Mindfulness Research in Social Work: Conceptual and Methodological

Recommendations Eric L. Garland

Mindfulness refers to a set of practices as well as the psychological state and trait produced by such practices. The state, trait, and practice of mindfulness may be broadly characterized by a present-oriented, nonjudgmental awareness of cognitions, emotions, sensations, and perceptions without fixation on thoughts of past or future. Research on mindfulness has proliferated over the past decade. Given the explosion of scientific interest in this topic, mindfulness-based therapies are attracting the attention of clinical social workers, who seek to implement these interventions in numerous practice settings. Concomitantly, research on mindfulness is now falling within the scope and purview of social work scholars. In response to the growing interest in mindfulness within academic social work, the present article outlines six conceptual and methodological recommendations for the conduct of future empirical studies on mindfulness. These recommendations have practical importance for advancing mindfulness research within and beyond social work.

KEY WORDS: evidence-based practice; meditation; mindfulness; randomized controlled trial; research methods

Mindfulness is linked with a set of cross-cultural principles and practices origi-nating in Asia more than 2,500 years ago that have parallel manifestations in numerous cultures around the world. With regard to its cur- rent academic usage, mindfulness refers to a psycho- logical phenomenon that is now being studied for its relevance to mental and physical health in fields such as medicine, psychology, and neuroscience. Across these fields, there is a growing body of litera- ture that attests to the efficacy of mindfulness-based therapies for a wide range of biobehavioral disorders. According to a search of PubMed and CRISP data- bases conducted on October 4, 2009, there were 1,614 peer-reviewed journal articles on mindfulness published in the scientific literature and 320 research grants on mindfulness funded by the National Insti- tutes of Health between 1998 and 2009. Indeed, there is mounting empirical evidence of the role of mindfulness in reducing stress and improving clinical outcomes across diverse conditions such as depres- sion (Teasdale et al., 2002), relationship difficulties (Carson, Carson, Gil, & Baucom, 2004), irritable bowel syndrome (Gaylord et al., 2011), crimi- nal recidivism (Himelstein, 2011), chronic pain (Rosenzweig et al., 2010), and addiction (Bowen et al., 2006; Garland, Gaylord, Boettiger, & Howard,

2010). Consequently, mindfulness-based interven- tions are becoming well-regarded for their therapeu- tic promise, as evidenced by recent publications in mainstream, respected academic outlets, such as the Journal of the American Medical Association (for exam- ple, Ludwig & Kabat-Zinn, 2008).

Given this burgeoning interest, mindfulness- based interventions are attracting the attention of clinical social workers who are increasingly imple- menting these treatments across diverse domains of practice. Concomitantly, research on mindfulness is now falling under the purview of social work scholars, many of whom seek to determine the comparative effectiveness of mindfulness-based interventions and apply the construct of mindful- ness to theories and models of social work practice. In response to the growing interest in mindfulness within academic social work, this article outlines six conceptual and methodological recommenda- tions for the conduct of future empirical research on mindfulness.

INCREASE PRECISION IN OPERATIONALIZATION OF THE CONSTRUCT OF MINDFULNESS To advance any field of scholarship, the precise operationalization of constructs is a necessary first

doi: 10.1093/swr/svt038 2013 National Association of Social Workers 439

step. Without such precision, empirical investiga- tions of putatively identical phenomena may result in widely divergent correlations between constru- cts of interest and inconsistent clinical outcomes across studies. Ultimately, imprecise operationaliza- tion of constructs presents a severe threat to validity that can undermine the quality of otherwise well-designed research studies (Shadish, Cook, & Campbell, 2002). Mindfulness research within and outside of social work has been rife with this prob- lem. An examination of Hicks (2009) edited vol- ume Mindfulness and Social Work clearly demonstrates this issue. According to Hick, mindfulness is an orientation to our everyday experiences (p. 1); to others in the edited volume, it is a specific and effective method of focusing the mind on the ess- ence of experience (p. 45), a way to mediate the development of professional self-concept (p. 93), an approach for increasing awareness (p. 125), an approach for performing all activities with full awareness (p. 154), and even a necessary condi- tion for an activist to become mature in her passion and mission to fight for justice (p. 178). This lack of conceptual clarity should be rectified and a uni- form, coherent set of definitions established, if mind- fulness research within social work is to advance.

