A Pilot Randomized Controlled Trial Among People Recovering from Mental Illness: A Tailored Mindfulness-Based Intervention versus Relaxation Training

ABSTRACT

Background

This study assessed the potential effectiveness, acceptability and feasibility of a tailored mindfulness-based intervention (MBI, REMIND 2.0) for personal recovery among people with mental illness during the COVID-19 pandemic.

Methods

In this pilot mixed methods randomized controlled trial, participants were assigned to either the MBI (n = 14) or the relaxation training (RT) (n = 14). Quantitative measures were used to assess primary outcomes, including personal recovery, mindfulness, self-compassion, resilience, and secondary outcomes, including depression, stress, anxiety, positive and negative moods, quality of life and general health at baseline (T0), post-intervention (T1) and one-month follow-up (T2). Quantitative interviews were conducted to explore the experiences and perceptions toward the MBI.

Results

Results indicated significant group and time interactions for all outcomes except anxiety and stress. MBI participants showed significant improvements in all outcomes at T1, which were maintained at T2, except for positive mood. RT participants showed a significant decline in resilience but significant improvements in all secondary outcomes at T1, but all outcomes significantly declined at T2, except for anxiety and stress. MBI participants were receptive toward the programme in all aspects of personal recovery.

Conclusions

The tailored MBI is a potentially effective, feasible and acceptable approach to facilitate personal recovery among people with mental illness. Differences between MBI and RT are discussed.

KEYWORDS:

• Mindfulness
• self-compassion
• mindfulness-based intervention
• relaxation
• mental illness
• mixed methods

Mindfulness-based interventions (MBIs) have been increasingly adopted in clinical settings due to their efficacy in enhancing both physical and mental health outcomes (Galante et al., 2023). In the Buddhist tradition, mindfulness promotes self-regulation by purposively focusing on the present moment to respond to internal (i.e., thoughts, emotions and sensations) and external (i.e., environmental) experiences (Baer, 2003). In particular, mindfulness skills involve observing, describing, acting with awareness, and accepting without judgment. Building upon this fundamental notion of mindfulness, MBIs tended to combine multi-components (e.g., cognitive therapy, self-compassion) according to various goals and targeted populations (Shapero et al., 2018). These included contemporary MBIs, such as mindfulness-based cognitive therapy (MBCT) and mindful self-compassion (MSC). These MBIs adhered to standardized protocols enabling clinical application and efficacy research (Crane et al., 2017).

Incorporating the self-compassion component in MBIs empowers participants with loving-kindness toward themselves and improves daily self-care, especially when experiencing suffering (Feldman & Kuyken, 2011; Gilbert, 2009). The Buddhist Four Noble Truths explain that suffering is an inevitable aspect of the human experience and provide guidance on how to cease its causes through qualities that include wisdom, kindness, equanimity and mindfulness (Aich, 2013; Crane et al., 2017). By practicing mindfulness and self-compassion, individuals can cultivate these qualities, gaining a sense of affective distance and utilizing awareness, understanding and acceptance to non-judgementally to shift away from internal conflicts to a compassionate attitude to face all experiences, including suffering. This affective distance also allows individuals to make conscious choices to respond rather than reacting automatically reacting (Desbordes et al., 2014). Recognizing the interconnectedness of all beings in the shared experience of suffering, individuals can improve self-care and develop the motivation to approach suffering with mindfulness and self-compassion to cultivate a compassionate attitude toward themselves and others (Hofmann et al., 2011).

Most MBIs in Western countries have been criticized as ethnocentric by predominantly reflecting Western values and ideologies (Kirmayer, 2015). While the underlying assumptions of mindfulness and self-compassion align with the universality of human sufferings, MBIs often overlook the cultural uniqueness and specific needs of diverse populations. Insufficient attention has been given to addressing the needs, characteristics and cultural of the targeted population in these MBIs (DeLuca et al., 2018). Instead, the focus has primarily been on meeting specific goals and diagnoses, neglecting the groups’ cultural values, needs and characteristics related that would effective engagement for participants from non-Western to maximize the benefits of MBIs (Rathod et al., 2018). Existing guidelines (e.g., Castro et al., 2010; Gearing et al., 2020) facilitate the development of further attuned MBIs for specific population. There is an immense need for a new generation of population-specific MBI, particularly for people recovering mental illness, that go beyond cultural attenuations and address their unique needs.

Drawing on insights from the findings of our pre-pilot study (Cheng et al., 2022), along with literature and expert input, a tailored MBI, named REMIND 2.0 was developed to facilitate personal recovery among community mental health service users. The development process involved three stages:

1. Conceptualizing the MBI based on contemporary MBI protocols and empirical evidence from the pre-pilot study.

2. Developing, revising, refining and standardizing the MBI.

3. Conducting a pilot study to assess feasibility and acceptability before a larger scale, fully powered randomized controlled trial (RCT).

The current pilot study explored the feasibility and acceptability of the tailored MBI (REMIND 2.0) for facilitating personal recovery among mental health service users. Additionally, we included an active relaxation training control group to address the confusion between mindfulness and relaxation practices, as observed in the pre-pilot trial and existing literature (Luberto et al., 2020). Hypotheses were not postulated due to the pilot nature of the current RCT.

Materials and method

Study design and timeline

The current study adopted a mixed-methods, exploratory, single-center, two-armed, parallel-group, pilot RCT of MBI for personal recovery among people with mental illness. The RCT took place from August 2020 to January 2021, including recruitment, intervention and data collection, at an Integrated Community Centre for Mental Wellness (ICCMW) in Hong Kong during the COVID-19 pandemic.

The current study adhered to the CONSORT extension to randomized pilot and feasibility trials (Eldridge et al., 2016) and the COREQ checklist (Tong et al., 2007). The flow of the current pilot RCT is depicted in Figure 1. The trial registration number for this study is ISRCTN13689468 (https://www.isrctn.com/ISRCTN13689468).

short-legendFigure 1.

Participants and procedures

Thirty-four participants were recruited by referral and advertisement in the ICCMW brochure. Participants had been screened for eligibility by research staff, ensuring both inclusion and exclusion criteria were met. A total of 28 eligible participants were enrolled in this trial.

Participant inclusion criteria were: aged ≥ 18; permanent Hong Kong residents with a period of stay ≥7 years; diagnosed with mental illness with standardized diagnostic criteria such as the Classification of Diseases or Diagnostic (ICD-10, World Health Organization, 2019) and Statistical Manual of Mental Disorders (DSM-5, American Psychiatric Association, 2013), for any length of illness; capable of giving informed consent; understand Chinese language to engage in the intervention or to complete written assessment; a current service recipient of the community-based mental health service (e.g. integrated mental wellness center, halfway house or vocational training center).

Participant exclusion criteria were: engagement in any mindfulness or meditation training concurrently; Established diagnosis of learning disability or cognitive impairment arising from any underlying medical condition resulting in significant functional impairment; Primary diagnosis of substance abuse, as it may increase and/or trigger symptoms of mental illnesses, and hence it has different mental illnesses trajectories; Mental state that precludes engagement (e.g., disorganized thinking and speech); Without informed consent; and Refused to be randomized, which were those who declined to be assigned to a group through the randomization process and expressed a preference for only a particular group.

After obtaining informed consent, participants completed a self-report questionnaire that included demographic and quantitative questions. Following the baseline assessment (T0), participants were required to complete two subsequent assessments immediately (T1) and one month after the MBI or RT (T2). Semi-structured interviews were conducted only among participants in the MBI group at T1 and T2. All procedures were approved by the Hong Kong Baptist University Research Ethics Committee.

Supermarket cash coupons were offered to participants as incentives after each follow-up assessment. The incentive amount gradually increased at each time point to minimize attrition. Participants had to complete the T0 and T1 assessment and attend at least five sessions of the MBI or RT to receive a HK$50 coupon. Participants who also completed the T2 assessment received an additional HK$100 coupon.

Both MBI and RT participants were expected to attend a minimum of five sessions to meet the fidelity requirements for minimum attendance. Missed sessions were made up during follow-up calls with facilitator, ensuring program integrity. This ensured that the participants received sufficient exposure to the intervention content.

Randomisation and blinding

Eligible participants were randomly assigned (1:1) to either the MBI or the RT group. The randomization process was conducted using SPSS Syntax and was stratified by age and gender. This process was carried out by a third-party administrator who was not involved in data collection or analysis. Each Participant was informed by their group assignment individually. While the participants were aware of their group allocation, quantitative data were collected through self-administered questionnaires. To ensure blinding, a third party masked group assignments with alphabetic codes, ensuring that the researcher remained unaware of the participants’ group allocation.

