Prompting to Practice: Daily Practice Mediates the Benefits of Supplement to a Mindfulness-Based Intervention

MBI, Mental Health, and Home Practice

There is a growing evidence base supporting the efficacy of MBIs aimed at adolescents and young adults for promoting mental health. Theoretically, MBIs are particularly effective because they dually support both top-down regulatory processes, such as attention allocation and emotion regulation, as well as bottom-up arousal processes, such as physiological stress responding (Davis & Hayes, 2011). In fact, meta-analytic evidence indicates that these interventions are effective for reducing problems with psychological distress (e.g., depression, anxiety, and stress) for adolescents and young adults (Dawson et al., 2020; Greeson et al., 2014; Miller et al., 1995; Zoogman et al., 2014). Although meta-analytic estimates of the effect sizes for MBIs in adult participants are moderate to large (Khoury et al., 2015), overall effects in younger participants are relatively small (Dawson et al., 2020; Zoogman et al., 2014).

Evidence evaluating the extent to which adolescent and young adult participants in an MBI adhere to home practice recommendations provides a possible explanation for these relatively weaker effects. A robust predictor of improvements from before to after an MBI is the amount of time that participants spend engaging in mindfulness practices between sessions (Goldberg et al., 2020; Lloyd et al., 2018), with both formal (i.e., time intentionally set aside for a longer practice) and informal (e.g., mindfully noticing sensations, thoughts, and feelings across the day) mindfulness practice being important (Birtwell et al., 2019). More specifically, there is evidence that practicing mindfulness increases levels of mindfulness, which then predict improvements in psychological functioning (Carmody & Baer, 2008).

Many MBI curricula encourage participants to establish a consistent mindfulness practice, in order to experience the benefits of mindfulness, and typically provide a specific set of instructions to participants for home practice between sessions (Broderick, 2013; Kabat-Zinn, 1990). However, MBI curricula vary dramatically in terms of what resources are actually provided to participants to support establishing a home practice (Lloyd et al., 2018), and many MBIs provide no structured resources beyond recordings and/or workbooks that participants can use (Broderick & Metz, 2009; Lloyd et al., 2018). In fact, many provide no structured resources, and simply recommend that participants practice at home between sessions. And, MBI participants report that they forget about home practice recommendations – both that they are supposed to practice and what they are supposed to practice (Lucas-Thompson et al., 2020) – between group sessions. Although adult MBI participants are relatively compliant with home practices recommendations (Brintz et al., 2024; Ribeiro et al., 2018), adolescent and young adult participants are not: some estimates suggest adolescent and young adult participants in MBI complete only about 25% of the recommended home practices (Quach et al., 2017). Therefore, efforts to increase home practice for adolescent and young adult participants in an MBI are needed, and if successful, should boost intervention effects.

The Role of Technological Supplements to MBI

Outside of the MBI literature, there is strong evidence for the benefits of just-in-time adaptive interventions as supplements to traditional behavioral interventions. Benefits are driven by delivering intervention content in daily life, tailoring it to individuals, and targeting moments of high need (Smyth & Heron, 2016). These intervention elements can be delivered via text message or smartphone applications, and in general are well-liked by participants (Heron & Smyth, 2010). Overall, there is evidence that they promote skill transfer from an intervention to daily life and, as a result, boost intervention effects (Heron & Smyth, 2010; Smyth & Heron, 2016), particularly if they are integrated with other intervention elements (e.g., a group-based intervention) rather than used as a stand-alone intervention (Heron & Smyth, 2010).

