Effect of Clinician Training in the Modular Approach to Therapy for Children vs Usual Care on Clinical Outcomes and Use of Empirically Supported Treatments

Key Points Question Is training in the Modular Approach to Therapy for Children (MATCH) associated with more use of empirically supported treatments, better clinical outcomes, and better service efficiency than usual care? Findings This randomized clinical trial found that training clinicians in MATCH was associated with high levels of adherence to empirically supported treatments (80.0%) compared with usual care (57.0%), but it was not associated with improved clinical outcomes or efficiency. Meaning These findings suggest that training in MATCH increases clinicians’ use of empirically supported treatments but does not necessarily improve clinical outcomes.

o Reduced ability to recruit at Kaupapa Māori and Pacific clinics was due to high 77 resource pressure at services, higher acuity of potential participants, fewer referrals 78 of participants in the eligible age range and logistic restrictions placed upon the 79 selection and number of clinicians able to participate in the study. 80  Due to budgetary restrictions and logistic considerations at the study sites, it was not 81 possible to extend recruitment past June 30 th 2015. 82  The recruitment of participants from Māori and Pacific providers has fallen well short of 83 our original study aims. Consequently, sub-analyses will no longer be conducted for our 84 ethnic groups. 85  A sample size of 200 participants will provide sufficient power to detect a meaningful 86 difference in our primary outcome measure among the entire cohort. 87 88 1.8 Main criteria for inclusion: Children and adolescents will be eligible for inclusion in the trial 89 if: 90  They are newly referred to CAMHS with a primary disorder that includes anxiety, 91 depression, trauma-related symptoms or disruptive behaviour; 92  They are 7 to 14 years of age on the date of consent; 93  They are able to provide written consent (or verbal assent) and have written 94 parental/guardian consent; and, 95  The child and their parent/guardian can speak English or there is a clinician who can 96 provide the necessary treatment, according to the treatment arm the participant is 97 randomised to, in the family's native language. 98 99 1.9 Main criteria for exclusion: Children and adolescent will be ineligible for inclusion in the 100 trial if: 101  They are currently receiving other treatment for their disorder from the CAMHS (or 102 another service); or, 103  They have a primary disorder of psychosis, severe intellectual disability, attention 104 deficit-hyperactivity disorder (where the primary reason for referral is inattention 105 and/or over-activity), autism or other pervasive developmental disorder, anorexia 106 nervosa or bulimia nervosa; or, 107  The young person is acutely suicidal; or, 108  They have a sibling that has previously been recruited into the study. 109 110 1.10 Intervention: Child and adolescent participants will receive MATCH therapy from a CAMHS 111 clinician (who usually provides clinical interventions in CAMHS) who has received MATCH 112 training from the MATCH trainers from the USA (and the clinician will be engaged in weekly 113 MATCH consultation sessions with MATCH trainers after the initial face-to-face training). In 114 summary, MATCH provides effective elements of the evidence-based treatments (EBTs) for 115 child anxiety, depression, trauma-related symptoms and disruptive behaviour in one protocol; 116 it caters for co-morbidity and provides an opportunity to address different disorders that may 117 emerge during therapy. MATCH is comprised of 33 modules (i.e. specific treatment procedures) 118 which can be organised in a flexible manner. 119 120 1.11 Duration of treatment: Duration of treatment will vary, based on the need of the client 121 and service-specific constraints (e.g. the need to treat rapidly in order to meet the demands of 122 the service  Mental health problems are common in children and adolescents (1)(2)(3)(4)(5), the impact is 185 considerable and it has been estimated that 50% of all adult mental health disorders 186 have their onset in adolescence (6).

187
 There are a number of evidence-based treatments (EBTs) for the most common mental 188 health problems, namely anxiety, depression, trauma-related symptoms and disruptive 189 behaviour (1). The Ministry of Health in New Zealand has repeatedly asserted the need 190 for EBTs in mental health generally (7-9), and in CAMHS specifically (9 (32). 248

The need for a trial in New Zealand 249
MATCH has been developed and evaluated in the USA, but before it can be introduced in New 250 Zealand CAMHS it should be formally tested here. The trial we are proposing will provide New 251 Zealand-specific assessment of effectiveness (clinical outcomes and service efficiencies) and a 252 specific investigation of the acceptability and effectiveness for Māori and Pacific people. New 253 Zealand has a growing child and adolescent population, especially Māori and Pacific population 254 (33,34), who are at increased risk of mental health problems (23-28  They have a sibling that has previously been recruited into the study. 469

