 21.69 (1.68) 22.67 (1.65) 22.2 (1.83)
Female 16 (23.88%) 8 (22.86%) 12 (40.00%) 36 (27.27%)
Heterosexual 53 (79.10%)* 11 (31.43%)* 20 (66.67%)* 84 (63.64%)
Race/Ethnicity
 Non-Hispanic White 10 (14.93%) 6 (17.14%) 6 (20.00%) 22 (16.67%)
 Non-Hispanic Black 26 (38.81%) 12 (34.29%) 9 (30.00%) 47 (35.61%)
 Multi-racial/Other 17 (25.37%) 4 (11.43%) 7 (23.33%) 28 (21.21%)
 Hispanic 14 (20.90%) 13 (37.14%) 8 (26.67%) 35 (26.52%)
Number of sessions attended 2.61 (1.314) 2.83 (1.543) 3.07 (1.856) 2.77 (1.511)
Size of groups attended 6.364 (1.066)* 7.914 (1.752)* 7.253 (1.436)* 6.977 (1.507)
Number of drinking days 7.43 (9.559)* 3.26 (4.494)* 5.43 (7.646) 5.87 (8.206)
Number of heavy drinking days 4.37 (7.232)* 1.26 (2.704)* 2.57 (4.446) 3.14 (5.867)
Number of drinks per drinking day^ 5.52 (4.83) 4.16 (3.42) 5.0 (5.70) 5.09 (4.77)
Maximum number of drinks in a day^ 8.91 (8.81) 6.51 (4.42) 9.17 (6.68) 8.42 (7.51)
Importance of cutting down drinking (0–10) 3.91 (4.751) 2.32 (3.282) 3.5 (5.277) 3.4 (4.565)
Open in a new tab

Note:

* =

significant difference between sites (p<.05). Unless otherwise noted, alcohol use items queried the prior 30 days.

^ =

abstinent participants treated as missing for this variable

Measures

Demographics

Demographics included age, sex assigned at birth (1=female, 0=not), race/ethnicity (Hispanic, non-Hispanic Black, non-Hispanic White, and Multiracial/other), sexual orientation (1=straight/heterosexual, 0=lesbian, gay, bisexual, or another sexual orientation). For race/ethnicity, participants were asked if they were Hispanic and then were asked which of the following groups best described them: Black, White, Asian, Native Hawaiian or Pacific Islander, American Indian or Alaskan Native, or Other. Participants could make multiple selections. For analysis, participants were categorized as Hispanic, non-Hispanic Black, non-Hispanic White, and Multiracial/Other.

Behavior coding

Coding process.

Trained, graduate-level research assistants familiar with coding group interventions coded for this study. All intervention sessions were complete before coding began. Biweekly coder meetings were held throughout the project to address progress and prevent coder drift. During meetings coders discussed difficult-to-code passages or session recordings encountered since the previous meeting. Discrepancies during coder meetings were resolved by consulting the appropriate coding manual and by input from an expert supervisor (J.M.H.). This was done to manage drift in future coding; coded session data were not changed. Intervention sessions were randomly selected for double-coding with the MISC 2.5 and AMIGOS for use in estimating inter-rater reliability.

Percentage of change talk.

Intervention sessions were evaluated using the CASAA Application for Coding Treatment Interactions (CACTI: (Glynn et al., 2012) coding software and the Motivational Interviewing Skill Code (MISC 2.5: (Houck et al., 2010). CACTI can be used with any mutually exclusive and exhaustive coding system, such as the MISC 2.5. Coders play digitally recorded sessions through CACTI and click on the appropriate button when they decide that an utterance corresponds to a particular code. CACTI continuously updates coding data for the session in a tab-delimited text file. These files can be combined easily into data files containing session-level behavior counts or session-level behavior sequences for analysis. The original source code for the free, open-source CACTI software can be found at https://github.com/jhouck/cacti-casaa. Executable versions are available at https://github.com/jling-NM/CACTI/releases. All MISC 2.5 codes were rated. The MISC summary variables of interest for this analysis were alcohol change talk (CT; client within-session statements supporting making/maintaining a positive behavior change or opposing a problematic health behavior) and sustain talk (ST; client within-session statements opposing a positive behavior change or supporting a problematic health behavior). These summary measures were computed using the alcohol-specific counts of client Desire, Ability, Reason, Need, Commitment, Other, and Taking Steps, both positive (i.e., CT) and negative (i.e., ST). We then used these summary measures to compute the summary measure of alcohol percent change talk (i.e., CT/(CT+ST)) for each session.

Group process.

