Web-based intervention to improve quality of life in late stage bipolar disorder (ORBIT): randomised controlled trial protocol

Current adjunctive psychological treatments for BD emphasise relapse-prevention by monitoring triggers and avoiding stress. However, this may be detrimental to self-esteem in late stage BD where relapse is often unrelated to discernible life events [5]. Indeed, having experienced more than 12 episodes of BD has been found to predict a negative response to cognitive behaviour therapy (CBT) [6]. In late stage BD, therefore, symptom-focused treatments may be less beneficial than approaches that recognise the unavoidability of suffering, emphasise redefinition of life goals, and prioritise QoL outcomes [7, 8]. These priorities are consistent with the “third wave” [9] or contextual cognitive-behavioural therapies [10], which typically prioritise mindfulness - an awareness of present experience and a non-judgemental stance towards that experience [11].

Empirical research into mindfulness based interventions (MBIs) for BD has primarily explored face-to-face delivery, with Mindfulness Based Cognitive Therapy (MBCT) [12] the primary target of investigation. Studies report benefits for BD symptoms and associated psychological variables, in particular anxiety and emotion regulation [13]. Beneficial effects of MBIs in other domains have been reported including QoL [14], well-being, stress and rumination [15]. Evidence for effects on relapse prevention is limited: one study found no benefits for bipolar relapse [16], while another found decreased depressive relapse in a bipolar subset of people with recurrent depression [17]. A recent meta-analysis on the efficacy of MBIs as an adjunctive treatment for BD showed significantly beneficial effects on depressive and anxiety symptoms in within-group analysis [18]. However, this significance was not observed in comparison with control groups, limiting conclusions due to the small number of controlled studies and small sample sizes. Furthermore, studies examining mindfulness as part of another treatment modality (e.g., Acceptance and Commitment Therapy; ACT) were excluded from their review, as were short-duration (< 3 week) interventions and self-help interventions delivered online [8]. A further limitation of studies to date is their focus on symptom/distress reduction, which is at odds with the main goals of CCBT approaches [19]. Mechanisms of action of such approaches are still unclear, as process variables (e.g., mindfulness, acceptance, self-compassion) have not been examined in mediational analyses. In summary, further clinical trials are needed to determine the efficacy of MBIs for improving QoL in BD and understanding their mechanisms of action.

Web-based psychological interventions have great potential to complement treatment as usual and overcome barriers to accessing psychological assistance for BD (e.g., cost, time, trust in professionals) [20]. Such interventions have demonstrated short- and long-term benefits for a range of mental disorders [21] and are acceptable to people with BD [22, 23]. There is a lack of consistent evidence regarding effectiveness of web-based psychological interventions for BD as most trials to date have focused on feasibility. Recent findings confirm that MBIs are effectively delivered via the web [24], including for mood disorder populations [25]. The benefits of low intensity web-based interventions are yet to be widely disseminated to people with BD, highlighting difficulties in translation into practice.

Our international team of researchers, clinicians and consumers developed and piloted a novel, guided self-help web-based psychological intervention to improve QoL in late stage BD [8]. The low-intensity MBI is specifically tailored for late stage BD, drawing from Acceptance and Commitment Therapy (ACT), Mindfulness Based Cognitive Therapy (MBCT) and Compassion-focused Therapy (CFT). The intervention was originally named ORBIT (Online, Recovery-oriented, Bipolar Individualised Tool), but to keep participants unaware of the present trial’s hypothesised superior condition, ORBIT now refers to the project as a whole. For this report, the active intervention is titled ‘Mindfulness for Bipolar 2.0’ and the active control condition is titled ‘Psychoeducation for Bipolar’, but: to minimise expectancy effects, the names and key features of each arm are not mentioned in recruitment or consent processes.

Published pilot data suggests the first iteration of Mindfulness for Bipolar was feasible, safe and potentially effective in improving QoL [8]. No further published studies have investigated web-based MBIs specifically tailored for late stage BD. However, in related severe and chronic mental illness populations, there is support for the efficacy of generic MBIs and MBCT [26, 27], as well as ACT [28, 29]. Evidence suggests that those with chronic mental illness benefit from mindfulness strategies to mitigate the distressing effects of symptoms (via mindful acceptance of internal experiences), and improve self-concept (via promoting an experience of self as observer, moderating self-evaluations and encouraging commitment to valued goals despite symptoms) [30]. Indeed, there is evidence that those with a significant BD history may specifically benefit from a bipolar-tailored version of MBCT [15, 31, 32].

Building on pilot study findings, we now describe the development of and protocol for a definitive RCT of the second iteration of the MBI (Mindfulness for Bipolar 2.0) for improving QoL in late stage BD.

The overarching aim of the present RCT is to assess the effectiveness and cost-effectiveness of a novel, web-based intervention in improving QoL in late stage BD vs. an active control.

