NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Ospina MB, Bond K, Karkhaneh M, et al. Meditation Practices for Health: State of the Research. Rockville (MD): Agency for Healthcare Research and Quality (US); 2007 Jun. (Evidence Reports/Technology Assessments, No. 155.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Meditation Practices for Health

Meditation Practices for Health: State of the Research.

Show details


The Practice of Meditation

Five broad categories of meditation practices were identified in the included studies: Mantra meditation (comprising TM®, RR, and CSM), Mindfulness meditation (comprising Vipassana, Zen Buddhist meditation, MBSR, and MBCT), Yoga, Tai Chi, and Qi Gong. One of the objectives of this review was to provide a descriptive overview and synthesis of information on meditation practices in terms of their main components, the role of spirituality, training requirements, and criteria for success. It is important to emphasize that the review on Topic I does not constitute a manual for any meditation practice. A more detailed explanation of any specific meditation practice described in this report should be sought in specialized texts or from master practitioners.

Given the variety of the practices and the fact that some are single entities (TM®, RR, and CSM, Vipassana) while others are broad categories that encompass a variety of different techniques or combination of practices (Yoga, Tai Chi, Qi Gong, MBSR, and MBCT), it is impossible to select components that might be considered universal or supplemental across practices. Though some statement about the use of breathing is universal among practices, this is not a reflection of a common approach toward breathing. The control of attention is putatively universal; however, there are at least two aspects of attention that might be employed and a wide variety of techniques for anchoring attention. The spiritual or belief component of meditation practices is poorly described in the literature and it is unclear in what way and to what extent spirituality and belief play a role in successful practice. The amount of variation in the described frequency and duration of practice make it difficult to draw generalizations about the training requirements for meditation practices. The criteria for successful meditation have also not been described well in the literature, though this may reflect the attitude that meditation is successful if one simply does it. At a clinical level, it might be argued that meditation is successful if it produces positive outcomes.


Providing a comprehensive review and summary of the scientific research on meditation practices requires the development of appropriate criteria by which to distinguish meditation practices from nonmeditation practices (what Ross327 has called a “demarcation criterion”). The development of such criteria is one of the most difficult yet important components of research on meditation practices,12 yet there is currently no consensus on a definition of meditation12 or on a way to classify the variety of meditation practices.12,37 Researchers have attempted to identify the components essential to the practice of meditation and to classify meditation practices in various ways:

  • any procedure that uses: (1) a specific, clearly defined technique, i.e., a “recipe” for meditation; (2) muscle relaxation in some moment of the process; (3) “logic relaxation”; (4) a self-induced state; and (5) a “self-focus” skill, or anchor;9
  • a discrete and well defined experience of "thoughtless awareness.12,125 Techniques that fail to provide the key experience of mental silence or thoughtless awareness, including techniques that use constant repetition of syllables, visualizations, or other thought forms are considered “quasi-meditation;12
  • techniques that seem to restrict awareness to a single, unchanging source of stimulation for a definite period of time;328
  • an exercise in which the individual turns attention or awareness to dwell upon a single object, concept, sound, image, or experience, with the intention of gaining greater spiritual or experiential and existential insight, or of achieving improved psychological well-being;34 and
  • a family of self-regulation practices that focus on training attention and awareness in order to bring mental processes under greater voluntary control and thereby foster general mental well-being and development and/or specific capacities such as calm, clarity, and concentration.10

Even if most investigators would agree that meditation implies a form of mental training that requires either stilling or emptying the mind to achieve a state of “detached observation,” few seem to consider this a sufficient demarcation criterion. Also, the general definitions offered above appear to be too narrow, excluding awareness-based forms of meditation such as Vipassana, MBSR, MBCT, and Zen Buddhist meditation.

Definitions usually focus on the phenomenological aspects of meditation practice and, with the exception of Cardoso et al.,9 rarely describe the necessary practical and physical components in sufficient detail to be translated into an operational definition of meditation. Further, though some investigators believe that research has shown meditation to be clearly distinguished from relaxation,3 as Manocha12 notes, there is sufficient evidence to show that “quasi-meditation,” techniques that do not cultivate a state of mental silence, do not differ from rest in terms of their physiological effects. Such results, if valid, make the development of clear demarcation even more important. This review has not evaluated whether meditation is indeed different from relaxation.

Whether defining meditation by one criterion or more, most investigators have looked for necessary and sufficient conditions with which to demarcate meditation practices from nonmeditation practices. Surprisingly, despite a persistent lack of consensus and the fact that demarcation criteria need not be bound by this approach,327 no author has examined alternative approaches to defining meditation despite some well-known developments in methods of demarcation in the philosophy of language329 and cognitive psychology,330 and more recent developments in ethics331 and evolutionary biology.332,333 Applying some of these techniques to meditation may prove fruitful.


To our knowledge, this is the first systematic examination of the components of and training for individual meditation practices.125 Classification of meditation practices is frequently based on the direction of attention,334 e.g., “opening up” versus “turning off,”328 positive versus negative,9 mindful versus concentrative,3 directive versus nondirective,335 etc. However, it has been suggested that the concentrative forms should not be viewed as opposites to the mindful or negative forms, but as the first step toward a progressive refinement of attention and concentration.335 Several practices in Yoga and Qi Gong will fall into both categories. Meditation practices may also be classified according to their historical origins with Indian and Chinese forms constituting two groups and clinically-based practices another.334

We employed consensus techniques to develop an extensive, though not exhaustive, list of 32 common meditation practices. The categories of meditation practices we have employed reflect only those practices identified in the English-language scientific literature and that satisfied the inclusion criteria for this review. There is a noticeable gap in the research that has been conducted on meditation practices—of the 32 practices identified in the Delphi process, only the 10 described here (TM®, RR, CSM, Vipassana, Zen Buddhist meditation, MBSR, MBCT, Yoga, Tai Chi, and Qi Gong) have been assessed in trials or using a before-and-after design. It is unlikely that our literature search failed to uncover a broad category of practices, though we may have missed certain practices, for example, techniques purportedly used by indigenous peoples of North America.34 However, given the comprehensiveness of our literature search strategy, it is unlikely that such practices, if subjected to scientific inquiry, would have been missed.

