What do we know about harm in mindfulness-based interventions?

Psychological therapy is one of the key pillars in the treatment of a wide range of mental illnesses, but the potential adverse effects of psychotherapies still remain less systematically studied than those of medications and medical devices and the field lacks agreed terminology. In this context, decisions about what to provide to whom may be more influenced by presumptions about what may happen than informed by the available data. A variety of outcomes have been identified as potential adverse effects of psychotherapy, such as a lack of clinical improvement, symptom deterioration, development of new symptoms, fear of seeking future treatment and maladaptive dependence on therapists.^{Reference Duggan, Parry, McMurran, Davidson and Dennis2}

Duggan et al have defined harm as any ‘sustained deterioration that is caused directly by the psychological intervention’.^{Reference Duggan, Parry, McMurran, Davidson and Dennis2} Using this definition, it is therefore important to distinguish between short-term discomfort, which can be considered part of therapeutic growth, and serious or long-term harms (serious adverse effects). What we encounter in mindfulness practice is often challenging, as we explore deeper through the many layers of our human experience, much of which we may be used to keeping shut away. Mindfulness therefore represents a journey of reacquaintance with difficult thoughts and emotions that we may have become disconnected from. Short-lived periods of discomfort should therefore be the rule, rather than the exception, for people engaging in mindfulness-based interventions. However, a recent review of clinical trials of such interventions concluded that adverse effects were overall rare (reported by 0–10% of all participants) and not found to be directly attributable to the mindfulness intervention.^{Reference Baer, Crane, Miller and Kuyken3} The review's authors therefore concluded that mindfulness can be uncomfortable without being harmful.

Why specific concerns about mindfulness for psychosis?

Historical concerns about use of mindfulness in psychosis can perhaps be linked to the over-reliance on early uncontrolled case study evidence which suggested that people developed psychotic episodes after taking part in meditation retreats.^{Reference Dyga and Stupak1} For example, Walsh & Roche^{Reference Walsh and Roche4} report three cases of apparent psychotic relapse in people diagnosed with schizophrenia who had taken part in intensive meditation retreats. The retreats were described as involving ‘many hours each day of sitting and walking meditation and total silence, without communication of any kind (even eye contact)’ and up to ‘18 hours of meditation a day’ over the course of a 2-week retreat. This bears little resemblance to the intensity and mode of delivery that would be typical of mindfulness-based interventions in mental healthcare settings. Furthermore, we know that sensory deprivation, social isolation, fasting and sleep deprivation are themselves risk factors for relapse in psychosis, and so these are likely to have been the riskiest aspects of the retreat rather than the meditation itself.

What does the current literature on mindfulness for psychosis show?

There have been at least a dozen randomised trials of mindfulness-based interventions for psychosis, with a combined sample size of over 1300, as well as several other trials of interventions in which mindfulness exercises are used (e.g. acceptance and commitment therapy; compassion-focused therapy). A recent meta-analyses of trial results indicated that interventions including mindfulness as a primary component show beneficial effects in various symptom dimensions, compared with treatment-as-usual and waiting-list conditions.^{Reference Jansen, Gleeson, Bendall, Rice and Alvarez-Jimenez5} This indicates that, on average, mindfulness is more likely to be beneficial than harmful in psychosis. Furthermore, examination of reports of study withdrawals showed no evidence of withdrawals arising from psychotic exacerbation during therapy, and trials have consistently shown a similar rate of withdrawals between mindfulness and control arms. Studies have also shown rates of treatment drop-out from mindfulness-based interventions to be no greater than for compared treatments. A series of studies have additionally tracked hospital admissions following mindfulness integrated with psychoeducation and found that mindfulness was associated with similar or lower hospital admission rates during and following therapy compared with control groups.

What can we take from this?

The belief that it is dangerous to invite people with psychosis to intentionally ‘turn towards’ their experiences remains present among clinicians and is perhaps connected to fears that it is dangerous to talk to people about their delusional beliefs or unusual experiences because this exacerbates their symptoms. There is a rich history of the ‘othering’ of psychotic experiences in psychiatry as something enigmatic and apart from normal human experience. However, patient advocacy groups have fought hard to overturn this fallacy, to be viewed not as less-than-human enigmas but as fully human beings. Turning towards our inner experiences may often be challenging, but there is no rationale for psychotic symptoms such as voices or delusions to be placed in a separate ‘category’ apart from normal human experience. In fact, there is now increasing evidence for the use of mindfulness in the treatment psychosis that reports safety, acceptability and clinical efficacy.

People with psychosis are among the most marginalised and discriminated against in society, and unfortunately experience significant inequalities both in their physical health outcomes, for example reduced life expectancy, and in the insufficient access they often have to evidence-based psychological therapies. A paternalistic attitude among clinicians about ‘protecting’ people from potentially harmful effects of psychological treatments can no longer be considered benevolent if these attitudes are unfounded and based on assumptions about the ‘otherness’ of psychotic experiences. We argue that early uncontrolled case studies reporting harmful effects of meditation for people with psychosis may have coalesced with ongoing clinician concerns about the viability of psychotherapeutic approaches in the treatment of psychosis. This may have led to an underdevelopment of the evidence base over the past 20 years, particularly with respect to a paucity of large, well-conducted randomised controlled trials of mindfulness for psychosis. There is, however, clear interest in the question of the effectiveness of mindfulness for psychosis, given that numerous meta-analyses and systematic reviews have been published in the field (at least 10 review articles since 2013). Existing data show that mindfulness for psychosis can be implemented safely, if delivered by experienced clinicians with appropriate mindfulness training and supervision, in both in-patient and community routine care settings. Adaptations to delivering mindfulness for psychosis include delivery in smaller groups, shorter duration of meditation, avoiding prolonged periods of silence, and using basic anchoring techniques and easily accessible and simple language.

For future research, we suggest a priority should be better systematic assessment of side-effects and adverse events of mindfulness for psychosis in large trials, in line with the recommendations of Baer et al,^{Reference Baer, Crane, Miller and Kuyken3} who noted that the potential harms of mindfulness are an under-researched topic. This will facilitate an improved understanding of the processes and mechanisms of psychotherapeutic change within mindfulness-based interventions for psychosis, to allow the field to develop and evolve according to evidence-based practice. In the interim, we recommend that people with psychosis are not denied access to mindfulness-based treatments purely on the basis of their diagnosis and/or symptom profile.