Literature Sources

Comprehensive searches were conducted in 17 relevant electronic databases up to September 2005. Other sources of potentially relevant studies included hand searches, reference tracking, contact with experts, and gray literature searches.

Study Selection

A set of strict eligibility criteria was used to include potentially relevant studies. They had to be comparative, be on any meditation practice, have more than 10 adult participants, provide quantitative data on health-related outcomes, and be published in English. The criteria of study methodology were modified to address each of the research topics of the review. Sources of secondary data (e.g., systematic reviews, narrative reviews, and book chapters) were used for topic I. Topics II to V included studies with a comparison/control group or control period: randomized controlled clinical trials (RCTs), nonrandomized controlled clinical trials (NRCTs) (topics III to V), prospective and retrospective observational studies with controls (topic II), case-control studies (topic II), uncontrolled before-and-after studies (topics II and V), and cross-sectional studies with controls (topic II).

Data extraction and Assessment of Study Quality

Trained research assistants extracted the data using a comprehensive and pretested data extraction form. One reviewer verified the accuracy and completeness of the data.

Studies included in the descriptive overview on the practice of meditation (topic I) were not assessed for methodological quality. For topics II to V, the methodological quality of RCTs and NRCTs was assessed using the criteria for concealment of allocation and the Jadad Scale. The quality of observational analytical studies (e.g., prospective and retrospective observational studies, case-control studies, and cross-sectional studies with controls) was assessed using the Newcastle-Ottawa Scales (NOS). The quality of the before-and-after studies was evaluated against four criteria adapted from the NOS.

Two independent reviewers assessed study relevance, extracted the data and assessed the methodological quality of the studies. Disagreements among reviewers were adjudicated by a third reviewer.

Synthesis of the Evidence

Data for topic I on the practice of meditation were synthesized qualitatively. A combination of qualitative and quantitative approaches was used to synthesize the data in Topics II to V. Details of individual studies were summarized in evidence tables including information on the article source, study design, study population (e.g., sample size, age, and gender), treatment groups, and outcomes. Meta-analyses using the standard inverse variance and random effects model were planned to derive pooled estimates from individual studies to support inferences regarding the magnitude and direction of the effect of meditation practices. Forest plots were used to display the individual and pooled results. An analysis of publication bias was also planned.

Topic I. The Practice of Meditation

Five broad categories of meditation practices were identified in the included studies: Mantra meditation (comprising the Transcendental Meditation^® technique [TM^®], Relaxation Response [RR], and Clinically Standardized Meditation [CSM]), Mindfulness meditation (comprising Vipassana, Zen Buddhist meditation, Mindfulness-based Stress Reduction [MBSR], and Mindfulness-based Cognitive Therapy [MBCT]), Yoga, Tai Chi, and Qi Gong. Given the variety of the practices and the fact that some are single entities (TM^®, RR, and CSM, Vipassana, MBSR, and MBCT) while others are broad categories that encompass a variety of different techniques (Yoga, Tai Chi, Qi Gong), 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 the 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 the successful practice of meditation. The amount of variation in the described frequency and duration of practice make it difficult to draw generalizations about the training requirements for meditation techniques. The criteria for successful meditation practice have also not been described well in the literature.

Topic II. State of Research on the Therapeutic Use of Meditation Practices in Healthcare

Eight hundred and thirteen studies provided evidence regarding the state of research on the therapeutic use of meditation practices. The studies were published between 1956 and 2005, with half of the studies published after 1994. Most of the studies were published as journal articles. Studies were conducted mainly in North America (61 percent). Of the 813 studies included, 67 percent were intervention studies (286 RCTs, 114 NRCTs and 147 before-and-after studies), and 33 percent were observational analytical studies (149 cohort and 117 cross-sectional studies).

Quality of studies. Overall, we found the methodological quality of meditation research to be poor, with significant threats to validity in every major category of quality measured, regardless of study design. The majority of RCTs did not adequately report the methods of randomization, blinding, withdrawals, and concealment of treatment allocation. Observational studies were subject to bias arising from uncertain representativeness of the target population, inadequate methods for ascertaining exposure and outcome, insufficient followup period, and high or inadequately described losses to followup.

Meditation practices. Mantra meditation practices such as the TM® technique and the RR were the most frequently studied meditation practices. Other mantra practices such as CSM, Acem meditation, Ananda Marga, concentrative prayer, and Cayce's meditation have been examined less frequently. The second category of meditation practices most frequently examined is Yoga. It includes a heterogeneous group of techniques such as Hatha yoga, Kundalini yoga, and Sahaja yoga. Mindfulness meditation, which includes MBSR, MBCT, and Zen Buddhist meditation, constitutes the third most studied group of meditation practices, Tai Chi the fourth, and Qi Gong the fifth. Finally, less than 5 percent of the studies on meditation have failed to explicitly describe the meditation practice.

