Engagement, clinical outcomes and therapeutic process in online mindfulness for psychosis groups delivered in routine care

Abstract

Objectives

There is growing evidence for the benefit of mindfulness‐based interventions (MBI) for people with psychosis. However, research is yet to evaluate the clinical benefit of delivering MBI groups online. We examine engagement, clinical outcomes, participant experience and therapeutic process of delivering therapy groups online in routine clinical practice.

Methods

The study used an uncontrolled pre‐post design to examine engagement, therapeutic benefits (depression, anxiety, beliefs about voices) and group process in a 12‐session online mindfulness group for individuals with a schizophrenia spectrum diagnosis with current distressing voices. Qualitative data on participant experience of online group therapy were analysed using Thematic Analysis.

Results

17/21 participants (81%) completed one of three consecutively run therapy groups. For completers there were significant reductions pre‐post in depression, anxiety, beliefs about voices and voice‐related negative affect, with medium to large effect sizes. There were individuals showing reliable and clinically significant improvements in each clinical outcome, and none showing reliable or clinically significant deterioriation. Participants’ rankings of the importance of different group therapeutic factors were very similar to those observed in face‐to‐face mindfulness for psychosis groups. Qualitative analysis of participant feedback identified three themes: ‘experience of online delivery’, ‘therapeutic benefits’ and ‘feeling connected to people in the group’.

Conclusions

Findings in relation to therapy engagement, clinical benefits, participant experience and group process offer encouragement that online delivery of mindfulness for psychosis groups may be a useful addition to mental health services for people with distressing voices.

Practitioner points.

• Eighty‐one percent completed online mindfulness‐based therapy groups in routine clinical practice.

• We found significant reductions in depression, anxiety, and beliefs about voices, with medium‐large effect sizes.

• There were indicators that therapeutic group process is comparable to in‐person mindfulness groups for psychosis.

• There was no indication of increased risk of harm associated with online delivery.

INTRODUCTION

There is growing evidence for the benefit of mindfulness and acceptance‐based approaches for people with schizophrenia (Hodann‐Caudevilla et al., 2020; Jansen et al., 2020; Louise et al., 2018). The most recent meta‐analysis published in 2020 of 10 studies (n = 1094) of mindfulness‐based interventions (MBIs) concluded that MBIs combined with treatment as usual (TAU) are effective for the treatment of schizophrenia when compared with both TAU control groups and active treatment control groups (Hodann‐Caudevilla et al., 2020). In pre–post comparisons, MBIs generated moderate to large effects in reducing overall schizophrenia symptomatology (g = 0.72) and small to moderate effects in reducing positive symptoms such as hearing voices (g = 0.32; Hodann‐Caudevilla et al., 2020). Group interventions have also been shown to have larger effect sizes (g = 0.46) than individual interventions (g = 0.08; Louise et al., 2018), suggesting that group delivery is more effective.

One of these approaches is Person Based Cognitive Therapy (PBCT, Chadwick, 2006), which integrates mindfulness theory and practice with a Cognitive Behaviour Therapy (CBT) approach to working with symptoms and with schemata and the self. PBCT has been evaluated both in pilot and full randomized controlled trials (RCTs) for people with distressing voices and persecutory delusions (Chadwick et al., 2016; Ellett et al., 2020) and in routine clinical practice in two studies (Dannahy et al., 2011; Jones et al., 2021). However, what is yet to be assessed is the clinical benefit of delivering the therapy online.

There is increasing evidence from a number of RCTs of the benefits of online MBIs for a range of mental health outcomes (e.g. Boettcher et al., 2014; Ly et al., 2014; Pots et al., 2016). A recent meta‐analysis of 15 studies across a range of populations (physical health [n = 5], anxiety and depression [n = 3], nonclinical [n = 7]), showed online MBIs to have small and significant effects on depression (g = 0.29), anxiety (g = 0.22), well‐being (g = 0.23) and mindfulness (g = 0.32), and moderate effects on stress (g = 0.51; Spijkerman et al., 2016). However, research is yet to examine the potential benefits of online MBIs for people with psychosis.

To our knowledge, this is the first study to evaluate a group‐based mindfulness intervention for people hearing distressing voices that is delivered entirely online and in routine clinical practice. This study reports on the findings of offering PBCT groups in one NHS Trust in the South of England. The study addressed the following research questions: when delivered online and in routine clinical practice, (1) how well do individuals engage with PBCT groups? (2) what clinical outcomes are achieved by individuals who complete PBCT therapy? and (3) what is the participant experience of delivering groups online?

