This chapter is adapted from Abotsie et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.
Design
The study was designed to answer the research question: Is a full-scale RCT of IPC for young people with mild depression delivered in non-specialist community services feasible? In this feasibility RCT, we planned to randomise 60 eligible young people in a 1 : 1 ratio to receive IPC-A or TAU. Participants were invited to take part in an assessment at baseline (pre randomisation) (see Supplementary Material 1) and followed up at 10 and 23 weeks. The feasibility trial aimed to recruit young people presenting with low mood who were receiving support from participating services in Norfolk and Suffolk, UK. A health economics component was included to inform the design of the economic evaluation in a future study. A process evaluation was incorporated to explore how the intervention is implemented across the counties. Qualitative data were collected through site profile questionnaires (SPQs) (see Supplementary Material 2), observations of IPC-A training workshops and supervision, video-/audio-recordings of treatment sessions (both IPC-A and TAU), interviews with participants (and parents) from the IPC-A and TAU arms and focus groups with youth mental health workers (YMHWs) and wider stakeholders.
Setting
The trial aimed to be conducted in two counties in England. While the sites are in the area served by one NHS mental health trust, tier 1/2 services or services for mild depression are not delivered by this mental health trust, as the severity of illness of young people is generally below the thresholds for NHS specialist CAMHS. Treatment at this level is delivered by a range of services locally. Within the two counties, some non-specialist mental health support services for CYP are provided by the County Councils. Teams delivering these services include Early Help teams, Young People’s teams, Family Support services, school nurses and NEET (not in education, employment or training) teams. In addition, non-specialist mental health support is provided by publicly commissioned independent counselling organisations. Such tier 1 or 2 services provide early interventions to CYPs with mild mental health problems and/or difficult family circumstances such as parental drug and alcohol dependency, parental poor mental health and domestic abuse. Most practitioners working in these services are not qualified mental health practitioners but may have some training in counselling, family work and social care. Staff delivering the IPC and TAU interventions would be employees of these organisations who support young people with mild depression as part of their usual job role. We believed having a variety of services involved across two counties would give a good balance of increasing the transferability of findings while keeping the scale of the study manageable within the available resources.
Eligibility criteria and recruitment procedure
Participant eligibility criteria
Inclusion criteria
- Age 12–18 years.
- Seeking help for low mood (as the primary presenting difficulty).
- Able to provide written informed consent or, for under-16s, written informed assent and parent/guardian consent.
- Of a level of illness where they would normally receive treatment from the service.
Exclusion criteria
- Learning disability necessitating non-mainstream schooling.
- Current psychotic disorder.
- Current substance dependence.
- Current significant suicidal ideation (Kiddie Schedule for Affective Disorders and Schizophrenia – Present and Lifetime version – ‘suicidal ideation’ threshold – ‘often thinks of suicide and has thought of a specific method’).
In line with the approach used successfully in the pilot, eligibility criteria were kept to a minimum to increase the external validity of the trial in the context of non-specialist services.
Excluded young people would be signposted to appropriate services. Young people would not be excluded based on insufficient English language skills. Interpreting/translation services and foreign language Revised Children’s Anxiety and Depression Scale (RCADS) were available.
There was not a numerical upper severity threshold for depression symptoms or suicidal ideation. However, we anticipated that the criterion ‘of a level of illness where they would normally receive treatment from the service’ would exclude young people presenting with significant suicidal ideation. Participants expressing significant suicidal ideation or planning at the time of screening remained ineligible for the study.
The planned recruitment period for the feasibility RCT was January–December 2020 (12 months). In order to recruit the target 60 participants, it would have been necessary to recruit an average of 5 participants per month across all sites; this was predicted to be lower in early stages of the study and higher in later stages, as more therapists completed training. The aim was to have six IPC therapists trained in each of the two counties who would each treat two to three young people with IPC. With IPC taking around 10 weeks (taking into account holidays), this means that, on average, each participating therapist would have one ongoing IPC case per half of the recruitment year.
Participants would be young people accessing participating services via the service’s standard referral pathways. Young people were triaged and assessed according to each service’s standard procedures. If the assessment identified low mood as a presenting difficulty, the case would be discussed with a clinical member of the research team (without identifying the young person) to ascertain likely suitability for the trial. The service was given the option of using the RCADS depression scale to help determine suitability, with a cut-off of 11 or over suggesting suitability (this cut-off was not an absolute).