The group session process was evaluated in two ways. First, we coded the group process using the Group Engagement subscale of the AMIGOS (Wagner & Ingersoll, 2018). Thus, all AMIGOS ratings are at the group level. We chose a priori to examine two ratings from this subscale: Climate (the extent to which group members are supportive and affirming of each other) and Cohesion (the degree to which group members evince bonding and solidarity), as we hypothesized these metrics would capture unique variance not included in constructs evaluated by the MISC 2.5 global ratings of MI integrity (Houck et al., 2010). Each AMIGOS Group Engagement rating is on a Likert-type scale ranging from 1 (low) to 5 (high); the instructions and rating form are available at https://motivationalinterviewing.org/sites/default/files/amigos_rating_form_v1.2.pdf. Second, we chose a priori to analyze the MISC 2.5 global scale, Empathy, as we hypothesized that this scale would not overlap with participants’ ratings of group climate and cohesion. Empathy focuses on the extent to which the clinician understands or makes an effort to grasp the client’s perspective and feelings (Moyers, Rowell, et al., 2016), and is rated on a Likert-type scale from 1 (low) to 5 (high).

Linking coding data to participants.

As in our prior work (D’Amico et al., 2015; D’Amico et al., 2017), participants were not identifiable in the session audio recordings. Participant behaviors observed during a given group session could be attributed to any participant who attended that session, and facilitator behaviors could influence each participant who attended that session. Therefore, ratings from each coded session were assigned to every individual who attended that particular group session. Participant-level mean ratings were then computed across the group sessions they attended for use in outcome analysis. For example, a participant who attended four group sessions that received percent change talk ratings of 50%, 73%, 45%, and 60% would receive a percent change talk score of 57%. Overall, there were 65 group sessions with 132 participants. The mean group size was 5.81 (SD = 2.40), with a range of 2–12.

Outcomes

Past month frequency and quantity of alcohol use, as well as frequency of cannabis and other illicit drug use, were the primary registered substance use outcomes for this trial (Tucker et al., 2020). This manuscript focuses on alcohol use outcomes given that the main study found effects on alcohol use at one year follow up (Tucker et al., 2023). In addition, given that AWARE specifically addressed both intentions to drink and frequency and quantity of drinking during the group sessions, we specifically focus on these outcomes as part of the group process.

Past month drinking.

Participants reported past 30 day frequency of drinking by indicating the number of days (0–30) they drank at least one full drink of alcohol and how many days they drank 5 or more drinks of alcohol in a row (i.e., within a couple of hours). For quantity of use, we assessed average number of drinks consumed on drinking days and the maximum number of drinks consumed on any day in the past 30 days.

Importance of cutting down on drinking

Importance of cutting down on drinking was a secondary outcome measure. This was assessed with a ruler modified from prior work (Boudreaux et al., 2012), asking how important it was to them to cut down or stop their use of alcohol. Those who did not use reported on their motivation to remain abstinent. Items were rated from 0=not at all to 10=extremely.

The intervention manual and deidentified data from this study will be available from the corresponding author on reasonable request one year after all aims of the project are completed.

Analysis plan

We estimated treatment site effects by using one-way analysis of variance for continuous measures, Fisher’s exact tests for binary characteristics, and chi-squared tests for categorical characteristics. For consistency with the study’s main outcome paper (Tucker et al., 2023), covariates in each outcome analysis included age, sex assigned at birth (female vs. not), race/ethnicity, sexual orientation, and treatment site, as well as the baseline measurement of the outcome variable. In terms of treatment site, preliminary analyses showed baseline differences in alcohol use across sites, with significantly greater drinking days and heavy drinking days at site 1 (7.43 and 4.37 days respectively) compared to site 2 (3.26, 1.26) and site 3 (5.43, 2.57) (see Table 1). We hypothesized that these differences in drinking behavior may be relevant to group processes and client within-session behavior. Therefore, rather than covary for site, we computed a new site variable comparing site 1 to the combination of the other two sites (site 1=1, sites 2/3=0). This new site covariate and its interaction with percentage change talk, Cohesion, Climate and Empathy were included in each respective outcome analysis. Examination of other baseline attributes (i.e., age, gender, sexual orientation, race) also revealed a significantly greater proportion of heterosexual participants at site 1 (79%) than at sites 2 (31%) and 3 (67%).

We examined long-term outcomes focusing on six and 12 months and models were estimated using full information maximum likelihood with the raw data which were not imputed or replaced. The proportion of the sample followed up at six months was 87%, 90.9% at 12 months, and 84.9% at both time points. All outcome measures were collected as integers. We used negative binomial regression for all count outcomes and ordinary least squares regression for other outcomes. Predictors of interest included MISC alcohol percent change talk, MISC Empathy, and AMIGOS Climate and Cohesion. Because the AMIGOS Climate and Cohesion scales were highly correlated, they are examined in separate models in our outcome analysis to allow us to test whether each AMIGOS group process measure provides additional predictive power beyond that of Empathy. Inter-rater reliability was estimated using the intraclass correlation coefficient (ICC model [3, 1], absolute agreement; (Shrout & Fleiss, 1979). Outcome analyses were conducted using MPlus v8.6, ICC analyses were conducted using SPSS v28.0.1.1.