However, these adaptive interventions have not historically been incorporated into MBIs. To fill this gap, we developed the first multi-modal adaptive supplement to an MBI (Lucas-Thompson et al., 2020). This supplement was designed based on an MBI for adolescents, Learning to BREATHE (L2B; Broderick & Metz, 2009) which has been implemented successfully in populations that vary based on age, including in university students (e.g., Mahfouz et al., 2018), most of whom are young adults. Rooted in the philosophy of mindfulness-based stress reduction (Kabat-Zinn, 1990), L2B is well-liked by participants, and predicts reductions in symptoms of depression and stress in adolescent and young adult samples (Bai et al., 2020; Broderick & Metz, 2009; Dvořáková et al., 2017; Eva & Thayer, 2017; Felver et al., 2019; Frank et al., 2021; Fung et al., 2019; Tang et al., 2020).

We created the supplemented MBI, Learning to BREATHE PLUS (L2B PLUS; Lucas-Thompson et al., 2020) to provide support between sessions of L2B. One key goal of L2B PLUS is to promote the establishment of a consistent daily home practice (Lucas-Thompson et al., 2020). In particular, intervention content is delivered throughout the day to promote home practice (e.g., 5 short messages a day, each day other than the day of the intervention meeting). Messages were designed to support both formal and informal mindfulness practices throughout the day. As noted, participants in an MBI report forgetting about home practice between sessions (Lucas-Thompson et al., 2020).

In addition, the behavior change literature indicates that two critical and modifiable predictors of behavior change are motivation to change and self-efficacy for making change (Fisher et al., 2003). For instance, low levels of motivation and/or self-efficacy (i.e., beliefs about one’s own ability to engage in a behavior or meet a goal) are major roadblocks to engaging in behavior change (e.g., establishing a mindfulness practice). However, theoretically, the information-motivation-behavioral (IMB) model that has been widely applied to behavior change emphasizes that motivation and self-efficacy can be increased, and that doing so is critical for creating and also maintaining behavior change (J. D. Fisher & Fisher, 1992; W. A. Fisher et al., 2003); empirically, there is meta-analytic evidence that the IMB model is a good theoretical framework for developing behavioral interventions (Chang et al., 2014). Other theoretical accounts for the mechanisms underlying behavior change similarly emphasize the importance of motivation and self-efficacy. For instance, authors of a recent systematic review of behavior change literature concluded that approach motivation (i.e., towards positive and pleasurable experiences) is a critical factor that underlies individual behavior change, at least until the behavior change has been established (Michaelsen & Esch, 2023); similarly, the health belief model frames self-efficacy as a key cognitive construct underlying health-relevant behavior change (Orji et al., 2012). Theoretical accounts of behavior change in the context of mindfulness have similarly emphasized motivation and self-efficacy, and the importance of applying lessons learned from the broader behavior change literature to the practice of helping individuals to establish a mindfulness practice (Schuman-Olivier et al., 2020). In terms of self-efficacy, positively framed messages are most effective at increasing self-efficacy (Egbert & Omosun, 2017). Therefore, the categories of messages included in L2B PLUS are (1) reminders of what was learned in the last group session and what to practice, (2) messages to increase motivation to practice mindfulness, which emphasize the benefits of regular mindfulness practice, and (3) positively-framed messages to support self-efficacy by emphasizing that participants already have the skills they need to establish a regular mindfulness practice.

Importantly, an initial pilot randomized controlled trial in young adults (the same as utilized in the current study) suggests that, relative to the traditional L2B program, L2B PLUS predicts greater pre-post increases in mindful attention and smaller pre-post increases in mental health symptoms (Lucas-Thompson et al., 2023). Although this evidence is promising, it is empirically unclear what mechanisms underlie outcome differences between L2B PLUS and L2B. As consistent home practice is a theoretical mechanism underlying the benefits of MBI (Lloyd et al., 2018; Vettese et al., 2009), and L2B PLUS was designed to increase rates of home practice, the goal of the current study was to investigate the extent to which differences in home practice mediate these condition differences.