The categories are only meant to be descriptive and conclusions have not been made on the basis of the broad categories, but at the level of individual practices. Despite this, it may be that we have not sufficiently distinguished between schools of Tai Chi and style of Yoga and that distinct techniques have been subsumed under one category or class of practice. This lack of specificity will have affected the results of our analysis in cases where, for example, two styles of Tai Chi (e.g., Wu and Chang) have been combined. However, this potential limitation should serve to highlight the need for more explicit descriptions of techniques and the need for studies on a wide range of techniques for similar health conditions.336

The broad categories we have employed can be criticized as being simplistic and as ignoring subtle differences among practices. However, the categorization of practices is a product of the typological divisions one makes. For example, we have chosen to class together TM® and RR, even though some may argue that there are sufficient differences between these two mantra meditation techniques to keep them separate. In addition, though Benson's68 original formulation of RR clearly falls within the category of mantra meditation, contemporary formulations of the technique are multifaceted and incorporate a body scan, which is a mindfulness meditation technique (Dr. Jeffrey Dusek, personal communication, December 2006). Some may contend that RR and TM® should not be classed together because during TM® one does not try to associate the mantra with the breath, or dissociate the breath from the mantra, but rather the mantra is favored. There may be other subtle differences between practices grouped together in the broad categories. The difficulty in categorizing practices and the dearth of detailed descriptions in the literature reinforce the need for detailed descriptions of all components of the interventions employed in efficacy and effectiveness studies.

Universal Components of Meditation Practices

The results of this review are similar to those of a study by Koshikawa et al.334 that examined the physical components of 12 different types of meditation practices (excluding Tai Chi and including Christian meditation practices, early Buddhist meditation, Ajikan meditation, and Hotei meditation) in order to determine what similarities, if any, existed among the practices. Using a survey methodology, the investigators questioned 12 experts (one for each type of meditation practice) regarding the environmental conditions required for practice, method of breathing, postures, body movement, use of mantra, object of attention, diet, and training requirements for mastery of the technique. Their results indicate that, though some meditation practices may share some features in common, no two practices are alike in all features and no features of practice are universal to them.


The complexity of meditation practices makes dissecting components difficult and questionable; components may be synergistic and imperfectly understood if artificially separated from the whole discipline within which they take place.328 For example, though we have noted that no practice of a meditative technique requires the adoption of a particular belief system, West34 has questioned the reliance of researchers on clinically standardized forms of meditation practice rather than examining meditation as a practice that may be inextricably bound up with belief systems and expectations and ignoring the use of meditation as a central component of the belief system and of the day-to-day life of the practitioner. Other researchers have noted that the specific components adopted in a given meditation practice depend on the desired outcome,116,128,129,334 a fact which may make finding the common components across several practices undesirable unless the same outcome can be achieved. In addition, different techniques are reported to have different effects, so even if subjective descriptions of two or more techniques make them appear similar, their similarity must still be rigorously assessed.

In addition, for some of the practices that involve movement (Tai Chi and some yogic and Qi Gong techniques), researchers face additional challenges in designing studies that can separate the effects of exercise from the effects of the meditation practice. As physical activity has been shown to produce beneficial effects in those same physiological and neuropsychological outcomes of interest in trials on meditation practices (e.g., blood pressure, mood, etc.), this type of research is particularly important if the benefits of meditation practices are to be accurately assessed.

Criteria of Successful Meditation Practice

No descriptions of meditation practices provided an explicit statement of the criteria for successful meditation practice beyond reference to the internal states of the practitioner. The criteria have generally been inferred from descriptions of the practice. For example, in TM® the practitioner attends a series of checkup meetings in which their technique is examined, implying that the adequacy of the technique is judged by an experienced practitioner. However, there is no statement that individual practitioners cannot assess the correctness of their technique themselves and no list of the components that an experienced teacher may be attending to in assessing the practice. The same is true of Zen Buddhist meditation. Because of this method, there may be some inconsistencies in the criteria for successful practice. However, this does not change the fact that there is a dearth of information on the determination of successful meditation practice and that this is an area in which future studies may improve.


Some overviews of meditation practices328 have provided descriptions of the training requirements for meditation. However, these descriptions have focused mainly on TM®, Zen Buddhist meditation, and some yogic practices, and they fail to capture the wide range of training practices suggested by our literature review. In addition, poor descriptions of the physical aspects of the meditation techniques and the requisite training hinder identifying the components that may be similar across practices and limit the proper construction of and comparison between studies on the effectiveness of specific meditation practices. Without a detailed knowledge of, for example, an adequate training period for a particular Hatha yoga technique versus that for a Tai Chi technique, such studies are already confounded by factors pertaining to the learning of a technique and not the effects of the technique per se. In addition, some investigators334 have found that if, as Naranjo335 has observed, the development of the attitude specific to meditation is essential and the hardest part of meditation to attain and this can only be realized through practice, then proper instruction seems paramount and a description of the proper duration and frequency of any given technique is crucial to designing and appraising such studies.

State of Research on the Therapeutic Use of Meditation Practices in Healthcare

We have summarized a vast body of evidence regarding a broad group of practices categorized under the umbrella term “meditation”. Some may argue that addressing a research question regarding the effects of “meditation on healthcare” would be as challenging as reporting on the effects of “medication on healthcare” (Personal communication, David Shannahoff Khalsa, May 2007). There were substantial variations among the studies in the description of the practices of meditation, the type of controls, the type of populations, and the outcomes reported.

