Design and sample size
This is a parallel-groups, superiority pragmatic RCT with 1:1 allocation to MBCT-SH or CBT-SH, with blinded assessments at all time points. Participants will be blind to the hypothesised direction of effects. Four hundred and ten people meeting eligibility criteria for major depressive disorder or mixed anxiety and depression will be randomly allocated to receive MBCT-SH or CBT-SH, along with six sessions of support from a PWP. Participants will complete measures at baseline (Time 0), 16 weeks post-randomisation (post-intervention, Time 1) and 42 weeks post randomisation (6-months follow-up, Time 2). In addition, 24 participants will be interviewed about their experiences of both self-help interventions (12 participants per arm) at Time 1 with a focus on better understanding barriers and facilitators to engaging in self-help interventions in IAPT services and investigating any negative experiences or effects of treatments.
The sample size was based on detecting a between-group effect size of 0.36 based on the difference between the reported between-group effect of CBT-SH (0.42)(20) and the reported between-group effect of MBCT-SH (0.78)(21). Recruiting 205 patients into each arm would provide 90% power to detect a between-group difference of 0.36 with a 5% alpha and a two-sided t-test whilst allowing for 20% attrition at post-intervention (as found in our pilot study). Therefore, a total sample size of 410 will be required.
The study design was informed through consultation with a PPI group chaired by a PPI co-applicant consultant (LL). Members of the PPI group had participated in a pilot study of the intervention or had lived experience of depression, mindfulness and/or CBT. The PPI group were instrumental in advising on a number of aspects of study design including the frequency and nature of support sessions and the study recruitment strategy and materials.
Participants
Participants will be recruited through ten IAPT services across England. A list of sites can be requested through the corresponding author.
Inclusion criteria are that participants: (1) be aged 18 years or over; (2) meet diagnostic criteria on the revised Clinical Interview Schedule (CIS-R)(22) for a primary diagnosis of a depressive episode, mixed anxiety and depression, or non-specified mild neurotic disorder; (3) score 10 or more on the PHQ-9(23) at their initial IAPT assessment (cut-off indicating a probable major depressive episode); and (4) have sufficient literacy skills to read and understand the self-help materials.
Exclusion criteria are that participants: (1) have severe symptoms of depression (a score 20 or more on the PHQ-9); (2) score of 4 on the CIS-R suicidality scale; and (3) express a strong preference (5/5) for one intervention over the other on the treatment preference question such that if randomised to the non-preferred intervention they would be likely to drop out of the intervention.
Measures
Demographics
A questionnaire will be completed to collect demographic information.
Diagnostic status
The Clinical Interview Schedule (CIS-R)(22) will be conducted at baseline (Time 0) to ascertain diagnostic status. The CIS-R is routinely used in primary care mental health and IAPT research(1) and has been validated for telephone completion(24).
Primary Outcome Measure
Depression symptom severity (PHQ-9)(23). The PHQ-9 is a 9-item self-report measure of depression symptom severity used in all IAPT services with good sensitivity and specificity. Items are rated on a four-point scale. Scores under 10 are considered sub-clinical, 10-14 mild, 15-19 moderate and 20+ severe. The PHQ-9 will be administered at Time 0, Time 1 and Time 2.
Secondary Outcome Measure
Generalised anxiety (GAD-7)(25). This is a 7-item measure of generalised anxiety used in IAPT. Items are rated on a 4-point scale and the measure has excellent psychometric properties(25). This will be administered at Time 0, Time 1 and Time 2.
Wellbeing (SWEMWS)(26). The short version of the Warwick Edinburgh Mental Wellbeing Scale consists of 7 questions rated on a 5-point scale designed to measure wellbeing. The scale has good psychometric properties and is used widely(27). This measure was added following advice from the Public and Patient Involvement (PPI) consultation panel. This will be administered at Time 0, Time 1 and Time 2.
Functioning (WSAS)(28). The Work and Social Adjustment Scale (WSAS) is a 5-item measure of daily occupational and social functioning that is used routinely in IAPT. This will be administered at Time 0, Time 1 and Time 2.
Mindfulness (FFMQ-15)(29). Mindfulness will be measured using 15-item version of the Five-Facet Mindfulness Questionnaire. This has excellent psychometric properties and is sensitive to change following MBCT(29). This will be administered at Time 0, Time 1 and Time 2.
