Desire to get drunk partially mediates effects of a combined text message-based alcohol intervention for young adults

Excessive alcohol consumption among young adults continues to be a significant public health problem. According to the 2021 National Survey on Drug Use and Health, 29.2 % of young adults aged 18–25 reported binge drinking (4+/5+ drinks on the same occasion for females/males) on at least one day in the past month (Substance Abuse and Mental Health Services Administration, 2022). Digital behavioral interventions help overcome barriers to treatment initiation by providing remote and timely support (Suffoletto, 2016) and could serve as the first step in care to assist at-risk young adults in reducing excessive alcohol consumption. Text messaging is a near-ubiquitous communication modality (Smith, 2011) that has proven to be useful in reducing alcohol consumption in young adults (Bendtsen et al., 2021) but intervention effects are small and variable. Understanding optimal behavior change techniques (BCT) and mechanisms of behavior change (MoBC) could help inform development of more effective text message interventions.

Optimal BCTs for text message alcohol interventions remain largely unknown. A meta-analysis examining BCTs in 93 digital alcohol interventions (72 % single-session; 0 text message) found larger effects occurred when ‘prompting commitment and/or goal review’ and ‘provision of normative information or feedback on performance’, yet smaller effects when ‘information on the consequences of alcohol consumption’ were provided (Black et al., 2016). A systematic review of 41 digital interventions (83 % web; 0 text message) found that the BCTs effective in reducing alcohol consumption included ‘behavior substitution’ and ‘problem solving’ whereas ‘self-monitoring’, ‘goal-setting’, and ‘review of goals’ were also effective but rarely used (Garnett et al., 2018). A recent systematic review of 52 remote digital interventions for alcohol and drug use (6 text message) found that promising BCTs were ‘avoidance/reducing exposure to cues for behavior’, ‘pros and cons’ and ‘self-monitoring of behavior’ (Howlett et al., 2021). There are two key limitations to existing BCT literature: (1) No prior systematic review has focused strictly on digital alcohol interventions for young adults, which is important because they differ substantially from older adults in reasons for drinking alcohol, drinking patterns, and alcohol use severity (Quigley and Marlatt, 1996); and (2) No study to our knowledge had compared different digital BCT effects on alcohol consumption head-to-head, therefore relative comparisons of BCTs are confounded by differences in procedures between trials.

Key MoBC for text message alcohol interventions also remain largely unknown. This is not surprising given that MoBC for behavioral interventions more broadly are just starting to be experimentally identified and verified (Nielsen et al., 2018). For example, a recent systematic review of 15 studies that examined mediators of cognitive behavioral therapy for alcohol and substance use disorder (10 interventions; 3 digital) found moderate evidence for coping skills as a key mediator, with some support for self-efficacy and craving, however there was significant variability in methodology and the majority of models were conditional on some trait of the therapy or individual (Magill et al., 2020). Advancing understanding of MoBC for alcohol interventions requires experiments intentionally designed to (1) test directional path models that can establish temporal dependencies between cognitive and behavioral processes (i.e. thoughts->actions); and (2) directly compare different BCT bundles that are theorized to affect MoBC differently (e.g. dismantled intervention) (Witkiewitz et al., 2022).

We seek to advance understanding of BCTs and MoBC for text message-based alcohol behavioral interventions focused on young adults. In prior work, we iteratively developed and tested a text message-based alcohol behavioral intervention that uses a combination of BCTs (i.e. COMBO) to reduce alcohol consumption among young adults immediately (i.e., at 3-months) and distally (i.e., at 6- and 9-months) (Suffoletto et al., 2015). In attempt to understand which bundle of BCTs is essential to elicit change, we disaggregated COMBO into the following BCT bundles: (1) TRACK (self-monitoring of drinking plans, desire to get drunk, and drinking quantity); (2) PLAN (feedback on drinking plans and desire to get drunk); (3) USE (feedback on drinking quantity); and (4) GOAL (prompts to set a personal goal to limit drinking, tips to achieve the goal, and feedback on goal success). In a recently completed randomized trial, we found that PLAN, USE, GOAL, and COMBO arms all had similar reductions in alcohol consumption and alcohol-related negative consequences relative to TRACK at 3- and 6-months post-randomization (Suffoletto et al., 2022). MoBC for these effective BCT bundles remains unknown.

In this study, based on our trial findings, we tested a conceptual path model (see Fig. 1) where PLAN, USE, GOAL and COMBO, relative to TRACK, reduce binge drinking and drinks per drinking day both directly and indirectly through reduced desire to get drunk. We focus on desire to get drunk as a key MoBC because it represents a key aspect of craving, which, according to the motivational model (Cox and Klinger, 1988) and incentive salience model (Cofresí et al., 2019), is a key driver of alcohol use in general (MacKillop et al., 2010). Desire is unique in that it is a dynamic construct that fluctuates over time and by context. Therefore, to capture variations in desire states and alcohol use over time, a repeated sampling strategy with short time intervals is needed to adequately assess the interplay between desire states and alcohol use and to establish whether any intervention might have an effect on this interplay. In pilot work we collected event-level data on desire and alcohol consumption and multilevel structural equation methods (MSEM) to find supportive evidence that desire to get drunk was associated with binge drinking in young adults and that different bundles of BCTs may reduce binge drinking partially through altered desire (Suffoletto et al., 2020).

In this study, we expand upon this pilot work to test the hypothesis that the causal effect of text message interventions on reducing alcohol consumption operate indirectly through a mediator of desire to get drunk. We used experience sampling methods (Trull and Ebner-Priemer, 2009) where data was measured on the two days per week when a given participant typically drank alcohol during exposure to the intervention, which allowed us to overcome recall biases (Shiffman et al., 2008), identify within-person, state-dependent relationships (Kuerbis et al., 2013), and establish the temporal order of path nodes. We used MSEM, which allowed us to simultaneously assess and disaggregate both within- and between-person relationships (Preacher et al., 2010). Findings could help determine intervention-specific MoBC to inform best practices for the next iteration of text message intervention design to help young adults reduce excessive alcohol consumption.

留言 (0)

沒有登入
gif