Results

The sample included 63 sessions attended by a total of 132 participants aged 18–25 (M=22.2, SD=1.83). Participants were primarily members of racial/ethnic minorities (83.3%) and male (71.2%). Table 1 provides demographic information and baseline measures of all outcome variables. Table 2 provides follow up information for drinking outcomes.

Table 2.

Drinking outcomes at follow-up

Measure Site 1 Site 2 Site 3 All
Mean (SD) Mean (SD) Mean (SD) Mean (SD)
6 month
Number of drinking days 6.054 (9.728) 2.484 (4.441) 4.000 (5.793) 4.591 (7.819)
Number of heavy drinking days 3.464 (6.809) 1.323 (4.293) 1.607 (3.381) 2.435 (5.563)
Number of drinks per drinking day^ 4.110 (3.356) 2.760 (1.758) 3.530 (3.007) 3.600 (2.935)
Maximum number of drinks in a day^ 7.784 (8.456) 4.286 (2.795) 7.020 (6.476) 6.641 (6.933)
Importance of cutting down drinking (0–10) 5.732 (4.011) 5.633 (4.021) 4.893 (3.852) 5.500 (3.956)
12 month
Number of drinking days 4.729 (8.252) 3.313 (5.921) 3.862 (6.626) 4.142 (7.282)
Number of heavy drinking days 2.034 (4.874) 0.531 (0.842) 1.655 (3.298) 1.542 (3.837)
Number of drinks per drinking day^ 3.870 (2.617)* 2.180 (1.006)* 3.170 (2.684) 3.170 (2.348)
Maximum number of drinks in a day^ 7.230 (9.168) 3.730 (2.604) 4.780 (3.335) 5.520 (6.572)
Importance of cutting down drinking (0–10) 5.518 (3.991) 5.344 (3.907) 5.143 (3.885) 5.379 (3.911)
Open in a new tab

Note:

* =

significant difference between sites (p<.05). Unless otherwise noted, alcohol use items queried the prior 30 days.

^ =

abstinent participants treated as missing for this variable

Site effects

In addition to the between-site differences on sexual orientation and number of drinking days at baseline, as previously mentioned, we also observed differences on mean size of group sessions attended and coding measures (percentage change talk, reflection:question ratio, percentage MI-consistent, percentage open question, percentage complex reflection, Empathy). For example, the facilitator had a higher percentage of open ended questions at site 1 compared to sites 2 and 3, and a higher reflections to questions ratio at site 1 compared to site 3. (See Table 3 for complete results).

Table 3.

Coding measures (session-level)

Site 1 (n=32) Site 2 (n=16) Site 3 (n=17) Overall (N=65)
Measure Mean (SD) Mean (SD) Mean (SD) Mean (SD)
%CT 0.581 (0.278)* 0.445 (0.172)*  0.630 (0.205)  0.556 (0.246)
R:Q 1.154 (0.263)* 1.190 (0.374)  0.810 (0.15)*  1.085 (0.315)
%OQ 0.611 (0.062)* 0.493 (0.044)*  0.399 (0.057)*  0.532 (0.104)
%CR 0.084 (0.035)* 0.108 (0.039)*  0.088 (0.037)*  0.091 (0.038)
%MIC 0.997 (0.003)* 0.995 (0.005)*  0.998 (0.002)  0.997 (0.004)
Empathy 3.190 (0.535)* 3.190 (0.544)*  2.650 (0.493)*  3.050 (0.571)
Climate 3.030 (0.782) 3.500 (0.632)  3.180 (0.809)  3.180 (0.768)
Cohesion 3.280 (0.634) 3.630 (0.619)  3.240 (0.903)  3.350 (0.717)
Open in a new tab

Note:

* =

between-site difference p < .05; %CT = percentage change talk; R:Q = ratio of reflections to questions; %OQ = percentage open questions; %CR = percentage complex reflections; %MIC = percentage MI-consistent

Inter-rater reliability

The ICC for percent change talk was 0.753. The ICCs for the AMIGOS measures of Climate and Cohesion were 0.250 and 1.0. There was restriction of range in the Climate ratings, which for both coders ranged only between 3–4; coder ratings agreed exactly 40% of the time and differed by 1 point 60% of the time. This is similar to the range restriction of global ratings in other studies involving closely-supervised interventionists(Miller et al., 2005; Moyers, Houck, et al., 2016). Because our hypotheses involved Climate and because no previous study has examined the association of this measure with intervention outcomes, the variable was retained for analysis.

Effects of percentage change talk

Six-month outcomes.

