Effectiveness and cost-effectiveness of humanistic counselling in schools for young people with emotional distress (ETHOS): study protocol for a randomised controlled trial

Meta-analysis of psychotherapeutic interventions with children and adolescents indicate that they are effective, with effect sizes comparing treatment to no treatment hovering in the 0.70 range [16, 17]. An effect size of 0.45 (95% confidence interval (CI) [0.37–0.53]) has been found for school-based interventions specifically [18]. Evidence primarily comes from trials of cognitive behaviour therapy (CBT), which is an established intervention for clinical presentations such as anxiety and depression in children and young people (e.g. National Institute for Health and Care Excellence [19]). However, many young people referred to school counselling services do not present with specific clinical disorders [20]. Rather, they are more likely to be experiencing psychological distress as a result of a range of life difficulties such as family issues, bullying, or academic problems.

There is a need for school-based interventions that can help young people identify, address, and overcome, the distress that arises from these life challenges; and that can minimise the likelihood of this distress developing into more severe psychological problems in adulthood. One potential intervention that may achieve this is school-based counselling [21, 22]. In the UK, school-based counselling provision tends to be based around a humanistic, person-centred model of practice [16, 23] with a focus on young people’s emotional difficulties and on a one-to-one basis with a counsellor [24]. Within the UK, the research to date suggests that it is perceived by children and pastoral care staff as a highly accessible, non-stigmatising, and effective form of early intervention for reducing psychological distress [20]. It has also been associated with positive change [20, 25]. For example, Cooper [20] and Cooper et al. [25] found significant reductions in psychological distress pre- and post-counselling (weighted mean difference (WMD) = 0.81, 95% CI [0.76–0.86]; and d = 1.49, 95% CI [1.29–1.69], respectively). Furthermore, secondary school pupils have reported that attending school-based counselling services positively impacted on their studying and learning [26]. School management has reported perceived improvements in attainment, attendance, and behaviour of young people who have accessed school-based counselling services [27]. Similarly, McElearney et al. [28] reported that school-based counselling interventions in Northern Ireland were effective for pupils who have been bullied. However, school-based counselling in the UK is heterogeneous, and with limited evidence of effectiveness.

In 2009, ‘school-based humanistic counselling’ (SBHC) was developed as a standardised form of school-based counselling [24]. The humanistic orientation of this approach reflects the predominantly person-centred/humanistic style of British school-based counsellors [20, 22]. However, it is based in evidence-based competences for humanistic therapies [29]. A competency framework, which forms a more extended basis for SBHC, has now been developed for those delivering humanistic counselling to 11–18 year-olds [30].

Four pilot studies of manualised SBHC against Pastoral Care As Usual (PCAU) (with the offer of counselling once they had completed participation in the trial), for young people experiencing moderate to severe levels of emotional distress have been conducted [24, 31‐33]. The first pilot study assessed the feasibility of conducting a trial of this nature, including likely recruitment and follow-up rates, and trial procedures [24]. The studies that followed aimed to contribute further data estimating the effectiveness of SBHC, and improve trial procedures including extending the intervention period, adding a 6-month follow-up, post-intervention time point, and aiming to include a more ethnically diverse sample [31‐33]. In three of these studies recruitment was through pastoral care referral, that is to say, via an already established team of professionals within the school tasked with supporting the emotional health and well-being of the school’s pupils [31‐33]. The total number of participants in each pilot study ranged from 32 to 64 and SBHC was delivered weekly for up to 10 weeks. In terms of feasibility, these studies found that recruitment levels were acceptable, with an average of 10.1 pupils recruited per school per academic year. Completion rates were also acceptable, with 76.2% to 100% of pupils completing the study to endpoint. All procedures were acceptable to the schools and young people involved, with no ethical concerns raised.

A pooled analysis of data across these four pilot studies suggests that SBHC brings about statistically significant, medium to large reductions in psychological distress as compared to PCAU, up to 12 weeks from assessment. On the primary outcome in each pilot study, the Young Person’s Clinical Outcomes in Routine Evaluation (YP-CORE) [34] the mean effect size (Hedges’ g) for counselling compared to PCAU at mid-point (6 weeks from baseline) was 0.47 (95% CI [0.09 to 0.88]; counselling n = 58, control n = 60) and at endpoint (3 months from baseline) it was 0.72 (95% CI [0.34 to 1.10], counselling n = 63, control n = 63). The mean effect size was not statistically significant at 6 months (g = 0.44, 95% CI [− 0.17 to 1.06]; counselling n = 23, control n = 23) or at 9 months from baseline (g = − 0.16, 95% CI [− 0.73 to 0.47]; counselling n = 21, control n = 24). However, these sample sizes were small.

