Meditation

What it is

Meditation here means mindfulness-based programmes, which teach you to pay attention on purpose to the present moment with a non-judgemental, accepting awareness [1]. The best-known are mindfulness-based stress reduction (MBSR) and mindfulness-based cognitive therapy (MBCT) [2]. The courses in two of the harm studies ran for 8 weeks [3][4]. Transcendental meditation, a separate technique, is also covered in this page [5].

How it might work

These are proposed mechanisms, not established ones. A 2024 review of 22 trials in people with schizophrenia spectrum disorders described attention training, emotion and stress regulation, decentering, self-compassion and cognitive restructuring as common mechanisms across the programmes [6].

What the research says

Preventing depression from coming back. This is the strongest finding. Pooled individual data from 9 trials (1,258 people with recurrent depression) found that MBCT lowered the risk of relapse over 60 weeks: hazard ratio 0.69 compared with usual care and other treatments, and 0.79 compared with active treatments [7]. A 2021 network meta-analysis of 14 trials (2,077 people) found a relapse rate about 27% lower than usual care (risk ratio 0.73), with no clear difference between MBCT and active treatments [2]. The two reviews disagree about active treatments: one found a small advantage for MBCT (HR 0.79), the other no clear difference [7][2].

Depression symptoms. A UK trial of 234 people whose depression had not fully gone after standard talking therapy found that MBCT plus usual care lowered PHQ-9 depression scores by 2.49 points more than usual care alone at 34 weeks (95% CI 1.09 to 3.89 points) [8]. A 2019 Cochrane review of women with breast cancer found a moderate short-term drop in depression with MBSR (SMD -0.54, 6 trials, high-certainty evidence). The drop was smaller but still present up to six months later, and there was no difference at up to two years [1]. A 2014 meta-analysis of 47 trials (3,515 people) found moderate evidence of lower depression at 8 weeks, a small effect of 0.30 (95% CI 0.00 to 0.59) [9].

Anxiety. The 2014 meta-analysis found moderate evidence of less anxiety compared with active controls: an effect of 0.38 at 8 weeks (95% CI 0.12 to 0.64) and 0.22 at 3 to 6 months [9]. In 36 studies of cancer patients, 20 of them without randomisation, anxiety fell by a medium amount compared with control groups (Hedges g 0.56) [10]. The 2024 Cochrane review found large reductions against waiting lists or no treatment (SMD -0.78, low certainty), but no difference from active treatments (SMD -0.06, moderate certainty) [5].

Stress. The 2024 Cochrane review found that perceived stress fell slightly more than with active comparators (SMD -0.24, moderate certainty) and more than with no-treatment controls (SMD -0.59, low certainty) [5]. The 2014 review rated the evidence on stress and distress as low [9].

Sleep. A 2019 meta-analysis of 18 trials (1,654 people) found better sleep quality than time-matched controls (effect 0.33 after treatment, 0.54 at follow-up), but no difference from specific sleep treatments [11]. In the breast cancer review, sleep quality improved slightly (SMD -0.38), with a confidence interval that includes no difference [1].

Chronic pain. The 2014 meta-analysis found moderate evidence of less pain (effect 0.33, 95% CI 0.03 to 0.62) [9]. A 2017 review of 38 trials found a small decrease in pain compared with all types of control, based on low-quality evidence [12].

Blood pressure. A 2017 meta-analysis of 19 trials found that non-transcendental meditation lowered systolic pressure by about 5 mmHg in readings taken without ambulatory monitors, and by about 4 mmHg in ambulatory monitoring [13]. Transcendental meditation lowered systolic pressure by about 2 mmHg compared with active controls, with moderate certainty [5]. Mindfulness trials showed larger but uncertain drops, and some estimates included no effect [5]. Heart attacks, strokes and deaths were reported in only one trial, so the effect on them is unknown [5].

Attention and focus. A 2020 meta-analysis of 34 studies found that a brief mindful attention exercise gave a small improvement in cognition overall (Hedges g 0.18), with a larger effect on complex thinking tasks (g 0.35); the authors noted many methodological flaws [14]. For longer programmes, the 2014 review found no clear effect on attention, or too little evidence to say [9]. Four small trials in ADHD (83 people) could not show whether meditation helps [15].