The interaction between alcohol percentage change talk and site (PCT × Site; B=1.195, p<.001) was significantly associated with drinks per drinking day at follow up. Examination of the interaction effect indicated that for sites 2 and 3, a higher percentage of change talk was associated with fewer drinks per drinking day (quantity). In contrast, for site 1 a higher percentage of change talk was associated with more drinks per drinking day. At six-month follow up there were no significant associations involving alcohol percent change talk for number of drinking days, number of heavy drinking days, maximum number of drinks in one day, or the importance of cutting down on drinking (all p > .05). For covariate effects, as expected, a greater number of baseline drinks per drinking day (B=.037, p=.044) was associated with more drinks per drinking day at follow up, a greater number of baseline drinking days was associated with more baseline drinking days at follow-up, (B=.057, p=.003), a greater number of baseline heavy drinking days was associated with more heavy drinking days at follow-up (B=0.108, p=.007), and a greater maximum number of drinks in one day was associated with a greater number of maximum drinks in one day at follow-up (B=.056, p=.004). In addition, self-reported Hispanic identity was also positively associated with drinks per drinking day at follow up (B=.690, p=.005).

12-month outcomes.

At 12 months, there were no significant associations for alcohol percent change talk and drinking outcomes. Only the association between number of baseline drinking days and number of drinking days at follow-up (B=0.089, p<.001) was significant. No other associations between covariates and outcomes were significant (all p > .05).

Effects of Empathy with Cohesion in the model

Six-month outcomes.

For number of drinking days, higher Cohesion scores (B=−.515, p=.021) were associated with fewer drinking days at six-month follow up. Only one covariate was significant whereby a greater number of drinking days at baseline (B=.053, p=.002) was associated with more drinking days at follow up.

For drinks per drinking day, higher facilitator Empathy scores (B=−.676, p=.031) were associated with fewer drinks per drinking day at six-month follow up. For covariates, both identifying as Hispanic (B=.838, p=.002) and being at site 1 (B=.274, p=.021) were positively associated with drinks per drinking day at follow up.

For maximum drinks in a day, higher facilitator Empathy (B=−.845, p=.026) was associated with fewer maximum drinks in a day at follow up. The covariates baseline maximum drinks (B=.057, p=.002) and being at site 1 (B=.373, p=.006) were positively associated with maximum drinks in a day at follow up.

For number of heavy drinking days, only covariates were significant. Specifically, more heavy drinking days at baseline (B=.104, p=.016) were associated with more heavy drinking days at follow up. There were no significant predictors of the importance of cutting down drinking at six months.

12-month outcomes.

For number of drinking days, greater facilitator Empathy (B=−1.587, p=.002) was associated with fewer drinking days at follow up. The covariate baseline drinking days (B=.080, p<.001) was positively associated with drinking days at follow up. Number of drinks per drinking day was not significantly associated with Empathy or Cohesion at 12 months. For covariates, only sex (B=−0.567, p=0.04) was significantly associated with this outcome, with fewer drinks per drinking day for female participants.

For maximum number of drinks in one day, higher interventionist Empathy (B=−1.295, p<.001) was associated with lower maximum drinks at follow up. The covariate baseline maximum drinks (B=.042, p=.017) was positively associated with maximum drinks at follow-up.

Greater facilitator Empathy was associated with a lower number of heavy drinking days (B=−2.313, p=.003); however, Cohesion was not associated with number of heavy drinking days. Neither Empathy nor Cohesion were associated with importance of cutting down drinking at 12 months. For covariates, only being at Site 1 was associated with a greater number of heavy drinking days (B=0.682, p=.008).

Effects of Empathy with Climate in the model

Six-month outcomes.

Only the covariate baseline number of drinking days (B=.059, p=.003) was significantly associated with the number of drinking days at follow up. For the number of drinks per drinking day, higher facilitator Empathy (B=−.702, p=.026) was associated with fewer drinks per drinking day at follow up. The covariates being at site 1 (B=.300, p=.018) and identifying as Hispanic (B=.832, p=.002) were associated with a greater number of drinks per drinking day at six months.

For maximum number of drinks in day, higher facilitator Empathy (B=−.877, p=.022) was associated with lower maximum number of drinks in a day at follow up. In terms of covariates, greater baseline maximum drinks in a day (B=.059, p=.001) and being at site 1 (B=0.367, p=.012) were associated with greater maximum number of drinks per day at follow up.

Only the covariate baseline heavy drinking days (B=.116, p=.008) was associated with number of heavy drinking days at follow up. There were no significant predictors of the importance of cutting down drinking at six months.

12-month outcomes.

For the number of drinking days, higher facilitator Empathy (B=−1.636, p=.002) was associated with fewer drinking days at follow up. Climate was not significantly associated with number of drinking days. The covariate baseline drinking days (B=.079, p<.001) was associated with the number of drinking days at follow up.