Intervention and the active control group

Mindfulness-based intervention (REMIND 2.0)

The REMIND 2.0 was a tailored MBI for facilitating personal recovery among people with mental illness in the community. It was an improved version of the REMIND 1.0, which was initially inspired by the MBCT and MSC protocols but with cultural considerations.

The content and practices of the REMIND 2.0 were improved based on the pre-pilot findings and comments from experienced meditators. There are some key features of the intervention:

• Health communication and cognitive approaches: the incorporation of health communication and cognitive approaches, such as the social cognitive theory (Bandura, 1986) and cognitive and meta cognitive learning strategies (Wiklund-Engblom, 2008). These approaches aimed to enhance the learning process and outcomes by creating an inclusive learning environment for participants with different educational levels and individual differences, including those whose mental states were disrupted by symptoms or psychotropic medication side-effects (Price et al., 2009; Trivedi, 2006).

• Use of physical objects or actions: Introduction of physical objects or actions can help beginners and amateurs stay present during practice sessions and improve their awareness.

• Integration of local images, phrases and stories: Incorporation of local images, phrases, and stories to make the content more relevant to participants.

• Assistance in describing experience: Participants received help describing their experiences through various given vocabularies and examples during practice sessions.

• Emphasis on mindfulness and self-compassion as a life attitude: The intervention emphasized practicing mindfulness and self-compassion as a life attitude, and relaxation and alleviation of emotional distress are the by-products of constant practice. Banners around the activity room, magnets for the participants to place at their homes and verbal reinforcement by the facilitator were used to reinforce these ideas.

• Integration of learnings and practices into daily routines: Participants were guided to discover effective methods for practicing mindfulness and self-compassion and integrate them into their daily routines throughout the intervention and follow-up calls with the facilitator.

• Independent practice: All practices could be done independently without a partner, allowing participants to engage in any practice at any time without the need to search for a partner.

As depicted in Table 1, the REMIND 2.0 consisted of eight weekly, 2-hour face-to-face sessions. Participants received weekly follow-up calls from the facilitator and a workbook to aid their learning progress and reduce attrition. Weekly assignments and short practices of 10 to 15 minutes were given between sessions.

Table 1. An outline of the mindfulness-based intervention, the REMIND program 2.0, for people in mental health recovery.

Week/ Session	Theme	Content	In-session practices	Home practices and assignments
One	Introduction to mindfulness and self-compassion	Definitions: mindfulness and self-compassion Benefits of mindfulness and self-compassion: empirical evidence	Breathing and sitting postures Grounding Counting breaths Body scan	Practice grounding, counting breaths or body scan
Two	Mindfulness: attention and awareness	Definition of mindfulness Master your emotions The wandering mind: becoming aware of the innate autopilot mode Gathering the scattered mind: pause and rediscover the intentional mode	Grounding Revisit a past incident with the five-finger method Butterfly hug Set goals to practice mindfulness in everyday life	The five-finger method Mindful eating Rearrange schedule for mindfulness daily practices
Three	Mindfulness: just be with the thoughts	Thoughts are thoughts, not facts Noticing the thinking patterns Spaciousness: from reacting to responding, the wise mind What” and “How” mindfulness skill Mindfulness and the snow globe metaphor: stirred or settled Myths about emotions	Grounding The “thoughts and feelings” exercise Novel object (e.g., crystal) mindfulness practice and reflection	Practice of the “what” and “how” skills in any daily activities
Four	Being aware of self-compassion: loving-kindness for self and others	What is/is not self-compassion? Connecting mindfulness and self-compassion Common Humanity: imperfection Loving-kindness to self and others Understand the concept of “backdraft” and how to handle it	Grounding Distinction between treating family members/friends and oneself in difficult time. Self-care Loving kindness phrases Self-compassion break Your core values	Continue daily mindfulness and self-compassion practices Reflect on your self-compassionate behaviors in daily life
Five	Practicing self-compassion:self-compassionate motives and motivation	Self-compassion vs. self-criticism Transforming self-criticism through motives, encouragement and motivation generate from self-compassion	Grounding Transforming self-criticism: self-motivate with self-compassion Affectionate breathing Becoming aware of your traits and strengths	Continue daily mindfulness and self-compassion practices Compose sentences to wish yourself well
Six	Fostering self-compassion: self-compassionate motives and motivation	The root causes of suffering Causes of unnecessary suffering? Resistance and avoidance Blending mindfulness and self-compassion together for facing challenges, resistance and avoidance Self-compassion with five senses	Grounding Becoming aware of your experience of resistance and avoidance Reflection: energy giving and taking activities Smiling practice	Continue daily mindfulness and self-compassion practices Describe and reflect on your experience of a situation that you resisted or avoided
Seven	Practice being with difficulties: interweaving mindfulness and self-compassion	Your alarm systems and action plans Strategies for meeting with difficulties and emotional distress Let go and adapt Hidden value in difficulties	Grounding Remembered wellness Silver linings: learning from difficulties, struggles and challenges in life	Continue daily mindfulness and self-compassion practices Consider mindfulness and self-compassion practices for your daily life Describe your self-compassionate behaviors
Eight	Being aware and compassionate in each “present moment” in the future	Things that you feel negative and things that you feel grateful for Gratitude and self-appreciation Tips for maintaining daily practice	Grounding Reflections and discussions of previous weeks’ materials Gratitude and self-appreciation

Relaxation training (RT)

The relaxation training (RT) group served as an active control group, closely mirroring the structure of the MBI. The RT also consisted of eight weekly, 2-hour face-to-face sessions, with follow-up calls, a workbook, home assignments, and practices for further learning and reduced attrition. The RT was led by the same MBI facilitator to ensure that the concepts of mindfulness and self-compassion were not included and to avoid the confounding effects (e.g., facilitator characteristics, personality and teaching style).

The content of the RT included evidence-based techniques, including (i) stress responses, (ii) sources of stress, (iii) correct breathing techniques, (iv) progressive muscle relaxation, (v) autogenic, (vi) visualization, (vii) combination techniques, and (viii) nutrition.

Primary quantitative measures

Personal recovery: recovery assessment scale (RAS)

Personal recovery was measured with the 24-item RAS, with five domains: personal confidence and hope, willingness to ask for help, goal and success orientation, reliance on others, and no domination by symptoms (Giffort et al., 1995). Higher scores indicate higher levels of perceived personal recovery. The Chinese RAS demonstrated good internal consistency (Cronbach’s α = 0.92) (Young et al., 2017).

Mindfulness: mindful attention awareness scale (MAAS)

Mindfulness was measured with the 15-item MAAS (Brown & Ryan, 2003), with a core structure focusing on the awareness or attention of the present moment. Higher scores indicate higher levels of mindfulness. The MAAS demonstrated good convergent and divergent validity and internal consistency (Cronbach’s α= .82 to .87).

Self-compassion: self-compassion scale (SCS)

Self-compassion was measured with the 24-item SCS, with six domains: self-kindness, self-judgment, common humanity, isolation, mindfulness, and over-identified (Neff, 2003). Higher scores indicate higher levels of self-compassion. The Chinese SCS demonstrated good internal consistency (Cronbach’s α = 0.84), test-retest reliability (Cronbach’s α = 0.89) and validity (Chen et al., 2011; Neff et al., 2008).

Resilience: Connor-Davidson Resilience Scale (CD-RISC)

Psychological resilience was measured with the 25-item CD-RISC, with five domains: personal competence, high standards and tenacity, trust or tolerance of negative affect and stress, acceptance of change and secure relationships, control and spirituality (Connor & Davidson, 2003). Higher scores indicate higher levels of resilience. The Chinese CD-RISC demonstrated satisfactory internal consistency reliability (Cronbach’s α = 0.91 to 0.97) (Ni et al., 2016; Yu & Zhang, 2007).

Secondary quantitative measures

Psychological distress: depression, anxiety and stress scales (DASS-21)

DASS-21 was used to measure depression, anxiety, and stress (Lovibond & Lovibond, 1995). Higher scores indicate higher levels of the affective states. The Chinese DASS-21 demonstrated good validity (adjusted GFI = 0.86) and internal consistency among the subscales (phi coefficients = 0.81 to 0.88) (Moussa et al., 2001).

Mood: positive and negative affect schedule (PANAS)

Mood was measured with the 20-item PANAS, with two subscales measuring the positive and negative affect (Watson et al., 1988). Higher scores indicate higher levels of affects. The Chinese PANAS demonstrated good internal consistency of both positive and negative affect subscales (Cronbach’s α = 0.85, 0.83) (Huang et al., 2003).