In addition, the preliminary findings of Pearce et al. [33] suggest that in the short-term, counselling compared to school-based pastoral care, primary and hospital care, and community-based services, had the greatest impact on within-school costs, reducing the amount of time/costs form teachers and pastoral care staff spent with these pupils. These initial results indicate that counselling may be a more appropriate and more effective resource.

These pilot studies indicate that a trial of SBHC is feasible and that there are initial indications of a short-term effect. However, a trial that is powered to detect clinically meaningful differences is required which can provide more comprehensive data on the effectiveness of SBHC, in particular its longer-term effects, cost-effectiveness and impact on educational outcomes; as well as identifying mechanisms of change.

Accordingly, this paper sets out the protocol for a randomised controlled trial (RCT) in a school setting, to determine the clinical and cost-effectiveness of SBHC compared to PCAU. The protocol adheres to the Standard Protocol Items: Recommendations for Intervention Trials (SPIRIT) Checklist which is available as an additional file (Additional file 1).

Participants are young people attending one of 18 secondary schools across London who meet the inclusion criteria, which includes being between 13 and 16 years of age at the time of assessment and experiencing moderate to severe levels of psychological distress as assessed by a score of ≥ 5 on the Strengths and Difficulties Questionnaire Emotional Symptoms (SDQ-ES) Scale [35] (see below, ‘Assessment’). In addition, young people need to be considered capable of comprehending the outcome measurement forms, with a guide English reading age of 13 years, want to participate in counselling and not be in current receipt of counselling or any other therapeutic intervention that may be impeded through participation in the trial. In order to increase the likelihood of participant attendance at research meetings, we ask that all potential participants have a school attendance record of at least 85% as assessed by the school. Young people are excluded from participation if they are unable to provide informed consent (not ‘Gillick competent’), their parent/carer has not provided informed consent, or the young person is planning to leave the school within the academic year. Additionally, the young person is not included in the trial if they are deemed at risk of serious harm to self or others at the time of assessment. We ask all young people if they are, in principle, willing to complete all outcomes measures at each research meeting, and for their counselling sessions to be audio-recorded (on the understanding that they may exercise their right to request that the recording device be turned off at any point during a session).

The study is being conducted in 18 schools across London, UK. All participating schools are located in urban areas, with 17 (94%) categorised as ‘major urban’ [36]. Six schools (33%) are located within the areas considered ‘most deprived’, and two (11%) schools in ‘least deprived’ areas of England (using the Index of Multiple Deprivation (IMD)) [37]. Over half (56%) of included schools fall between these two categories. The percentage of pupils eligible for free school meals, across included schools, ranges from 11 to 53% (median 33%), with over 25% of pupils eligible for free schools meals in 11 (61%) included schools. Nine (50%) schools have a pupil population of between 700 and 1100 students, five schools’ (28%) population of pupils exceeds 1100, and three (17%) schools are smaller than 700 students. Five (28%) schools are single-sex schools, with three of these single-sex girls’ schools (17% of all included schools). Schools that already had counselling provision, or planned to incorporate counselling provision during the trial period; or whose first language was not English, have not been included in the trial.

SBHC is based on competences for humanistic counselling with young people aged 11–18 years [30] and follows a clinical practice manual developed for the trial (Kirkbride, 2016, unpublished manuscript). It is based on the theory that distressed young people have the capacity to successfully address their difficulties if they can talk them through with an empathic, supportive, and qualified counsellor. School-based humanistic counsellors use a range of techniques to facilitate this process, including active listening, empathic reflections, inviting young people to access and express underlying emotions and needs, and helping clients to reflect on, and make sense of, their experiences and behaviours. Young people are also encouraged to consider the range of options that they are facing, and to make choices that are most likely to be helpful within their given circumstances. As part of the intervention, young people participating in the trial are asked to complete a sessional outcome measure in accordance with the recommendations outlined in the Children and Young People Practice Research Network (CYP PRN) report that highlights the clinical utility of employing a regular feedback tool in counselling [38]. Increasingly, the use of such a tool is becoming standard practice within the field and within Children and Young People’s Increasing Access to Psychological Therapies services (CYP IAPT). Our sessional measure is the Outcomes Rating Scale (ORS) [39] which assesses the following areas of life functioning: personal well-being, family and close relationships, social relationships and general well-being; and is integrated into the therapeutic dialogue.