The field of scientific research on meditation practices does not appear to be organized under a shared theoretical framework, but instead consists of distinct groups of investigators working within different approaches of treatment theory (e.g., physiological, cognitive, behavioral, and cognitive-behavioral) that fail to engage each other meaningfully.

The majority of studies on meditation practices identified in this review have been conducted in Western countries and published as journal articles within the past 15 years. The majority of research in meditation practices has been conducted as intervention studies (67 percent), with 49 percent being RCTs or NRCTs. A similar bibliometric analysis on the clinical application of Yoga has revealed an increase in publication frequency over the past three decades with a substantial and growing use of RCTs.337

We identified and excluded from the review a considerable number of multiple publications (n = 108). In some instances, the same study was published in two separate journal articles without full cross reference, a practice of redundant publication that has been considered scientific misconduct.338340 Including redundant publications in systematic reviews and meta-analyses increases the risk of overestimating the effect size. The problem of redundant publications has not yet been sufficiently explored in the scientific literature; therefore, it is unknown how the proportion of redundant publications in meditation research compares to other areas of scientific inquiry. Authors of future studies on meditation should avoid redundant publications and must adhere to guidelines for good publication practices.340

Quality of the Evidence

Overall, we found the methodological quality of meditation research to be poor, with significant threats to validity in every major category of quality regardless of study design. Observational studies accounted for 33 percent of all the studies in the review. This type of study is open to several forms of systematic error such as selection bias, detection bias, and attrition bias. Intervention studies that used designs with pre-post treatment comparisons within the same group (known as single group before-and-after studies or uncontrolled trials) are not as rigorous as designs that use between-group comparisons because they do not allow investigators to determine whether the results are due to the meditation practice or to other factors. Studies with stronger designs such as RCTs and NRCTs allow a greater sense of confidence in study results; however, we found the quality of reporting to be poor for most of the intervention studies included in the review. This finding is not unique to the area of meditation research, and quality of reporting is a frequent problem in other areas of complementary and alternative medicine (CAM) research.341

The publication of the Consolidated Standards of Reporting Trials (CONSORT)342 statement in 1996 was aimed at the improvement of the quality of research reports of RCTs. It is unknown how the quality of reporting of RCTs of meditation practices has changed after the dissemination of CONSORT in the CAM community, but it is noteworthy that only 20 percent of the RCTs identified in the review described how the randomization was carried out, 8 percent were described as double-blind, and 4 percent described how they concealed the allocation.

The lack of double-blind RCTs has been a controversial topic not only in meditation research, but also in other areas of CAM,343 surgical interventions,344 and behavioral treatments.345 Some authors have called for a “paradigm shift,” suggesting that the quality of CAM research should be evaluated by other methodological standards. Some commentators have argued that the placebo-controlled trial is not a valid or fair method for evaluating CAM treatments.346348 Specifically, it is claimed that the scientific techniques of treatment protocols, randomization, double-blind conditions, and use of placebo controls distort the “holistic” therapeutic milieu of CAM. However, the notion of “holistic” interventions as opposed to “conventional medicine” may be an artificial misconception. Just as CAM does, traditional interventions provide treatments within a symbolic healing context by using “nonspecific” therapeutic attention and expectations.349

There is little argument against the idea that RCTs provide the least biased method for finding a reliable answer on the effectiveness of any therapeutic intervention, including CAM practices such as meditation. Based on empirical evidence and theoretical considerations, there are some basic characteristics that should always be considered when evaluating the quality of an RCT: randomization, blinding, handling of patient attrition in the analysis, and allocation concealment.40,42,350,351

Some authors have supported the idea that “those who insist that the evidence to support complementary and alternative medicine can legitimately be softer than in mainstream medicine will have to reconsider their position. Double standards in medicine existed for many years; undoubtedly they still exist today, but hopefully their days are numbered.”352

Double-blinding of the instructor and participant to treatment in meditation studies is often infeasible, a consequence of the fact that instructors must apply a specially learned skill in a particular therapeutic context (e.g., with clinically depressed patients). However, double blinding is still possible because the difficulties in blinding the experimenter can be circumvented by blinding the participants using a sham procedure such as similar attention control intervention or a placebo with a different mode of administration (as it has been done in psychotherapy research) and by blinding them to the hypothesis, and by blinding the outcome assessors to the nature of the intervention and the hypothesis. In cases where the comparison is an active treatment and blinding of participants to the treatment is impossible, it may still be possible to blind participants to the research hypothesis to minimize expectancy bias. Therefore, research on CAM should adhere to the same methodological requirements for all clinical research, and randomized, placebo-controlled clinical trials should be used for assessing the efficacy of CAM treatments whenever feasible and ethically justifiable.353

When double blinding was assessed using an individual components approach, we found that, although the vast majority did not use double blinding to hide the identity of the assigned interventions (97 percent), a small but promising percentage reported the use of double-blind procedures. The idea that it is possible to design high-quality trials in meditation and implement double-blind procedures by selecting appropriate control groups is gaining support in meditation research. We agree with other researchers that the implementation of these trials in any area of CAM research, including meditation, require much more preparation than trials of pharmacological interventions, and components such as blinding procedures, selection of credible placebos, and consistency of inherently individualized interventions are challenging issues that need extensive evaluation.343

Our conclusion here is that the idea that “due to the nature of meditation, it is impossible to double blind meditation practices” has been used as an excuse to justify the overall low quality of research that characterizes this body of evidence. However, the over emphasis of the “double-blinding” issue does not hide the fact that 95 percent of the studies failed to describe how they concealed the allocation to the interventions under study or the fact that overall, only 20 percent of the trials described the procedures of randomization, and that only half described study dropouts.