METHOD

Design

The study used an uncontrolled pre–post design to examine the therapeutic benefits of 12 sessions of group PBCT delivered online in routine clinical practice. As the therapy groups formed part of routine clinical practice, NHS Research Ethics Committee approval was not required (UK Policy Framework for Health and Social Care Research; Department of Health, 2017). The project was registered with the local NHS Audit Department and we obtained informed consent from all participants. While participating in the therapy, all individuals received TAU from secondary care community mental health teams, typically consisting of regular outpatient appointments with a care coordinator and psychiatrist, and medication.

Participants

Twenty‐one people were screened and offered a place in a group. Four people did not engage and are not included in the present study, which describes outcomes and process for those 17/21 people who attended at least six sessions. The sample consisted of six females and 11 males, all were of a White British ethnicity. Mean age was 41.73 (SD = 12.52), 12 were single, and five were married. In terms of employment status, one student, two retired, 12 unemployed, and two employed. All participants had a diagnosis of a schizophrenia spectrum disorder (confirmed by Consultant Psychiatrist) with current distressing voices. At screening, all 17 participants reported experiencing subjectively distressing, controlling voices, and this profile was confirmed by baseline mean scores on the Beliefs about Voices Questionnaire – Revised (BAVQ‐R; Chadwick et al., 2000) subscales for malevolence, omnipotence, and resistance (see Table 1).

TABLE 1.

Clinical outcomes following therapy

	Baseline Mean (SD)	Post‐therapy Mean (SD)	t statistic	Effect size (Cohen's d)	Reliable change (N)	Clinically significant change (N)
Depression	25.7 (7.3)	20.0 (6.0)	3.1**	0.8	6/14	6/14
Anxiety	18.0 (5.5)	13.8 (5.2)	2.5*	0.8	5/14	5/14
Malevolence (0–18)	16.5 (3.3)	13.5 (4.0)	3.0**	0.8	5/14	5/14
Omnipotence (0–18)	16.3 (4.9)	13.1 (4.9)	2.8**	0.7	3/14	3/14
Benevolence (0–18)	8.5 (2.8)	9.0 (3.5)	0.9	0.1	2/14	1/14
Engagement (0–24)	10.3 (4.3)	9.9 (5.0)	1.0	0.09	1/14	1/14
Resistance emotion (0–12)	9.2 (2.1)	7.7 (3.1)	2.5*	0.6	4/14	4/14
Resistance behaviour (0–15)	13.8 (3.3)	13.3 (3.1)	0.8	0.1	2/14	2/14

*sig at .05.

**sig at .01.

Measures

Patient Health Questionnaire‐9

This is a nine‐item self‐report measure of depression symptom severity (Kroenke et al., 2001). Items are rated on a 4‐point scale from 0 (not at all) to 3 (nearly every day), with a score range of 0–27. In terms of cut offs, scores under 10 are considered sub‐clinical, 10–14 mild, 15–19 moderate, and 20+ severe. The scale has good levels of reliability (Cronbach's alpha = 0.89), sensitivity (0.92), and specificity (0.78; Gilbody et al., 2007).

Generalized anxiety disorder assessment‐7

This is a 7‐item self‐report measure of generalized anxiety (Spitzer et al., 2006). Items are rated on a 4 point scale from 0 (not at all) to 3 (nearly every day), yielding a score range of 0–21. Scores under 5 are considered sub‐clinical, 5–9 mild, 10–14 moderate, and 15+ severe. The scale has good levels of reliability (Cronbach's alpha = 0.92), sensitivity (0.89), and specificity (0.82).

Beliefs about Voices Questionnaire – revised

This is a 35‐item measure of people's beliefs about their auditory hallucinations, and their emotional and behavioural reactions to them (Chadwick et al., 2000). Each item is rated on a 0 (disagree) to 3 (strongly agree) scale. Six subscales were computed: malevolence, omnipotence, benevolence (all score ranges 0–18), engagement (range 0–24), resistance emotion (range 0–12), and resistance behaviour (range 0–15).