Quality of life: World Health Organization Quality of Life-BREF (WHOQOL-BREF)

The overall quality of life and general health were measured by two global items from the WHOQOL-BREF (World Health Organization, 1996). Higher scores indicate higher quality of life or general health. The Hong Kong Chinese WHOQOL-BREF demonstrated good reliability (ICC ranged from 0.73 to 0.84) across domains (Leung et al., 2005).

Feasibility

Feasibility was assessed through the recruitment, retention, dropout and compliance rates and the time required to reach the target sample size. The recruitment rate was determined by the number of eligible participants who provided informed consent and were randomized into a group. The retention rate was determined by the number of participants who attended the post and follow-up assessments compared to those who completed the baseline. Dropout and compliance rates were determined by the number of participants who were absent or attended the sessions and assessments.

Acceptability

Acceptability of the intervention was assessed using a global quantitative item of the overall satisfaction on a 6-point Likert scale (0 extremely unsatisfied, 5 extremely satisfied). Additionally, the qualitative interviews responses were considered “acceptable” if most participants responded positively to the questions related to the content and perceived effectiveness of the MBI.

Qualitative semi-structured interviews

Individual semi-structured interviews were conducted at T1 and T2 to gain insights into the participants’ experiences, the change process, benefits and barriers associated with MBI and recovery. Both interviews followed the standardized protocols, with open-ended questions developed to guide the interview process. Each interview lasted between 45 to 60 minutes and was conducted in person. All interviews were digitally audiotaped and transcribed.

Data analysis

Quantitative data were analyzed using SPSS 29.0. Missing data were handled on an intent-to-treat (ITT) basis using the Expectation-Maximisation (EM) algorithm. Descriptive statistics of demographic characteristics and total scores were summarized using mean, standard deviation, frequency, and percentage. ANOVAs were employed to estimate the interaction, between and within effects across time and groups. Greenhouse-Geisser correction and post-hoc tests with Tukey and Games-Howell were used when appropriate. The results should be interpreted cautiously, as no a priori sample size calculation was carried out for this pilot study.

Qualitative data from interviews were transcribed and analyzed with thematic analysis (TA), following the guidelines of Braun and Clarke (2022). NVIVO 12.0 was used to assist the analysis process. Data were inductively coded into themes and subthemes with repeated revision and comparison to ensure coherence, exhaustive, and mutual exclusivity. Reflections were undertaken to minimize biases throughout the process. The first author (DC) completed the TA, and the coauthor (DY) discussed with the first author to refine the preliminary results. Data triangulation was performed with the quantitative and qualitative findings before finalization to enhance confidence and minimize uncertainty.

Results

Feasibility

Recruitment rate

To attain the desired sample size of 28 eligible participants, 34 potential participants were approached. Consequently, an 82.35% recruitment rate was achieved over a 16 days recruitment period (see Figure 1).

Retention rate

All participants completed the follow-up assessments, except one from the MBI group, who declined the qualitative interviews due to physical illness but completed the quantitative assessments T1 and T2. All 84 quantitative assessments were completed, and the retention rate was 100% for both the MBI and RT groups.

Dropout and compliance rate

MBI session attendance rate was 95.54%, in which a participant was absent from two sessions and another three were absent from one session. RT session attendance rate was 92.86%, in which a participant was absent from three sessions, and another five were absent from one session. All the absent participants who attended the sessions via Zoom videoconferencing were due to slight symptoms and would like to stay home as a precautionary measure during the pandemic – no participant dropout from either group.

Acceptability

The mean overall satisfaction score of the MBI was 4.5 (SD= 0.65) and the RT was 4.43 (SD= 0.65). All participants were receptive to the content of the MBI. They generally perceived a positive influence on their personal recovery during and after the MBI. These are supported by the qualitative findings from the thematic analyses.

Demographic characteristics

The demographic characteristics of the participants are presented in Table 2. There were no significant differences in baseline demographic characteristics between MBI and RT. Participants were mainly female (MBI: 92.9%; RT: 85.7%), aged 27 to 68 years (MBI: M= 54.64, SD= 10.93; RT: M= 55.79, SD= 10.19). All participants completed at least secondary education. Most participants lived in private housing (MBI: 92.9%; RT: 92.9%), and all lived with their families. Most participants were diagnosed with depression (MBI: 64.3%; RT: 64.3%), of which four stated comorbidities with anxiety disorders (MBI: n= 1; RT: n= 3). Others were diagnosed with anxiety disorders (MBI: 28.6%; RT: 14.3%) and psychotic disorders (MBI: 7.1%; RT: 21.4%). The majority had greater than 10 years of the onset of mental illness (MBI: 64.3%; RT: 71.4%), and none were hospitalized. Although most had experience in formal or informal mindfulness or meditation (MBI: 78.6%; RT: 64.3%), they did not practice at all, and only a few rarely practiced (MBI: n= 3; RT: n= 1).

Table 2. Demographic characteristics of participants in mindfulness-based intervention (MBI) and relaxation training (RT) at baseline.

Baseline Characteristics	MBI		RT
	n	%	n	%
Period of stay
Born in Hong Kong	11	78.6	12	85.7
7 years and above	3	21.4	2	14.3
Gender
Female	13	92.9	12	85.7
Male	1	7.1	2	14.3
Education level
Junior secondary	3	21.4	2	14.3
Senior secondary	5	35.7	8	57.1
Tertiary	6	42.9	4	28.6
Employment status
Unemployed	2	14.3	6	42.9
Full-time	4	28.6	2	14.3
Part-time	2	14.3	1	7.1
Homemaker	6	42.9	5	35.7
Income source
Salary	4	28.6	3	21.4
Family support	10	71.4	10	71.4
Government subsidies	0	0	1	7.1
Marital status
Single	2	14.3	5	35.7
Married	11	78.6	8	57.1
Divorced	1	7.1	1	7.1
Living arrangement
Alone	0	0	1	7.1
With Family	14	100	13	92.9
Housing type
Public housing	1	7.1	1	7.1
Private housing	13	92.9	13	92.9
Diagnosis
Depression	9	64.3	9	64.3
Anxiety	4	28.6	2	14.3
Psychotic disorders	1	7.1	3	21.4
Year of onset
Greater than 10 years	9	64.3	10	71.4
Within 10 years	5	35.7	4	28.6
Number of hospitalisations
None	14	100	14	100
Experience in mindfulness or meditation practice
No	3	21.4	5	35.7
Yes	11	78.6	9	64.3
Mindfulness or meditation practice habit
No	11	78.6	13	92.9
Yes, rarely	3	21.4	1	7.1

Note. N = 28 (n = 14 for each group). Participants in the MBI group were on average 54.64 years old (SD = 10.93). Participants in the RT group were on average 55.79 years old (SD = 10.90). Participant demographic characteristics did not differ by group.

Primary quantitative outcomes

The descriptive and between- and within-subject ANOVA results are presented in Tables 3 and 4 respectively. For personal recovery, the between-subject results showed significant main effects of time [F (1.39, 36.12) = 24.35, p < .001, η_P² = 48], group [F (1, 26) = 18.97, p < .001, η_P² = .42], and time by group interaction [F (1.39, 36.12) = 60.39, p < .001, η_P² = .70]. Among the MBI group, the within-subject results showed the mean recovery scores differed significantly between time points [F (1.22, 15.86) = 60.14, p < .001, η_P² = .82]. Post-hoc tests revealed the mean recovery scores increased by 24.14 (p < .001) at T1 but then decreased by 0.21 at T2 (p= .99) and overall increased by 23.93 (p < .001) from T0 to T2. Among the RT group, results showed the mean recovery scores differed significantly between time points [F (2, 26) = 10.01, p < .01, η_P² = .44]. Post-hoc tests revealed the mean recovery scores decreased by 2.93 (p= .37) at T1, further decreased by 5.36 at T2 (p < .01) and overall decreased by 8.29 (p < .01) from T0 to T2.

Table 3. Means, standard deviations and between-subject analyses of variance in quantitative outcomes at baseline (T0), post-intervention (T1) and follow-up (T2).