Therefore, syntheses of the results from studies included in this review should be interpreted with caution due to the serious threats to theinternal validity of the included studies.

Types of Interventions

Although a relatively small group of meditation practices have been studied in the scientific literature, they vary in many respects. There was a remarkable heterogeneity across the studies regarding the description of the characteristics and implementation of the practice even within the same type of meditation. Differences in theoretical assumptions underlying the practices of meditation may explain why studies conducted on similar meditation practices often differed in the potential benefits that were assessed. Some authors have declared that meditation poses a considerable challenge for the principles of evidence-based medicine.17 Meditation is a complex and multifaceted intervention, difficult to standardize, and for which specific effects are hard to distinguish.17 It is important, therefore, that investigators make an effort to avoid the excessive heterogeneity that characterizes this field, by clearly defining and reporting the intervention procedures of the meditation practice under scrutiny.

Types of Control Groups

Control groups are essential for the valid evaluation of the effects of meditation practices; however, the problem of the inadequacy of control groups in meditation research is not new.354 Almost half of the RCTs and NRCTs have used WL or no treatment approaches for the control group rather than a comparator that would more fully control for the variety of influences that may bias the results including expectancy effects, social interactions, attention given by instructors, and time spent in the practice. Some authors have argued that the use of WL as a control group is clearly inappropriate as no one expects to improve while they are waiting to begin treatment. This situation may create a negative expectation of improvement that may spuriously amplify the difference in treatment effect between the intervention and the control.345 Therefore, caution should be exercised when interpreting studies comparing the effectiveness of meditation practices to no treatment or WL.

A wide array of active control groups were used in the intervention studies on meditation practices. Active controls included exercise and other physical activities, states of rest and relaxation, educational activities, PMR, cognitive behavioral techniques, pharmacological interventions, psychotherapy, biofeedback techniques, reading, hypnosis, therapeutic massage, acupuncture, and other meditation techniques.

The results of this review show that the control groups employed in meditation research are many and various, and it is unknown how comparable they are across studies. Meditation practices are disparate with regard to specific components, and there is the potential for well-designed studies to employ disparate control groups. Authors of future studies need to design control groups with a clear vision of the research question and the hypothesized mechanism and full consideration of how threats to validity may be best addressed for a given meditation practice.

Types of Study Populations

The vast majority of studies on the effects of meditation practices have been conducted in healthy populations as compared to clinical populations. It can be argued that studies of healthy individuals are useful to assess how meditation practices prevent certain clinical conditions and enhance wellness and well-being. However, studying the therapeutic effects of meditation practices in a healthy population does not provide a clear picture of their effectiveness as therapeutic interventions in healthcare. Clinical studies of meditation practices have addressed conditions with high mortality and morbidity rates, or burden of disease including hypertension, cardiovascular disorders, substance abuse, anxiety disorders, cancer, asthma, chronic pain, type II DM, and fibromyalgia. The first three conditions were among the six leading sources of premature death and disability in the United States in the mid-1900s and are projected to continue to be so to the year 2020, as measured by disability-adjusted life years (DALYs).355,356

Types of Outcome Measures

Studies varied widely in their use of outcome measures. Outcomes of physiological functions, particularly cardiovascular measures, were the most frequently studied. Psychosocial outcomes (i.e., psychiatric and psychological symptoms, measures of personality and positive outcomes) and outcomes related with clinical events were also frequently assessed. Compared to physiological and psychosocial outcomes, little has been explored on cognitive and neuropsychological functions. Some authors have argued that relatively gross outcomes such as physiological measures have taken prominence in meditation research.354 However, considering that the close interdependence of the mind and body should be taken into account when evaluating the responses to meditation practices, more subjective and experiential variables354 are paramount to evaluate the effects of these mind-body techniques.

Evidence on the Efficacy and Effectiveness of Meditation Practices

We have summarized the evidence regarding the efficacy and effectiveness of meditation practices for the three most studied conditions in the scientific literature: hypertension, cardiovascular diseases, and substance abuse.

We conducted a series of direct and indirect meta-analyses comparing a variety of meditation practices versus a comparison group in hypertensive patients. We provided pooled estimates for the following comparisons: TM® versus HE, TM® versus PMR, RR versus BF, Qi Gong versus WL, Yoga versus NT, Yoga versus HE, and Zen Buddhist meditation versus blood pressure checks.

A few studies of poor methodological quality were available for each comparison, mostly reporting nonsignificant results (TM® had no advantages over HE to improve measures of SBP, DBP, body weight, heart rate, stress, anger, self-efficacy, cholesterol, dietary intake, and level of physical activity in hypertensive patients; RR was not shown to be superior to BF in reducing blood pressure in hypertensive patients; Yoga did not produce clinical or statistically significant effects in blood pressure when compared to NT; Zen Buddhist meditation was not better than blood pressure checks to reduce SBP in hypertensive patients; Yoga was not better than physical exercise to reduce body weight in patients with cardiovascular disorder. When indirect meta-analysis was used, we did not find differences between MBSR and Yoga to control anxiety symptoms in cardiovascular patients. It is unknown whether these are truly “negative findings” (i.e., one cannot say that there is evidence of no effect) or if there is a lack of power to detect a statistically significant result due to the low number of studies included in the meta-analyses (i.e., we can say that there is no evidence of effect).

A few statistically significant results favoring meditation practices were found: both TM® versus PMR, and Qi Gong versus WL for DBP and SBP, Zen Buddhist meditation versus blood pressure checks for DBP, and Yoga versus HE to reduce stress. The positive results from these meta-analyses need to be interpreted with caution, as biases, such as expectancy bias, cannot be excluded.