The Current Study

Home practice is an important predictor of improvements in outcomes during an MBI (Lloyd et al., 2018), but rates of compliance with home practice recommendations are very low (Quach et al., 2017). L2B PLUS (Lucas-Thompson et al., 2020) was designed to supplement a traditional group-based MBI (L2B) and, in part, promote home practice by providing intervention content to remind participants about home practice, motivate them to engage in home practice, and encourage self-efficacy for building a new mindfulness habit. L2B PLUS predicts greater benefits for young adult participants in terms of mindfulness and mental health (Lucas-Thompson et al., 2023), but it is unclear the extent to which differences in home practice mediate these condition differences. The goal of the current study was to fill this gap in the literature, by examining daily home practice as a mediator of L2B PLUS vs. L2B differences in mindful attention, psychological distress, depressive symptoms, and emotion regulation. The initial investigation of the benefits of L2B PLUS relative to L2B focused on pre- to post-intervention changes in key outcomes (Lucas-Thompson et al., 2023). In the current study, we focus on testing mediation in the context of intensive repeated measurements, as participants reported home practice, as well as mindful attention, psychological distress, depressive symptoms, and emotion regulation, each day of the intervention period in a daily diary. In addition, given the importance of consistency of home practice to theoretical benefits of mindfulness (Kabat-Zinn, 1990), we utilized these intensive repeated measurements to examine condition differences in consistency (versus erraticism) of home practice.

Method

Data Analytic Plan

We first conducted intercept only, unconditional random effects models to estimate intraclass correlations (ICCs) within individuals so that we could determine the amount of clustering present; all ICCs were well above .05 (Table 1). Therefore, ICCs indicated that multilevel modeling was necessary to account for the dependencies in the data (Hox, 1998); in addition, the use of multilevel modeling is supported by design effects in the current study, particularly an interest in level 2 effects (Lai & Kwok, 2015). Therefore, mixed effects models were conducted in Stata, with observations nested within individuals. Because of the repeated measurements, we first explored whether there were time trends in the variables that needed to be accounted for in the models using an augmented Dickey-Fuller test in the tseries package in R (Trapletti et al., 2019). Results indicated that the variables were stationary (ps<.001; the null hypothesis in these analyses is that there is a time trend, and the alternative hypothesis is that there is stationarity), so we conducted analyses across the entire intervention period.

OPEN IN VIEWER

Table 1.

Descriptive Statistics for Key Study Variables

​	M a	SD a	ICC
Daily mindfulness practice	2.01	1.96	.41
Mindful attention	3.76	1.12	.55
Psychological distress	2.55	.77	.34
Difficulties in emotion regulation	.79	.73	.52
Depressive symptoms	4.98	1.67	.35

Next, we tested for baseline equivalence in outcomes (mindful attention, psychological distress, and difficulties in emotion regulation) based on condition using ANOVAs; when there were baseline differences in these variables, we included them as covariates in our final models.

To test our primary hypotheses, we first estimated direct effects of condition (L2B PLUS vs. L2B alone) on the mediator (daily mindfulness practice) and each of the dependent variables (levels of mindful attention, psychological distress, difficulties with emotion regulation, and depressive symptoms) in mixed effects models in Stata 15. Then, we ran cross-level mediation models to estimate the indirect effects of condition on outcomes through daily mindfulness practice. In Krull & MacKinnon’s terminology (Krull & MacKinnon, 2001), we specifically conducted a 2→1→1 mediation model, as we estimated the indirect effects of a level 2 effect (condition) on level 1 variables (mindfulness, psychological distress, and depressive symptoms) through a level 1 variable (mindfulness practice). We used the ml_mediation package (Ender, 2014) in Stata 15 to estimate indirect effects. Because this research question was focused on condition differences (a time-invariant, between-person variable), the raw values of scores on the mediator and dependent variable across all assessments were utilized in these analyses (i.e., raw values were not decomposed into between-person and within-person elements). In addition, we ran heterogeneous mixed-effects location scale models (Hedeker et al., 2009) using MIXREGLS (Hedeker & Nordgren, 2013) to model not just random intercepts (mean associations, or ‘location’) but also random variances both between (heterogeneity) and within (erraticism) people (‘scale’).