Potential suitable young people were invited to participate, and those who expressed an interest met with the trial’s research practitioner, who carried out informed consent procedures and screened the young person to ensure they met the above criteria.
Consent
The Chief Investigator retained overall responsibility for obtaining informed consent but delegated this duty to the study research practitioner who was trained in obtaining valid informed consent according to the ethically approved protocol, principles of Good Clinical Practice and Declaration of Helsinki. The informed consent process included a discussion with the potential participant (and his or her parent/carer if under 16) about the objectives of the study, what he or she would be asked to do if they chose to participate and the possible risks and benefits of participation. Potential participants (and their parent/carer if applicable) were provided with written information and given at least 48 hours to read and consider the information before being asked for consent. Young people and their parents/carers were given the opportunity to ask questions and have these answered in full.
If the young person wished to participate following this process, they were asked to complete a consent form (if 16 or over) or assent form (if under 16) to document the informed consent/assent process and their willingness to participate. For young people under 16, in addition to the child’s assent to participation, the consent of a parent or carer (adult with parental responsibility) was required for the young person to be included in the study. Consent to participate in an interview as part of the process evaluation was sought during the main consent procedures. However, it was not a requirement that a young person consented to a process evaluation interview in order to be included in the study.
We did not include individuals who did not have capacity to give their consent/assent to participation. During the consent process, it was made completely and unambiguously clear that the participant was free to refuse to participate in all or any aspect of the study, at any time, without giving a reason and without incurring any penalty. The participant’s continued willingness to participate was confirmed at each study contact before commencing any research procedures. Participants were free to withdraw from the study at any time up until the time of data analysis without giving a reason and without prejudicing his or her further treatment. Data collected up to the point of withdrawal would be used if the participant (and their parent/carer in the case of participants under 16) consented to this. Every effort was made to ensure that vulnerable young people were protected and participated voluntarily in a safe environment free from coercion or undue influence.
As the reading ages and levels of understanding of potential participants varied and did not necessarily mirror chronological age, and in line with patient and public involvement (PPI) feedback, instead of preparing separate information sheets for children aged 13–15 years and young people aged 16–18 years, we created an ‘easier to read’ version of the information sheet and a ‘detailed’ version. All young people were provided with both versions of the participant information sheet and could choose to read the version they found more accessible or to read both. Members of the Study’s Youth Advisory Panel reviewed the information sheets to ensure the format and language used were appropriate for the target age group.
In addition, informed consent for staff participation was sought prior to the training workshops. All staff members trained in the intervention were given a verbal explanation of the objectives of the study, what he or she would be asked to do if they chose to participate and the possible risks and benefits of participation. Staff were provided with a written information sheet and had the opportunity to ask questions and have these answered in full before deciding whether to participate. If the staff member decided to participate following this process, they were asked to complete a consent form to document this process.
Staff involved in delivering TAU were invited to consent to participation on an individual basis when they were assigned to work with a young participant allocated to the TAU arm. This consenting process was completed by an unblinded member of the research team.
Sample size
We planned to randomise 60 eligible young people in total. The target sample size was not based upon estimation of efficacy but was in keeping with published suggestions32 and was believed to be practically possible within the limits of the project. Further, it was anticipated that a sample of this size would enable the assessment of the rates of recruitment and retention to a reasonable degree of precision. Assuming an attrition rate of around 20%, a sample of 60 would provide a 95% confidence interval (CI) of width 20% (i.e. ± 10%). For a recruitment rate of around 50%, the interval width would be around 25% (i.e. ± 12.5%).
Intervention and control arms
Intervention
Interpersonal counselling for adolescents is a brief, manualised psychological intervention derived from IPT. IPC helps clients to identify the reciprocal interaction between their current depressive symptoms and interpersonal relationships, with a focus on one of four domains: grief, relationship disputes, big changes and loneliness and isolation. The therapist works with the client to identify effective strategies to deal with their interpersonal problems, which should improve depressive symptoms.
Interpersonal counselling for adolescents (IPC-A) is an adapted form of IPC designed to suit the needs of adolescents. The intervention is delivered over three to six (30- to 60-minute) sessions, depending on participant needs. IPC-A is based on the manual developed by Weissman et al.,33 with minor modifications to make it suitable for young people. IPC-A arm participants would also have access to standard health and care provision throughout their participation; the extent to which provision of IPC-A altered use of these services would be monitored using the modified Client Service Receipt Inventory (CSRI).