For the majority of the comparisons, meta-analyses were derived from only two open-label trials; therefore, performance bias and detection bias may have contributed to an overestimate of the treatment effect. In some instances, the appropriateness of the comparison group was questionable (e.g., Qi Gong versus WL)

Other reviews have summarized the evidence on the effects of Tai Chi in hypertension, and on TM® for hypertension, cardiovascular diseases, and substance abuse. Differences in selection criteria and review methods preclude a direct comparison of the results among the reviews.

Wang et al.29 assessed the evidence on the effects of Tai Chi in hypertension and concluded that Tai Chi produces benefits in cardiovascular function. The review included evidence from two studies published in the non-English literature, and another study in a population of normal elderly, not individuals diagnosed with hypertension. Differences in the selection criteria of study participants and language of publication may explain the differences in the findings between Wang et al.29 and our review.

Walton et al.357 reviewed the literature on the effectiveness of TM® in the treatment or prevention of cardiovascular diseases and concluded that TM® produced reductions in blood pressure, carotid artery intima-media thickness, myocardial ischemia, left ventricular hypertrophy, mortality, and other relevant outcomes. The authors adopted a qualitative approach for the synthesis of the evidence. The Walton review357 did not report on the use of systematic literature searches or on the assessment of the methodological quality of the evidence, but adopted a methodological approach where significant findings were emphasized within studies. Differences between Walton's conclusions and the results reported in our review may be due to differences in the methodological approaches to synthesize the evidence. We conducted comprehensive searches of the scientific literature and assessed the methodological quality of the trials. Our synthesis of the evidence combined a qualitative approach with quantitative meta-analytic methods that assessed mean treatment effects in relation to the between-study variability of treatment effects. Furthermore, differences in the selection criteria (i.e., type of participants, diagnostic criteria, publication year) for the inclusion of studies may also explain differences in the conclusions of the reviews.

Canter et al.25 conducted a systematic review on the effects of TM® for blood pressure. Six trials were identified but only one evaluated the effect of TM® in hypertensive individuals, whereas the others were conducted in adults with normal blood pressure and adolescent populations. The authors concluded that there was insufficient good quality evidence to conclude whether or not TM® has a positive effect on blood pressure.

Evidence on the effects of TM® on substance abuse has been summarized in two reviews.22,358 Alexander et al.358 conducted a meta-analysis of 19 studies to provide a single estimate of treatment effect. The review included a variety of study designs such as cross-sectional studies, “retrospective studies,” “longitudinal studies,” and “experiments with random assignment.”358 Effect sizes across studies were provided for categories of study designs (“well-designed” studies, cross sectional studies, and general population studies). Gelderloos et al.22 conducted a review of 24 studies of TM® for preventing and treating substance abuse. The authors concluded that “taken together”, the studies demonstrate an improvement in psychosocial outcomes. The review did not use a systematic approach to select and appraise the literature and made no distinctions among the variety of study designs that were considered.

Other systematic reviews have synthesized the evidence on the efficacy and effectiveness of meditation practices for conditions other than hypertension, cardiovascular diseases, and substance abuse. However, it was beyond the scope of this report to examine conditions other than hypertension, cardiovascular diseases and substance abuse. Other systematic reviews have examined the effects of Tai Chi for a variety of medical diseases,359 chronic conditions,29 rheumatoid arthritis,360 improvement of aerobic capacity,361 and elderly populations.362 Studies on the effects of Yoga in depression145 and anxiety,363and MBSR on health status measures18 and a variety of medical conditions7 have been also reviewed. Finally, other reviews have assessed the effects of a variety of meditation practices such as Qi Gong in Chinese cancer patients,364 RR in adult patients,365 meditation therapy programs for anxiety disorders,19 and the effects of TM® on cognitive function23 and psychological health.366

It is expected that systematic reviews have heterogeneity in their results when they bring together studies that are both clinically and methodologically diverse.367 Statistical and clinical heterogeneity constituted a frequent and considerable problem when pooling the results, and, in some cases, it precluded an effort to summarize data across the studies. Clinical heterogeneity was due to differences across the trials in the characteristics of study populations, the implementation of the meditation practice, outcome measurement, and followup period. Clinical heterogeneity may have explained why trials with different types of participants, interventions, or outcomes showed different effects. When statistical heterogeneity exists, pooled results are uncertain or conditional.368

The poor methodological quality of the trials limits the strength of inference regarding the observed treatment effects reported in this review and constitutes a possible shortcoming of the meta-analysis.367 The lack of description of the methods of allocation concealment, randomization, description of withdrawals and dropouts per treatment group, the absence of double blinding the interventions, and the use of incompatible or inappropriate control groups undermine the results of many clinical studies. Therefore, researchers are advised against making firm statements regarding treatment effects based on the quantitative summaries reported in this review.

Some factors have impeded the scientific progress regarding the efficacy and effectiveness of meditation practices in healthcare. Few studies have described the meditation practices or control procedures in sufficient detail, which prevents a sensible analysis of the observed differences in treatment effects for some classes of meditation practices. Other limitations include insufficient information regarding the characteristics of the trainer's competence and experience, the lack of an accurate assessment of participants' expectancy, compliance and motivation, and the paucity of descriptions of the statistical power of the intervention effect.