The trial is a prospective, parallel group, rater-blind, superiority RCT with a one-to-one allocation ratio comparing the Mindfulness for Bipolar 2.0 intervention with a structurally equivalent, validated and therapeutically credible active control condition (Psychoeducation for Bipolar). The primary endpoint is immediately post-treatment (upon completion of the 5-week intervention). Follow-up time points are at 3-months and 6-months. Both interventions are guided self-help, and not intended to replace treatment as usual. The study setting is online, and participation in both arms occurs through a secure server.

Participants will be sequentially allocated to intervention arms using a one-to-one ratio by predetermined permuted block randomisation with a block size of 10. The permutation sequence within blocks will be randomly generated by SAS Version 9.4, overseen by an off-site statistician, and coded into the website so randomisation is automated and free of potential allocation bias. Participants are unaware of the primary hypothesis of the study, and have no prior knowledge of the two intervention types, but are unavoidably aware of the intervention to which they have been randomised. To reduce expectancy effects for superiority of the intervention over the control condition, the trial is framed as a comparison between two interventions designed to improve QoL. Assessors will remain unaware of treatment condition, and assessors who become aware of the treatment condition will, if possible, be replaced. Success in keeping assessors unaware of the treatment condition will be quantified by asking interviewers to guess allocation upon completion of the RCT, and assessing whether they can do so at greater than chance levels. Statistical analyses will be conducted blind to treatment allocation by an off-site statistician.

The study will be administered and conducted through a single-site, with recruitment primarily via open social media sites (e.g., International Bipolar Foundation Facebook site). Recruitment will also be facilitated in four English-speaking countries via advertisements disseminated online through public and mental health sites and social media, listservs and traditional advertisements (offline) posted in clinical services.

As summarised in the flow diagram (see Fig. 1), individuals interested in participating will be invited to visit the study website (https://​www.​orbitonline.​org/​) where they can sign up, view and provide their consent to the participant information and consent form, provide their contact details (including details of their medical health professional), and respond to initial screening questions regarding eligibility (i.e., received a diagnosis of BD from a mental health professional, experienced 10 or more mood episodes).

An assessor from the research team will then contact potential participants via telephone to conduct the first component of the baseline assessment (T0) to confirm eligibility criteria and assess comorbidities. Potential participants deemed eligible except for current episode status/psychotic symptoms/active suicidality (see exclusion criteria below) will be offered a later assessment (one month following initial phone call) to determine eligibility at that time. Eligible participants will complete the second component of the baseline assessment (online questionnaires) (T0), followed by randomisation. An online coach will be assigned to the participant (contact initiated via a ‘welcome’ message). Participants will be asked to complete post-intervention (5-weeks, T1), 3-month (T2) and 6-month (T3) post-baseline assessments, involving a telephone interview and online questionnaires (see Table 1 for schedule of assessments). Participation in each assessment (phone and online components) will be compensated with USD$25 gift vouchers.

Assessors (masters-level in psychology) will be trained on study measures. Inter-rater reliability will be established for key observer-rated outcome measures (i.e. Montgomery-Asberg Depression Rating Scale, MADRS; Young Mania Rating Scale, YMRS), telephone assessment interviews will be recorded (except for participants who decline), and inter-rater checks will be conducted every three months. If rater drift is identified, assessors will be retrained.

Online questions will be made mandatory where possible to reduce missing data. Comprehensive data cleaning (range and other distributional checks, comparison with published norms where relevant) will be conducted prior to analyses.

To ensure ready translation, minimally restrictive inclusion/exclusion criteria have been set. Inclusion criteria are as follows:

Exclusion criteria are as follows:

Risk management procedures have been developed through our experience with other online interventions and websites for BD [22, 39‐43] and through consultation with the CREST.BD Community Advisory Group. Procedures pay particular attention to suicide risk, distinguish between solicited and spontaneously reported events, and are reviewed weekly during the trial by the study’s executive committee. To minimise impact of any adverse events, being under the care of a medical practitioner is an inclusion criterion, and participants consent to this professional potentially being contacted by the researchers as part of a ‘red flag decision tree’ (detailed below). The informed consent statement explains that the medical practitioner or local emergency department remains the participant’s first point of contact. Both the pre-registration page and informed consent statement highlight participants’ central role in their own safety and wellbeing, emphasising that the website does not act as an emergency service, and that the website is not monitored in real time. Links to emergency resources will be provided on the website (e.g., unsuicide. wikispaces.com ).

Participants will be withdrawn from the trial if their participation compromises clinical care as determined by the study’s executive committee. On a case-by-case basis, this might include hospital admission, acute hypomania or mania, psychotic symptoms or active suicidal intent.