Inclusion and exclusion criteria -CAMHS clinicians 470
Criteria for inclusion: Clinicians will be eligible for inclusion in the trial if: 471  They are involved in the assessment and clinical treatment of children and adolescents 472 and their families at participating CAMHS. We anticipate that most clinicians will be 473 nurses, social workers and psychologists (with some occupational therapists, 474 psychotherapists and medical practitioners). 475  The clinician is full-time (or at least 0.6 FTE); and, 476  The clinician anticipates working at the CAMHS for at least 15 months after the MATCH 477 training is provided. 478 479 Exclusion criteria: Clinicians will be ineligible to inclusion in the trial if: 480  They provide non-clinical support, rather than assessment or treatment. For example, 481 cultural advisors or youth workers who provide non-clinical interventions. 482

Recruitment: 483
Recruitment will be via CAMHS clinicians. Eligible participants will be invited to participate by a 484 clinician whom they have met. 485

Screening eligibility check: 486
Potential participants will be identified by CAMHS clinicians during their standard initial 487 assessment (or intake assessment). Those children and adolescents with primary disorder that 488 includes anxiety, depression, trauma-related symptoms or disruptive behaviour aged 7 to 14 489 years old will be invited to participate. 490

Participant information and consent: 491
Child and adolescent participants and their parents/caregivers will be given written information 492 sheets and consent forms. Written consent will be collected from all parents/caregivers. 493 Written consent (or verbal assent for younger participants) will be collected from all child or 494 adolescent participants. Consent or assent will be sought from children and adolescents 495 depending on their developmental stage. This will be done on a case-by-case basis and 496 determined by the clinician. Those who do not want to participate will receive the care that is 497 usually provided in that service. 498

Baseline assessments: 499
The following baseline data will be collected before the child/adolescent starts treatment: 500  Demographics; 501  SDQ; 502  BPM; 503  TPA; 504  DAWBA; and, 505  Child Health Utility (CHU9D). 506 507 The following information will be collected from the child's/adolescent's CAMHS clinician: 508  Child's/adolescent's medication use. 509 510 Self-rated assessments will be read out to children and young people who may have difficulty 511 filling out these forms without this assistance. 512

Randomisation: 513
Randomisation will occur on two levels: 514 515  Randomisation of clinicians: All clinicians from participating CAMHS teams will be invited 516 to participate. Consenting CAMHS clinicians will be block randomised (by service/team) 517 in a 1:1 ratio to MATCH or UC. The block size will vary across sites depending upon the 518 number of individuals likely to be recruited from each site. Clinicians will be stratified 519 on the basis of previous evidence-based therapy training (i.e. those with versus those 520 without accredited training in cognitive behavioural therapy or behavioural parent 521 training). 522 523  Randomisation of child and adolescent participants: After the eligibility check, and once 524 consent/assent and collection of baseline assessments have been completed, 525 participants will be randomised in a 1:1 ratio to receive MATCH or UC stratified by sex 526 and ethnicity. 527 528 Allocation concealment will be assured by using centralised computer generation of the 529 randomisation sequence. 530 531 We will use web-based randomisation procedures to determine treatment allocation. 532

Blinding: 533
The assessors of the outcome measure/s (i.e. those completing the BPM and DAWBA with 534 participants) will be blind to treatment allocation. Those research assistants administering the 535 assessments (i.e. BPM and DAWBA) will not have access to any data that may unblind them. 536 Due to the nature of the intervention (i.e. MATCH) and UC, it will be obvious to clinicians which 537 treatment they are providing. Assessor blinding will be maintained by ensuring that the 538 research assistants (who will collect assessment data) will be unaware of treatment allocation. 539

Study intervention: 540
MATCH (intervention) consists of a manual and a training package (i.e. six days of block training 541 and a year of weekly telephone/Skype consultation sessions from a MATCH trainer to support 542 clinicians in the use of MATCH). MATCH was specifically designed to combine the effective 543 elements of the EBTs for anxiety, depression, trauma-related symptoms and disruptive 544 behaviour in one protocol, cater for co-morbidity and provide an opportunity to address 545 different disorders that may emerge during therapy. MATCH is comprised of 33 546 modules/specific treatment procedures which can be organised in a flexible manner. 547 Children/adolescents and their families are also given an integral role in defining the goals of 548 therapy. Clinicians use a web-based system (specifically developed for this study) to monitor 549 progress and adapt therapy (The TrACY eMonitor) in consultation with a child/adolescent and 550 their family until a problem is resolved. 551 552 Usual care/UC (control) group: Usual care will be the treatment that is usually provided to a 553 child/adolescent at a CAMHS (e.g. case management, therapeutic group work and 554 psychotherapy). Information on what UC was provided to each child/adolescent participant 555 will be collected. A random sub-set of audio recorded therapy sessions will also be reviewed 556 for content of EBT. 557