Subjective assessment of therapeutic group factors

We assessed the subjective relative importance of seven nonspecific group factors (catharsis, hope, universality, interpersonal learning, cohesion, altruism, self‐discovery) and one specific skill (learning mindfulness). Statements supplied by Yalom (1995) relating to seven group therapeutic factors (e.g. altruism: ‘helping others and being important in their lives’) and one statement relating to mindfulness (‘using mindfulness to cope with distressing thoughts, images or voices’) were presented to participants at the end of therapy. Participants were asked to rank order the statements from most (1) to least (8) important aspects of their experience in the group.

Qualitative data

All participants provided feedback on their experience of undertaking the group therapy online. Participants were asked two questions: (1) what was your experience of attending the group online and (2) how did you feel being part of the group? Participant responses were transcribed verbatim and analysed using thematic analysis (Braun & Clarke, 2006). The analysis process comprised six phases, consistent with good practice guidelines (Braun & Clarke, 2006), and two additional checks were used in order to assess the quality of the analysis. First, as an initial reliability check, the two people involved in coding conducted a consensus review and appraisal of themes across the dataset. Second, to assess inter‐rater reliability, the sample quotations were independently allocated to the list of themes by two raters. This revealed a Kappa value of 0.85, indicative of an excellent level of agreement.

Procedure

Individuals were referred for therapy by their care coordinator. All 21 service users referred were screened by the same Clinical Psychologist on the telephone and offered a place in a group. All participants were then emailed a Qualtrics link to complete baseline questionnaires online prior to the start of therapy. All individuals were offered a space in one of three consecutive 12‐week PBCT groups that were delivered solely online using Zoom. Groups took place between August 2020 and July 2021. Each group session lasted 90 min and was facilitated by two Clinical Psychologists, who received weekly supervision from a therapist with expertise in mindfulness for psychosis. The therapy manual is published elsewhere (Chadwick, 2006) and has been used in previous in‐person mindfulness for psychosis research (Chadwick et al., 2016; Ellett et al., 2020). At the end of therapy, all participants were sent a second Qualtrics link by email to complete the post‐therapy clinical outcome measures online and were asked to provide feedback about their experience of attending the group online. Thus, throughout the entire patient journey, there was no in‐person contact.

RESULTS

Engagement with online group therapy

Of the 21 people offered a place in a group, four did not complete therapy (one from group 1, two from group 2, and one from group 3), indicating an 81% therapy completion rate. We report data from those 17 people who completed therapy (pre‐defined as attending at least six therapy sessions, consistent with previous research [Ellett et al., 2020]). All 17 participants who completed therapy completed pre‐group questionnaires, and 14 completed post‐therapy questionnaires, indicating 82% data completeness. The majority of the sample (n = 15, 88%) attended between nine and 12 therapy sessions (range 6–12); mean attendance across all three groups was 10 sessions and the modal number of sessions attended was 12. One participant was loaned an iPad with data enabled from the service to facilitate access to the group online.

Routine clinical outcomes and adverse events

Table 1 summarizes clinical outcomes with effect sizes following the 12‐week group therapy. There were significant reductions pre‐ to post‐therapy in depression and anxiety with large effect sizes. On the BAVQ‐R, malevolence and omnipotence reduced significantly with large effect sizes, and resistance emotion reduced significantly with a medium effect size. Reliable and clinically significant change was calculated for each clinical outcome – for the BAVQ‐R, this was achieved by comparison with scores reported from a sample of 117 voice hearers (Ellett et al., 2017). As can be seen in Table 1, there were reliable and clinically significant changes in all clinical outcomes, ranging from n = 6 (depression) to n = 1 (engagement). There were no reliable or clinically significant deteriorations in any of the clinical outcome measures.

In keeping with recommendations for monitoring harm in mindfulness for psychosis research (Ellett & Chadwick, 2021), we report adverse events for all 21 service users offered a space in a group. Over the timeline of the groups (i.e. initial screening to post‐group assessment), one of the 21 service users had a psychiatric admission – and in fact, online therapy delivery allowed the person to continue attending and complete therapy. Over the timeline of the groups, no service user attempted suicide.

Qualitative analysis of participant experience

All 17 service users who completed a group were positive about the experience and reported meaningful benefits. Three key themes were identified: experience of online delivery, in which participants identified both ‘practical and psychological benefits’ and ‘challenges’; therapeutic benefits which included ‘relating mindfully to voices and paranoia’ and ‘being myself more fully’, and finally ‘feeling connected to the people in the group’ (see Table 2).

