Mindfulness, defined as bringing moment-to-moment awareness of one’s thoughts, emotions, and bodily sensations in a non-evaluative and accepting way, was introduced as a secular practice in mainstream settings in the late 1970’s when Jon Kabat-Zinn developed the mindfulness-based stress reduction (MBSR) program to help relieve chronic pain symptoms (1, 2). Mindfulness-based cognitive therapy (MBCT), an adaptation which includes elements of cognitive therapy, was created in the early 2000’s as a treatment for depression (3). MBSR/MBCT are mindfulness programs traditionally formatted as eight-week courses with weekly group sessions guided by a trained mindfulness instructor (1, 3). Some of the strongest evidence for these mindfulness-based programs (MBPs) is in the context of recurrent depression (4, 5) and chronic pain (6), but there is also scope for investigating several adaptations of these programs in wider populations.
A population-based strategy can be adopted to investigate the effects of MBPs in the general population. According to this approach, most conditions fall on a continuum in that there are few ‘highs’ and ‘lows’ and many values that fall somewhere in between (7). In the context of mental well-being, you may find that a smaller proportion of the population experiences a ‘high’ well-being state and a ‘low’ well-being state and that a larger proportion experiences a range of different states between the two ends of this distribution (see Figure 1). Melzer et al. (2002) (8) examined this approach in the context of common mental health conditions and found that symptoms were widely distributed and followed a continuous model. To date, the majority of mindfulness research seems to pertain to the effects of MBPs on these ‘lows’ or ‘highs’ which dismisses the proportion of the population that may be at a moderately elevated risk. It is important that future research examines the effects of MBPs across the entire population distribution of mental well-being to understand the extent to which these programs can produce entire population shifts in therapeutic outcomes. This work will also pave the way to better understand how community wide factors (e.g. mindfulness experience, history of anxiety or depression) within the general population may impact the direction and magnitude of these potential population shifts.
Research that has investigated the effects of MBPs across the entire general population is currently underexplored. In a previous meta-analysis conducted by Khoury and colleagues (2015) (9), the overall effects of included studies (n = 29) demonstrated large improvements on stress and moderate improvements on anxiety, depression, and quality of life in pre-post analyses of MBSR in the context of ‘healthy individuals,’ yet only four studies specifically indicated general population samples. Other reviews (10-12) have included non-clinical samples but it is unclear how they operationalize this definition in regards to symptom count within the general population. Without clear definitions of participant samples, it is difficult to know where on the population distribution curve of specific outcomes (e.g. well-being, stress, anxiety, depression, etc.) MBPs demonstrate effects. Moreover, without this information, we cannot fully understand the how, why, and for whom these programs work best. The authors of this review have defined the “why” as the mediators and mechanisms through which change is produced and the “how” (e.g. sustained mindfulness practice) as the explanatory medium through which these mechanisms are targeted. The “for whom” is understood as the moderators (e.g. individual characteristics) that impact both how and why the treatment works.
Figure 1 illustrates a conceptual example of a population distribution of mental well-being pre-post a MBP in the general population. This figure is inspired by Geoffrey Rose’s population-based strategy (7) and the findings of Melzer et al. (2002) (8). The authors of this review define the general population as the full distribution curve rather than the largest proportion of the population (e.g. the blue shaded region). The general population may include a smaller proportion of ‘lows’ and ‘highs’ in terms of individuals that have a low mental well-being and high mental well-being. The largest proportion of the general population may include individuals with state values that range between these highs and lows. It should be noted that this figure is not based on actual data and is only used to help describe a more nuanced framework of this approach, which recognises that mental well-being may fall on a continuum and that, through increased mindfulness training, individuals may experience increased well-being.