To that end, the following operationalizations of mindfulness are offered. First, mindfulness is a state, a naturalistic mindset characterized by an attentive and nonjudgmental metacognitive monitoring of moment-by-moment cognition, emotion, percep- tion, and sensation without fixation on thoughts of past and future (Garland, 2007; Lutz, Slagter, Dunne, & Davidson, 2008). Mindfulness is meta- cognitive in the sense that it involves a meta-level of awareness that monitors the content of con- sciousness while reflecting back upon the process of consciousness itself (Nelson, Stuart, Howard, & Crowley, 1999). Mindfulness is naturalistic in that it is a basic and inherent capacity of the human mind, although people differ in their ability and willingness to actualize this state (Brown, Ryan, & Creswell, 2007; Goldstein, 2002).

Second, mindfulness is a practice (or, more accu- rately, a set of practices) designed to evoke and fos- ter the state of mindfulness. The practice of mindfulness involves repeated placement of atten- tion onto an object while alternately acknowledg- ing and letting go of distracting thoughts and emotions. Objects of mindfulness practice can include the sensation of breathing; the sensation of

walking; interoceptive (Craig, 2003) and proprio- ceptive (Brodal, 2004) feedback about the bodys internal state, movement, and position; visual stim- uli such as a candle flame or running water; mental contents such as thoughts or feelings; or the quality of awareness itself (Lutz et al., 2008). These prac- tices are taught and trained in mindfulness-based interventions.

Third, mindfulness is a trait or disposition that may be developed over time through the repeated practice of engaging in the state of mindfulness. This trait may be characterized as the propensity toward exhibiting nonjudgmental, nonreactive awareness of ones thoughts, emotions, experi- ences, and actions in everyday life (Baer, Smith, Hopkins, Krietemeyer, & Toney, 2006). As a trait, mindfulness is roughly normally distributed (Walach, Buchheld, Buttenmllerc, Kleinknechtc, & Schmidta, 2006). People vary in the extent to which they exhibit mindful dispositions, yet this dispositional- ity can be strengthened through training. People who participate in mindfulness-based interventions evidence increases in trait mindfulness, which med- iates the effects of training on clinical outcomes (Carmody & Baer, 2008).

Thus, integral to mindfulness is the notion of state by trait interaction, that is, recurrent activation of the mindful state via mindfulness practices leaves lasting traces that may accrue into durable changes in trait mindfulness (Garland, Fredrickson, et al., 2010), possibly mediated through neuroplasticity and experience-dependent alterations in gene expression (Garland & Howard, 2009). Indeed, recent research suggests that mindfulness practice can lead to increases in grey matter density in parts of the brain that subserve emotion regulation, learning, memory, and the ability to shift ones perspective (Holzel et al., 2011). More research is needed to determine whether such neurobiologi- cal changes index the development of trait mind- fulness over time resulting from mindfulness training.

USE RANDOMIZED CONTROLLED DESIGNS WHEREVER POSSIBLE A large number of social work studies use nonex- perimental and quasi-experimental research designs that are subject to severe threats to internal validity (Shadish et al., 2002). Despite the presence of these threats, authors often overstep the data by making causal claims from what are, at best, descriptive or

440 Social Work Research Volume 37, Number 4 December 2013

correlational findings. Studies that attempt to test causal hypotheses (for example, hypotheses of therapeutic efficacy) using suboptimal research designs weaken the portfolio of social work research and lower the esteem of the profession in interdisciplinary venues.

This is an especially serious problem when it comes to research on mindfulness, which is still met with skepticism within many academic circles as a New Age or mystical practice that amounts to little more than a placebo. Hence, it is essential to use research designs in mindfulness research that can control for the effects of matura- tion, social desirability, expectancy, and placebo effects.