For maximum number of drinks in one day, higher facilitator Empathy (B=−1.226, p<.001) was associated with lower maximum number of drinks in one day. Climate was not significantly associated with maximum number of drinks. The covariate baseline maximum drinks was significantly associated with maximum number of drinks in a day at 12 months (B=.040, p=.02).

Greater facilitator Empathy was associated with a lower number of heavy drinking days (B=−2.348, p=.003); however, Climate was not associated with heavy drinking days. For covariates, only being at Site 1 (B=0.719, p=.009) was associated with greater heavy drinking days. There were no significant predictors of the number of drinks per drinking day or the importance of cutting down drinking at 12 months.

Discussion

The current study is the first to examine the effects of the group process for emerging adults on drinking outcomes using several different measures, including actual change talk during the group sessions and observed measures of facilitator Empathy and observed measures of group Cohesion and Climate (versus group member ratings). This is also the first study to test the AMIGOS scale with clinical trial data to understand how both cohesion and climate are associated with long-term drinking outcomes among participants experiencing homelessness who received a group MI intervention. Briefly, effects of covariates (e.g., number of baseline maximum drinks in a day at baseline was associated with maximum number of drinks per day at six months) were as expected. However, it was important to control for these to understand effects of the group process on subsequent drinking above factors that we know are often correlated with outcomes.

Overall, we found that a higher percentage of change talk was associated with fewer drinks per drinking day at the six-month follow up, but only at sites 2 and 3. In contrast, for site 1, a higher percentage of change talk was associated with more drinks per drinking day. Notably, even though we found reductions in drinking for all participants at all three sites in the main outcome study, there were significant differences in alcohol use between these sites at baseline, with site 1 participants reporting a greater number of both drinking days and heavy drinking days than participants at sites 2 and 3. Furthermore, anecdotally, although MI integrity did not vary by site and the facilitator at site 1 had a higher percentage of open ended questions and reflections to question ratio compared to the other sites, the level of disorder within the group sessions at site 1 was much greater than at sites 2 and 3 as indicated by the digital recordings. As space was limited at all the drop-in centers but most specifically for site 1, the groups at site 1 were conducted in a room located in an area with considerable external noise. Perhaps at least partly due to this, participants at site 1 tended to be not as engaged in terms of listening and responding to the facilitator’s reflections and open ended questions during the group session. Participant discussion tended to be more “off task” than at the other two sites. For example, when the facilitator reflected on the participants’ discussion of consequences from use, one participant said if he had something to do the next day, he would not drink because he needed to be responsible. Immediately after, before the facilitator had time to respond, another participant then said that having sex would keep them from drinking, and the conversation got off track for several minutes. In addition, as the facilitator was reflecting, participants at site 1 would often talk over them, so it was difficult for group members to hear the reflection of group member change talk. These factors (environmental factors such as space and noise, but also individual personalities) may have contributed to the site differences we found in associations of change talk with drinking behavior. Further work is needed to understand the group process more generally, and also for this developmental age group.

In addition to the change talk during the group, we also assessed three different observer group measures of Empathy, Cohesion and Climate. Interestingly, these each explained unique variance in drinking outcomes. First, as found in individual sessions (e.g., Magill et al., 2018; Moyers, Houck, et al., 2016), higher facilitator Empathy scores were associated with fewer drinks per drinking day at the 6-month follow up, lower number of drinking days and lower number of heavy drinking days at the 12-month follow-up, and lower maximum drinks at both the 6- and 12-month follow ups in the model that included Climate scores. In addition, higher Empathy scores were associated with a lower number of drinks per drinking day at the 6-month follow up, lower maximum drinks at the 6- and 12-month follow up, and a lower number of both drinking days and heavy drinking days at the 12-month follow up in the model that included Cohesion scores. Furthermore, higher Cohesion scores in the group were associated with fewer drinking days at the 6-month follow up. These findings emphasize the important role of the facilitator’s skill in relating to the group and making an effort to understand participants’ perspectives and experiences. Results also suggest that group bonding and solidarity are important influences on group participants’ drinking behavior. Interestingly, a 2018 meta-analysis study showed that these two group processes may interact whereby group leader theoretical orientation moderated the association between cohesion scores and outcomes such that when the group leader had an interpersonal and supportive style, the association between cohesion and positive outcomes was higher (Burlingame et al., 2018).

Overall, findings highlight that for group MI, it is important to include both standard MI measures, such as percent change talk, and group process-specific measures, such as climate and cohesion, as these can help provide a more in-depth understanding of what is happening in the group session. This can be particularly important with populations who may be marginalized, like young people experiencing homelessness, as often they experience discrimination and feel shamed (Henriques et al., 2022; Toolis & Hammack, 2015), which may make them more wary of participating in groups addressing sensitive issues, such as substance use.