To uncover how and why specific treatments such as MBPs have an effect on a particular outcome, Kazdin (13) argues that one must elucidate the mechanisms of change. These variables explain how a treatment (T) impacts change in a specific outcome (O) by illuminating the ‘interim processes’ between the treatment and the outcome (TàO). The ultimate goal with uncovering a mechanism of change is therefore to understand this relationship (TàMechanismsàO) from ‘beginning to end’ (13). There are methodological requirements put forth that help elaborate a mechanism of change. These requirements involve establishing the mediator (M), which demonstrates the relationship (TàO) statistically and is therefore an intermediate step in establishing a causal relation and mechanism of change [see Figure 2a for simple mediation]. Once all or most of the requirements are met and the relationship has been demonstrated by a series of studies, then one can argue for a mechanism of change (13) [see Kazdin (2007) for full criteria]. It is important to note that the distinguishing factors between a mediator and mechanism are discrete in that both attempt to uncover the same processes yet a mediator is established at an earlier stage of research whereas a mechanism is uncovered after extensive replication studies. Mediators can be understood as unrefined mechanisms that require further investigation to determine whether the methodological requirements are met and the finding has been supported in numerous studies.
In addition to understanding how MBPs effect change, it is also important to ask for whom they are more and less effective. Moderators are variables that effect the presence or absence, as well as the ‘direction or strength,’ of the relationship between the treatment (T) and outcome (O) [TàO] (13) [See Figure 2b for simple moderation]. In the context of mindfulness research, these variables are of interest to not only help refine the mediators and mechanisms that explain how and why MBPs work but also to help uncover for whom these programs work best. By elucidating these characteristics along with mediators and mechanisms, MBPs can be thoughtfully adapted and applied to specific populations and contexts (14, 15). It is important to note that there are different methods that can be used to evaluate these mentioned pathways of change and each method has their own strengths and limitations.
The existing literature on the how and why MBPs work in the general population is sparse and typically does not meet all the methodological requisites set out above by Kazdin (2007) (13). Gu et al. (2015) (16) identified twenty papers, in a systematic review and meta-analysis, that examined mediators of MBCT and MBSR on mental health and wellbeing outcomes in clinical samples (e.g. adults with symptoms of depression) and non-clinical samples (e.g. students and individuals with limited or no previous mindfulness training). The main results indicated strong evidence for cognitive and emotional reactivity, moderate evidence for mindfulness and repetitive negative thinking, and weak evidence for self-compassion and psychological flexibility as potential mechanisms through which MBPs produce change (16). However, it is unclear how representative these clinical and non-clinical samples are of the general population and most of the included studies have methodological shortcomings when it comes to establishing a mechanism of change according to Kazdin’s requirements (13, 16). For example, most studies did not report on temporal ordering to establish that change in the proposed mediator occurred before change in the outcome variable of interest (16). Research on the for whom MBPs work best in the general population is also limited. A comprehensive meta-analysis found that therapist training in mindfulness, rather than clinical training, moderated clinical outcomes (10). However, there were few studies that reported on this variable and it is still unclear how each sample (clinical versus non-clinical) is defined in the context of the general population.
This lack of evidence on the how, why, and for whom MBPs work in the general population seems to be attributed to a failed agreement on how to best define and operationalize population samples. One aim of this review is to help reach some consensus on these definitions and operationalizations and to further explore the extent to which relevant studies have met the requirements put forth by Kazdin (2007) (13) in establishing moderators, mediators and mechanisms of change.
In addition to exploring these variables in MBPs, there is also an opportunity to explore the how, why, and for whom in mindfulness-based practices. Kazdin (2007) (13) argues that one way to identify possible mediators and mechanisms of change is by conducting dismantling trials that isolate individual components of a treatment to explore how they bring about change in isolation rather than in combination of each other. Applying this idea to secularized mindfulness programs, one way to dismantle a MBP (e.g. MBCT) is to isolate individual meditation practices (e.g. body scan, mindful movement, etc.) within the program and to explore pathways of change (mediators, mechanisms, and moderators). Mindfulness practices are regarded as a core component of MBPs (3, 15) and conceptual frameworks (17) along with early research (18-20) suggests that these individual meditation practices can effect change on outcomes without being nested within a formal MBP. However, future research needs to address the change processes to better understand the particular aims and skills of individual mindfulness meditation practices.
Past research (18, 19) seems to highlight practices that are derived from the general framework of mindfulness-based programs, but the Crane et al. (2017) (15) definition is not implemented which makes it difficult to pin-point whether or not these practices originate from MBSR/MBCT and their adaptations. Sauer-Zavala et al. (2012) (20) compared three different meditation practices (sitting meditation, mindful yoga, and body scan) and found greater improvements in psychological well-being in mindful yoga compared to the other practices, and more change in emotion regulation difficulties in the mindful yoga condition compared to the body scan condition. Graduate students who led the practices followed scripts from MBSR/MBCT curricula. This helps clarify the effects of these practices within these frameworks, but it is important to note that practice adherence was not measured. Without this measurement, it is difficult to fully understand how change is produced (20).