Measure	MBI		RT		Δ			ANOVA
	M	SD	M	SD	M	SE	Effect	F ratio	ηp^2
Personal recovery
T0	78.14	10.88	83.50	12.79	5.36	4.49	G	18.97***	.42
T1	102.29	7.33	80.57	10.71	21.72	3.47	T	24.35***	.48
T2	102.07	8.54	75.21	8.66	26.86	3.25	G x T	60.39***	.70
Mindfulness
T0	50.43	10.41	51.71	7.83	1.28	3.48	G	53.41***	.67
T1	74.07	5.84	48.64	6.89	25.43	2.41	T	41.60***	.62
T2	73.43	6.37	44.29	5.82	29.14	2.31	G x T	99.57***	.79
Self-compassion
T0	69.14	10.81	75.86	12.81	6.72	4.48	G	31.00***	.54
T1	103.21	5.00	74.93	12.56	28.28	3.61	T	78.25***	.75
T2	102.57	6.42	66.29	10.18	36.28	3.22	G x T	140.01***	.84
Resilience
T0	56.00	13.58	57.50	10.70	1.50	4.62	G	15.59***	.38
T1	71.43	8.92	56.50	10.55	14.93	3.69	T	11.67***	.31
T2	71.07	8.63	45.14	6.36	25.93	2.87	G x T	37.82***	.59
Depression
T0	14.00	2.08	11.79	1.19	2.21	0.64	G	39.19***	.60
T1	6.00	1.36	10.29	1.07	4.29	0.46	T	236.02***	.90
T2	5.71	1.07	11.93	1.21	6.22	0.43	G x T	174.35***	.87
Anxiety
T0	8.71	0.83	8.64	1.01	0.07	0.35	G	.75	.03
T1	4.50	1.16	5.21	1.63	0.71	0.54	T	212.08***	.89
T2	4.00	1.18	4.50	2.03	0.50	0.63	G x T	1.52	.06
Stress
T0	13.86	0.95	13.00	8.8	0.86	2.37	G	.46	.02
T1	6.14	1.41	6.30	1.33	0.16	0.52	T	303.52***	.92
T2	6.79	2.16	6.50	2.47	0.29	0.88	G x T	1.17	.04
Positive mood
T0	24.29	2.70	24.43	4.16	0.14	1.33	G	15.35***	.37
T1	32.86	3.21	27.00	4.13	5.86	1.40	T	49.37***	.66
T2	30.50	2.53	22.79	3.77	7.71	1.21	G x T	26.54***	.51
Negative mood
T0	29.93	5.78	29.57	5.40	0.36	2.11	G	13.56***	.34
T1	19.36	3.97	26.29	5.17	6.93	1.74	T	69.00***	.73
T2	19.43	3.55	31.50	4.67	12.07	1.57	G x T	55.02***	.68
Quality of life
T0	2.93	0.48	2.93	0.48	0	0.18	G	7.74*	.23
T1	4.21	0.58	3.79	0.43	0.42	0.19	T	72.40***	.74
T2	4.14	0.54	3.29	0.47	0.85	0.19	G x T	10.80**	.29
General health
T0	2.93	0.83	2.79	0.89	0.14	0.33	G	4.54*	.15
T1	3.79	0.58	3.43	0.76	0.36	0.26	T	18.33***	.41
T2	3.86	0.54	3.00	0.39	0.86	0.18	G x T	4.01*	.13

Note. N = 28 (n= 14 for each group). MBI= Mindfulness-based intervention; RT = Relaxation training control group. G = group; T = time; Δ = difference. *p < .05; **p < .01; ***p < .001.

Table 4. Means, standard deviations and within-subject analyses of variance in quantitative outcomes at baseline (T0), post-intervention (T1) and follow-up (T2).

Measure	MBI		ANOVA		RT		ANOVA
	M	SD	F ratio	ηp^2	M	SD	F ratio	ηp^2
Personal recovery
Δ T0-T1	24.14***	11.59	60.14***	.82	−2.93	7.03	10.01**	.44
Δ T1-T2	−0.21	4.34			−5.36**	5.30
Δ T0-T2	23.93***	10.75			−8.29**	8.41
Mindfulness
Δ T0-T1	23.64***	9.12	97.18***	.88	−3.07	5.98	15.47***	.54
Δ T1-T2	−0.64	3.32			−4.36***	2.95
Δ T0-T2	23.00***	7.91			−7.43***	5.58
Self-compassion
Δ T0-T1	34.07***	9.54	171.83***	.93	−0.93	5.41	18.27***	.58
Δ T1-T2	−0.64	3.25			−8.64***	6.78
Δ T0-T2	33.43***	9.17			−9.57***	7.27
Resilience
Δ T0-T1	15.43***	9.76	30.56***	.70	−1.00**	6.14	18.84***	.59
Δ T1-T2	−0.36	2.95			−11.36	9.61
Δ T0-T2	15.07***	10.45			−12.36***	8.94
Depression
Δ T0-T1	−8.00***	1.52	242.27***	.95	−1.50***	0.52	40.33***	.76
Δ T1-T2	−0.29	1.14			−1.64***	0.93
Δ T0-T2	−8.29***	2.02			0.14	0.77
Anxiety
Δ T0-T1	−4.21***	0.89	199.45***	.94	−3.43***	1.45	65.32***	.83
Δ T1-T2	−0.50	1.02			−0.71	1.27
Δ T0-T2	−4.71***	0.99			−4.14***	1.61
Stress
Δ T0-T1	−7.71***	1.27	206.26***	.94	−6.71***	1.20	114.64***	.90
Δ T1-T2	0.64	1.34			0.21	2.22
Δ T0-T2	−7.07***	2.02			−6.50***	2.07
Positive mood
Δ T0-T1	8.57***	3.96	53.93***	.81	2.57**	2.65	16.59***	.56
Δ T1-T2	−2.36***	1.82			−4.21***	2.83
Δ T0-T2	6.21***	3.40			−1.64*	2.79
Negative mood
Δ T0-T1	−10.57***	4.07	86.73***	.87	−3.29***	1.94	24.63***	.66
Δ T1-T2	0.71	1.33			5.21	3.17
Δ T0-T2	−10.50***	4.18			1.93*	3.15
Quality of life
Δ T0-T1	1.29***	0.61	56.21***	.83	0.86***	0.36	24.02***	.65
Δ T1-T2	−0.07	0.27			−0.50**	0.52
Δ T0-T2	1.21***	0.58			0.36*	0.50
General health
Δ T0-T1	0.86**	0.77	16.59***	.56	0.64**	0.74	6.16**	.32
Δ T1-T2	0.07	0.62			−0.43**	0.51
Δ T0-T2	0.93***	0.62			0.21	0.80

Note. N = 28 (n= 14 for each group). MBI= Mindfulness-based intervention; RT = Relaxation training control group. G = group; T = time; Δ = difference. *p < .05; **p < .01; ***p < .001.

For mindfulness, the between-subject results showed the main effect of time [F (1.26, 32.69) = 41.60, p < 0.001, η_P² = .62], group [F (1, 26) = 53.41, p < 0.001, η_P² = .67], and the time by group interaction were significant [F (1.26, 32.69) = 99.57, p < 0.001, η_P² = .79]. Among the MBI group, the within-subject results showed the mean mindfulness scores differed significantly between time points [F (1.19, 15.53) = 97.18, p < .001, η_P² = .88]. Post-hoc tests revealed the mean mindfulness scores increased by 23.64 (p < .001) at T1 but then decreased by 0.64 at T2 (p= .99) and overall increased by 23 (p < .001) from T0 to T2. Among the RT group, results showed the mean mindfulness scores differed significantly between time points [F (1.39, 18.06) = 15.47, p < .001, η_P² = .54]. Post-hoc tests revealed the mean mindfulness scores decreased by 3.07 (p= .23) at T1, further decreased by 4.36 at T2 (p < .001) and overall decreased by 7.43 (p < .01) from T0 to T2.

For self-compassion, the between-subject results showed significant main effects of time [F (1.64, 42.62) = 78.25, p < 0.001, η_P² = .75], group [F (1,26) = 31.00, p < 0.001, η_P² = .54] and the time by group interaction [F (1.64, 42.62) = 140.01, p < 0.001, η_P² = .84]. Among the MBI group, the within-subject results showed the mean self-compassion scores differed significantly between time points [F (1.18, 15.37) = 171.83, p < .001 η_P² = .93]. Post-hoc tests revealed the mean self-compassion scores increased by 34.07 (p < .001) at T1 but then decreased by 0.64 at T2 (p= .85) and overall increased by 33.43 (p < .001) from T0 to T2. Among the RT group, results showed the mean self-compassion scores differed significantly between time points [F (2, 26) = 18.27, p < .001, η_P² = .58]. Post-hoc tests revealed the mean self-compassion scores decreased by 0.93 (p= .90) at T1, further decreased by 8.64 at T2 (p < .01) and overall decreased by 9.57 (p < .001) from T0 to T2.