Staff trained as IPC-A therapists received two full days of initial training. Prior to delivering IPC-A to trial participants, trainees needed to achieve adequate scores on audiotaped ratings of two therapy sessions for each of two cases, write an adequate reflective log of the two cases and attend supervision regularly.
Following successful completion of the training, therapists received clinical supervision weekly, with a maximum of three therapists per group. We intended that supervision would be provided in a group format to allow therapists to explore the theory and practice of IPC through engaging in shared discussions of real-world cases. Each supervision session lasted up to 1.5 hours. There were a number of trained IPC supervisors in the local area who expressed an interest in supervising the delivery of the intervention within the trial. A further two supervisors were trained to supervise IPC: an IPC therapist and IPT practitioner who were trained as IPC leads to supervise trainees in accordance with the treatment manual,28 who would have overall responsibility for co-ordinating the provision of clinical supervision.
Control
The control arm received TAU, the standard support provided by services. At present, there is not a standard TAU in these non-specialist sectors. There is great variation between and within services about what is offered to young people with low mood, for example, psychoeducation, non-directive counselling and/or behavioural activation. There is usually little specific training and supervision for these interventions. Staff received the normal management and supervision that they normally receive for their casework. Hence, treatment was not at a lower standard than before the trial. As therapeutic approach between and within services varies, supervision also varies (as there is not a standard systematic TAU, there is not a standard systematic supervision as usual). Standard supervision for TAU is less frequent and less intense than the systematic supervision for IPC-A. This may affect quality of therapy but may also increase cost (which would be captured in the health economics analysis). Crucially, the trial compared the whole IPC-A package (including training and supervision) against the whole TAU package.
Participants were not denied access to any treatment option available as part of current provision. However, staff providing individual support to TAU participants did not attend any IPC-A training and did not receive any IPC-A supervision to minimise contamination. Staff trained as IPC-A therapists were required to agree not to discuss any aspect of their training or supervision with colleagues not trained in IPC-A. The interventions that constitute TAU for this group were monitored via the modified CSRI and process evaluation.
Although the practitioners who delivered TAU were not qualified mental health professionals (as in the IPC-A arm), they were able to consult with or offer a joint appointment with a mental health professional [e.g. primary mental health worker (PMHW) or clinical psychologist] or signpost/refer the young person to other local services.
Further note on contamination: The relationship between RCT study design in mental health, ethics and contamination is complex. In this case, TAU is the treatment (varied within and across teams) that is currently given as standard to young people with mental health problems. Given the fact that there is no evidence that IPC-A is better than TAU, it is acceptable for this TAU to be given in the study. Contamination would mean that the presence of the trial would mean that young people receiving care in the TAU arm would be receiving IPC-A interventions, which they would not have received if the trial did not exist. This could happen through TAU therapists attending IPC-A training and/or IPC-A therapists in the team talking about IPC-A practice. Such contamination may improve outcomes in the TAU arm, causing a type 2 error. It is ethical to try to avoid contamination because there is no evidence that IPC-A is better than TAU, and we are doing this study to investigate if there could be a difference – we, therefore, need to ensure that TAU really is TAU.
Randomisation and blinding
Randomisation was co-ordinated remotely by the Norwich Clinical Trials Unit (CTU). Participants were randomised in a 1 : 1 allocation ratio, using a stochastic minimisation algorithm to minimise imbalance between groups in baseline symptom severity, gender and study site. Allocation was managed by the Data Management Team at Norwich CTU via a web-based system; it was not accessible by anyone outside of the team, including the research team, trial therapists and participants; thus, allocation concealment was maintained.
Blinding
Research practitioners collecting follow-up data were blind to the participant’s treatment allocation. A second unblinded member of the research team received the outcome of the randomisation via an automated notification from the system set-up and managed by the CTU and passed details of allocation to the clinical service. Given the nature of the intervention, it was not possible for participants and those involved in delivering the intervention to remain blind. Following allocation, all participants in the study and therapists were asked not to reveal the group to which the participants were randomised to the research practitioner. Participants were reminded at the beginning of each contact with the research practitioner post randomisation not to disclose their allocation. Any potentially unblinding data were stored separately in a secured database to which the research practitioner did not have access. As the study’s Chief Investigator and participants’ responsible clinicians were unblind to treatment allocations, no emergency unblinding procedures were required for this study.