‘Pastoral care’ in UK schools is a generic, umbrella term used to describe targeted services offered by the school to support the well-being of pupils on roll. Such services vary across schools and while there is no one approach, the existence of such services within schools is a standard practice and seen as inextricably linked to teaching, learning, and the curriculum. For many UK schools, in particular those with limited financial resources, pastoral care is the only provision on offer to accommodate pupils’ well-being needs as targeted interventions, such as counselling, are seen as too costly. Despite the differences in pastoral care services across UK schools, there are also some similarities in the models and approaches employed. For example, form tutors may be tasked with monitoring any emotional difficulties as they arise for pupils in their form group, policies and guidance will be in place relating to bullying and behaviour, schools will have established links to local external services or agencies, for example Child and Adolescent Mental Health Services (CAMHS). School pastoral care services in the UK do not routinely offer counselling, and differ from humanistic counselling in several ways: (1) pastoral care is offered by any member of staff, whatever their qualifications, (2) pastoral care is advice-giving and directive, (3) pastoral care does not involve a collaborative contract, with an agreement on boundaries, confidentiality, or therapeutic goals, (4) pastoral care is most often provided ‘ad hoc’, and is not time limited while the pupil is on roll.

PCAU within the current study, therefore, consists of the included schools’ pre-existing, ‘as usual’ systems for supporting the emotional health and well-being of young people within their particular school, which may involve a personal tutor, or a school inclusion lead meeting regularly with the young person to speak about their difficulties. PCAU does not consist of humanistic counselling, but additional school and non-school-based interventions for young people may be provided, such as a short placement in a learning support unit, some one-to-one support or group work from a school nurse, learning mentor, educational psychologist or behaviour support team, and interagency meetings (for example, the creation of a Common Assessment Framework). The intensity and length of support will depend upon the specific needs of each young person. The number of interventions and the specific form of support is recorded by the pastoral care team in a pastoral care log (see below). Participants randomly assigned to the PCAU condition will be offered the opportunity to access SBHC 6 to 9 months after their assessment.

Adherence to the SBHC model is assessed at two levels: (1) counsellor level and (2) client level.

Counsellor-level assessment uses a young-person-adapted version of the Person Centred and Experiential Psychotherapy Rating Scale [41], the PCEPS-YP, developed specifically for this trial. ‘Calibration tapes’ have been developed, using a sample of recordings which have been rated by national experts, using the PCEPS-YP. Calibration tapes provide a target rating for use in the training and standardisation of subsequent ratings. The national experts for SBHC are based at the Metanoia Institute, Edge Hill University, the University of Nottingham, and the University of Strathclyde.

An independent sub-team of auditors will then rate 20-min recorded segments of counselling sessions which will be selected at random. All counsellors will be sampled. For each counsellor, auditors will assess sessions from four clients, spread across the duration of the counsellor’s involvement in the trial. This will include two recordings per client (and. therefore eight recordings in total, per counsellor). The 20-min segments will be randomly selected from one session in the first half of the counselling work with each selected client (excluding the first session), and one session from the second half of their work with that client (excluding the last session). This sampling strategy will yield a total of 144 recordings to be rated in the trial.

At the client level, adherence to SBHC competences are audited in supervision using a short version of the PCEPS, which has been adapted for work with young people (PCEPS-YP-S). To audit sessions, supervisor and counsellor listen to a segment of a minimum of 10 min in length together, per client, with each of them completing, comparing and discussing their ratings.

Adherence of supervision of SBHC practice will be monitored using an ETHOS supervision adherence form developed for the trial. An independent sub-team of auditors will rate 20-min recorded segments of supervision sessions which will be selected at random. Every counsellor, working with each supervisor, will be sampled. Auditors will assess two sessions for each counsellor the supervisor works with, spread across the duration of the counsellor’s involvement in the trial (and, therefore, 36 recordings will be assessed in total).