Although we focused on L2B versus L2B PLUS differences, we conducted supplementary analyses to examine potential difference in practice and outcomes based on original condition (L2B, L2B + text messages, L2B + just-in-time support (JIT), and L2B + texts + JIT) using the same analytic strategy discussed above.

Results

Means and standard deviations of key variables are presented in Table 1. Fifty-nine participants contributed a total of 1250 daily diaries, with an average of 21.2 observations provided by each participant (range: 1–52). Across those 1250 days of daily diaries, 16% included reports of no home practice of mindfulness, 30% of one mindfulness practice, 25% of two mindfulness practices, and the remaining 29% of more than two mindfulness practices that day. There were no baseline differences in mindful attention (p = .43, M_L2B = 3.23, SD _L2B = 0.72; M_L2BPLUS = 3.04, SD _L2BPLUS = 0.72), psychological distress (p = .08, M_L2B = 32.25, SD _L2B = 3.99; M_L2BPLUS = 31.07, SD _L2BPLUS = 3.95), depressive symptoms (p = .47, M_L2B = 43.75, SD _L2B = 8.86; M_L2BPLUS = 45.16, SD _L2BPLUS = 7.31), or difficulties in emotion regulation (p = .55, M_L2B = 2.98, SD _L2B = 0.71; M_L2BPLUS = 2.82, SD _L2BPLUS = 0.63); therefore, we did not include baseline levels of these characteristics in our primary models.

There were significant differences in daily mindfulness practice across the intervention period between L2B PLUS versus L2B alone participants, such that L2B PLUS participants reported completing significantly more daily mindfulness practices than those in L2B alone (Figure 2, a paths). There were no direct effects of condition on mindful attention, psychological distress, difficulties in emotion regulation, or depressive symptoms (Figure 1, c’ paths). However, greater daily mindfulness practice was associated with significantly greater levels of mindful attention and significantly lower levels of psychological distress, difficulties in emotion regulation, and depressive symptoms (Figure 1, b paths). Finally, the indirect effects of condition on mindful attention, psychological distress, difficulties in emotion regulation, and depressive symptoms were all significant (Figure 1).OPEN IN VIEWER

In addition, there were condition differences in not just the levels of daily mindfulness practice, but also variances. Of note, there was significant variability in the consistency of daily mindfulness practice (b = .43, SE = .10, p < .001), and participants who had higher mean rates of practice were significantly more erratic in their rates of practice (b = .82, SE = .07, p < .001). Although there were no significant condition differences in consistency of mindfulness practice (b = .18, SE = .17, p = .28), these location scale models indicated that participants in the L2B PLUS group were significantly more heterogeneous (i.e., dissimilar from each other) in their rates of daily mindfulness practice than those in the L2B alone group (b = 1.40, SE = .43, p < .001).

Discussion

The goal of the current study is to investigate the extent to which, relative to the standard group L2B, L2B PLUS multi-modal supplement increases rates of home practice, which then promote greater mindfulness as well as lower psychological distress, depressive symptoms, and difficulties with emotion regulation difficulties. Results indicated that, compared to participants in L2B alone, participants in L2B PLUS practiced significantly more but were also more dissimilar from each other in rates of practice, and that there were significant indirect effects of condition on outcomes through rates of daily practice. These results provide initial support both for the benefits of L2B PLUS for home practice as well as the benefits of home practice for mindfulness, mental health, and emotion regulation in the context of an MBI.