Data collection
Participants were young people accessing participating services via the service’s standard referral pathways, as detailed above. Young people were triaged and assessed according to each service’s standard procedures. If this assessment identified low mood as a presenting difficulty, the case would be discussed with a clinical member of the research team (without identifying the young person) to ascertain likely suitability for the trial. Potentially suitable young people (and/or parents/carers) were invited to participate. If they expressed an interest, consent was given to the service to pass on their details to the research team.
Those who expressed an interest met with the trial’s research practitioner who carried out informed consent procedures and screened the young person to ensure they met the eligibility criteria.
Face-to-face, telephone, video call and/or internet-delivered quantitative assessment took place at baseline, 5, 10 and 23 weeks; questionnaires were completed online by the participant with support from the research practitioner. Young people were invited to take part in qualitative interviews at the end of treatment; up to 20 were planned to take part in these. Staff and stakeholders were invited to take part in focus groups.
The following participant data were collected at baseline (face-to-face, telephone and/or video call interview, internet-delivered questionnaires):
- demographic characteristics of the young person
- Kiddie Schedule for Affective Disorders and Schizophrenia, depression section34,35
- RCADS36
- Family Assessment Device37
- Cambridge Friendships Questionnaire38
- Employment, Education or Training in previous 4 weeks (NEET status)
- Short Warwick-Edinburgh Mental Well-being Scale (SWEMWBS)32
- modified CSRI39
- Child Health Utility Index 9D (CHU-9D).40
Follow-up assessments
The following participant data were collected at 5-week follow-up (online with telephone support):
- RCADS36
- Family Assessment Device37
- Cambridge Friendships Questionnaire.8
The following participant data were collected at 10- and 23-week follow-up (face-to-face, telephone and/or video call interviews, ± internet-delivered questionnaires):
- RCADS36
- Family Assessment Device37
- Cambridge Friendships Questionnaire38
- Employment, Education or Training in previous 4 weeks (NEET status)
- SWEMWBS 32
- modified CSRI39
- CHU-9D40
The proposed primary outcome measure for a future effectiveness trial was the RCADS, which is a continuous self-rated questionnaire of depressive and anxiety symptoms, with six subscales, including depression. The RCADS is used as the primary outcome measure for emotional disorders in CAMHS in England, as recommended in the Department of Health Children and Young People’s Improved Access to Psychological Therapies (CYP-IAPT) programme. The results from this feasibility study could potentially be benchmarked against results from countrywide CAMHS services. The RCADS is also used as the primary measure in routine English IPT for adolescent practice – the depression scale is used at each session as part of routine IPT-A. We extended this to IPC-A in the pilot36 and weekly RCADS depression was a useful part of therapy and certainly acceptable to young people and therapists; and it was a highly useful primary outcome scale in the research evaluation. The original Chief Investigator was part of a review of adolescent depression measures published in 2015 and found the RCADS to have good psychometric properties.36
We used the observer-administered K-SADS at baseline to test for presence of Diagnostic and Statistical Manual of Mental Disorders depressive disorders. While not an outcome measure, we used this to help us to describe the sample, in particular what proportion of participants have MDD. The K-SADS is the gold-standard diagnostic interview schedule in adolescents, with excellent validity and reliability.34
The schedule of enrolment, interventions and assessments in accordance with Standard Protocol Items: Recommendations for Interventional Trials guidelines is shown in Figure 1.
FIGURE 1 Standard Protocol Items: Recommendations for Interventional Trials figure. FAD, family assessment device; Cambridge FQ, Cambridge Friendship and Relationship Quotient. (Figure not reproduced here; see the original article.)
Attendance data
Information about gender of the therapist, attendance/non-attendance at planned therapy sessions and location of sessions was collected by therapists in both treatment arms.
Process evaluation
Please refer to Chapter 4, Methods.
Analysis
Statistical analysis plan
Recruitment and retention rates were estimated with 95% CIs. Assuming sufficient information, time until drop-out was analysed using ‘time-to-event’ methods to identify baseline factors likely to be related to drop-out. The proposed primary outcome measure for the definitive RCT was the RCADS depression score at 10 weeks. Although the proposed study was not designed to assess efficacy, the mean between-group difference was estimated using a general linear model including baseline RCADS depression score and treating therapist as a random effect. A 95% CI was constructed to assess whether the treatment benefit was feasibly greater than the minimal clinically significant difference, that is, whether or not it was included within the CI. A similar approach was undertaken for the secondary outcome measures. The rate of completion of each outcome measure was reported. If appropriate, depending on the proportion of missing values, multiple imputation will be undertaken and between-group differences re-estimated as a sensitivity analysis. Further parameters, such as within-group variation, needed for the design of a subsequent full-scale trial, will also be estimated.