School pastoral care teams are the initial point of contact for entry into the trial. Prior to any assessments of eligibility, pastoral care staff, briefed and trained on the details of the trial, identify which young people in their school they deem to be potentially eligible using the trial eligibility criteria. Pastoral care staff approach, and invite, potentially eligible young people for a pre-screening meeting to discuss the project, and if the young person expresses an interest in participating in the trial, and is willing to ask their parents/carers for signed consent, the pastoral care teacher sends a standardised trial letter to their parents/carers with information about the trial and consenting procedures. This method of identifying young people who may benefit from counselling reflects a relatively commonplace procedure in pastoral care referral to counselling services in UK schools. At this stage, it is explained to parents/carers that not every young person who is put forward to take part in the trial is eligible, and that assessment procedures are not a judgement about the young person or their problems, but an assessment of trial eligibility. Once parents/carers ‘opt in’ informed consent is obtained pastoral care staff make an initial referral to the ETHOS research team, on the basis of likelihood that the young person they have identified will meet all eligibility criteria during the assessment meeting. Informed parent/carer consent is obtained either in writing, or via the telephone with a member of the pastoral care staff or an ETHOS researcher acting as a proxy to obtain consent in this way. Consent obtained by proxy is either audio-recorded, or witnessed by a third party.

On receiving parent/carer consent, the young person is then invited to an assessment meeting with an ETHOS assessor. Assessment meetings are held in a confidential and secure environment within the school. The assessor talks the young person through the aims and nature of the trial, and explains to them the assessment and intervention procedures using an Information Sheet for Young People developed for the trial, during which time the young person has the opportunity to ask any questions they have about their potential involvement. During this meeting, the assessor makes a judgement as to whether they believe that the young person is of sufficient maturity and understanding to consent to participate in the research study, using Competence to Consent Guidelines developed for the trial. This judgement is made prior to any other assessment of eligibility; and the young person’s written, informed consent is obtained prior to any study procedures being carried out, or outcome measures being administered. The young person’s written, informed consent is required in addition to their parent’s/carer’s.

Where consent has been obtained from both the young person and their parent/carer, each young person is assessed for their eligibility into the trial by an assessor. All assessors are provided with comprehensive training in the study and the purposes and principles of the assessments. The primary measure for determining suitability for the trial is the Strength and Difficulties Questionnaire (SDQ) [35]. The SDQ measures psychological difficulties along 25 attributes, divided equally between five sub-scales: emotional symptoms, conduct problems, hyperactivity/inattention, peer-relationship problems, and prosocial behaviour. Respondents are required to respond to each of the 25 attributes by reflecting on the last 6 months. For each sub-scale, the score ranges from 0 to 10, with higher scores indicating greater difficulties. Young people completing the SDQ at the assessment meeting for the current trial are required to score ≥ 5 on the emotional symptoms sub-scale (SDQ-ES). Risk assessment is carried out following a discussion with the young person, and using the YP-CORE item 4 (‘I’ve thought of hurting myself”) which has a score range of 0 to 4, with higher scores indicating greater psychological distress [34]. If a young person scores ≥ 1 (on YP-CORE item 4) the assessor asks, in a sensitive and gentle manner, whether they feel that they are at risk to themselves or others, and makes a professional judgement using guidelines for assessing and managing risk developed for the trial. It is explained to the young person that, as there is an equal chance of being allocated to PCAU for 6 to 9 months prior to counselling, it is important for the research and pastoral care teams to know whether they need more immediate attention. If the young person is assessed as being in immediate danger or at risk of serious harm, procedures for onward disclosure and referral are followed. Assessors review the remaining criteria for participating in the trial with the young person before confirming their eligibility. All young people are asked to complete a demographic questionnaire in the first part of the assessment meeting. Figure 1 presents a flow chart of the progress of participants through the trial.

Consenting young people are allocated to one of two groups, SBHC or PCAU, via remote access to the central randomisation procedure that is hosted by an in-house vb.​net application with a Standard Query Language (SQL) server backend at the Clinical Trials Unit, Manchester Academic Health Science Centre, University of Manchester (MAHSC-CTU). This is restricted access and users other than the system owner do not have access to future allocations. Sequence generation is, therefore, concealed from both the assessor and young person, as well as from the rest of the core research team. The randomisation ratio is 1:1 using the method of permuted blocks within school strata with adjacent block sizes themselves varying randomly within pre-specified limits.