Evidence on the Role of Effect Modifiers for the Practice of Meditation

The role of effect modifiers in the practice of meditation is a topic that has so far been neglected in the scientific literature. Evidence from RCTs and NRCTs regarding the interaction of meditation practices with other variables in populations of patients with hypertension, cardiovascular disorders, or substance abuse is scarce. A few studies conducted exploratory post hoc analyses (i.e., a subgroup analysis, multiple regression, or analysis of variance) that were intended to be hypothesis generating. Due to the small sample sizes in the studies, there were small numbers of subjects in each of the variable subgroups, lowering the power to detect any relationship with the outcomes produced by the practice of meditation. The lack of evidence on the role of effect modifiers has been pointed out by other authors.17,354 Variables that may be important for the therapeutic effect of meditation practices include individual characteristics of the meditator, characteristics and training experience of the instructor, and the role of motivation and expectancy.17,354

Evidence on the Physiological and Neuropsychological Effects of Meditation Practices

We have summarized the evidence from RCTs, NRCTs and before-and-after studies regarding the physiological and neuropsychological effects of meditation practices. Our meta-analysis revealed that the most consistent and strongest physiological effects of meditation practices in healthy populations occur in the reduction of heart rate, blood pressure, and LDL-C. The strongest neuropsychological effect is in the increase of verbal creativity. There is also some evidence from before-and-after studies to support the hypothesis that certain meditation techniques decrease visual reaction time, intraocular pressure, and increase breath holding time. Though over half of the combined effect estimates are not statistically significant, the potential clinical significance of these estimates must be carefully considered. However, all of the studies included in the meta-analyses were of low methodological quality and, for this reason, the results should be interpreted cautiously.

Of the 311 studies reporting physiological and neuropsychological outcomes, only 53 (17 percent) were eligible for meta-analysis. Though small, this proportion is even smaller when one considers the 813 studies pertaining to research on the therapeutic use of meditation practices included in topic II. Some investigators have claimed that there are many empirical studies that have shown that meditation practices are effective at treating stress-related states,4 including reducing heart rate, breathing, and blood pressure. In addition, previous literature reviews have noted the seemingly large number of research papers that purport to show the therapeutic benefit of meditation practices.25,113,157,359,369 This review has shown that there are startlingly few scientific studies that could be statistically combined to provide evidence on the physiological and neuropsychological effects of meditation practices. While other investigators have noted the need for rigorous meta-analyses of the therapeutic use of meditation practices,3,369 to our knowledge there are only two previous English-language meta-analyses361,369 that examine the physiological effects of meditation practices and none examining the cognitive or neuropsychological effects. However, the two meta-analyses cover neither the range of meditation practices examined here nor the breadth of outcomes.

The clinical and methodological diversity of the studies make estimating the effects of meditation practices difficult. This difficulty is reinforced when one considers that 25 of the 44 (57 percent) outcome measures examined in the analyses had levels of heterogeneity that suggest important clinical differences between the studies. In addition, 8 of these 25 (32 percent) had heterogeneity measures greater than 80 percent, making overall effect estimates unwise because the implied clinical among study populations would render the overall estimates spurious.

The overall low methodological quality of the studies indicates that most suffered from methodological problems that may produce overestimations of the treatment effects or compromise the generalizability of the study results. Empirical evidence has demonstrated that trials “that were not double blinded yielded larger estimates of treatment effects compared with trials in which authors reported double blinding (odds ratios exaggerated, on average, by 17 percent)”.40,370 Though difficult to do in studies on meditation practices, appropriate blinding is a special source of concern where an expectation of the efficacy of the practice under study on the part of the subject and assessor may bias outcome measures.

The low rate of reporting of withdrawals and dropouts and the reasons for dropping out are also of concern because this makes the assessment of the comparability between the intervention and control groups difficult. An additional concern is that patients who drop out may differ in important ways from those who complete the meditation regimen (e.g., being favorably predisposed to meditation practice), but, without adequate reporting, these differences remain hidden and their effects on outcomes remain unknown.

Regarding the predominant use of healthy subjects in the included studies, though of benefit for ascertaining the physiological and neuropsychological effects of meditation practices in this group, the use of healthy subjects limits the generalizability of the findings and provides information that is unlikely to be of use to clinicians who normally treat patients with specific health conditions.

Finally, the results of this meta-analysis indicate that research on the effects of meditation practices has been hindered by the use of weak study designs, specifically before-and-after studies (also known as single group pretest-postest designs and uncontrolled trials). Although the before-and-after study is simple and practical, it has been argued that results from such study designs be considered circumstantial evidence,371 that is, hypothesis generating for further research using more rigorous study designs. The lack of a concurrent control group and the resulting inability to control for temporal trends, regression to the mean, and sensitivity to methodological features make it difficult to ascertain the true causal effect of a meditation practice. Clinical outcomes—whether good or bad—may be a result of factors other than the practice of meditation. For this reason, the estimates of the physiological and neuropsychological effects of meditation practices that are made on the basis of single-group studies should be considered carefully.

Strengths and Limitations

This evidence report is a systematic and comprehensive review of the indexed scientific literature available on the effectiveness of meditation practices supplemented by a search for relevant gray literature, abstracts from scientific meetings, dissertations and theses, reference lists, and trial registries. As noted previously, the need for rigorous meta-analyses of the therapeutic use of meditation practices has been recognized by other researchers.3,369 To our knowledge, there has been no other meta-analysis of the effectiveness of meditation practices that covers the range of meditation techniques examined here or the breadth of health outcomes. In addition, the relatively large number of included studies reported in dissertations (10 percent of all studies) may have reduced the potential effects of publication bias (i.e., the tendency for studies with positive outcomes to be published more frequently). We were also able to identify and exclude from the review a significant number of multiple publications that may have also affected the results of our meta-analyses and their conclusions.