From 1990 through the early 2000s, many stud- ies on mindfulness used randomized wait-list con- trol groups (for example, Astin, 1997; Davidson et al., 2003; Shapiro, Schwartz, & Bonner, 1998; Speca, Carlson, Goodey, & Angen, 2000). This type of research design is capable of controlling for history and maturation threats to validity as well as creating statistically comparable groups at baseline, and thus it represents a significant advance over nonrandomized or quasi-experimental studies with comparison groups. Yet, wait-list controlled designs remain vulnerable to threats to validity stemming from expectancy and placebo effects, which can be substantial (Shapiro, 1981).

Given these concerns, for much of the past dec- ade, mindfulness researchers have used randomized controlled trial designs in which participants are randomly assigned to either a mindfulness-based intervention or a credible, therapeutically active control condition. Perceived intervention credibil- ity can be measured with self-report scales, such as Borkovec and Naus (1972) Attitudes Towards Treatment Questionnaire, and statistically con- trolled (if necessary) in analyses of covariance. Such scales contain items assessing the extent to which the research interventions are perceived to be logi- cal treatments for the targeted clinical condition and how confident participants are that they will reduce their symptoms. Ideally, participants would perceive control treatments to be equally credible to experimental mindfulness interventions (for example, Garland, Gaylord, et al., 2010; Gaylord et al., 2011). Expectancy effects can also be mini- mized through careful advertising of the research. For example, a flyer that contains the statement We are conducting research on mindfulness-

based treatments for cocaine addiction is in- herently flawed, as it suggests the treatment of interest or preference and potentially introduces expectancy effects that may confound study re- sults. In contrast, research advertisements should conceal the identity of the experimental and control treatments. For instance, the same flyer would minimize expectancy effects by stating, We are conducting research to compare the eff- ectiveness of two forms of treatment for cocaine addiction: a mindfulness-based treatment and a support group.

Moreover, the presence of significant main effects of time on clinical outcome variables sug- gests that the control condition may have been therapeutically active; yet, the presence of a signifi- cant Treatment Time interaction term in the hypothesized direction indicates that the experi- mental mindfulness treatment led to significantly larger therapeutic change over time than the con- trol treatment. For example, in a randomized con- trolled trial of psychosocial treatments for irritable bowel syndrome, Gaylord et al. (2011) found that participants in a mindfulness training intervention and a conventional support group experienced sig- nificant reductions in abdominal pain; yet, relative to those in the support group, participants in the mindfulness training intervention experienced sig- nificantly greater reductions over the course of training.

The use of credible, therapeutically active con- trol groups may eliminate confounds introduced by expectancy and placebo effects as well as other nonspecific therapeutic factors such as atten- tion by a caring professional, group dynamics, soc- ial support, empathy, and the therapeutic alliance (Castonguay, Goldfried, Wiser, Raue, & Hayes, 1996; Duncan, Miller, & Sparks, 2007). When a study of a mindfulness-based intervention identi- fies significant clinical outcomes within the con- text of this rigorous research design, it may provide evidence against the Dodo bird verdict (for a re- view, see Budd & Hughes, 2009).However, it should be noted that a study comparing a mindfulness-based intervention to a no-treatment control is asking a substantively different question than a study com- paring a mindfulness-based intervention to an active placebo control condition. In the former case, the design allows one to measure the efficacy of participation in a mindfulness-based interven- tion; whereas in the latter case, the design allows

Garland / Mindfulness Research in Social Work 441

one to measure the efficacy of the active ingredient in a mindfulness-based intervention, that is, the practice of mindfulness itself. It should be noted that many active control conditions are not merely placebo controls, but instead are legitimate, estab- lished treatments. For example, Kuyken et al. (2008) compared mindfulness-based cognitive therapy (MBCT) to maintenance of antidepressant medication as a means of preventing depression relapse. Thus, studies that use an active control group can ascertain whether mindfulness training is more effective than alternative treatments, in contrast to studies with a no-treatment control condition that answer the more basic question, Is participation in a mindfulness-based interven- tion associated with positive clinical outcomes?