For resilience, the between-subject results showed significant main effects of time [F (2, 52) = 11.67, p < 0.001, η_P² = .31], group [F (1, 26) = 15.59, p < .001, η_P² = .38], and the time by group interaction [F (2, 52) = 37.82, p < .001, η_P² = .59]. Among the MBI group, results of within-subject repeated ANOVA showed that the mean resilience scores differed significantly between time points [F (1.12, 14.59) = 30.56, p < .001, η_P² = .70]. Post-hoc tests revealed the mean resilience scores increased by 15.43 (p < .001) at T1 but then decreased by 0.36 at T2 (p= .96) and overall increased by 15.07 (p < .001) from T0 to T2. Among the RT group, results showed the mean resilience scores differed significantly between time points [F (2, 26) = 18.84, p < .001, η_P² = .59]. Post-hoc tests revealed the mean resilience scores decreased by 1 (p= .91) at T1, further decreased by 11.36 at T2 (p < .01) and overall decreased by 12.36 (p < .01) from T0 to T2.

Secondary quantitative outcomes

For depression, between-subject results showed significant main effect of time [F (1.61, 41.81) = 236.02, p < .001, η_P² = .90], group [F (1, 26) = 39.19, p < .001, η_P² = .60] and the time by group interaction [F (1.61, 41.81) = 174.35, p < .001, η_P² = .87]. Among the MBI group, the within-subject results showed the mean depression scores differed significantly between time points [F (1.43, 18.57) = 242.27, p < .001, η_P² = .95]. Post-hoc tests revealed the mean depression scores decreased by 8 (p < .001) at T1, further decreased by 2.9 at T2 (p = .74) and overall decreased by 8.29 (p < .001) from T0 to T2. Among the RT group, results showed the mean depression scores differed significantly between time points [F (2, 26) = 40.33, p < .001, η_P² = .76]. Post-hoc tests revealed the mean depression scores decreased by 1.5 (p < .001) at T1 but increased by 1.64 at T2 (p < .001) and overall increased by 0.14 (p = .88) from T0 to T2.

For anxiety, the between-subject results showed a significant main effect of time [F (2, 52) = 212.08, p < .001, η_P2 = .89] but the effects for group [F (1, 26) = .75, p = .40, η_P² = .03] and the time by group interaction [F (2, 52) = 1.52, p = .23, η_P² = .06] were not significant. Among the MBI group, the within-subject results showed the mean anxiety scores differed significantly between time points [F (2, 26) = 199.45, p < .001, η_P² = .94]. Post-hoc tests revealed the mean anxiety scores decreased by 4.21 (p < .001) at T1, further decreased by 0.5 at T2 (p = .25) and overall decreased by 4.71 (p < .001) from T0 to T2. Among the RT group, results showed the mean anxiety scores differed significantly between time points [F (2, 26) = 65.32, p < .001, η_P² = .83]. Post-hoc tests revealed the mean anxiety scores decreased by 3.43 (p < .001) at T1, further decreased by 0.71 at T2 (p = .16) and overall decreased by 4.14 (p < .001) from T0 to T2.

For stress, the between-subject results showed significant main effect of time [F (1.58, 41.01) = 303.52, p < .001, η_P² = .92] but not for group [F (1, 26) = .46, p = .31, η_P² = .02], and time by group interaction [F (1.58, 41.01) = 1.17, p = .50, η_P² = .04]. Among the MBI group, the within-subject results showed the mean stress scores differed significantly between time points [F (1.42, 18.47) = 206.26, p < .001, η_P² = .94]. Post-hoc tests revealed the mean stress scores decreased by 7.71 (p < .001) at T1 but increased by 0.64 at T2 (p = .26) and overall decreased by 7.07 (p < .001) from T0 to T2. Among the RT group, results showed the mean stress scores differed significantly between time points [F (2, 26) = 114.64, p < .001, η_P² = .90]. Post-hoc tests revealed the mean stress scores decreased by 6.71 (p < .001) at T1 but increased by 0.21 at T2 (p = .98) and overall decreased by 6.5 (p < .001) from T0 to T2.

For positive mood, the between-subject results showed significant main effects of time [F (2, 52) = 49.37, p < .001, η_P² = .66], group [F (1, 26) = 15.35, p < .001, η_P² = .37], and the time by group interaction [F (2, 52) = 26.54, p < .001, η_P² = .51]. Among the MBI group, the within-subject results showed the mean positive mood scores differed significantly between time points [F (1.33, 17.25) = 53.93, p < .001, η_P² = .81]. Post-hoc tests revealed the mean positive mood scores increased by 8.57 (p < .001) at T1 but decreased by 2.36 at T2 (p < .001) and overall increased by 6.21 (p < .001) from T0 to T2. Among the RT group, results showed the mean positive mood scores differed significantly between time points [F (2, 26) = 16.59, p < .001, η_P² = .56]. Post-hoc tests revealed the mean positive mood scores increased by 2.57 (p < .01) at T1 but decreased by 4.21 at T2 (p < .001) and overall decreased by 1.64 (p = .13) from T0 to T2.

For negative mood, the between-subject results showed significant main effects of time [F (2, 52) = 69.00, p < .001, η_P² = .73], group [F (1, 26) = 13.56, p < .001, η_P² = .34], and the time by group interaction [F (2, 52) = 55.02, p < .001, η_P² = .68]. Among the MBI group, the within-subject results showed the mean negative mood scores differed significantly between time points [F (1.16, 15.04) = 86.73, p < .001, η_P² = .87]. Post-hoc tests revealed the mean negative mood scores decreased by 10.57 (p < .001) at T1 but increased by 0.71 at T2 (p = .99) and overall decreased by 10.5 (p < .001) from T0 to T2. Among the RT group, results also showed that the mean negative mood scores differed significantly between time points [F (2, 26) = 24.63, p < .001, η_P² = .66]. Post-hoc tests revealed the mean negative mood scores decreased by 3.29 (p < .001) at T1, but then increased by 5.21 at T2 (p < .001) and overall increased by 1.93 (p = .11) from T0 to T2.

For quality of life, the between-subject results showed significant main effects of time [F (2, 52) = 72.40, p < .001, η_P² = .74], group [F (1, 26) = 7.74, p < .05, η_P² = .23], and the time by group interaction [F (2, 52) = 10.80, p < .01, η_P² = .29]. Among the MBI group, results of within-subject results showed the mean quality of life scores differed significantly between time points [F (1.30, 16.96) = 56.21, p < .001, η_P² = .83]. Post-hoc tests revealed the mean quality of life scores increased by 1.28 (p < .001) at T1 but then decreased by 0.07 at T2 (p = .71) and overall increased by 1.21 (p < .001) from T0 to T2. Among the RT group, results showed the mean quality of life scores differed significantly between time points [F (2, 26) = 24.02, p < .001, η_P² = .65]. Post-hoc tests revealed the mean quality of life scores increased by 0.86 (p < .001) at T1 but decreased by 0.50 at T2 (p < .05) and overall increased by 0.36 (p = .06) from T0 to T2.

For perceived health, the between-subject results showed significant main effects of time [F (2, 52) = 18.33, p < .001, η_P² = .41], group [F (1, 26) = 4.54, p < .05, η_P² = .15], the time by group interaction [F (2, 52) = 4.01, p < .05, η_P² = .13]. Among the MBI group, the within-subject results showed the mean perceived health scores differed significantly between the time points [F (2, 26) = 16.59, p < .001, η_P² = .56]. Post-hoc tests revealed the mean perceived health scores increased by 0.86 (p < .01) at T1, further increased by 0.07 at T2 (p = .97) and overall increased by 0.93 (p < .001) from T0 to T2. Among the RT group, results showed the mean perceived health scores differed significantly between time points [F (2, 26) = 6.16, p < .01, η_P² = .32]. Post-hoc tests revealed the mean perceived health scores increased by 0.64 (p < .05) at T1 but decreased by 0.42 at T2 (p < .05) and overall increased by 0.21 (p = .71) from T0 to T2.

Post-intervention qualitative interviews: immediately after the intervention (T1)

Four main themes and related subthemes emerged across interviews with the MBI participants at T1 (see Table 5).

Table 5. Main themes and subthemes at post-intervention and one-month follow-up interviews with the mindfulness-based intervention participants.

