Stepped Care Versus Group Cognitive Behaviour Therapy for Paediatric Anxiety in Primary Care

Stepped Care Versus Group Cognitive Behaviour Therapy for Paediatric Anxiety in Primary Care: a Multisite Randomised Controlled Non-inferiority Trial

Objectives: Over the past decade in Sweden, a novel approach to primary healthcare units, known as first-line mental health (FLMH), has been introduced to enhance mental health services for children and adolescents. Through a structured and collaborative method involving experts, clinicians, and patients, a trans-diagnostic psychological intervention for anxiety was formulated based on cognitive behavioural therapy (CBT), designed to align with the FLMH care framework. This study seeks to assess the efficacy of the CBT intervention "Step-by-Step" in a randomised, single-blind, parallel-group, multisite non-inferiority trial. The objective is to investigate whether this treatment is not less effective than the evidence-based Cool Kids program.

Rationale for the Study: Childhood anxiety is prevalent and linked to suffering and impairment across various life domains. While cognitive behavioural therapy (CBT) is a proven treatment for anxiety disorders in children, there is a pressing need to enhance access to such interventions. Introducing CBT for childhood anxiety within primary care settings offers an avenue for early and readily available treatment. However, a gap exists in the availability of CBT interventions specifically designed for and assessed within primary care contexts.

BACKGROUND: Anxiety is the primary reason for contacting the First Line Mental Health (FLMH) in Sweden. Urgently needed is the increased accessibility of evidence-based interventions (EBIs) for child and adolescent anxiety. Cognitive behavioural therapy (CBT) has proven efficacy in child anxiety treatment, though predominantly within psychiatric settings. The FLMH focuses on early interventions for mild to moderate mental health problems in children and adolescents, posing a challenge in identifying EBIs adapted to their needs. The FLMH services often struggle with fitting full-length EBIs into primary care or, more commonly, shorten interventions without empirical support for maintained efficacy. A structured co-creative approach was used to develop a CBT intervention, "Step-by-step," for FLMH. Preliminary testing indicated acceptability, appropriateness, and feasibility, with positive feedback leading to minor revisions in the treatment manual.

OBJECTIVE: The overarching aim of this research project is to increase the availability of evidence-based psychological treatments for children with anxiety treated within primary care.

PRIMARY RESEARCH QUESTION: Is Step-by-step an effective and resource efficient treatment in reducing anxiety severity compared to an evidence-based and well-established treatment (Cool Kids).

SECONDARY RESEARCH QUESTIONS: Is Step-by-step more suitable for certain subgroups of patients in terms of baseline characteristics? Is Step-by-step non-inferior to the Cool Kids regarding children's depressive symptoms, and quality of life? Is Step-by-step non-inferior to the Cool Kids in reducing family accommodation and increasing parental self-efficacy? Are changes in family accommodation and parental self-efficacy mediators of treatment effects? Are there differences in mental health care utilisation between treatments (three-year follow up of health care registers)? Do Step-by-step and Cool Kids differ in terms of treatment adherence, patient satisfaction, and therapists' fidelity? Do Step-by-step and Cool Kids fit the needs of patients?

PROJECT DESCRIPTION: Half of the participants will be randomised to receive Step-by-step, and half of participants will be randomised to receive Cool Kids. All participants will be followed up 12 weeks after treatment initiation (post assessment), and one year after the post assessment. The primary endpoint is set to the post assessment. Participants will be asked to consent for linkage with registers, including retrieving information about utilization of health care up to three years after the post assessment.

CONTROL GROUP JUSTIFICATION: The researchers opted for Cool Kids as the comparator in order to assess and compare the effectiveness of the new treatment, Step-by-Step, against an established evidence-based treatment. Further, the Cool Kids is the most implemented treatment for childhood anxiety in Sweden. Finally, waitlist would not be an ethical choice because there are efficient treatment options available for anxiety in children.

POWER ANALYSIS: The researchers found one previous paediatric anxiety treatment trial employing a non-inferiority design with the Pediatric Anxiety Rating Scale (PARS) as the primary outcome. In this trial, researchers used six points as non-inferiority margin. However, given that the standard deviation of the PARS typically varies between 5-6 and treatment effects for clinician rated outcome (CBT vs. controls) on average are equivalent to a Cohen's d of 0.94, a 6 points margin is likely too large.

Preliminary results from a study exploring the minimal clinical individual difference (MCID) on PARS benchmarked against the Clinical Global Impression Scale (CGI) suggested a change of 6 points as MCID between pre- and post-treatment. We used simulated data with a pre- post within subject design to explore different non-inferiority margins' effect on the proportions of individuals meeting the MCID criteria. Compared to a change on the PARS corresponding to approximately 70% meeting the MCID criteria (i.e., parallel to remission rates in previous trials of the Cool Kids), a change of 3 points less, would correspond to approximately 50% meeting the MCID criteria.