In the same sample as utilized for the current study, there is evidence that L2B PLUS predicts greater pre-post benefits for mindfulness and mental health than L2B alone (Lucas-Thompson et al., 2023). L2B PLUS was designed to increase rates of home practice (Lucas-Thompson et al., 2020), in part because of arguments that home practice is a mechanism underlying MBI benefits (Lloyd et al., 2018; Vettese et al., 2009), and that the amount of time participants engage in mindfulness during sessions predicts gains in outcomes during an MBI (Goldberg et al., 2020). The results of the current study suggest that providing support in daily life via reminders, motivational messages, self-efficacy messages, and/or just-in-time support for practicing mindfulness during stress increases rates of home practice relative to the standard group program without these resources.

In addition, the results of the current study are consistent with arguments that home practice is one mechanism underlying the benefits of MBI (Lloyd et al., 2018; Vettese et al., 2009) as well as evidence for pre-post benefits of L2B PLUS relative to L2B alone for mindfulness, mental health (Lucas-Thompson et al., 2023), and mindfulness stress buffering (Lucas-Thompson et al., 2024). There were significant indirect effects of condition on mindful attention, psychological distress, depressive symptoms, and difficulties in emotion regulation through rates of reported daily practice. A systematic review indicates that there is marked variability in the resources that are provided to MBI participants, and many MBIs provide no or very few structured resources (Lloyd et al., 2018). It is likely not surprising, then, that rates of compliance with home practice recommendations in MBI are often very low (Quach et al., 2017). Providing intervention content and just-in-time support in daily life seems to benefit target outcomes of MBI (mindfulness, mental health, and emotion regulation) through increases in home practice by providing these supports; essentially, L2B PLUS is likely supporting skill transfer from the group program to daily life better than L2B alone. This evidence for the benefits of this multi-modal supplement is also in keeping with the larger evidence that just-in-time adaptive elements promote skill transfer and, as a result, boost intervention effects (Heron & Smyth, 2010; Smyth & Heron, 2016).

Interestingly, though, participants in the L2B PLUS condition were more heterogeneous (i.e., dissimilar from each other) in their rates of daily practice than were those in the L2B condition. There weren’t significant condition differences in erraticism, suggesting that L2B PLUS participants had higher but similarly inconsistent levels of practice as those in L2B alone. Together, these findings suggest that there is additional work to do to refine L2B PLUS to support a consistent mindfulness practice from day to day; it is also possible that rates of home practice are inconsistent, in part, because we measured practice during the early stages of habit formation. For instance, it can take several months to establish a new habit (Wood, 2019). It will be critical for future research to examine condition differences in consistent mindfulness practice over a longer period of time, during potential habit maintenance rather than formation. Also, the greater heterogeneity in the L2B PLUS vs. L2B alone condition suggests that there are unique subgroups of individuals who respond differently to the L2B PLUS intervention components, a possibility that should be explored in future research. A key benefit of just-in-time adaptive interventions is the tailoring of intervention content to individuals (Guan et al., 2024; Smyth & Heron, 2016); therefore, future research should aim to better understand the heterogeneity in mindfulness practice in this group to inform appropriate and effective tailoring of L2B PLUS to individuals. For instance, this pattern may reflect individual differences in responsiveness to the L2B PLUS intervention, differences which may be a function of genetic, physiological and/or psychological characteristics. The vantage sensitivity framework posits that differences in sensitivity to environmental stimuli may be at the root of individual differences in responses to psychological interventions (de Villiers et al., 2018). There is also evidence that dose-response relationships are moderated by the extent to which individuals engage with a health intervention (Lippke et al., 2016), and that both engagement in digital interventions and appraisals of intervention content are predicted by personality (Khwaja et al., 2021). Therefore, this greater heterogeneity in the L2B PLUS group could be caused by individual differences in responses to or engagement with the PLUS components, as well as individual differences in feelings about or reactions to the digital content that was delivered. An important step for future research is to evaluate the factors that underlie these group differences in heterogeneity in order to continue to refine the PLUS components.