Main Themes		Subthemes
Post-intervention
Theme 1	Misconceptions about mindfulness from previous experience
Theme 2	Cultural relevance	(i) Content: language and context (ii) Sharing: social inclusion and collective intelligence (iii) Religion: consideration and incorporations
Theme 3	Integrate knowledge from REMIND 2.0 for everyday living	(i) Facilitating factors (ii) Difficult but challenging factors
Theme 4	Perceived benefits of REMIND 2.0 on personal recovery	(i) Transformation of self and intrapersonal relationship (ii) From emotion regulation to symptomsattenuation (iii) Transformation in interpersonal relationships (iv) Mindful compassionate hope for future
One-month Follow-up
Theme 1	Recollection of REMIND 2.0 content with personal experience
Theme 2	Living everyday with mindfulness and self-compassion	(i) Reinforcing factors (ii) Challenging yet possible
Theme 3	Continued to thrive through the personal recovery journey	(i) Self and intrapersonal relationship (ii) Emotion and symptom regulation (iii) Interpersonal relationships (iv) Hopes, goals and plans

Theme 1: Misconceptions About Mindfulness From Previous Experience

More than half of the participants had prior experience with MBIs that exclusively focused on mindfulness without incorporating self-compassion. The length of those MBIs varied, with most participants attending a single session or fewer than eight sessions that involved practices only. Some participants had experience with eight-session interventions such as MBCT and MBSR. Mindfulness was vaguely explained but focused on breathing, stretching, walking and eating.

Participants expressed a need for a clear understanding of the concepts and rationale behind all the practices. Initially, they perceived mindfulness as an airy-fairy concept, associating it with relaxation exercises for instant relief or simply doing things at a slower pace. They struggled relate mindfulness to their own lives and everyday experiences. A few participants found the practices ineffective for managing emotion, except for breathing and stretching that provided instant relief. As a result, they stopped practicing. Also, the absence of handouts and materials for reviewing what they had learnt during these interventions was mentioned. They highlighted a sense of distance in the relationships between the facilitator and group members, and among the group members. Despite these challenges, participants remained curious about the popularity of mindfulness and decided to give it another try. Nearly all participants intended to learn more about mindfulness and self-compassion for the purpose of self-care in the current MBI.

Theme 2: Cultural Relevance of REMIND 2.0

Cultural relevance was more apparent to the participants in three aspects: (i) content: language and context, (ii) sharing: social inclusion and collective intelligence, and (iii) religion: considerations and incorporations.

Content: Language and Context

The content of the current MBI was generally considered helpful in comprehending abstract and unfamiliar concepts among the participants. They appreciated that the content was presented in the local Hong Kong Chinese Language, which aided their comprehension and recall. Specifically, a few participants mentioned that the rhymed short phrases recited at the end of each practice and displayed on banners in the classroom were effective in helping remember the key content. Further, almost all considered the value of examples and metaphors that were relevant to the local context. These examples helped them to self-reflect, visualize the scenarios and better connect the information to their own lives, making it more relatable and applicable to their everyday experiences.

Notably, the concept of self-compassion initially was awkward to most participants due to its contradiction with traditional values that prioritize the needs of others over one’s own. The current MBI addressed these misconceptions and emphasized the importance of self-compassion in taking better care of oneself and others. This clarification made the concept more acceptable and appealing, and it became a goal they aspired to achieve.

Sharing: Social Inclusion and Collective Intelligence

All participants recognized the significance of the group dynamics in their transformative process in the current MBI. They felt a strong sense of connectedness and understanding among the group members, as they shared similar experiences of emotional distress, goals, and demographic backgrounds. The MBI provided a supportive and secure environment that fostered a sense of inclusion and belong, akin to a family.

The participants also acknowledge the collective intelligence that emerged through the discussions and sharing of experiences and resources with the group. This collective intelligence enhanced their self-understanding, promoted a sense of common humanity, reduced feelings of isolation, increase acceptance, and reduced self-judgment, as well as facilitated their learning. The group dynamic also encouraged participants to attend the current MBI and engage in regular mindfulness and self-compassion practices.

Religion: Considerations and Incorporations

Initially, some participants had concerns about the potential conflicts between mindfulness practice and their religious beliefs. However, these concerns were addressed through the explanations provided in the workbook and discussion with the facilitator and other group members. As a results, participants generally came to believe that mindfulness and compassion were compatible with their religious beliefs. They combined the practice with their religion’s teachings, which allowed for a more focused religious practice. This integration helped participants to understand life better, maintain their practice in everyday life, improve their emotional well-being and feel more at ease. Some participants also found their own practice helped address concerns from their friends about religious compatibility and even invited them to practice together.

Theme 3: Integrate Knowledge from REMIND 2.0 for Everyday Living

The Integration of knowledge refers to participant’s ability to apply mindfulness and self-compassion concepts and practices in their daily lives. While participants focused on facilitating factors that helped them integrated the knowledge into their daily life, they also acknowledged some challenges.

Facilitating Factors

Participants generally believed the MBI curriculum gradually assisted them in integrating mindfulness and self-compassion into their daily lives. The content provided theoretical knowledge and practical applications suitable for incorporation into various daily scenarios. Home assignments and practices were considered essential cornerstones for experimenting with different practices and gradually realizing mindfulness and self-compassion on a daily basis. Participants reported dedicating 20 to 30 minutes to daily practice, which facilitated self-reflections, exploration of suitable methods, and integration of the knowledge and practice in their daily routines.

Individual support from the facilitator was also helpful in integrating knowledge into daily life. Most participants viewed the facilitator as embodying the qualities of mindful awareness, non-judgmental, open-mindedness and self-compassion, which served as an inspiration. Individual follow-up calls allowed for personalized assistance tailored to each participant’s background, expectations, needs, abilities and individual differences, such as physical and mental health conditions, religion, social roles, lifestyle and interests. For example, the facilitator provides personalized assistance to each participant in exploring further and identifying suitable practice methods.

Supplementary materials, such as workbooks, banners with overview sentences, and objects like raw crystals for object mindfulness practice, acted as reminders and helped in integrating knowledge into daily life. The gratitude journal also served as a reminder to attend positive events, and a few participants aimed to write down positive vocabulary and experiences to share with others.

Difficult but Challenging Factors

Some participants explored the difficulties but reframed them as “challenges” when integrating mindfulness and self-compassion into their everyday lives. Some found it challenging to apply these practices during emotional distress or stressful circumstances, as they tended to react automatically without much self-control. However, most participants recognized these circumstances as valuable opportunities for practice.

Fatigue and busy lifestyles were also seen as potential obstacles to practicing mindfulness and self-compassion. Nevertheless, participants emphasized the importance of maintaining mindful awareness and self-compassion even in these challenging situations and found ways to adapt their practice, such as in a comfortable, static position.

Theme 4: Perceived Benefits of the REMIND 2.0 on Personal Recovery

Almost all participants believed that the current MBI exceeded their initial expectations. They developed a deeper understanding of mindfulness and self-compassion and learnt how to apply these concepts and practices in their daily lives. This newfound knowledge led to personal enlightenment and growth. They also experienced an expansion of their social support network through this MBI. They hoped to continue developing these experiences and incorporating mindfulness and self-compassion into their ongoing recovery journey.

Transformation of Self and Intrapersonal Relationship

Considering the self, most participants described the development of their “wise minds,” which involved increased self-awareness, reflection, and a focus on balancing their logical and emotional minds. They reported making more effective responses rather than reacting impulsively and improving their ability to articulate their experiences. This newfound self-awareness led to an increase self-compassion, with participants becoming less judgmental and more accepting of themselves. They were more receptive to change and less clinging to their experiences, especially during difficult times.

Participants also cultivated friendly relationships with themselves by being present and compassionate to themselves and attending to their needs, caring for themselves and seeking help when necessary, as they would do for their loved ones. They became more disciplined, courageous, and motivated to overcome challenges.

From Emotion Regulation to Symptoms Attenuation

Participants developed alternative insights into how to view, respond to, and relate to their emotions and symptoms. Rather than suppressing or avoiding their emotions, they learnt to process them by meeting and observing their internal experiences in the present moment. This allowed them to either let go of these experiences or respond effectively through self-care and compassion. They were able to openly meet and accept their emotions, understanding them as part of common humanity. A few participants also learnt to be present with their internal experiences without ruminating or spiraling down, recognizing them as independent and impermanent events. Even in the moments of uncontrollability, they could regulate their emotions by anchoring themselves to the present moment and implementing emergency strategies learnt from the MBI.