Compared with other treatments. A broad 2013 meta-analysis of 209 studies found mindfulness therapy moderately effective overall, with larger effects against waiting lists (g 0.53) than against other active treatments (g 0.33) [16]. A 2014 review found no evidence that meditation programmes were better than any active treatment such as drugs, exercise or other therapies [9]. An umbrella review of 44 meta-analyses (30,483 people) found that mindfulness programmes beat passive controls, but effects were smaller and less often significant against active controls [17].

Doses used in studies

The courses in the two harm studies ran for 8 weeks [3][4]. The 2014 review measured effects at 8 weeks and again at 3 to 6 months [9]. Relapse trials followed people for up to 60 weeks [7], and the UK depression trial measured outcomes at 34 weeks [8]. The 2024 Cochrane review included trials lasting at least 12 weeks [5]. Daily practice time is not given here because the abstracts read did not report it [3].

Safety and interactions

Unpleasant experiences are common in studies. In a survey of 434 US adults who had meditated, 32% reported at least one adverse effect. The most common were anxiety, traumatic re-experiencing and emotional sensitivity. About 10% had an effect lasting a month or more, and 1.2% had impairment lasting that long [18]. In 96 people on MBCT, 83% reported at least one meditation-related effect and 58% reported a negative one; lasting bad effects occurred in 6 to 14% [3]. Among 1,232 regular meditators, 25.6% reported particularly unpleasant experiences [19]. In two uncontrolled mindfulness courses (158 people), 3 to 7% reported harm from the course [4]. A qualitative study described 59 types of meditation-related experience, ranging from minimal and transient to severe and enduring [20].

Retreats and trauma. People who had attended a meditation retreat were more likely to report unpleasant experiences [19]. In the same survey, childhood adversity was linked to a higher risk of these effects [18].

Psychosis. Case reports describe psychotic disorders in people who meditated, and a 2022 review says meditation may trigger or worsen psychotic states [21]. In its 28 cases, 14 had triggers such as insomnia, lack of food, stress or drug use, and a causal link is hard to prove [21]. A 2024 systematic review of 22 trials in people with schizophrenia spectrum disorders reported possible benefits, such as fewer hospital admissions, but found the programmes too varied to interpret clearly [6].

Trial harms. Reporting of adverse effects in trials is inconsistent [17]. The 2025 UK depression trial reported no trial or treatment-related serious adverse events [8].

Upper limits and toxic doses. None of the studies read set an upper limit or toxic dose [17].

Medicine interactions. None of the studies read tested interactions with medicines [17].

Pregnancy and breastfeeding. The studies read did not report on pregnancy or breastfeeding [17].

Other conditions. No studies read addressed kidney, liver or bleeding disorders [17]. Among the 28 case-report patients, 14 had acute psychosis and 7 had schizophrenia [21].

Caveats and open questions

Blinding is not possible. Participants know whether they are meditating, so the breast cancer review rated all trials at high risk of performance and detection bias [1]. Many trials compare meditation with waiting lists or usual care, which can make benefits look larger than they are against active treatments [17].

Effects are smaller against active treatments. The 2024 Cochrane review found little or no difference from active treatments for anxiety or depression [5]. The 2014 review found no evidence that meditation programmes were better than any active treatment [9].

Results are inconsistent. Blood pressure results and several mental-health results showed substantial heterogeneity between trials [5].

Long-term effects are unclear. In women with breast cancer, anxiety and depression showed little or no difference up to two years after MBSR [1]. More well-designed, large trials are needed [12].

Substance use is uncertain. For alcohol and drug use, a Cochrane review found the evidence uncertain compared with no treatment, and only a slight reduction in days of substance use compared with other treatments [22].

Funding and conflicts. Several key reviews and trials were run by researchers who teach or run mindfulness programmes, and some receive fees for workshops [7][8]. All but two authors of the 2016 relapse review were investigators on the original trials [7]. One author of the 2024 Cochrane review is also a mindfulness teacher, which the review declares [5].

Outcomes. Most benefits were measured with self-report scales, and heart outcomes were rarely measured [5].