Systematic review of the safety of mindfulness-based interventions for psychosis

May combine two claims.

Overall, reporting of harm was inconsistent across studies and the quality of data collection and reporting varied.

“Overall, reporting of harm was inconsistent across studies and the quality of data collection and reporting varied.”

May combine two claims.

The review included 39 studies with a total of 2684 participants.

“The review included 39 studies, with a total n of 2684 participants across studies.”

• Also stated in: § Characteristics of Included Studies · plain

  “This systematic review identified 39 relevant studies.”

• Also stated in: § Characteristics of Included Studies · plain

  “The total number of participants across the studies was 2684.”

May combine two claims.

The percentage of studies reporting on each harm index and the prevalence of harm varied greatly across indices.

“The percentage of studies reporting on each index of harm, and the prevalence of harm, varied greatly across each index.”

• Also stated in: § Prevalence of Harm Indices · plain

  “The percentage of studies reporting on the different indices of harms varied substantially, with 0% of studies reporting on side effects of interventions and 92% reporting on study dropout.”

May combine two claims.

Quality scores ranged from 7 to 22 with a mean score of 16 (SD = 4).

“The scores ranged from 7 to 22 with a mean score of 16 (SD = 4).”

May combine two claims.

In no study was a reported death attributed by the authors to either trial participation or suicide, though not all studies made an explicit comment on likely cause.

“In no study was a reported death attributed by the authors to either trial participation or suicidethough not all studies made an explicit comment on this question of likely cause.”

May combine two claims.

Of the 11 studies reporting a data collection method, 10/11 (91%) used review of clinical or insurance records and one used outcome assessments.

“Of these, 10/11 (91%) used review of clinical or insurance records (depending on country) and one study used outcome assessments to collect harm-related data.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

LE conducted the re-run of the search in December 2023, identifying one additional study that met eligibility criteria.

“LE conducted the re-run of the search (using the same databases and search terms) in December 2023, which identified one additional study that met eligibility criteria and was therefore included in the review.”

May combine two claims.

Non-completion of treatment was reported in 14/39 (36%) of included studies; prevalence was 30/501 (6%) in the MBI group and 41/501 (8%) in the control group.

“| Non-Completion | 14/39 | 30/501 | 41/501 |”

May combine two claims.

Study dropout was reported in 36/39 (92%) of included studies; prevalence was 131/1101 (12%) in the MBI group and 141/1096 (13%) in the control group.

“| Study Dropout | 36/39 | 131/1101 | 141/1096 |”

May combine two claims.

Utilisation of crisis services was reported in 2/39 (5.3%) of included studies; prevalence was 2/47 (4.3%) in the MBI group and 9/44 (20.5%) in the control group.

“| Utilisation of | 2/39 | 2/47 | 9/44 |”

May combine two claims.

Adverse events recorded was reported in 19/39 (49%) of included studies; prevalence was 8/604 (1%) in the MBI group and 11/601 (2%) in the control group.

“| Adverse Events | 19/39 | 8/604 | 11/601 |”

May combine two claims.

Symptom deterioration was reported in 2/39 (5.3%) of included studies; prevalence was 1/20 (5.0%) in the MBI group and 2/20 (10.0%) in the control group.

“| Symptom | 2/39 | 1/20 | 2/20 |”

May combine two claims.

Hospitalisation was reported in 14/39 (36%) of included studies; prevalence was 62/406 (15%) in the MBI group and 159/477 (33%) in the control group.

“| Hospitalisation | 14/39 | 62/406 | 159/477 |”

May combine two claims.

Death was reported in 9/39 (23%) of included studies; prevalence was 6/362 (1.7%) in the MBI group and 4/354 (1.1%) in the control group.

“| Death | 9/39 (23%) | 6/362, 1.7% | 4/354, 1.1% |”

Classified as probably not a substantial claim.

The absence of thorough reporting on harm precludes a balanced analysis of benefits versus harms.