This is not to say that nonexperimental research designs have no place in mindfulness research. To the contrary, much can be learned about the associations between trait mindfulness and related constructs using cross-sectional or longitudinal research. However, it is imperative that authors draw careful conclusions that do not overstep the data. For instance, a prospective observational study of 339 individuals undergoing a mindfulness-based stress and pain management course found that participants reported significant improvements in trait mindfulness, positive reap- praisal coping, catastrophizing, and perceived stress over eight weeks of training (Garland, Gaylord, & Fredrickson, 2011). Further, the association between increases in trait mindfulness and decreases in stress was partially mediated by increases in posi- tive reappraisal but not by decreases in catastroph- izing. Although these findings are potentially clinically useful, one cannot conclude that mind- fulness training caused the observed changes. At best, one can only conclude that these changes occurred while participants were engaged in a mindfulness training program. Findings such as these are relevant only to the extent that they are interpreted with great precision; otherwise, they will remain unpublished or, worse, be pub- lished and tarnish the reputation of mindfulness researchers within and beyond academic social work.

Several key observational, quasi-experimental, and experimental research studies that represent the broad scope of research on mindfulness as a state, trait, and practice are presented in Table 1.

INVESTIGATE THERAPEUTIC MEDIATORS AND USE DISMANTLING DESIGNS Social work, as an applied field, is often myopically focused on clinical outcomes to the exclusion of more basic forms of scientific research. However, asking the question By what processes does this treatment work? is often a key step in refining and optimizing an intervention (Kazdin & Kendall, 1998; Kraemer, Wilson, Fairburn, & Agras, 2002). For instance, if an initial study reveals that increases in trait nonreactivity mediate the therapeutic effect of mindfulness training on chronic pain (compare, Garland, Gaylord, Palsson, et al., 2012), mindful- ness interventions tested in future clinical research projects might emphasize techniques designed to increase nonreactivity in order to boost treatment effect sizes. Thus, examining therapeutic mediation could enable social work researchers to determine how mindfulness-based interventions might be targeted most effectively to the populations and to identify problems of greatest interest to clinical social workers in the field. Therapeutic mediation can be tested by a number of statistical methods, in- cluding canonical regression procedure (Baron & Kenny, 1986), bootstrapping (Preacher & Hayes, 2004), structural equation modeling (Kline, 1998), or latent growth curve approaches (Preacher, Wichman, MacCallum, & Briggs, 2008), among others.

Moreover, by establishing the mediators of treat- ment, one can assess whether a treatment is inter- nally consistent with the theoretical orientation in which it is grounded (Hayes, Strosahl, & Wilson, 1999). This is important, both for maintaining the- oretical coherence and preventing reductionism and subsequent dismissal by researchers operating from other theoretical orientations. As one promi- nent example of this issue, mindfulness practices have been construed by some as relaxation tech- niques, believed to reduce stress via evocation of a relaxation response (Benson, Beary, & Carol, 1974). However, mindfulness meditation has been shown to produce significantly different cardiovas- cular and autonomic effects than relaxation training (Ditto, Eclache, & Goldman, 2006), findings that argue against the reductionistic construal of mind- fulness practice as a mere relaxation technique. Fur- ther, a randomized controlled trial demonstrated that whereas both mindfulness practice and relaxation training led to reduced distress, mindfulness practice alone led to significant decreases in ruminative

442 Social Work Research Volume 37, Number 4 December 2013

Table 1: Select Key Studies Representing the Broad Scope of Research on Mindfulness as a State, Trait, and Practice

Study Sample Design Operationalization of Mindfulness

or Related Phenomena Pertinent Results

Bowen et al. (2009)

168 adults with substance use disorders

RCT of MBRP versus standard substance use treatment services

Trait mindfulness assessed by the FFMQ; acceptance assessed by the AAQ

MBRP participants, relative to the control group, reported significantly fewer days of drug and alcohol use. Relative to the control group, MBRP led to significant increases in acceptance and the Acting with Awareness subscale of the FFMQ.

