How group factors affect adolescent change talk and substance use outcomes: implications for motivational interviewing training.

Motivational interviewing (MI) is an empirically-based counseling approach in which a clinician uses a collaborative, non-confrontational, and non-judgmental style to resolve a client's ambivalence to changing their behavior (Miller & Rollnick, 2012; Rollnick, Miller, & Butler, 2008). Client change talk (statements arguing for change; Maybe I should stop using marijuana) and sustain talk (statements in opposition to change; I don't think I need to stop using marijuana) are often described as opposite sides of the client's ambivalence to change. Clinicians are encouraged to elicit and promote client change talk (CT) and reduce instances of sustain talk (ST) through the strategic use of open-ended questions, reflections, affirmations, and summaries (Miller & Rollnick, 2012). When clients express CT about the target behavior, it often indicates their readiness to change, whereas clients who express ST are often more ambivalent about change. In fact, client CT has been posited as an active ingredient of successful MI interventions (Baer et al., 2008; Barnett et al., 2014; Moyers, Martin, Houck, Christopher, & Tonigan, 2009). Most research studies examining CT and ST are limited to evaluating individual sessions. Several research studies with adults show that clients who express CT report reduced substance use between 12 and 34 months later (Bertholet, Faouzi, Gmel, Gaume, & Daeppen, 2010; Walker, Roffman, Stephens, Wakana, & Berghuis, 2006). In the few studies that have evaluated adolescent CT in individual sessions, higher frequency of youth CT remarks has been associated with fewer substance use days (Amrhein, Miller, Yahne, Palmer, & Fulcher, 2003; Baer et al., 2008) and drinks per week (Bertholet et al., 2010). Conversely, youth who communicate more ST remarks are more likely to report poorer drinking outcomes (Vader, Walters, Prabhu, Houck, & Field, 2010) and fewer days of abstinence at follow-up (Baer et al., 2008). Research examining individualized sessions (Amrhein et al., 2003; Baer et al., 2008; Gaume, Bertholet, Faouzi, Gmel, & Daeppen, 2013) has also shown that the subtype of CT may be important in predicting substance use outcomes. Subtypes of CT include statements indicating Desire (e.g., “I want to quit doing drugs”), Ability (e.g., “I can do it...this is doable”), Reasons (e.g., “I hate the way cigarettes smell”), Need (“I need to stop”), Commitment (e.g., “I stopped seeing him so I wouldn't smoke”), and Taking Steps (“This week, I won't go to any parties”). Each of these subtypes can also be a subtype of ST if it is expressed in opposition of change. For example, Desire ST would indicate not wanting to change (e.g., “I do not want to quit”) and Reason ST would offer reasons why one continues to use (e.g., “I like smoking because it relaxes me”). Dismantling which subtype of CT is associated with behavior change may help clinicians narrow in on which statements to reflect and emphasize during the busy group process. Research with adults who report AOD use has demonstrated that the Commitment CT subtype is particularly important in predicting AOD use outcomes (Amrhein et al., 2003). In one individual-based MI study with emerging adults, Desire, Ability, and Need subtypes of CT were associated with improved alcohol outcomes, whereas Desire, Ability, and Need subtypes of ST were significantly associated with poorer alcohol use outcomes (Gaume et al., 2013). Another study evaluating individual MI among homeless adolescents showed that youth who expressed Reasons in favor of changing AOD use had higher reductions in AOD use days at the 1-month follow-up, whereas youths’ expressions of Desire or Ability ST were predictive of fewer days abstinent at one and three-month follow-ups (Baer et al., 2008). In group interventions, CT and ST both by the individual and by other group members may have a unique influence on the group process. Groups typically include youth who vary in their readiness to change, severity of AOD use, problems, and willingness to speak up in a group of their peers (D'Amico, Osilla, & Hunter, 2010; Wagner & Ingersoll, 2012). This heterogeneity may either improve youth outcomes or lead to iatrogenic effects. Only two studies have evaluated CT and ST in group MI sessions. They found that group CT was associated with lower alcohol intentions, expectancies, past month drinking, and past month heavy drinking at 3-month follow-up (D'Amico et al., in press) and improved marijuana outcomes at 12-month follow-up (Engle, Macgowan, Wagner, & Amrhein, 2010). In contrast, group ST was related to decreased motivation to change, greater alcohol expectancies and marginal increases in marijuana use (D'Amico et al., in press). Only one study examined a subtype of CT or ST in adolescent group sessions. Engle and colleagues (2010) measured Commitment CT and found that this type of talk was associated with less frequent marijuana use, however, they did not examine other subtypes of change and sustain talk. The current study moves the field forward by examining whether all subtypes of CT and ST in the adolescent group setting influence AOD outcomes. Research on individual sessions has shown that specific session content such as decisional balance exercises can evoke CT when discussing the cons of using (LaBrie, Pedersen, Earleywine, & Olsen, 2006), and also may elicit ST when discussing the pros of using substances (Miller & Rose, 2013). To date, there is no research examining session content and group-level CT and ST. This is important to understand because group treatment is often used with youth (Kaminer, 2005) who typically receive multiple group sessions with various session content, often with rolling admission with youth attending different sessions. Ultimately, knowing the types of session content that may be positively associated with greater CT in the group setting could help clinicians better understand how to deliver more effective AOD groups and provide broader implications for how MI could be used when treating other target behaviors. We address these important questions in the current study by evaluating the change and sustain talk remarks of a 6-session group MI intervention trial, Free Talk, among at-risk youth who received a first-time AOD offense and were involved in the California Teen Court system (D'Amico, Hunter, Miles, Ewing, & Osilla, 2013; D'Amico et al., 2010). We first examined how session content was associated with CT and ST, and then we assessed whether different subtypes of CT/ST were associated with AOD outcomes three months later. Based on what we know from the available literature, we tested the hypotheses that session content explicitly focused on evoking CT (e.g., decisional balance; rulers) would be associated with more CT versus session content that utilized MI but was more focused on providing information (e.g., discussion of how AOD use affects the brain); we also expected that Reason and Commitment subtypes of CT/ST would be more strongly associated with AOD outcomes than would other subtypes of CT/ST.