The Change Interview(19) will be used to guide the qualitative component of the study. This is a widely-used semi-structured interview designed to explore participants’ experiences of psychological interventions and has been successfully used in previous studies by members our team. Questions about participants’ experiences of the effects of their allocated intervention on personally-relevant outcomes (i.e. not necessarily restricted to depressive symptoms) have been added at the end of the interview following advice from the study’s lived experience consultation panel. The Change Interview will be conducted by the PPI consultant co-applicant on the study (LL) with the expectation that this will facilitate participants to be open about their experiences of the study and interventions. This will be conducted at Time 1.
Health Economic Measures
Service use. An adapted version of the Adult Service Use Schedule (AD-SUS) will be used to measure individual-level all-cause hospital and community-based health and social care resource use. The AD-SUS was developed in previous research for use with people with common mental disorders(30,31) and adapted for the purpose of this study. The baseline (Time 0) AD-SUS will cover the period 3 months prior to randomisation. At the Time 1 and Time 2 follow-up interviews, the AD-SUS will cover the period since last interview.
Health-related quality of life (EQ-5D-5L)(32). The EQ-5D is a five-dimension, five-level, generic, preference-based measure of health-related quality of life covering mobility, self-care, usual activities, pain/discomfort and anxiety/depression. We will use the recently developed five level version, to maximise sensitivity(33). This measure will be administered at Time 0, Time 1 and Time 2.
Intervention Evaluation Measures and Tools
Intervention expectation form. This will be used to assess expectation of benefit and treatment credibility at Time 0.
Lasting negative effects questionnaire. This will be used to ask participants about any lasting negative effects of their allocated intervention at Time 2.
PWP rating scale. This will be used for participants to rate the quality/helpfulness of the sessions between the participant and their PWP at Time 1.
Weekly diaries. These record the extent to which participants are engaging with the self-help course each week during intervention delivery.
Engagement questionnaire (end of treatment). This records the extent to which participants engaged with the self-help course during the entire course of the intervention and will be administered at Time 1.
Engagement questionnaire (follow-up). This records the extent to which participants continued to engage with the self-help course following the end of the intervention and will be administered at Time 2.
Session attendance. Number of PWP sessions attended (0-6) and duration of each support session.
Treatment completion. This is defined as attending at least 50% of the PWP sessions (i.e. attending at least 3 sessions).
Procedure
See Figure 1 for the SPIRIT Schedule of enrolment, interventions, and assessments. Patients with a diagnosis of depression will be sought through IAPT services. The study will also be advertised in General Practitioner surgeries and elsewhere. Self-referrers will be guided to the study via their local IAPT service. Informed consent will be obtained from each participant. Potential participants will be given a copy of the study participant information sheet and will have the opportunity to discuss the study in person with the research assistant (RA) before signing the consent form (a copy of which can be obtained from the corresponding author).
Once the participant has consented to participate in the trial, the participant will complete the full set of baseline measures with an RA present in person or by phone. Measures will be completed online or on paper, depending on participant preference. Participants who do not meet eligibility criteria at the baseline assessment will be referred back to the person who conducted their initial assessment for usual care to be offered by the service.
At the end of the baseline assessment, eligible participants will be randomised to either the MBCT-SH or CBT-SH arm. Participants will then be given their allocated self-help workbook. Randomisation will be stratified by centre and PHQ-9 score (mild or moderate) using random block length. Eligible participants will be randomly allocated using the Sealed Envelope(34) online service. The study statistician will use Sealed Envelope to set up and test the randomisation procedure incorporating stratification by site and PHQ-9 severity category (mild or moderate) using random block length and 1:1 allocation. The statistician will not have any further involvement in the randomisation process. The RA will randomise participants by completing the online form with participant’s details. This will immediately show whether the participant is assigned to the MBCT-SH or CBT-SH arm and participants will be given their self-help workbook. Participants will not be told the hypotheses in relation to the arm to which they have been randomised.
Participants will then guide themselves through their allocated intervention over the 8-week course with six PWP support sessions. A maximum of 16 weeks is given for the intervention period to allow participants to complete their allocated 8-week course to take account of breaks for holidays, sickness etc. PWP support sessions may be offered by phone or face-to-face, depending on usual practice in the service and participant preference. This mirrors the usual way in which self-help interventions are offered in IAPT – i.e. offering a self-help workbook alongside a limited number of PWP support sessions.
Participants will complete measures online (with a postal option) at 16 weeks post-randomisation (post-intervention, Time 1) and 42 weeks post-randomisation (6-month follow-up, Time 2). Participants will have the option to choose whether to complete the Time 1 and Time 2 measures with an RA present in person or by phone (who will be blind to group allocation) or on their own. In the event that an RA is required to be present during Time 1 or Time 2 assessments and becomes unblinded the assessment will be completed by another, blinded RA. Where Time 1 and Time 2 assessments are not completed, participants will be contacted at weekly intervals for up to one month to remind them to complete these assessments, unless participants have discontinued from the study.
