Twenty-one trials (sample size range 37 to 172 participants) were included in the review. Follow-up ranged from the end of treatment to 15 months. Twelve trials scored 3 on the Jadad scale. Four trials reported the use of two or more therapists. A number of trials reported independent assessment of adherence and competence of therapist (details unclear). About half of the trials used a power calculation. Intention to treat was used in 16 trials. Descriptions of concomitant treatments and adherence to treatment manual appeared to be poorly reported.
Medium effect sizes were reported in general.
Non-clinical populations (four trials): For mindfulness-based stress reduction in non-clinical populations, mental health was improved in all four trials. Physical health was improved in two trials.
Clinical populations with physical illnesses (11 trials): For mindfulness-based stress reduction in clinical populations with physical illnesses, six of nine trials showed significant improvements in mental health compared to control group. Two of six trials reported significant improvements in physical health.
Clinical populations with psychiatric disorders (six trials): For combined mindfulness-based stress reduction/cognitive therapy in clinical populations with psychiatric disorders (six trials), there were higher improvements in the mindfulness interventions where this was compared with active control in both mental and physical outcomes, but the higher effect was not sustained at four weeks follow-up.
Specifically, compared with control, mindfulness-based stress reduction was associated with reduced perceived stress and/or psychological distress (seven trials), reduced depressive symptoms (10 trials), and reduced anxiety symptoms (six trials).
The authors stated that publication bias could not be ruled out.