Control groups. The number of control groups used in the 668 controlled studies ranged from one to four. The majority of the studies utilized an active, concurrent control. Among the RCTs and NRCTs, the practice of exercise and other physical activities constituted the most frequent active comparator followed by conditions involving states of rest and relaxation, health education, and progressive muscle relaxation. Almost half of the RCTs and NRCTs included comparison groups consisting of participants assigned to waiting lists, or participants that did not receive any intervention. The vast majority of observational studies used comparison groups consisting of individuals that had not been exposed to any type of meditation practice.

Study population. The majority of studies on meditation practices have been conducted in healthy populations. The three most studied clinical conditions are hypertension, other cardiovascular diseases, and substance abuse. Other diseases that have been frequently examined include anxiety disorders, depression, cancer, asthma, chronic pain, type II diabetes mellitus, and fibromyalgia.

Outcome measures. Physiological functions, particularly cardiovascular outcomes, were the most frequently reported outcome of interest in meditation research. Psychosocial outcomes, outcomes related to clinical events and health status, cognitive and neuropsychological functions, and healthcare utilization outcomes have also been evaluated in studies of meditation practices.

Topic III. Evidence on the Efficacy and Effectiveness of Meditation Practices

We summarized the evidence from RCTs and NRCTs on the effects of meditation practices for the three most studied clinical conditions identified in the scientific literature: hypertension (27 trials), other cardiovascular diseases (21 trials), and substance abuse disorders (17 trials).

A few studies of overall poor methodological quality were available for each comparison in the meta-analyses, most of which reported nonsignificant results. TM^® had no advantage over health education to improve measures of systolic blood pressure and diastolic blood pressure, body weight, heart rate, stress, anger, self-efficacy, cholesterol, dietary intake, and level of physical activity in hypertensive patients; RR was not superior to biofeedback in reducing blood pressure in hypertensive patients; Yoga did not produce clinical or statistically significant effects in blood pressure when compared to nontreatment; Zen Buddhist meditation was no better than blood pressure checks to reduce systolic blood pressure in hypertensive patients. Yoga was no better than physical exercise to reduce body weight in patients with cardiovascular disorders. When the relative effectiveness of a variety of meditation practices was assessed using indirect meta-analysis, we found that there were no significant differences between MBSR and Yoga to control anxiety symptoms in cardiovascular patients. Meta-analysis of the effects of meditation practices for substance abuse was not possible due to the diversity of practices, comparison groups, and outcome measures reported in each of the studies reviewed.

The results of the three highest quality trials (Jadad score = 3/5) examining, respectively, Mindfulness meditation, RR, and Yoga are inconclusive with respect to the effectiveness of meditation pratices.

The study comparing Mindfulness meditation with usual care (NS) for alcohol and cocaine abuse found little indication that Mindfulness meditation enhanced treatment outcomes for substance abuse patients. The study comparing RR with PMR and rest groups for alcohol abuse found generalized effects for BP, but not for the other outcome measures (anxiety, HR, and GSR). The RR and PMR groups did not exhibit increased BP as observed in control subjects. RR and PMR produced significant changes in tension. The study comparing Yoga with exercise for alcohol abuse found a significantly greater recovery rate for the Yoga group.

Statistical and clinical heterogeneity among the trials constituted a frequent and considerable problem when pooling the results, and in some cases, it precluded summarizing data across the studies. The poor methodological quality of the trials limits the strength of inference regarding the observed treatment effects reported in this review. The lack of description of the methods of allocation concealment, randomization, description of withdrawals and dropouts per treatment group, the absence of appropriate blinding, and the use of incompatible or inappropriate control groups undermine the validity of the results of many clinical studies.

Topic IV. Evidence on the Role of Effect Modifiers for the Practice of Meditation

The role of patient or meditation characteristics as effect modifiers in the practice of meditation is a topic that has so far been neglected in the scientific literature. Few studies have systematically examined factors such as dose, duration, or other specific features of meditation as moderators of the effects on outcomes. Evidence from RCTs and NRCTs regarding the interaction of meditation 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. No conclusions on the role of effect modifiers can be drawn from the analysis of the individual studies. Individual patient data is required to appropriately examine this issue.

TOPIC V. Evidence on the Physiological and Neuropsychological Effects of Meditation Practices

The physiological and neuropsychological effects of meditation practices were evaluated in 311 studies. The majority of studies have been conducted in healthy participants. 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 cholesterol. 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 practices decrease visual reaction time, intraocular pressure, and increase breath holding time. As found in studies included for topic III, the overall low methodological quality of the studies indicates that most of the studies suffered from methodological problems that may result in overestimations of the treatment effects or compromise the generalizability of the study results. Particularly, the lack of a concurrent control group in the before-and-after studies results in an inability to control for temporal trends, regression to the mean, and sensitivity to methodological features. Therefore, results from meta-analyses of the physiological and neuropsychological effects of meditation practices should be interpreted cautiously.

The very small number of trials available for each comparison precluded testing for publication bias.