Twenty-four participants will be invited to take part in the qualitative Change Interview after their post-intervention (Time 1) quantitative assessment is completed. Participants will be interviewed on a first come, first served basis with 12 participants interviewed in each of two groups: (1) MBCT-SH intervention completers, and (2) CBT-SH intervention completers.
Therapy Protocols
MBCT Self-Help
The MBCT-SH workbook ‘The Mindful Way Workbook'(35), written for clinical populations, presents MBCT as a self-help package. MBCT-SH participants will be given the workbook and will be asked to guide themselves through the self-help course within a 16-week time period (the time period determined in our pilot). As is routine at Step 2, participants will be offered six PWP sessions to answer questions and provide encouragement.
PWPs currently train in CBT-SH. To match training between arms we will offer PWP training in MBCT-SH. The training involves PWPs: (1) attending an MBCT group as a participant and guiding themselves through the MBCT-SH workbook, or completing the MBCT course using the workbook as a guide, and (2) attending a two-day experiential mindfulness skills workshop. As is standard in delivering MBIs, PWPs will be encouraged to maintain their own personal mindfulness practice throughout the study, and this will be recorded using the PWP Mindfulness Practice Record. Fortnightly MBCT-SH group supervision for PWPs will be provided.
CBT Self-help
The CBT-SH workbook 'Overcoming Low Mood and Depression'(36) has evidence demonstrating its effectiveness in reducing depression symptom severity(37). This workbook is often used at Step 2 in IAPT.
Matched to the MBCT-SH condition, participants allocated to CBT-SH will be given a copy of their workbook and will be encouraged to guide themselves through within 16 weeks alongside six PWP sessions to answer questions and provide encouragement.
As in the MBCT-SH condition, fortnightly CBT-SH group supervision for PWPs will be provided.
Intervention Fidelity and Adherence
The same PWPs will deliver both interventions in order to minimise therapist effects, as some PWPs achieve substantially better outcomes than others. In order to minimise therapeutic drift and therapy contamination the PWP protocols are detailed and there will be fortnightly supervision of PWPs. PWPs will be asked to audio record at least one complete MBCT-SH case and at least one complete CBT-SH case. A random 10% sample of recordings from each PWP will be rated for fidelity to the therapy protocols by a clinical psychologist trained in CBT and MBCT and who is independent of the PWP training and supervision.
Participants will be prompted each week by the research assistant to complete weekly diaries (online/paper versions) to record the amount of workbook read and time spent engaged in intervention tasks.
(Serious) Adverse Events Monitoring
A protocol for identifying and independently assessing serious adverse events will ensure that such events are addressed in a timely fashion and responded to, including if a serious adverse event is classified as potentially study-related. Serious adverse events and their classification will be reported to the Data Safety Monitoring Board (DSMB) and Trial Steering Committee (TSC) and action will be taken as deemed necessary.
Serious and other adverse events will be discussed in intervention supervision (with CS or FJ) and in service-based clinical supervision in the relevant IAPT service. Where deemed in the best interests of participants, the study intervention may be discontinued and other treatment options may be recommended by the IAPT service.
Planned Data Analysis
The primary analysis will be a quantitative analysis of the primary outcome at the primary endpoint (Time 1) using an intention-to-treat approach (ITT) where participants are analysed as per their randomisation allocation regardless of treatment received.
Secondary analyses will consist of ITT analyses of the primary outcome at the follow-up time point (Time 2) and all secondary outcomes. A per protocol analysis will also be carried out for those participants receiving an adequate dose of their allocated intervention, defined as attending at least 50% of the PWP sessions (i.e. at least 3 sessions).
A descriptive summary of all measures will be provided by group (MBCT-SH & CBT-SH) and time point (Time 0, Time 1 and Time 2) as appropriate. Comparisons between groups for all measures will be carried out using independent t-tests and Chi-square testing for continuous and categorical data, respectively. Unstandardised effect sizes for the primary outcome and secondary outcomes will be estimated using linear mixed models with treatment group (MBCT-SH vs CBT-SH), time and a treatment group by time interaction entered as fixed factors; site, baseline PHQ-9 and baseline value of the outcome will be entered as covariates. Individual participants will be included in the analysis as random effects. Contrasts will be used as appropriate to estimate effects at different time points. A non-significant group by time interaction will imply common treatment effects at each time point.