Adverse events will be monitored via questions at post-test and both follow-up time points, and any adverse events incidentally reported through participant contact with coaches, forum posts, or assessments will also be recorded. Severity of any adverse events, and causal relationship to the trial will be assessed through participant self-report and logged. Adverse events will be reviewed at monthly executive meetings and 6-monthly meetings of the independent trial committee (who will also explore patterns of adverse events by condition). Serious adverse events suspected or known to be related to participation in the trial will be reported to Swinburne University of Technology Human Research Ethics Committee.

Clinical risks emerging will be addressed through a detailed ‘red flag decision tree’ which captures the various points at which the researchers could become aware of risk information (particularly suicidality), namely, baseline and follow-up (post-intervention, 3-month, 6-month) assessment interviews, posted comments on the website (e.g., forum content) and communications with the online coach (see below). Red flag events are those where there is immediate risk of harm (e.g., scoring high on suicide items, forum posts or coach emails suggesting active suicidality), as contrasted with other concerns highly prevalent in this population (e.g., comments about waxing and waning symptoms, connectedness or distress). Only in the former case have participants consented for researchers to break confidentiality and communicate with their clinician or emergency services, and these events are considered red flags for action of some kind. The decision tree distinguishes between red flag information suggesting immediate risk of harm for which real-time intervention is feasible/recommended (e.g., when active suicidality is identified during a phone assessment), and when it is not (e.g., when the research team becomes aware of active suicidality mentioned in a forum post from 48 h previous). All trial staff will be comprehensively trained on protocol procedures.

To minimise attrition and non-adherence the intervention is brief, with a 5-week structured participation time, with each new module released sequentially. Participants will complete modules at their own pace: content within modules is organised into ‘chunks’ allowing exploration in briefer or longer individual sessions, depending on personal preferences. Suggested guidelines on using the program to maximise benefit will be provided, including spending a minimum of 1–2 h each week on content and ongoing skills-practice (with daily practice recommended for mindfulness skills and mood monitoring).

Both interventions comprise four modules, with new module content delivered sequentially each week over four weeks. The fifth week is positioned as an opportunity to consolidate skill development. Participants can return to previous modules across the program at any time. To support ongoing generalisation of skills, participants retain access to the program for the 6-month follow-up period.

To maximise engagement, the interventions follow best practice in persuasive system design. Features known to impact engagement are included: (1) dialogue support (praise from coach and forum moderator, email reminders), (2) social support (social facilitation through discussion threads in moderated forums) and (3) primary task support (best practice principles for modularisation of content, personalisation/monitoring of progress, prompted self-monitoring, and rehearsal) [44]. Website presenters include experienced clinical academics and consumers with lived experience of BD. Topics within modules are organised primarily around videos (2–3 min in length), followed by text, reflective exercises, audio exercises and additional information PDFs. Participants can comment publically on videos (via comments feed) and complete exercises which are private but can be shared with others users of the website.

Coaching support improves adherence to web interventions [40]. Both interventions therefore include personal coaching support through asynchronous email contact with a trained online coach. Participants can send as many messages as they like to their coach, and will receive one response per week. Completed activities can be shared with the coach to facilitate skill development. A moderated forum is included to develop a sense of online community. The forum is moderated by trained consumers with lived experience of BD, who will also seed discussion threads with encouraging tips.

The schedule of assessments is provided in Table 1. Assessments are performed as close as practicable to specified time points. All clinician-rated assessments are conducted by telephone by expertly trained assessors, unaware of treatment allocation. All self-report assessments are completed online via a secure, encrypted online survey platform (Qualtrics). Participants are contacted via email to arrange a time for phone assessments and to prompt completion of online assessments. If an email response is not received within two days, an SMS reminder is sent to the participant. Assessors will monitor the completion of questionnaires on a regular basis and make telephone calls to facilitate completion where such support may be required. The MINI is conducted via telephone at baseline to confirm eligibility, establish comorbid diagnoses and record demographic variables.

Sample size was determined by power analysis using G*Power 3. Analyses were conducted on the primary endpoint (Brief QoL.BD score) at T1 relative to baseline. Based on our pilot study (intent-to-treat d = 0.52), and Brief QoL.BD change in comparable RCTs [22, 51], a between-group effect size of d = 0.4 was conservatively estimated. This small-to-moderate effect size is comparable to effect sizes found for adjunctive CBT on a range of outcomes in BD [78], and therefore can be considered a clinically important difference given the new intervention is low intensity, low cost, and high access. A sample size of 200 (100 in Mindfulness for Bipolar 2.0 and 100 in Psychoeducation for Bipolar) would provide at least 80% power (1-β) to detect an effect of this size at α = 0.05 (two-tailed). Attrition at immediate follow-up was conservatively estimated at 33% based on attrition rates in the pilot study and our earlier RCT of a web-based intervention for BD [39], so N = 300 randomised participants are required to generate the required sample size of 200 for the primary analysis.

Recruitment is ongoing. It is anticipated that the trial will be completed (T3) by September 2019.