It is important to note some limitations of this work. First, participants were from drop-in centers in southern California, and therefore may not be representative of other young people experiencing homelessness across the nation. Second, as with other group work, we cannot link the specific participant who delivered the change talk with behavior change. Instead, each group session received an overall change talk score. Although this limits our ability to know which individual provided the change talk during the group session, as noted in our other group process research, it indicates that individual young people do not necessarily have to produce change talk themselves to benefit from change talk in the group and to make positive changes in their behavior (D’Amico et al., 2015; D’Amico et al., 2017). Future work might consider using personal microphones to better capture individual change talk (especially in the context of a chaotic group) and link this change talk to that individual’s own change behavior. Finally, there was restriction of range in the AMIGOS climate rating, limiting our power to detect its effects. Future studies involving a larger number of groups or recruiting interventionists across a broader range of skill may be necessary to evaluate more fully the influence of climate.

In sum, the current study moves the field forward by highlighting the importance of measuring multiple factors in the group process to understand intervention outcomes. Both what is ‘uttered’ and what is observed by rating the group process provide different methods of evaluation of the group process and allow us to better bridge the gap between research and practice. In the main study, drinking outcomes improved at the 12-month follow up for all participants, regardless of site (Tucker et al., 2023). However, our findings on group process tell a more in-depth story of how sessions “went” in terms of increasing behavior change, and overall feelings of solidarity and support. Researchers and clinicians often wonder-- what makes groups “work” (or not)--and what do we need to do to make the group setting more effective to help people make healthy choices and change risk behavior? Our results highlight the importance of change talk, as in other individual and group studies, and also show that group cohesion and facilitator empathy are important in predicting change for group members. All of these factors are related to the skill of the facilitator. For example, a meta-analysis showed that greater motivational interviewing skills are associated with greater client change talk in individual settings (Magill et al., 2018), and Miller and Rose highlight the importance of the facilitator as part of the relational hypothesis of MI’s efficacy which emphasizes a direct relationship between therapist style and client outcome (2009). Furthermore, group MI studies have indicated that increasing overall change talk in the group setting can lead to positive behavior change (D’Amico et al., 2015; D’Amico et al., 2017). Studies must continue to bridge the gap between research and practice to provide a clearer understanding of how the group process may affect outcomes for young people, particularly those who may be at higher risk for poor outcomes.

Supplementary Material

SUPPMATERIAL

Acknowledgments

This work was supported by grant R01AA025641 from the National Institute on Alcohol Abuse and Alcoholism (PI: Tucker). Trial registration: ClinicalTrials.gov Identifier: NCT03735784. https://clinicaltrials.gov/ct2/show/record/NCT03735784. The authors wish to thank the AWARE facilitators for intervention delivery, the RAND Survey Research Group for its assistance with data collection, the three drop-in centers for their support of this research, and the individuals who participated in the study.

Availability of data and material:

The intervention manual and deidentified data from this study will be available from the corresponding author on reasonable request one year after all aims of the project are completed. Requestors of data will be asked to complete a data-sharing agreement that provides for (1) a commitment to using the data only for research purposes and not to identify any individual participant; (2) a commitment to securing the data using appropriate computer technology; and (3) a commitment to destroying or returning the data after analyses are completed.