The assessment of the methodological quality for all study designs is also a strength of this review. Methodological quality may be defined in various ways.372 Our approach to the methodological quality of the studies on meditation practices focused on an assessment of the internal validity of the studies, as recommended by several researchers.42,373375 Various criteria to assess methodological quality of studies are available in the scientific literature,376 and there is no consensus on which quality assessment tool can be recommended without reservation.50 For the assessment of the methodological quality of RCTs, we have chosen two assessment tools that have well-established face validity, and for which a relationship with bias has been proven in empirical studies.40,350 The selection of the Jadad scale has relative merit since it uses a simple and easy to understand approach that incorporates the most important individual components of internal validity: randomization, blinding, and handling of patient attrition. Based on empirical evidence and theoretical considerations, these aspects should always be assessed when evaluating the quality of an RCT.350

The most important dimension of methodological quality is internal validity, defined as the confidence that the design, performance, and report of a trial prevent or reduce bias in the outcomes.372 We have not addressed in our approach other important aspects of good research practice—those contributing to studies' external validity and adherence to ethical procedures. Although such factors are important and help to put study findings in context, they may not be directly related to internal validity, but may contribute indirectly to it. It is unknown how factors related with external validity may bias study results, and, therefore, research syntheses' findings. Certainly, the external validity of a trial is a very important concept that it is worthy of consideration in future reviews; however, it was not covered in our methodological assessment.

We have adopted a model for quality assessment of research on meditation based on stringent criteria of research methodology. Evaluation of CAM treatments, including meditation, requires a stringent and systematic approach.352,377 The Jadad scale is the most commonly used quality scale for RCTs in pharmacological and nonpharmacological reviews.378 The decision to use both the Jadad scale and the concealment of allocation approach reflects our emphasis on using the same methodological standards to assess the quality of research in meditation as applied to other areas of CAM research.

We did not make any decisions in terms of inclusion or exclusion of studies in the review or in the meta-analyses based on the overall Jadad score. We also analyzed the methodological quality of the RCTs by the individual components of the scale (i.e., percentage of studies that satisfied the Jadad criteria).

Though no reliable and valid instruments have been developed for the assessment of observational studies and before-and-after studies, the instruments used here serve to indicate important potential methodological weaknesses, tempering the conclusions that may be drawn, and highlighting areas in which future research might improve.

Despite its strength, the use of nonstandardized quality assessment instruments may be questioned. However, the assessment criteria were not used to produce an overall quality score or to exclude studies from the review, but only to draw out commonalities in potential methodological problems. Because of the potential methodological weaknesses of the studies and the use of weak study designs, the question of how meditation achieves its effects remains almost as open to debate as it did over 25 years ago.379

It is unlikely that all of the meditation research meeting our inclusion criteria has been identified and acquired. In particular, a number of Indian journals have not been indexed and are difficult to acquire, particularly Yoga specialty journals. We did not contact either any religious/spiritual organization to acquire information regarding unpublished studies. Nevertheless, it is likely that the vast majority of publications that satisfy our inclusion criteria have been examined and that the general trends reported in this review are sufficiently representative of the research on meditation practices.337

Peer reviewers have provided references to potentially relevant studies that were not identified during the development of this report. To increase the transparency of this report, we have collated the references of these studies following the “References and Included Studies” section. Despite the comprehensiveness of our search strategies for the literature search, there are inevitable gaps in literature retrieval, especially with respect to gray literature when conducting systematic reviews. The impact of the potentially relevant studies identified by the peer reviewers should be weighed against the number of studies that were actually retrieved and included.

The restriction of included studies to English-language publications is of special concern in this topic because of the origin of many of these techniques in non-English speaking countries. In light of a recent bibliometric study on Yoga that reported that there is a large amount of research by Indian researchers,337it is possible that there is a substantial evidence base on Yoga that remains untapped. In addition, it is likely that a significant amount of the research on Tai Chi29 and Qi Gong has been published in the Chinese language. However, despite this potential weakness, some research has shown that compared to language inclusive meta-analyses, language restricted meta-analyses did not differ with respect to the estimate of benefit of the effectiveness of an intervention, and there is no evidence that language restricted meta-analyses lead to biased estimates of intervention effectiveness.380

This review may be also be criticized for ignoring important differences between meditation practices and techniques by using categories for studies using “single entity” practices, e.g., TM®, RR, and CSM, and for those practices that are made up of a broad array of techniques, e.g., Yoga, Tai Chi, and Qi Gong. Thus while the meta-analytic techniques used here may be appropriate for standardized “single entity” practices, such an approach, when used to combine complex interventions, may produce spurious or misleading results. For example, one of the problems of combining the results of studies that use different yogic techniques is that “fine grained” descriptions of many of these techniques are not reported. This lack of reporting increases the possibility of pooling the results for yogic practices that were putatively designed to have different effects.

To address this potential problem, we have used measures of heterogeneity to help identify those groups of studies that may differ in important clinical characteristics as well as examining the descriptions of the techniques employed in the studies. The combining of results was based on these “fine grained” descriptions; however, poor reporting of meditation practices employed in studies leaves open the possibility that such combinations may have occurred. In addition, caution should be taken in concluding that the effects of complex or composite interventions are due to the practice of meditation rather than to other main components of the treatment such as physical exercise.

The approach adopted here of combining the results of only two studies may be considered inappropriate by some researchers because it is unlikely that only two studies provide strong evidence with respect to the general direction or effect size of the intervention. Also, if the results of two studies differ in direction of effect, at least one more study is needed to help strengthen the evidence regarding the true direction of the effect. However, it must be remembered that one of the principal reasons for conducting a meta-analysis is not only for summarizing the discrepant results of a large number of studies but also for overcoming the imprecision resulting from small sample sizes. By combining several studies with small samples, the overall estimate provides a more precise estimate of effect than either of the studies on their own. Thus, combining only two studies can provide an informative picture of the likely effect of an intervention.