Standardised (Cohen’s d) effect sizes for each outcome will be calculated by dividing the between-group unstandardized effect by the baseline pooled standard deviation.
95% confidence intervals will be calculated for all unstandardised estimates. Group differences in dichotomous outcomes at the different time points will be analysed in a similar way but using multilevel logistic regression models and baseline PHQ-9 scores. Baseline balance will be presented in the descriptive table broken down by study arm. No adjustment will be made for differences between co-variates at baseline. Outliers will be removed if and only if they look erroneous. Scores at the extreme will not be removed if they are deemed to be true. The suitability of the assumption of approximate normality will be explored by plotting the residuals from this model. If normality is violated then transformations and non-parametric testing will be employed. A sensitivity analysis will be carried out by carrying out the final analysis on the primary outcome, with and without any individual cases what were involved in violations of the protocol.
Additional exploratory analysis
In the event that there is no evidence to support the primary hypothesis of MBCT-SH superiority over CBT-SH on our primary outcome, an additional exploratory analysis will be carried out to explore non-inferiority of MBCT-SH compared to CBT-SH. The analysis will be based on detecting a between-group non-inferiority margin of 2 points on the PHQ-9 with a one-sided α = 0.025. To operationalise this, a two-sided 95% confidence interval will be created around the effect size (MBCT-SH – CBT-SH) and we will conclude non-inferiority if the upper limit of the CI is wholly below 2 PHQ-9 points for both the PP and ITT analyses. The non-inferiority limit was set through consultation with service users and clinicians and looking at the literature (38,39)
Data entry accuracy and missing data
We aim to minimise missing data and data entry inaccuracies at the point of collection. The Qualtrics online survey software used to collect all data will automatically flag any unanswered questions, giving participants to chance to answer these. If a participant would prefer not to answer a question they can leave it unanswered for a second time and the software will proceed onto the next page. At the point of analysis data will be summarised to look at patterns of missingness. Missing data will be replaced using multiple imputation as appropriate. Missing data will be assessed and if more than 5% of data is missing multiple imputation will be carried out followed by a sensitivity analysis. The sensitivity analysis will compare the results for a complete case analysis to the imputed data analysis. Multiple imputation will be carried out under the assumption the data is missing at random (MAR). Anonymised data will be stored on the Qualtrics platform and on password-protected NHS and university computers. Personal data will be stored securely in locked filing cabinets in NHS research offices and on password protect NHS computers.
Planned interim analysis and stopping rules
No interim analysis has been planned. The trial will be paused or stopped if deemed necessary by the DSMB.
Multiple testing
There shall be no multiple testing.
Economic Evaluation
The economic evaluation will be conducted covering the period from Time 0 to Time 2 and will take the NHS/personal social services perspective preferred by NICE(40). Resource use data from the AD-SUS will be combined with unit costs to calculate the total costs of the MBCT-SH and CBT-SH groups. The cost of the two interventions, MBCT-SH and CBT-SH, will be calculated using a micro-costing approach(41) based on PWPs’ salary, including relevant on-costs and overheads. Data on the number and duration of PWP contacts in the MBCT-SH and CBT-SH arms will be recorded using a proforma completed by PWPs. Data on indirect time, including preparation and supervision, will be collected directly from the PWPs. All other health and social care services will be costed using nationally applicable published unit costs, such as the NHS Reference Costs for hospital costs, the British National Formulary for medication and the PSSRU Unit Costs of Health and Social Care.
Costs and outcomes will be compared and presented in terms of mean differences and 95% confidence intervals obtained by non-parametric bootstrap regression to account for the non-normal distribution commonly found in economic data(42). Cost -effectiveness will be assessed through the calculation of incremental cost-effectiveness ratios and will be explored in terms of quality-adjusted life years (QALYs) calculated from the EQ-5D-5L and using the area under the curve approach(43). Uncertainty will be explored using cost-effectiveness planes and cost-effectiveness acceptability curves based on the net-benefit approach(44,45). These curves are an alternative to confidence intervals around ICERs and show the probability that one intervention is cost-effective compared to the other, for a range of values that a decision maker would be willing to pay for an additional unit of an outcome. All economic analyses will include relevant baseline variables to provide a more relevant treatment-effect estimate(46). The primary analysis will include those with complete data needed to be included in the economic evaluation. Sensitivity analyses will explore the impact of missing data.
Qualitative Data Analysis
Qualitative data will be analysed using Thematic Analysis(47) in order to identify facilitators and barriers to treatment completion for each intervention arm. Analysis will be informed by the structure of the Change Interview(19), but it will also be able to identify inductively-derived themes reflecting participants’ experiences.