References

  1. Arrow K, Yap K, & Chester A (2021). Group climate in online group cognitive behaviour therapy predicts treatment outcomes. Clinical Psychologist, 25(2), 153–163. 10.1080/13284207.2020.1829944 [DOI] [Google Scholar]
  2. Bonsaksen T, Lerdal A, Borge F-M, Sexton H, & Hoffart A (2011). Group climate development in cognitive and interpersonal group therapy for social phobia. 15(1), 32–48. 10.1037/a0020257 [DOI] [Google Scholar]
  3. Boudreaux ED, Sullivan A, Abar B, Bernstein SL, Ginde AA, & Camargo CA (2012). Motivation rulers for smoking cessation: A prospective observational examination of construct and predictive validity. Addiction Science and Clinical Practice, 7, 1–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Burke CW, Firmin ES, & Wilens TE (2022). Systematic review: Rates of psychopathology, substance misuse, and neuropsychological dysfunction among transitional age youth experiencing homelessness. American Journal of Addiction, 31(6), 523–534. 10.1111/ajad.13340 [DOI] [PubMed] [Google Scholar]
  5. Burlingame GM, Theobald McClendon D, & Yang C (2018). Cohesion in group therapy: A meta-analysis. Psychotherapy, 55(4), 384–398. [DOI] [PubMed] [Google Scholar]
  6. D’Amico EJ, Houck JM, Hunter SB, Miles JNV, Osilla KC, & Ewing BA (2015). Group motivational interviewing for adolescents: Change talk and alcohol and marijuana outcomes. Journal of Consulting and Clinical Psychology, 83(1), 68–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. D’Amico EJ, Houck JM, Tucker JS, Ewing BA, & Pedersen ER (2017). Group motivational interviewing for homeless young adults: Associations of change talk with substance use and sexual risk behavior. Psychology of Addictive Behaviors, 31(6), 688–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Elliott R, Bohart AC, Watson JC, & Murphy D (2018). Therapist empathy and client outcome: An updated meta-analysis. Psychotherapy, 55(4), 399–410. 10.1037/pst0000175 [DOI] [PubMed] [Google Scholar]
  9. Engle B, Macgowan MJ, Wagner EF, & Amrhein P (2010). Markers of marijuana use outcomes within adolescent substance abuse group treatment. Research on Social Work Practice, 20, 271–282. [Google Scholar]
  10. Flückiger C, Del Re AC, Wampold BE, & Horvath AO (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psychotherapy, 55(4), 316–340. 10.1037/pst0000172 [DOI] [PubMed] [Google Scholar]
  11. Glynn LH, Hallgren KA, Houck JM, & Moyers TB (2012). CACTI: Free, Open-Source Software for the Sequential Coding of Behavioral Interactions. PLoS ONE, 7(7), e39740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Harris T, Rice E, Rhoades H, Winetrobe H, & Wenzel S (2017). Gender differences in the path from sexual victimization to HIV risk behavior among homeless youth. Journal of Child Sexual Abuse, 26(3), 334–351. 10.1080/10538712.2017.1287146 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Henriques E, Schmidt C, Pascoe R, Liss K, & Begun S (2022). Counter-narratives of structural oppressions, stigma and resistance, and reproductive and sexual health among youth experiencing homelessness. Qualitative Health Research, 10, 1447–1463. 10.1177/10497323221110694 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hogue A, Henderson CE, Ozechowski TJ, Becker SJ, & Coatsworth JD (2021). Can the group harm the individual? Reviewing potential iatrogenic effects of group treatment for adolescent substance use. Clinical Psychology: Science and Practice, 28(1), 40–51. [Google Scholar]
  15. Houck JM, Hunter SB, Benson JG, Cochrum LL, Rowell LN, & D’Amico EJ (2015). Temporal variation in facilitator and client behavior during group motivational interviewing sessions. Psychology of Addictive Behaviors, 29(4), 941–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Houck JM, Moyers TB, Miller WR, Glynn LH, & Hallgren KA (2010). Motivational Interviewing Skill Code (MISC) version 2.5. In. Retrieved from http://casaa.unm.edu/download/misc25.pdf.
  17. Johnson JE, Burlingame GM, Olsen JA, Davies DR, & Gleave RL (2005). Group climate, cohesion, alliance, and empathy in group psychotherapy: Multilevel structural equation models. Journal of Counseling Psychology, 52(3), 310–321. 10.1037/0022-0167.52.3.310 [DOI] [Google Scholar]
  18. Ladd BO, Tomlinson K, Myers MG, & Anderson KG (2016). Feasibility and reliability of a coding system to capture in-session group behavior in adolescents. Prevention Science, 17(1), 93–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Los Angeles Homeless Services Authority. (2020). 2020 greater Los Angeles youth count presentation.
  20. MacKenzie KR (1981). Measurement of group climate. International Journal of Group Psychotherapy, 31(3), 287–295. 10.1080/00207284.1981.11491708 [DOI] [PubMed] [Google Scholar]
  21. Magill M, Apodaca TR, Borsari B, Gaume J, Hoadley A, Gordon REF, Tonigan JS, & Moyers TB (2018). A meta-analysis of motivational interviewing process: Technical, relational, and conditional process models of change. Journal of Consulting and Clinical Psychology, 86(2), 140–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Miller WR, Moyers TB, Arciniega L, Ernst D, & Forcehimes A (2005). Training, supervision and quality monitoring of the COMBINE study behavioral interventions. Journal of Studies on Alcohol, 66(15), 188–195. http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2005-10269-026&site=ehost-live wrmiller@unm.edu [DOI] [PubMed] [Google Scholar]