As a result, most participants reported greater acceptance and regained control over their symptoms as they felt more relaxed and relieved and experienced fewer extreme emotions, including depression, agitation, frustration, and anger. A participant mentioned that the doctor had reduced the dosage of psychotropic drugs in the recent medical follow-up session, which the participant attributed, at least indirectly, to their knowledge and practice of mindfulness and self-compassion.

Transformation in Interpersonal Relationships

Based on the transformations mentioned in the previous sections, participants generally observed improvements, particularly within their families. Due to the COVID-19 pandemic, most participants spent more time with their family members and less time with their friends. This shift resulted in some challenges as a few of them had more frequent conflicts with their families. They were aware that family members tend to take each other for granted without awareness.

Nevertheless, all participants were able to approach and relate to their families with compassion. They became less judgmental and more accepting, forgiving, empathetic, and appreciative of their families. They also became mindful of the impact of their reactions had on their relationships and learnt to pause before reacting, leading to fewer arguments. Some participants mentioned a shift from criticizing or attempting to solve problems for others to being present, mindfully listening, communicating constructively, and resolving issues collaboratively. Improved self-awareness led to better descriptions and expressions of self, resulting in fewer misunderstandings and improved communication. Most importantly, the practice of mindful compassionate self-care also contributed to more fulfilling relationships.

Some individuals maintained close relationships with their friends through phone and instant messaging, allowing them to exchange compassionate and positive responses. They even shared tips on mindfulness and self-compassion with their family and friends, hoping to improve together.

Mindful Compassionate Hope for Future

Participants in the MBI expressed their commitment to practicing mindfulness and self-compassion daily after the MBI. Some participants also desired to cultivate gratitude and acceptance, reduce judgmental thoughts, and become more mindful. They recognized that self-compassion motivated self-care through mindfulness and expressed a determination to become their own closest friends to be present with themselves with more compassion and understanding, especially during difficult times.

Almost all expressed a desire to increase their self-compassion to better care for others and contribute to their community. They believed in the interconnectedness of all beings and anticipated further improvement in their overall well-being and relationships through continued practice. They embraced an optimistic outlook, being open-minded to the possibilities in life and living their daily lives contentedly, even in the face of challenges, alongside their community.

Follow-up qualitative interviews: 1-month after the intervention (T2)

Three main themes and related subthemes were identified across the follow-up interviews with the REMIND 2.0 MBI participants at T2 (see Table 5).

Theme 1: Recollection of REMIND 2.0 Content with Personal Experience

Overall, the participants demonstrated a strong ability to recall extensive information related to mindfulness and self-compassion. They could recall not only the concepts, exercises, and practices but also the metaphors presented in the content of the MBI.

Participants had two logical sequences when recalling the MBI content. One began with mindful awareness of their internal and external experiences, followed by their understanding of cognition and completed with concepts of self-compassion, self-smoothing and self-care. The other began with recalling the concept of self-compassion, followed by their understanding of cognition and mindfulness practice as a way of self-care, and ended with self-compassion to self-smooth and self-care. Both sequences interweaved mindfulness, understanding of cognition and self-compassion.

Interestingly, several participants showed a natural tendency also to describe how they had fit those content, including concepts, exercises and practices, in their everyday life while recalling the MBI content. They incorporated those content into their daily activities and social interactions. They also found the content memorable as it was applicable to their needs and helpful, which they practiced daily and extensively discussed and shared within the group during and after the MBI sessions.

Theme 2: Living Everyday with Mindfulness and Self-compassion

Almost all participants incorporated mindfulness and self-compassion in their daily lives. They engaged in both formal mindfulness practices and integrate mindfulness into their routine activities. The practice of self-compassion motivated them to care for themselves while remaining attuned to their experiences and showing compassion toward themselves. Some participants dedicated 10 to 30 minutes, three times per week or once a day, to formal mindfulness practice. They found that this practice helped them to be able to practice mindfulness and self-compassion in their daily routines.

Reinforcing Factors

All participants identified factors that reinforced their commitment to living mindfully with self-compassion. They found that self-compassion served as a motivation for them to continue to practice as a form of self-care, and mindfulness increased their awareness of the importance of self-care and caring for others.

Staying connected with others in their MBI group provides social support, which helped them sustain their daily practice. Through sharing experiences and resources, interacting with group members strengthened their commitment to practice and facilitated positive exchanges and encouragement.

Participants found the MBI workbook helpful in reinforcing the concepts and practices of mindfulness and self-compassion. Some participants integrated these practices into daily habits, recognizing their value in enhancing their favorite activities.

Most importantly, the participants wished to maintain the benefits they experienced from practicing mindfulness and self-compassion and continue to make further improvements. The visible improvements they observed in themselves, as well as feedback from others, such as their family members, served as reinforcement for their commitment to the practice.

Challenging yet Possible

Whilst participants viewed the integration of mindfulness and self-compassion into daily life as feasible, they also acknowledged the challenges that arose. Personal circumstances played a significant role, with some participants noting that emotional turmoil required additional mental effort. Physical exhaustion, laziness, or busyness also demanded mental effort to engage in the practice.

Besides, environmental distractions were seen as potential obstacles, but participants found ways to overcome them by practicing in different locations. The flexibility of mindfulness practices, which can be done anywhere and at any time, made it more feasible for participants to continue their practice. Despite the challenges, most participants viewed them as opportunities for self-care and further practice.

Theme 3: Continued to Thrive Through the Personal Recovery Journey

Participants maintained positive impacts from T1 and continued to thrive in their personal recovery journeys. They also developed new plans to support their on-going, drawing on the knowledge and skills gained from this MBI.

Self and Intrapersonal Relationship

Participants observed ongoing evolution of their sense of self and intrapersonal relationships as a result of the MBI. They developed a more balanced perspective that enabled them to effectively respond to internal and external experiences. Increased self-awareness and reflection facilitated the development of “wise mind,” which allowed them to better balance emotional and rational thought processes. They reported improved concentration and intentional responses instead of automatic reactions, which were the outcomes of their daily practice.

In terms of self-relationship, participants worked on cultivating self-compassionate, treating themselves as they would treat a good friend. They became more present, understanding, and accepting of their experiences, reducing self-judgment. Some participants replaced self-criticism with compassionate affirmations. Developing a compassionate relationship with themselves empowered them to face challenges and prioritize self-care on a daily basis.

Participants recognized the importance of seeking help from their social network or professionals when needed. They found that practicing mindfulness and self-compassion helped them effectively communicate their experiences and needs to receive appropriate care.

Emotion and Symptom Regulation

Participants found that mindfulness practice helped them decenter from their emotions, leading to improved regulation of emotions and symptoms. They began to see their emotions and symptoms as temporary and not defining their identity. Regular practice enhanced their mastery of their emotions and symptoms by allowing them to pause, step back from negative emotional experiences, and respond effectively with self-compassion and the wisdom gained through practices. Rather than avoiding or resisting emotions, they learned to acknowledge, coexist with, and let go of them.

By continuously practicing mindful self-compassion, participants reported more effective emotion and symptom regulation, increased self-awareness, and improved emotional regulation. They experienced reduced symptoms associated with their psychological conditions and felt less disturbed by these symptoms. Despite these improvements, participants emphasized the importance of medication as part of self-care. Nevertheless, a participant mentioned that pro-re nata psychotropic medications were no longer needed for emotion regulation.

Participants generally reported feeling emotionally more stable, calmer, and relaxed, with reduced stress levels. They mentioned improvements in sleep quality, vitality and a decrease in somatic symptoms. Negative emotions such as agitation, frustration, anger, rumination, depression and anxiety were reduced, while positive emotions such as happiness, satisfaction, warmth, and excitement increased in their daily lives.

Interpersonal Relationships

Participants maintained and slightly improved their interpersonal relationships, building upon the personal changes they experienced through continuous mindful self-compassion practice. For instance, listening and staying present mindfully rather than repeatedly dwelling on past grievances and expectations while avoiding being judgmental and defensive are essential. They related the mindful self-compassion skills to and found them helpful to maintain these attitudes in daily interactions, continuously leading to steady friendships and positive dynamics in family relationships.

Participants in the study have become more aware of how their responses impacted others, and they reflected and thought of effective responses, which usually had positive outcomes. Accepting and non-judgmental, considering others’ perspectives, and recognizing our common humanity supported their effective responses. This also helped them to let go of past grievances and expectations radically, as well as to reduce misunderstandings and improve relationships among family members. Participants felt more mutual understanding and closer to their family members, and could express affection openly.

Participants emphasized the importance of understanding and effective communication with family members, while aligning with realistic expectations and being more aware of the underlying positive intentions among family members. They were also becoming more grateful and appreciative of their family. They are working together with their family to enhance their well-being through sharing what they have learnt from the mindfulness and self-compassion practice.