TABLE 2.

Themes and illustrative quotations

Theme	Subtheme	Illustrative quotes
Experience of online delivery	Practical and psychological benefits	“Doing the group on zoom was a lot better than I thought it would be, better than face to face” “When you're talking, you're looking at the one person who's talking and it's more personal” “Not having to travel has been good with the online format” “Good to have people's names displayed on the online format” “Using the headphones has helped me to concentrate” “It has been helpful to be able to turn the camera off at times during the group when my voices were bad” “I don't feel as vulnerable joining the group online”
	Challenges	“I experienced technical difficulties – my sound kept going off and I had to exit zoom and then rejoin” “I had trouble joining the zoom meeting 1 week, and I was really paranoid that they'd deliberately stopped me joining the group”
Therapeutic benefits	Relating mindfully to voices and paranoia	“Mindfulness has been helpful. I’ve learned how to respond differently to voices and I’m a lot less stressed by them” “Voices don't consume and overtake me – through mindfulness, I’ve learned that there's another way” “I was able to sit in the garden and I noticed feeling under threat. The fact that I was able to do this is an achievement”
	Being myself more fully	“I have learned that I can do things for me and it's ok to do things that are for me” “Voices no longer define who I am” “I have developed a sense of acceptance of my voices; they are there but they don't define me. I am more than voices” “I’m starting to get more enjoyment out of doing things again – gardening and my hobbies, and I went to the pub. I’m putting less energy into the voices and into other things that pop up like paranoia…I feel like it's all cooling down a bit” “I have started going to a yoga class. It has been a struggle, but I have done it” “I have spent 30 years punishing myself every day; starting this group and learning mindfulness has allowed me to experience good days and the positive experiences I have in my life”
Feeling connected to the people in the group		“We have all been open to giving things a go and bringing new things into our life. The group is a platform for a new beginning” “It was helpful to hear about other peoples’ experience of hearing voices and knowing that I’m not the only one” “The group was open and I felt accepted” “We all supported and accepted each other in the group”

Subjective importance of therapeutic group factors

Table 3 shows the mean ranking of each of the eight group therapeutic factors. As can be seen, universality was the most important therapeutic factor as ranked by participants. The joint second most important factors were learning mindfulness and instillation of hope.

TABLE 3.

Subjective importance of therapeutic group factors

Group factor and linked statement	Mean rank
Universality: ‘Learning I'm not the only one with my type of problem’	2.5
Learning mindfulness: ‘Using mindfulness to cope with distressing thoughts, images or voices’	3.5
Instillation of hope: ‘Seeing that others had solved problems similar to mine’	3.5
Interpersonal learning: ‘Feeling more trustful of groups and other people’	4.3
Group cohesion: ‘Belonging to and being accepted by a group’	4.6
Self‐understanding: ‘Discovering and accepting previously unknown or unacceptable parts of myself’	5.1
Altruism: ‘Helping others and being important in their lives’	5.8
Catharsis: ‘Learning how to express my feelings’	6.4

DISCUSSION

The study presents the first empirical data examining engagement, clinical benefits, participant experience, and group process in a group mindfulness‐based therapy for psychosis delivered solely online and in routine NHS mental health services in the UK. In terms of engagement, the majority (81%) of participants offered a place in a group completed therapy by attending at least six of the 12 group sessions. This compares well with completion rates from face‐to‐face delivery of mindfulness for psychosis groups for people with distressing voices, both in routine care (81% in Dannahy et al., 2011; 59% in Jones et al., 2021) and a full RCT (72% in Chadwick et al., 2016). Participant feedback identified online delivery as facilitating access, and this was observed in the present study where service users were able to continue attending sessions even when admitted to hospital or when leaving the geographical area. While online completion of measures in the present study was again strong, it is noticeable that three therapy completers did not complete post‐group measures. It may be that further work is needed to understand barriers and facilitators to online data collection. Overall, these data suggest that levels of engagement and therapy completion in online mindfulness for psychosis groups are at least comparable to that observed in face‐to‐face delivery in routine clinical practice.