Finally, a main weakness of this report is the lack of assessment of the appropriateness of controls. The need for appropriate controls, described by some researchers as the most difficult conundrum for designing research trials in meditation,381 is closely related to the difficulties in designing rigorous double-blind meditation trials. Though some controls may be adequate to compare the relative effectiveness of two different interventions (e.g., rest meditation versus quiet rest), such controls may not be adequate placebo controls needed to assess the effects of meditation interventions.381 Though we are unaware of assessment tools developed to specifically address this issue as it pertains to meditation practices, the comprehensive categorization given in this report of the kinds of controls used in meditation research provides future researchers with a starting point for examining the appropriateness of controls for various therapeutic meditation practices.

Future Research

Future research in practices of meditation has several challenges. First, there is a need to develop a consensus on a working definition of meditation applicable to a heterogeneous group of practices. The application of consensus techniques, such as the Delphi method used in this report, is one approach to refine operational criteria and to standardize terms with the goal of achieving consistency among the characterizations of meditation practices. The validity and reliability of any operational definition applied to diverse meditation practices should be thoroughly investigated. Another area of future inquiry consists of systematically comparing the effects of different meditation practices that research shows have promise.

We have assessed the quality of meditation research from studies that have been published between 1956 and 2005. Half of them have been published after 1994. We did not set any restrictions in terms of the year of publication of the included studies, and it is possible that the standard for a rigorous study in the earlier years of research might be different before 1994 than that of the past 15 years. Future reviews should examine how the quality of studies on meditation practices has evolved over time and particularly whether guidelines such as CONSORT have improved the reporting of RCTs.

We have analyzed the evidence of the therapeutic effects of meditation practices for the three most studied conditions identified in the scientific literature. Evidence of the effects of meditation practices for other conditions frequently reported in the scientific literature (i.e., a variety of mental health problems such as anxiety disorders and depression, and musculoskeletal conditions such as fibromyalgia and chronic pain)) should be evaluated in systematic reviews in the near future. Further reviews should address the effects of meditation practices as strategies to enhance wellness and well-being in healthy population.

In light of the few intervention studies that provided direct comparisons of meditation practices or that used similar control groups, special attention should be paid to developing studies that provide a more accurate assessment of the efficacy and effectiveness of meditation practices, both against standard therapies and against each other. The appropriate selection of controls is also paramount if progress is to be made with respect to determining the effects of meditation practices. Future research should be directed toward investigating the unique challenges that the studies on meditation practices present in designing appropriate controls. In addition, more research should be done on the “dose response” of meditation practices to determine what may be effective study durations and to help standardize courses of therapeutic meditation.

As noted earlier, blinded allocation to meditation treatments may be difficult, but it is not impossible. There are many ways in which to circumvent the difficulties in blinding the experimenter many of which rely on “creative” (i.e., nonstandard methods). These suggestions follow other proposed modifications of the traditional double-blind methodology such as the “dual-blinding” approach (a methodology where the subject and an external evaluator, but not the practitioner, are blind to treatment) 382 Given the strength of the RCTdesign in providing estimates of effectiveness, it appears important to develop research in this domain instead of trying to change the instruments with which the quality of research is assessed.

NCCAM is striving to elevate CAM research to a higher standard, and we think that creative solutions to the difficulties of conducting randomized, double-blind controlled trials should be applied to meditation research.

Key methodological issues in the study of meditation using an evidence-based approach should be further explored through the analysis of important factors such as the impact of publication bias in meditation research (e.g., positive outcome bias, time to publication bias, empirical evidence of relationships between study quality and effect estimates in meditation research, the impact of language bias in systematic reviews of meditation practices, the impact of year of publication of primary studies on pooled estimates in meditation research, trends of quality of primary studies and systematic reviews in meditation, and use of quality assessment tools in meditation research). The effect of report of funding and disclosure of conflict of interest and positive outcomes also merits formal evaluation.

Because of the difficulty of determining causation using uncontrolled before-and-after designs, it is recommended that these study designs be avoided in future research on the effectiveness of meditation practices. Researchers should aim to employ designs and analytic strategies that optimize the ability to make causal inferences (in some cases this may require the use of uncontrolled before-and-after designs). Although it is important to suggest conducting more high quality studies based on the standards for RCTs, it is also important to develop alternative study designs and analytic tools that can incorporate the special features of meditation practices to fully investigate the possible effects of these practices. As well, future studies would benefit from having larger samples with concurrent controlled designs, using disease-specific measures and providing clearer descriptions of intervention components. The quality of reporting of meditation research would be improved by a wider dissemination and stricter enforcement of the CONSORT guidelines within the CAM community.


The field of research on meditation techniques and their therapeutic applications has been clouded by confusion over what constitutes meditation and by a lack of methodological rigor in much of the research. Further research needs to be directed toward distinguishing the effects and characteristics of the many different techniques falling under the rubric “meditation.” The single and multimodality meditation practices included in this report were categorized for pragmatic reasons, but specific attention must be paid to developing definitions for these techniques that are both conceptually and operationally useful. Such definitions are a prerequisite for scientific research of the highest quality. Research of higher quality is vital to respond appropriately to the many persistent questions in this area. The dearth of high-quality evidence highlights the need for greater care in defining and choosing the interventions and in choosing controls, populations, and outcomes that permit comparison of studies across techniques regarding their therapeutic effects. More care in these choices will allow effects to be estimated with greater reliability and validity. More randomized trials that draw on the experience of investigators or consultants with a strong background in clinical and basic research should be conducted. As a whole, firm conclusions on the effects of meditation practices in healthcare cannot be drawn based on the available evidence. However, the results analyzed from methodologically stronger research include findings sufficiently favorable to emphasize the value of further research in this field. It is imperative that future studies on meditation practices be more rigorous in design, execution, and analysis, and in the reporting of the results. Greater importance should be placed on the reporting of study methods and providing detailed descriptions of the training of the participants, qualifications of meditation instructors, and on reporting the criteria and methods used to determine a successful meditation practice.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...