  23. Miller WR, & Rose GS (2009). Toward a theory of motivational interviewing. American Psychologist, 64(6), 527–537. https://doi.org/2009-13007-002 [pii] 10.1037/a0016830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Moyers TB, Houck JM, Rice SL, Longabaugh R, & Miller WR (2016). Therapist empathy, combined behavioral intervention and alcohol outcomes in the COMBINE research project. Journal of Consulting and Clinical Psychology, 84(3), 221–229. 10.1037/ccp0000074 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Moyers TB, Martin K, Manuel JK, Miller WR, & Ernst D (2009). The Motivational Interviewing Treatment Integrity (MITI) Code. Version 3.1. http://casaa.unm.edu/download/miti.pdf
  26. Moyers TB, & Miller WR (2013). Is low therapist empathy toxic?. Psychology of Addictive Behaviors, 27(3), 878–884. 10.1037/a0030274 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Moyers TB, Rowell LN, Manuel JK, Ernst D, & Houck JM (2016). The Motivational Interviewing Treatment Integrity Code (MITI 4): Rationale, preliminary reliability and validity. Journal of Substance Abuse Treatment, 65, 36–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Norcross JC, & Lambert MJ (2018). Psychotherapy relationships that work. Psychotherapy, 55(4), 303–315. [DOI] [PubMed] [Google Scholar]
  29. Osilla KC, Ortiz JA, Miles JNV, Pedersen ER, Houck J, & D’Amico EJ (2015). How group factors affect adolescent change talk and substance use outcomes: Implications for motivational interviewing training. Journal of Counseling Psychology, 62(1), 79–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Pedersen ER, Tucker JS, & Kovalchik S (2016). Facilitators and barriers of drop-in center use among homeless youth. Journal of Adolescent Health, 59, 144–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Rotheram-Borus MJ, Song J, Gwadz M, Lee M, Van Rossem R, & Koopman C (2003). Reductions in HIV risk among runaway youth. Prevention Science, 4, 173–187. [DOI] [PubMed] [Google Scholar]
  32. Ryum T, Hagen R, Nordahl H, Vogel P, & Stiles T (2009). Perceived group climate as a predictor of long-term outcome in a randomized controlled trial of cognitive-behavioural group therapy for patients with comorbid psychiatric disorders. Behavioural and Cognitive Psychotherapy, 37(5), 497–510. 10.1017/S1352465809990208 [DOI] [PubMed] [Google Scholar]
  33. Santa Maria D, Padhye N, Yang Y, Gallardo K, & Businelle M (2018). Predicting sexual behaviors among homeless young adults: Ecological momentary assessment study. JMIR Public Health Surveillance, 4, e39. 10.2196/publichealth.9020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shrout PE, & Fleiss JL (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. [DOI] [PubMed] [Google Scholar]
  35. Slesnick N, Dashora P, Letcher A, Erdem G, & Serovich J (2009). A review of services and interventions for runaway and homeless youth: Moving forward. Child Youth Serv Rev, 31(7), 731–742. 10.1016/j.childyouth.2009.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Thompson RG, Aivadyan C, Stohl M, Aharonovich E, & Hasin DS (2020). Smartphone application plus brief motivational intervention reduces substance use and sexual risk behaviors among homeless young adults: Results from a randomized controlled trial. Psychology of Addictive Behaviors, 34(6), 641–649. 10.1037/adb0000570 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Thompson RG, Elliott JC, Hu MC, Aivadyan C, Aharonovich E, & Hasin DS (2017). Short-term effects of a brief intervention to reduce alcohol use and sexual risk among homeless young adults: Results from a randomized controlled trial. Addiction Research Theory, 25(1), 24–31. 10.1080/16066359.2016.1193165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Toolis EE, & Hammack PL (2015). The lived experience of homeless youth: A narrative approach. Qualitative Psychology, 2(1), 50–68. 10.1037/qup0000019 [DOI] [Google Scholar]
  39. Tucker JS, D’Amico EJ, Pedersen ER, Rodriguez A, & Garvey R (2020). Study protocol for a group-based motivational interviewing brief intervention to reduce substance use and sexual risk behavior among young adults experiencing homelessness. Addiction Science & Clinical Practice, 15, Article 26. 10.1186/s13722-020-00201-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Tucker JS, D’Amico EJ, Rodriguez A, Garvey R, Pedersen ER, & Klein DJ (2023). A randomized controlled trial of a brief motivational interviewing-based group intervention for emerging adults experiencing homelessness: 12-month effects on substance use and sexual risk behavior. Journal of Substance Use and Addiction Treatment, 152, 209114. https://doi.org/0.1016/j.josat.2023.209114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tucker JS, D’Amico EJ, Ewing BA, Miles JNV, & Pedersen ER (2017). A group-based motivational interviewing brief intervention to reduce substance use and sexual risk behavior among homeless young adults. Journal of Substance Abuse Treatment, 76, 20–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Wagner CC, & Ingersoll KS (2012). Motivational interviewing in groups. Guilford Press. [Google Scholar]
  43. Wagner CC, & Ingersoll KS (2018). Development and initial validation of the Assessment of Motivational Interviewing Groups–Observer Scales (AMIGOS). International Journal of Group Psychotherapy, 68, 69–79. 10.1080/00207284.2017.131558769 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SUPPMATERIAL
NIHMS1967383-supplement-SUPPMATERIAL.docx (84.5KB, docx)

Data Availability Statement

The intervention manual and deidentified data from this study will be available from the corresponding author on reasonable request one year after all aims of the project are completed. Requestors of data will be asked to complete a data-sharing agreement that provides for (1) a commitment to using the data only for research purposes and not to identify any individual participant; (2) a commitment to securing the data using appropriate computer technology; and (3) a commitment to destroying or returning the data after analyses are completed.

RESOURCES