“However, the absence of thorough reporting on harm precludes a balanced analysis of benefits versus harms.”

May combine two claims.

Jansen et al. (2020) noted that no study in their review reported any serious adverse events and only 44% of studies reported on adverse events of any kind.

“Again, in relation to MBIs for psychosis specifically, Jansen et al. (2020) noted that no study in their review reported any serious adverse events and only 44% of the studies reported on adverse events of any kind.”

• cites Jansen et al., 2020 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The appropriateness of the loaned definition of serious adverse events has been called into question because it includes outcomes unlikely in psychotherapy.

“However, the appropriateness of the loaned definition of SAEs has been called into question as this includes, for instance, birth defects which appear to be unlikely implications of psychotherapy interventions.”

• cites Baer et al., 2019 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

There is marked variation in the terms used to define harm, with a precise definition of adverse events lacking.

“There is marked variation in the terms used to define harm, with a precise definition of adverse events (AEs) lacking.”

• Also stated in: § Introduction · plain

  “Vaughan et al. (2014), who found that adverse events lacked a standardised definition”

• cites Vaughan et al., 2014 sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Ellett and Chadwick (2021) recommend that data from randomised controlled trials be prioritised when drawing generalisable conclusions regarding harm.

“The first recommendation is for data from randomised controlled trials to be prioritised when drawing generalisable conclusions regarding harm.”

• cites Ellett and Chadwick, 2021Not checked yet

  Not checked in the free run.

May combine two claims.

A recent systematic review of cognitive behavioural therapy for insomnia found that only 32.3% of studies addressed adverse events and 7.1% met criteria for adequate reporting.

“a recent systematic review reporting on adverse events in cognitive behavioural therapy for insomnia, which found that only 32.3% of studies addressed adverse events in any way, with an even smaller 7.1% of studies meeting criteria for adequate reporting of adverse events (Condon et al., 2021).”

• cites Condon et al., 2021 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Psychopharmacological trials were significantly more likely to report on harm than psychotherapy trials.

“Vaughan et al. (2014), who found that adverse events lacked a standardised definition as well as finding that psychopharmacological trials were significantly more likely to report on harm than psychotherapy trials.”

• cites Vaughan et al., 2014 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

A recent systematic review and meta-analysis found no difference in reported adverse events in MBSR and MBCT compared to control conditions, but did not include trials of MBIs for psychosis.

“This review found no difference in reported adverse events in MBSR and MBCT compared to control conditions, however they did not include trials of MBIs for psychosis in their review (Wong et al., 2018).”

• cites Wong et al., 2018 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

There are two primary problems with assessment of the safety of psychological interventions in trials: lack of a clear definition or operationalisation of harm, and lack of consistent monitoring and reporting on harm.

“This points to two primary problems with respect to assessment of the safety of psychological interventions in trials.Firstly, the lack of a clear definition or operationalisation of harm that is relevant and comprehensively covers the range of potential adverse outcomes within the emotional, cognitive, social, functioning, and physiological domains.”

• Also stated in: § Introduction · plain

  “Secondly, the lack of consistent monitoring and reporting on harm in research trials with a view to ultimately being able to conclusively decide how safe an intervention is.”

Classified as probably not a substantial claim.

Ellett and Chadwick (2021) recommend indices of harm: hospitalisation, crisis service use, serious adverse events, adverse events, study drop out, noncompletion of therapy, symptom deterioration and side effects of therapy.

“The remaining items are recommended indices of harmhospitalisation, crisis service use, SAEs, AEs, study drop out, noncompletion of therapy, symptom deterioriation and side effects of therapy”

• cites Ellett and Chadwick, 2021Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The protocol was pre-registered on PROSPERO on 31 August 2021 (ID: CRD42021274649).

“The protocol was pre-registered on PROSPERO on 31 st August 2021 (ID: CRD42021274649).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data extraction was conducted by BV, SB and LE; 38/39 (97%) studies were independently double extracted by SB and BV.