A statistical analysis plan (SAP) was written in accordance with Norwich CTU guidance and approved by the independent Data Monitoring Committee (DMC) prior to the formal statistical analysis (see Supplementary Material 3).
Health economic analysis
As this was a feasibility study, it was not considered possible to demonstrate the cost effectiveness of the intervention because the study was not powered to demonstrate effectiveness. However, we collected information to inform the design of an economic evaluation alongside a future definitive trial. The intention was to generate information to inform any future study. Examples would include: the likely cost of the intervention and key components of resource use. It also evaluated the use of the CHU-9D instrument as a potential instrument to use in the future study.40
The resources required to provide the interventions (IPC and TAU) were recorded. These included training; ongoing clinical supervision; and staff time required to provide the intervention. Each session offered (and its location) in both arms was recorded. Recording of these events was included in the study design in the form of a record sheet. These were combined with appropriate unit cost data to provide an estimate of the cost of providing IPC-A. It is also important to measure any resources related to participants’ mental health in both the intervention and control groups. This was conducted by means of a modified CSRI conducted at baseline, 10 and 23 weeks. Modifications included adjustments to the following aspects of the CSRI. The time frame for recall was adjusted. This was for the last 3 months at baseline and ‘since we last met’ at the 10- and 23-week follow-up. This effectively gave a time frame of the last 10 weeks for the 10-week follow-up and the last 13 weeks for the 23-week follow-up. To reduce burden on participants, the a priori aim was to make the modified CSRI as simple as possible but to still capture relevant and important service use. For this reason, questions were omitted on the following categories: usual living situation; and contacts with criminal justice system. Questions were limited to the following: contacts with specified types of care providers in the school setting; use of inpatient services; use of other secondary care services; contacts with primary care and community health and social services, residential social care services; and use of services by the participants' family. For drugs, named relevant drugs were specified that may have been prescribed for mental health reasons. We also asked about use of other prescribed medicines. For simplicity, no details were recorded of time and travel costs to access services. All demographic data were also deleted from the CSRI, as relevant data were requested elsewhere in the study. Any modifications made were done so in consultation with other ICALM investigators. The CSRI was collected by means of a face-to-face interview, by telephone/teleconference or videocall/video conference, depending on wishes of participants and safety issues arising from COVID-19 infection risk or wider pandemic measures.
The original intention was to analyse resource use data by study arm to highlight any potential areas of differences between trial arms in use of NHS and social care services, including emergency department attendances. However, due to very low numbers, this was not considered to be informative. For this reason, analysis focused on the whole sample, looking at occurrences of missing data, types of resource use reported, patterns over the three time periods and overall performance of the instrument. The measure of health-related quality of life used in the study was the CHU-9D.40 One important outcome of the feasibility study was an assessment of the suitability of this instrument for use in a future full-scale trial. This was assessed by means of descriptive analysis of the pattern of responses at different time points.
Process evaluation
Please refer to Chapter 4, Data analysis, for details of the analysis for process evaluation.
Progression criteria
The primary intended output of the research was the design of a subsequent trial. The Trial Steering Committee (TSC) planned to assess the trial against the following criteria and make recommendations regarding the suitability of the proposed design for the full-scale trial.
- Recruitment rate is at least 80% of target.
- At least 70% of those randomised to receive the intervention attended at least three therapy sessions within the 10-week treatment window.
- Follow-up assessments are completed by at least 80% of participants at 10 weeks and 70% of participants at 23 weeks.
- At least 80% of IPC treatment sessions reviewed meet treatment fidelity criteria.
- Contamination of the control arm can be sufficiently limited for individual randomisation to be justified.
- The mean RCADS depression scores of the IPC-A and TAU groups at 10 weeks are indicative of a clinically significant difference in depression (3 points).
Governance
Throughout the duration of the study, research processes have been conducted according to the UK Policy Framework for Health and Social Care. A Trial Management Group met on a monthly basis where all aspects of the study were discussed and any issues were identified and resolved. The TSC and DMC were provided with progress reports throughout the duration of the study. Both the committees provided statements of support for the study to be extended beyond the original 2-year timeline due to the suspension of the study during COVID-19.
A provisional SAP was completed at the beginning of the study. Due to the number of changes throughout the study, the SAP was amended following suggestions by the DMC and was approved prior to the analysis being conducted.