An additional important area for future research to explore is the extent to which these differences in engagement and, therefore, outcomes are driven by particular components of the L2B PLUS intervention. Although not a focus of the current study, in part because of relatively small sample sizes in the individual PLUS sub-conditions, supplementary analyses did indicate that it was those in the L2B + JIT group who had significantly greater practice rates than those in L2B alone; there were no significant differences in practice between the other sub-conditions and L2B alone. However, the indirect effects from sub-condition to practice to outcomes were not significant. It is possible that this pattern reflects limited power to detect mediation given the small sub-condition cell sizes; it is also possible that it is the combination of JIT and intervention messages that has the strongest indirect effect from practice to outcome. This latter argument is in keeping with evidence that those randomized to the full PLUS program experienced greater decreases in mental health symptoms than those randomized to individual elements (Lucas-Thompson et al., 2023). However, that differences in practice relative to L2B alone were most evident in the group that received only JIT messages could reflect that participants benefit from the most from targeted support during stress, which is described by adolescents are being a primary underminer of mindfulness practice (Moran et al., 2026). In adolescent samples in particular, there is evidence that stressors predict lower levels of mindful attention (Lucas-Thompson et al., 2022, 2023, 2024). Adolescents in particular seem to struggle with mindfulness practice during stressful periods; thus, just-in-time support for practicing under stress should theoretically improve overall rates of home practice. However, those in the L2B + JIT group also received the fewest intervention prompts each day; therefore, it is also possible that this pattern suggests important information about an ideal daily dosage for PLUS intervention content, which other groups may exceed. Again, additional investigation in larger samples of the active ingredients of the PLUS program that is the most helpful for establishing a consistent daily practice routine will be important in future research.

Despite the practical and scientific contributions of the current findings, there are important limitations to note. First, the sample was relatively small, and lacked heterogeneity in race, ethnicity, and socioeconomic status; in addition, all participants were students at a university who self-enrolled in a mindfulness-based course described as targeting stress management. The latter characteristic of the current study likely reduces generalizability to adolescent populations who may have less intrinsic interest in mindfulness for stress management. It is important in future research to test differences in L2B PLUS and L2B in larger, more diverse samples, including in broader samples of adolescents, who have varying degrees of interest in or motivation to practice mindfulness. In addition, participants self-reported all primary outcomes for the current study. It is possible that part of the observed condition difference in home practice rates was due to demand characteristics, if L2B PLUS participants felt stronger pressure to report high rates of home practice than L2B participants. However, all participants received the same home practice messages at each group session of the intervention. In addition, rates of mindfulness practice were relatively low overall, suggesting that participants felt comfortable reporting low rates of practice in general. Nonetheless, the assessment of home practice was also a potential limitation of the current study, with a focus on overall number of mindfulness practices each day rather than other operationalizations, such as the minutes of time spent in practice, measuring practice separately by formal and informal mindfulness practice, and/or measuring the ways that people use mindfulness (Li et al., 2016). There is evidence that the amount of time that participants spend in both formal and informal mindfulness practice is important for overall program benefits (Birtwell et al., 2019; Goldberg et al., 2020), so it will be critical in future research to aim to replicate the findings in the current study using different and more nuanced measures of home practice.

In conclusion, the results of the current study add to the growing literature suggesting that the multi-modal supplement to L2B is an important extension to the traditional L2B, and that the delivery of intervention content and just-in-time support in daily life is translated into greater home practice, better gains in mindful attention, better mental health, fewer emotion regulation difficulties, and better abilities to remain mindful during stressful periods and thereby use mindfulness as a buffer against mental health challenges during stress (Lucas-Thompson et al., 2020, 2023, 2024). In addition, these results add to the literature about the role that home practice plays in MBI, and suggest that greater home practice is related to higher levels of mindful attention and lower levels of mental health problems and emotion regulation difficulties. Although it is important to replicate these results in larger, more diverse samples with more-nuanced assessments of home practice, these findings make an important contribution to the literature on the science and practice of MBI, and indicate that multi-modal supplements delivered in daily life will likely benefit program participants.