Given the information above, we decided to set the non-inferiority margin to 3 points. The non-inferiority analysis will be based on the PARS scores at the primary endpoint (12 weeks after treatment initiation).

We estimated the statistical power for different sample sizes that the Step-by-step is not worse than the Cool Kids at the 12-week post assessment. We considered one time-point and a response rate of 90%. The estimated power was the observed rejection proportion over 10,000 generated simulated datasets. The data were generated based on the information available in a pilot study on 30 patients. Based on the simulation, the estimated enrolment of participants is set to 154 participants. In other words, if the Step-by-step is truly non-inferior to the Cool Kids program, 154 patients will give an estimated probability of more than 90% that the upper limit of a one-sided 97.5% confidence interval will be below the non-inferiority limit of 3 points on the PARS. No specific non-inferiority margin is defined for the secondary outcome measures.

SCREENING AND RECRUITMENT: At first visit (face-to face, by video or telephone) at the FLMH-unit, the therapist typically collects information about the medical history (including potential previous mental health problems) of the child, assess the severity of mental distress and the global function of the child. On this occasion, eligible children and parents/guardians will receive oral and written information about the study and will be given an opportunity to ask questions about participation. If interested in participation, informed consent will be obtained from all parents/guardians (digitally). If not interested in participation, patients will be offered regular care at the FLMH-unit. In cases where the child needs contact with specialised psychiatric care and/or other organisations than FLMH (i.e., social services or school) families will be referred according to standard routines at the FLMH units.

Children with parents/guardians' consent will be interviewed with the PARS by a FLMH-therapist. During the PARS interview, children and parents/guardians will complete additional baseline assessment via the secure online data collection tool. After the PARS interview, the therapist will assess adherence to the inclusion criteria, and carefully evaluate potential exclusion criteria to ensure that participants meet the required eligibility standards for enrolment in the study. Upon confirming a participant's eligibility for inclusion in the study, the therapist will proceed to open a sequentially numbered envelope. These envelopes have been meticulously prepared in advance by a researcher and contain the randomised assignment.

RANDOMISATION, ENROLMENT AND MASKING: Participants will be randomised at a 1:1 ratio to Step-by-step or Cool Kids. Randomisation and masking procedures will be conducted by an external researcher not involved in the recruitment. For each semester, all FLMH-units will receive 14 sealed envelopes (numbered between 1-14) with seven Step-by-step and seven Cool Kids allocations. We will use a block-randomisation, with block sizes randomly varying of 2 and 4. The randomisation will be generated using a computer random number generator by the researchers. Assessors conducting post- and follow-up assessments will be blind to treatment allocation. The outcome measure (PARS) is identical for both groups, ensuring that the assessors remain blind. At post- and follow-up assessments, participants will be reminded not to reveal their arm allocation. To measure blinding integrity, all assessors will record whether the participating families inadvertently reveal their group allocation, and subsequently guess each participant's treatment allocation.

END OF TRIAL: The trial will end when the registry data (at the 3-year follow up) has been collected.

STATISTICAL ANALYSES: Data will be analysed according to intention-to-treat principles and according to per protocol as adequate in non-inferiority trials. Data analysis on the continuous outcome measures will be performed with mixed-effect regression analyses for repeated measures. The model will include fixed effects of time, treatment group, and an interaction effect of treatment group by time, as well as random intercept to account for individual differences. Ordinal data will be analysed with ordinal regression, binary data will be analysed with logistic regression. Within- and between-group effect sizes will be reported as Cohen's d.

DATA MANAGEMENT All aspects of data management of the trial will comply with the General Data Protection Regulation and good clinical practice.

Inclusion Criteria:

• The child is 7-12 years old (Confirmed by the parent/guardian and from medical records)
• The child suffers from mild to moderate anxiety. The term "mild to moderate anxiety" has not been explicitly defined but is a term aimed to specify a demarcation between the FLMH and the psychiatric specialist services. Adequate level of care for the patient is decided based on an overall clinical assessment. The Children Global Assessment Scale (CGAS) is often used to guide the decision on level of care, where a CGAS value above 60 clearly indicates primary care services, a value below 50 clearly indicates specialist services. For values between 50 and 60 the overall clinical evaluation of symptom severity and impaired function guide the decision on level of care. (Confirmed by the assessor at the FLMH-unit).

Exclusion Criteria:

• Other psychiatric or developmental disorders than anxiety, or social issues that primarily require other interventions, i.e., where an intervention targeting mild to moderate anxiety is not adequate. (Confirmed by assessor at the FLMH-unit).
• The child has another ongoing psychological intervention. An ongoing psychological treatment for any other psychiatric disorder. (Confirmed by the parent/guardian).