In relation to clinical benefits, we found significant reductions in depression, anxiety, and beliefs about voices (specifically malevolence, omnipotence) from pre–post therapy, with medium‐large effect sizes. There was also a significant reduction in voice‐related negative affect on the BAVQ‐R resistance‐emotion subscale (measuring voice‐related fear, low mood, anger, and anxiety). For participants providing pre–post data, reliable and clinically significant improvement was observed in 43% for depression and 36% for anxiety, and there were no reliable or clinically significant deteriorations in any of the clinical outcome measures. In a recent pilot RCT of mindfulness groups for psychosis with persecutory delusions (Ellett et al., 2020), 64% of participants receiving mindfulness showed a clinically significant pre–post reduction in depression. Reduced depressive symptomatology is emerging as a robust and important clinical benefit of mindfulness groups for people with diagnoses of schizophrenia spectrum disorders. Finally, in relation to the monitoring of possible harm in mindfulness for psychosis (Ellett & Chadwick, 2021), it is worth noting the low incidence of adverse events, therapy dropouts, and the complete absence of clinically significant deterioration in depression, anxiety, and beliefs about voices.

Collectively, our findings in relation to both engagement and clinical outcomes are consistent with a growing body of evidence showing the benefits of online MBIs for a range of mental health outcomes (Spijkerman et al., 2016). We add to this literature by showing empirically the clinical benefits of online delivery specifically for people with psychosis in routine clinical practice. Importantly, given that individuals with psychosis are known to be at particular risk of digital exclusion (Ennis et al., 2012), we made the a priori decision to provide loan iPads if needed. It is noteworthy that in the present study, one participant required a loan iPad. This is consistent with recent quantitative research showing that digital exclusion in people with psychosis has declined, but still remains (Robotham et al., 2016). Future research might usefully examine both facilitators and barriers to accessing online MBIs for people with psychosis in routine clinical practice.

The qualitative analysis identified a number of novel practical and psychological benefits of online delivery, including not having to travel (for some participants, a trigger for voices and paranoia), being able to turn off the camera when psychotic experiences were overwhelming, and using headphones to aid concentration. There were also some unique challenges, including technology glitches, which for one person triggered paranoia. The therapeutic benefits highlighted by participants, in particular relating mindfully to voices and paranoia, are consistent with findings from face‐to‐face mindfulness for psychosis groups (Abba et al., 2008). It is also particularly encouraging, given the online format, that individuals reported feeling connected to other people in the group. Data relating to the subjective importance of group therapeutic factors are again resonant of findings from face‐to‐face psychosis groups. In both the present study and a study of face‐to‐face mindfulness groups for people with distressing voices (Chadwick et al., 2009), not only did universality emerge as the subjectively most important group process, but universality, learning mindfulness, and instillation of hope were ranked as the three most important of the eight therapeutic factors.

There are some limitations of the study that warrant consideration. While all service users referred were offered a group, the present sample lacked ethnic diversity, reflecting wider problems in many UK NHS mental health services. The lack of a control group opens the possibility that observed changes were due to non‐specific factors (e.g. time or group participation), and the pre–post design means that we did not address other important feasibility outcomes (e.g. barriers and facilitators). Additionally, the lack of randomization in the study means that the results are more susceptible to bias. The small sample size renders conclusions as tentative and findings need to be replicated. Finally, there was no follow‐up assessment; therefore, the study is silent about whether the improvements shown would be maintained. A larger scale RCT to assess the efficacy of group mindfulness therapy delivered online is now warranted.

Overall, our findings in relation to engagement, clinical benefits, participant experience, and group process offer real encouragement that online delivery of mindfulness for psychosis groups may be a useful addition to mental health services that seek to be inclusive and increase patient choice.

CONFLICT OF INTEREST

All authors declare no conflict of interest.

AUTHOR CONTRIBUTION

Lyn Ellett: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Laura Dannahy: Conceptualization (equal); Data curation (equal); Methodology (equal); Project administration (equal); Writing – review & editing (equal). Paul D J Chadwick: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Methodology (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal).

Ellett, L. , Dannahy, L. , & Chadwick, P. (2022). Engagement, clinical outcomes and therapeutic process in online mindfulness for psychosis groups delivered in routine care. Psychology and Psychotherapy: Theory, Research and Practice, 95, 467–476. 10.1111/papt.12382

DATA AVAILABILITY STATEMENT

Outcome data that support the findings of this study are available from the corresponding author upon reasonable request.