Allocation of each participant to SBHC or PCAU is recorded in a separate Assessment Log used only by the assessors. Blinding of participants, providers and assessors is not possible, so this trial employs a ‘tester blind’ design wherein testers are blind to the young person’s allocation for the duration of the trial. To help ensure the success of the blind, a different tester is employed at 6 weeks/mid-point, 12 weeks/endpoint and 24 weeks/follow-up for each participant. The success of the blind is assessed by asking testers to report whether the participant or member of the pastoral care staff have revealed what group the participant is allocated to, on the predesigned Case Report Form (CRF) developed for the trial. Participants and school staff are asked not to reveal the group to which participants have been assigned, as far as is possible.

It is necessary to unmask the allocation of participants to testers conducting the mid-point/6-week meeting to allow the administration of the Working Alliance Inventory Short Form (WAI-S) [42] in the SBHC group (see below, ‘Measures’). Therefore, three separate CRFs have been developed at the mid-point/6-week meeting to ensure that the remaining measures are administered under blind conditions: CRF 1 contains all measures for administration to both groups and is used in the first part of the meeting, CRFs 2 and 3 contain measures only applicable to the SBHC or PCAU group, respectively, and are used in the second part of the meeting. Once CRF 1 has been completed by the participant, the participant’s allocation is revealed to the tester who opens a sealed envelope containing information as to which group the participant is in, and which CRF (2 or 3) should, therefore, be administered. Only testers conducting the mid-point/6-week test have access to the contents of the envelope, which is destroyed following use. The only measures completed following the unmasking of the tester are those relevant to the participant’s specific allocation.

Once eligibility has been established, the second part of the assessment meeting requires participants to complete a battery of measures. In addition, eligible participants attend research meetings at 6 (mid-point), 12 (endpoint) and 24 (follow-up) weeks following their assessment meeting. Pastoral care staff at each included school work with the core research team to schedule research meetings with participants, and follow-up with participants if they are absent (e.g. due to sickness or a school trip). Research meetings are scheduled to fit into the each school’s individual daily timetable. A summary of the assessments, per time point, are presented in Fig. 2 and described below.

Sample size was calculated to take into account clustering and participants lost to follow-up. Firstly, without either clustering or participants lost to follow-up, for 90% power to detect a standardised mean difference (SMD) of 0.5, 86 participants would be required per arm (172 in total). The effect size was determined by pooling findings on the primary outcome from four previous studies and making a conservative estimate [24, 31‐33]. Intra-class correlation coefficient (ICC) for counsellors was estimated from prior data as 0.05 [24, 31‐33]. On average, we estimated that nine young people could be seen per counsellor (4.5 on average per assessment phase); if there is 20% loss to follow-up, this leaves a mean of 7.2. The ICC and average number of young people together leads to a design effect of 1.31, which, when multiplied by the pre-cluster sample size, gives 1.31 × 86 = 113 (rounded up). Hence, after loss to follow-up has taken place, 113/7.2 = 16 (rounded up) practitioners are required per arm. Add 1 [57] to give 17 counsellors. To find the number before loss to follow-up, we calculated 17 × 9 = 153 participants required per arm and 153 × 2 = 306 in total.

Names of participants on the consent forms (personal data) are stored separately from anonymised data in locked filing cabinets and only accessible by named personnel. A separate code (‘pre-randomisation code’) links names to Participant ID numbers used in data files, which are password protected and only accessible by named personnel.

Statistical analysis will be carried out by the ETHOS principal statistician and economist according to a statistical analysis plan agreed in advance. The principal statistician will be blind to the randomisation, with the exception of analysis of the WAI-S which is only completed by participants in the SBHC group. Allocation for all other measures will be coded as a non-identifiable variable in the clinical effectiveness analysis to minimise potential bias. The principal economist cannot be blind to allocation due to the nature of the data being analysed (e.g. number of counselling sessions attended).

Written protocols for managing risk, and monitoring and reporting AEs and serious adverse events (SAEs) is followed for all trial participants throughout the trial (from the point of enrolment into the trial, to the follow-up/24-week time point). An adverse event (AE) is defined as any negative psychological, emotional or behavioural occurrence, or sustained deterioration in a research participant. In the current trial, we have included arrest by police; running away from home; excluded from family home; school exclusion; significant decrease in school attendance; significant deterioration in behaviour, including threatening violence, exhibiting violent behaviour or serious injury to another person, exposure to violence or abuse; significant increase in emotional difficulties; self-harm (if not a presenting issue), or escalating self-harm (when it is a presenting issue); a complaint made against the counsellor, or an issue with the counsellor, resulting in discontinuation of counselling; suicidal ideation; suicidal intent; hospitalisation due to drugs or alcohol, or for psychiatric reasons; and death, including suicide.