“Data extraction was conducted by BV, SB and LE.38/39 (97%) studies were independently double extracted by SB and BV.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Title and abstract screening and full text review were conducted independently by BV and SB, with consensus for conflicts reached through discussion.

“For the original search conducted in 2021, both title and abstract screening and full text review were conducted independently by BV and SB and consensus for conflicts was reached through discussion and reference to the inclusion and exclusion criteria.”

Probably not a substantial claim: may not pass S3 (load-bearing).

BV systematically re-checked the data extraction of numerical data related to harm for 38 of the 39 studies to reduce the risk of errors.

“BV systematically re-checked the data extraction of numerical data related to harm for 38 of the 39 studies to reduce the risk of errors.”

Probably not a substantial claim: may not pass S3 (load-bearing).

An amendment to the protocol was submitted to PROSPERO on 28 February 2022 because not enough studies reported on some indices of harm to allow meta-analyses of risk differences.

“Subsequently, an amendment to the protocol was submitted to PROSPERO on 28 th February 2022.This amendment reflected the identification, following the data extraction process, that not enough studies had reported on some indices of harm such that meta-analyses of risk differences were possible.”

Probably not a substantial claim: may not pass S2 (checkable).

If these harm indices are systematically monitored and reported on in trials, they would provide sufficient information for clear conclusions related to harm to be drawn.

“which, if systematically monitored and reported on in trials, would provide sufficient information for clear conclusions related to harm to be drawn.”

• cites Ellett and Chadwick, 2021Not checked yet

  Not checked in the free run.

May combine two claims.

Heterogeneity in MBIs makes it harder to interpret data on noncompletion of therapy, compounded by studies setting different cutoff rates of attendance to define completion.

“This heterogeneity in MBIs also makes it harder to interpret data on the index, noncompletion of therapya challenge that is compounded by studies setting different cutoff rates of attendance to define completion.”

May combine two claims.

The findings from the meta-analyses need to be interpreted cautiously given heterogeneity in definition and reporting of adverse events and the low number of studies reporting on some indices.

“The findings from the meta-analyses need to be interpreted cautiously, given the heterogeneity in definition and reporting of adverse events, and given the low number of studies reporting on some indices (e.g.symptom deterioration and use of crisis services).”

May combine two claims.

Study drop-out can occur for a variety of reasons including dissatisfaction or harm experienced during therapy, circumstantial reasons, or early success of therapy, and it will not always be possible in randomised studies to ascertain which applies.

“For example, study drop-out can occur for a variety of reasonsdissatisfaction/harm experienced during therapy, or circumstantial reasons, or even early success of therapy (Lilienfeld, 2007) -and it will not always be possible in randomised studies to ascertain which applies.”

• cites Lilienfeld, 2007 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Even for indices where data were clearest (e.g. hospitalisation), the findings must remain provisional, and for some indices (e.g. symptom deterioration and side effects) it is too soon to draw even tentative conclusions.

“Nonetheless, even for indices where the data were clearest (e.g.hospitalisation), the findings of this review must remain provisionaland for some indices (e.g.symptom deterioration and side effects) it is too soon to draw even tentative conclusions.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

The picture in relation to MBIs for psychosis reflects a wider tendency for psychotherapy researchers to have been more attentive to assessment of benefits than to assessment of harm.

“Indeed, this picture in relation to MBIs for psychosis reflects a wider tendency for psychotherapy researchers to have been more attentive to assessment of benefits than to assessment of harm.”

May combine two claims.

Many possible causes of dropout are unrelated to harm, which limits confidence in attributing dropout rates to an intervention.

“Also, as noted above, there are many possible causes of dropout that are unrelated to harm, and this limits confidence in attributing dropout rates to an intervention”

May combine two claims.

A broad range of control conditions was employed across the 39 studies, and the review did not compare harm indices by type of control condition.

“Finally, a broad range of control conditions was employed across the 39 studies (the most common being treatment as usual) and we did not compare harm indices by type of control condition employed.”