Carmody & Baer (2008)

174 adults with a wide range of stress, chronic pain, and anxiety issues

Prospective observational study of MBSR

Trait mindfulness assessed by the FFMQ; time spent in formal mindfulness practice

Increases in trait mindfulness were significantly associated with time spent in formal mindfulness practice. Increases in trait mindfulness mediated the effects of time spent in mindfulness practice on psychological symptoms, stress, and well-being.

Feldman, Greeson, & Senville (2010)

190 female college students Random assignment to 15 minutes of MT, LKM, or PMR

State mindfulness (decentering) assessed by the TMS; frequency of or reactivity to repetitive thoughts

MT participants reported significantly greater state mindfulness (decentering) relative to the other two conditions. Relative to the other conditions, a 15-minute session of MT reduced negative reactions to repetitive thoughts.

Garland, Gaylord, et al. (2010)

53 alcohol-dependent adults in long-term residential treatment

RCT of MORE versus an addiction support group

Thought suppression (that is, a construct that is the opposite of mindfulness) assessed by the WBSI

MORE participants, relative to those in the support group, experienced significantly larger decreases in stress and thought suppression. Among MORE participants, decreases in thought suppression were associated with decreased fixation on alcohol cues and increased heart rate variability recovery from stress and alcohol cues.

Gaylord et al. (2011)

75 female patients with irritable bowel syndrome

RCT of MT versus a support group

Trait mindfulness assessed by the FFMQ MT participants, relative to those in the support group, experienced significantly decreased abdominal pain and increased quality of life. The effect of MT on these clinically significant outcomes was mediated by increases in trait mindfulness (nonreactivity).a

Holzel et al. (2011)

33 healthy adults either participating or waiting to participate in a MBSR course

Quasi-experiment comparing MBSR to wait-list control group

Trait mindfulness assessed by the FFMQ; changes in brain structure assessed by structural magnetic resonance imaging

MBSR participants, relative to those in the wait-list control group, reported significant increases in trait mindfulness and exhibited significant increases in grey matter concentration in left hippocampus, cingulate cortex, cerebellum, and temporo-parietal junction.

Kuyken et al. (2010)

123 patients treated with antidepressants who had 3 depressive episodes

RCT of MBCT versus continued antidepressants

Trait mindfulness assessed with the KIMS; self-compassion assessed with the SCS

The therapeutic effects of MBCT were mediated by increases in trait mindfulness and self-compassion over the course of treatment. MBCT moderated the association between cognitive reactivity and depression.

Teasdale et al. (2002)

100 patients in remission or recovery from major depression

RCT of MBCT versus treatment as usual (for example, doctor)

Metacognitive awareness assessed by the MACAM

MBCT participants, relative to the control group, experienced significantly fewer occurrences of depression relapse. MBCT led to significantly increased metacognitive awareness.

Notes: MBRP = mindfulness-based relapse prevention; AAQ = Acceptance and Action Questionnaire (Hayes et al., 2004); FFMQ = Five-Facet Mindfulness Questionnaire (Baer et al., 2006); LKM = loving-kindness meditation; KIMS = Kentucky Inventory of Mindfulness Skills (Baer, Smith, & Allen, 2004); MACAM = Measure of Awareness and Coping in Autobiographical Memory (Moore, Hayhurst, & Teasdale, 1996); MBCT = mindfulness-based cognitive therapy; MBSR = mindfulness-based stress reduction; MORE = mindfulness-oriented recovery enhancement; MT = mindfulness training; PMR = progressive muscle relaxation; RCT = randomized controlled trial; SCS = Self-Compassion Scale (Neff, 2003); TAU = treatment as usual; TMS = Toronto Mindfulness Scale (Lau et al., 2006); WBSI = White Bear Suppression Inventory (Wegner & Zanakos, 1994). a These mediational data are presented in Garland, Gaylord, Palsson et al. (2012).