An Adverse Events Reporting Form is used by all individuals in contact with participants, who are trained to recognise and respond, in an ethical and timely way, to risk and any issues relating to safeguarding. The overall safety of participants is the responsibility of the trial’s chief investigator (CI). However, in practice the CI relies on all research staff, counsellors, supervisors and school staff to ensure that AEs are identified and addressed in an appropriate and timely manner. Thus, safety is a shared responsibility. Individuals completing the form are asked to consider whether the AE is serious, defined as any AE which is life threatening or results in death, and whether it may be a result of participating in the trial. The severity of each AE is also assessed, according to its intensity, duration and the degree of impairment to the young person (or, when relevant, another person such as in case of risk to others). Severity is graded as ‘mild’, ‘moderate’, ‘severe’, ‘very severe’, or ‘extremely severe’.

The CI has responsibility for reviewing and signing the AE Reporting Forms, for ensuring that the relevant school staff member is aware of the occurrence of any AEs or SAEs at their school; and the ETHOS clinical lead (CL) in instances where the AE or SAE has occurred in a participant in the SBHC group. The CI is also responsible for reporting the AE or SAE to the Chair of the Data Monitoring and Ethics Committee (DMEC) (see ‘Governance and oversight’ below). In the case of SAEs, or those deemed related to participating in the trial, expedited reporting procedures are followed, which includes a reporting timeframe of one week from receipt of the AE Reporting Form.

A panel of young people (drawn from the Young Person’s Advisory Group at the National Children’s Bureau, NCB) and a panel of parents and carers (drawn from the Parent and Carers Advisory Group at NCB) has advised on the development of an appropriate engagement strategy to keep young people (and parents and carers where appropriate) abreast of project developments. The aim is to ensure that the trial explores issues of relevance to young people, and minimises participant attrition. Representatives from both panels have been invited to join the Trial Steering Committee (see below). Panel members include young people who have received training from the NCB Research Centre in a range of research skills (e.g. appraising the academic literature, reviewing engagement materials, co-producing research summaries). We have aimed to involve these panel members at all stages of the study where possible, and look to include them in our dissemination activities where we know that peer feedback is a highly effective way of ensuring that research findings reach the intended audience.

A Trial Steering Committee (TSC) has been established with an independent chair, and representatives from different, relevant professional groups (including a counselling academic, an educationalist, a representative from the ESRC (funder), an independent statistician, and an independent economist), a representative young person and a representative parent/carer, together with members of the research team. The role of the TSC is to monitor the scientific integrity of the trial, the scientific validity of the trial protocol, assessment of the trial quality and conduct as well as for the scientific quality of the final trial report. Decisions about the continuation or termination of the trial or substantial amendments to the protocol are the responsibility of the TSC.

A Data Monitoring and Ethics Committee (DMEC) has also been convened under the direction of an independent chair. The DMEC comprises a clinician, a separate statistician from the TSC, who is also independent of the Trial Management Group (TMG) (see below) and similarly, a separate and independent economist. The role of the DMEC is to review accruing trial data and to assess whether there are any safety issues that should be brought to the participants’ attention, whether any safety amendments should be made or if there are any reasons that the trial should not continue. Open reports are provided to the DMEC by the TMG and closed reports are provided by the CTU. The DMEC Chair reports the DMEC’s recommendations to the Chair of the TSC and may request additional reports or information if required.

A TMG has been established and includes those individuals responsible for the day-to-day management of the trial including the CI, project manager, principal statistician and economist, and all co-researchers. Notwithstanding the legal obligations of the lead organisation (University of Roehampton) and the CI, the TMG has operational responsibility for the conduct of the trial including monitoring overall progress to ensure that the protocol is adhered to, and taking appropriate action to safeguard the participants and the quality of the trial if necessary.

ISRCTN Registry, ISRCTN10460622 (11 May 2016). The study commenced recruitment in September 2016 and recruitment due for completion in February 2018.