Hopes, Goals and Plans

All participants expressed their intentions to continue practicing mindful self-compassion daily as they recognized its benefits for themselves and others. They understood the importance of self-care and viewed daily practice as a means to better care for their own well-being. They also planned on sharing their experiences and working on developing self-compassion and gratitude as part of their daily routine.

Participants believed in the interconnectedness of the community and expressed their hopes that the wider community could also benefit from practicing mindfulness and self-compassion. They saw it as a way to improve overall well-being and foster a more compassionate and supportive society. They recognized the positive impact of their own practice on their relationships and were hopeful that by spreading their experiences and knowledge, they could inspire and encourage others to practice mindfulness and self-compassion. They mentioned the MBI offered them social support and found solace and purpose in their participation, which both empowered them to face life and the future with a positive outlook. The new friendships they formed during the MBI provided them with companionship and a share journey of growth and exploration.

Regular practice of mindfulness and self-compassion had gradually improved the lives of the participants, and they found motivation and hope through their daily practice. They recognized the transformative power of mindfulness and self-compassion and were optimistic about the positive changes it had brought to their live. At the same time, they also acknowledged the inevitability of future challenges and opportunities for personal growth. They expressed a willingness to embrace these challenges and saw them as opportunities for further development and learning.

Discussion

The current pilot RCT was conducted to assess the feasibility and acceptability of an MBI (REMIND 2.0) developed to facilitate personal recovery in people with mental illnesses in the community. The MBI was compared to the RT control group. Results indicated that the MBI group showed improvements in all primary outcomes, including personal recovery, mindfulness, self-compassion, and resilience at T1 that were maintained at T2, while the RT group showed declines in all primary outcomes at T1 that continued to decline at T2. Both MBI and RT groups effectively improved secondary outcomes, including depression, anxiety, stress, positive and negative moods, quality of life, and perceived health. Notably, compared to RT, MBI had more significant positive effects in both primary and secondary outcomes, except for stress and anxiety.

Results from the qualitative interviews provided further insights into the positive impacts of the MBI on personal recovery, psychological and physical well-being. Participants highlighted various aspects of the MBI that contributed to their positive experiences, such as its design, content, materials, facilitators, personalized guidance, and social support. The MBI helped participants cultivate mindfulness and self-compassion, leading to reductions in anxiety and stressed and an increase in self-compassion, mindfulness and open-mindedness. Participants reported improved emotional stability, symptom management, and intra- and interpersonal relationships, as well as renewed hope and goals. Studies have shown that MBI can enhance hopefulness (Astuti et al., 2020), which hope was considered as crucial in the change process and promoting personal and functional recovery among people with mental illness (Acharya & Agius, 2017; Coşkun & Şahin Altun, 2018; Hayes et al., 2017).

MBI participants showed more significant improvement in most outcomes, the MBI were not superior to RT in reducing anxiety and stress. Previous evidence have shown that while MBIs promote positive outcomes, they are not significantly more effective than RT in areas such as anxiety (Jain et al., 2007; Lancaster et al., 2016). A systematic review also found that online interventions focused on mindfulness or relaxation were comparatively effective in reducing anxiety (Mikolasek et al., 2018). Other systematic reviews showed the superiority of the effects and maintenance benefits among the MBI group compared to active controls on psychiatric symptoms and resilience, but did not differ from active controls on anxiety at post-intervention and stress at follow-up, in both general adult and mental health population (Goldberg et al., 2018, 2022; Joyce et al., 2018).

MBI participants showed significant improvements in all outcomes at T1, which were maintained at T2, except for positive mood. The sustained benefits could be explained by the participants actively incorporated mindfulness and self-compassion into their daily lives. They engaged in comprehensive practices that allowed them to practice anywhere and anytime, reframed difficulties as growth opportunities, and received continued social support from the group. Social support was found to be essential in sustaining the participants’ practice during and after the MBI, as studies have supported (Tollstedt, 2017; Van Aalderen et al., 2014; Wyatt et al., 2014). Yet the decline in positive mood may be attributed to the loss of weekly group activities. During the COVID-19 pandemic, the general population in Hong Kong experienced significant increases in unhappiness, depression, stress, and anxiety, aggregating to the prolonged fear, restrictions, and societal issues encountered in daily life (Cheung et al., 2021; Zhao et al., 2020). Further, the intention of mindfulness was not to directly increase positive mood in the short term. Instead, it aimed to help individuals become more aware of their experiences and cultivate acceptance in the long term (Quaglia et al., 2014).

In contrast to MBI, RT participants only showed significant improvements in all secondary outcomes that related to psychological distress, mood and quality of life at T1. However, at T2, all outcomes significantly declined, except for anxiety and stress. This might be explained by their distinct approaches in achieving similar effects on anxiety and stress. MBI, on the other hand, might have a broader influence on a range of outcomes and might have longer-term effects compared to RT. Although both MBI and RT are mind-body approaches, the differences in their effects can be attributed to the differences in their theoretical bases and practices and resulting in distinct outcomes (Luberto et al., 2020). MBI emphasizes being aware of habitual, automatic reactions that cause suffering and reducing reactivity to increase effective responses. RT emphasizes eliciting relaxation responses to counteract chronic stress and related diseases to promote health outcomes (Esch et al., 2003). The MBI practices focus on being present at the moment and accepting experiences without judgment. They foster awareness of internal and external experiences, such as thoughts, feelings, emotions and body sensation. In contrast, RT practices focus on inducing a relaxed state through muscle control, breathing, or imagination. They prioritizes the voluntary control of muscles, breathing, or imagination of scenarios. Rausch et al. (2006) found that meditation and RT were more effective than treatment-as-usual in reducing cognitive, somatic and state anxiety. RT demonstrated potential effectiveness in reducing physiological arousal-related outcomes. Previous research indicated that MBIs may offer a broader range of psychological benefits compared to RT. This could be attributed to a complex mechanism that extends beyond physiological changes (Smith et al., 2017; Wolf & Serpa, 2015).

With distinct theoretical basis and objectives, MBI aims to liberate the mind and alleviate suffering in the long term, while RT focuses on physiological body control through relaxation techniques to provide short-term relief and assess mental resources. MBI empowers participants by influencing their way of living and attitude, while RT empowers participants with practical, on-the-spot techniques to achieve the relaxation state. Interestingly, the current results have shown that both MBI and RT appeared to have potential overlaps in the effects on anxiety and stress. However, MBI had more holistic and enduring positive impacts compared to RT.

Limitations and future research implications

The first limitation of the current study was inherent to most psychological intervention RCTs – it was single-blinded, as double-blinding was not feasible. Nevertheless, the researcher was blinded to group allocations until the completion of the analysis.

Another limitation was the single center and small sample of the current pilot study raises concerns about the generalizability of the findings. Participants primarily consisted of middle-aged women, with at least secondary education, residing in private housing. Previous studies indicated a higher likelihood of females engaging in mindfulness or meditation interventions (Bodenlos et al., 2017; Upchurch & Johnson, 2019). MBIs studies in the Hong Kong community mental health settings also involved predominately female participants (Lam et al., 2020; Ting et al., 2020). Future research could explore such gender disparities.

Despite the small sample, the qualitative data saturation was reached. The current study provided quantitative estimates and in-depth qualitative information for further development of MBI for mental health service recipients. Due to resource constraints, only MBI participants were interviewed in the current study. Future interviews could include interviews with participants who have experienced either or both MBI and RT to investigate their experiences and understanding of both interventions. This would provide a more comprehensive understanding of the interventions from the participants’ perspectives.

Conclusions

The REMIND 2.0 MBI was designed to address the needs of individuals in mental health recovery and promote personal recovery. The findings from this pilot phase support the feasibility and acceptability of REMIND 2.0 among mental health service users. Although there was no significant difference in reducing anxiety and stress compared to the active RT controls, the MBI group demonstrated greater significant improvement across the recovery, psychological and quality of life outcomes. The results highlight the potential effectiveness of REMIND 2.0 and warrant future fully powered definitive RCTs to comprehensively investigate its effectiveness.

Acknowledgments

We extend our heartfelt appreciation to the entire team, staff and participants from all the service units, who actively supported this research. We are truly grateful for their unyielding valuable contributions, particularly during the arduous COVID-19 pandemic.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Participants in this study were assured that their sensitive and personal information would be kept confidential. Raw research data would not be publicly accessible, and individual data was aggregated and summarized.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.