Research on the mediators and mechanisms through which mindfulness-based practices produce change is in its early stages. Kok and Singer (2017) (21) suggest that each meditation practice has a ‘phenomenological fingerprint’ or rather a unique subjective experience. For instance, it was observed that the body scan meditation brought about greater changes in interoceptive awareness (the ability to use the body as a barometer for one’s thoughts and emotions) compared to a breathing, loving-kindness, and observing-thought meditation. Britton and colleagues (2018) (22) tested two ‘active ingredients’ of MBCT by dismantling this program and creating two structurally equivalent curricula that either implemented focused attention (FA) elements or open-monitoring (OM) elements and compared these curricula to a formal MBCT intervention in individuals with mild to severe affective disturbances. One interesting result indicted that individuals that took part in the OM program demonstrated greater improvements in nonreactivity whereas participants in the FA program experienced more change in attentional control. However, given that there was no ‘business as usual’ control, it is difficult to understand whether changes in outcome were as a result of practice-specific effects or non-specific effects (22). Overall, future research needs to further explore potential mediators and mechanisms (e.g. interoceptive awareness, nonreactivity, and attentional control) of mindfulness-based practices within the general population.
Several moderators of brief mindfulness interventions (ranging between one single session to two week sessions) were examined by Schumer, Creswell, and Lindsay (2017) (19). Type of mindfulness training exercise (focused breathing, multiple exercises, body scan, or other) was explored as a potential moderator and the results suggest that the ‘other’ practice type accounted for more change in negative affectivity outcomes (g = .30, p <. 001, n = 27) compared to the other practice types. This implies that the characteristics of these ‘other’ practices may predict greater improvements in outcome. However, without a clear description of these ‘other’ practices, it is difficult to understand which qualities or skills of these practices may account for such change (19). Moderation analyses conducted by Blanck et al. (2018) (18), suggest that total amount of practice and delivery mode (face to face versus online) may have an effect on outcome yet the results were non-significant. Taken together, there is evidence that demonstrates positive effects of these different mindfulness practices on a range of outcomes (18-21) but the evidence-base on the how, why, and for whom these practices work best is still in its infancy and requires clear delineation of both the mindfulness practices and conceptual framework, with clear definitions of key constructs (population, moderator, mediator, mechanism, treatment, and outcome). So far, and to the knowledge of the authors, there is no review that has comprehensively examined these pathways of change in practices derived from MBSR/MBCT and their adaptations. In light of the limited evidence on the change processes of both MBPs and the practices derived from these programs across the general population, the rationale for this review is to provide a summary of the current evidence to inform future research studies and stakeholders in the field who would be interested in adapting these programs and practices in contexts for the general population.
Objectives
The primary aim of the current scoping review is to examine the how, why and for whom MBPs and practices operate in the general population. Due to the novelty and breadth of this research topic, the authors decided to follow the stages and methodology of a scoping review.
The objective of this scoping review is to explore the following questions: (a) Through which mediators, mechanisms, and moderators do MBPs and practices produce change in the general population of adults? (b) Which design approaches (e.g. RCT, quasi-experimental, dismantling trials, etc.) are used to evaluate mediation, moderation, or a mechanism of change? (c) Which data analytic approaches (e.g. multiple regression, path analysis, etc.) are used to evaluate mediation, moderation, or a mechanism of change? (d) To what extent have the identified papers met the formal recommendations laid out by Kazdin (2007) (13) in establishing pathways of change? The key elements of these research questions include the general population (population), mediators, mechanisms, and moderators (concept), and MBPs and practices (context). Several established and accepted frameworks (23-25) will be used to drive the methodology and stages of the scoping review. This review protocol adheres to the PRISMA-P checklist along with the PRISMA extension for Scoping Reviews (PRISMA-ScR) checklist (26, 27). See Additional File 1 and 2 for completed check-lists.