Outcome measures: 558
The outcome measures have been chosen carefully to ensure that they have robust 559 psychometric properties and to reduce the burden for children, their families and clinicians. 560 561 Primary outcome measures: 562 Clinical outcome measure. The Brief Problem Monitor (BPM) is based on the Brief Problem 563 Checklist (42) and will be collected weekly. The BPM is a 19-item assessment using data from 564 parents and children to measure internalising, externalising, hyperactivity and total problems. 565 It was developed from the widely used Child Behavior Checklist (CBCL) and Youth Self Report 566 (YSR) (43). The BPM has sound psychometric properties (42). It can be administered by phone 567 or face-to-face and takes approximately five minutes to complete, making it a practical and 568 robust measure of the trajectory of change in clinical symptoms over time. 569 570 Delivery of EBT. All MATCH and UC therapy sessions will be audiotaped. A randomly selected 571 subset approximately 5-10 % (both MATCH and UC) will be assessed by the research team 572 (blind to treatment allocation) for EBT content using the methods and the coding system 573 developed for the initial trial of MATCH (44). We will double-code a subset of the sample to 574 check inter-rater agreement (i.e. κ>0.70). 575 Efficiency of services. Comparisons will be made between MATCH and UC for the following: 576 Total therapy input by clinicians (in minutes), contact with the CAMHS (in weeks), number of 577 therapy sessions and number of missed therapy sessions. 578 579 Secondary outcome measures: 580  The Strengths and Difficulties Questionnaire (SDQ) (38) will be collected monthly over 581 the phone. The SDQ for children aged 4 to 16 years is widely used in both research and 582 practice in New Zealand and elsewhere (38). It is was the preferred tool for parents and 583 children in a New Zealand study of child/youth mental health outcome measures (39). It 584 produces a total score with five subscales: emotional symptoms; conduct problems; 585 hyperactivity; peer relationship problems; and prosocial behaviour. Total scores of 17 or 586 above suggest likely cases with mental health disorders (40). It has satisfactory internal 587 consistency, test-retest reliability, and inter-rater agreement (41). 588  The Top Problems Assessment (TPA) has been developed to allow the child and family to 589 identify and provide severity ratings for the three top problems of greatest concern to 590 them and track progress in addressing these problems. There is good evidence for 591 reliability, validity, and sensitivity to change (45). It is brief to administer once the top 592 problems are identified (takes <2 minutes to complete on a weekly basis) and can be 593 administered by phone. 594  The Development and Well-Being Assessment (DAWBA) (46, 47) will be used to 595 determine the number of diagnoses pre-to post-intervention. 596  Child Health Utility (CHU9D) is a measure of health-related quality of life designed 597 specifically for children. There are 9 self-rated items that take fewer than 5 minutes to 598 complete. CHU9D consists of a descriptive system and a set of preference weights, 599 giving utility values for each health state described by the descriptive system. It allows 600 for the calculation of quality adjusted life years (QALYs) for use in cost utility analysis 601 (49). 602 603 Other measures 604 We will collect data on the treatment provided in usual care, the child's/adolescent's 605 psychotropic medication use and clinician rated serious adverse events (SAEs). The Therapist 606 Satisfaction Inventory (TSI) (50) will be used to determine clinicians' satisfaction with therapy. 607 Brief questionnaires will be used to determine treatment satisfaction amongst child/adolescent 608 participants, their parents/caregivers as well as CAMHS managers/team leaders. In-depth individual interviews will be conducted with: 634 1. Clinicians who work in Kaupapa Māori (KM) and Pasifika (P) services and delivered 635 MATCH in the trial. These clinicians may or may not be of Māori or Pasifika ethnicity 636 themselves. However, because they work in Kaupapa Māori and Pasifika services they 637 are well versed in kaupapa, tikanga and culturally appropriate delivery of care (n = 2 for 638 KM; n = 2 for P). 639 2. Team leaders of the Kaupapa Māori and Pasifika services (n = 3) 640 3. Clinicians, working in mainstream services, who delivered MATCH during the trial and 641 have self-identified as Māori or Pasifika ethnicity (n = 1). 642 With informed consent, all interviews will be digitally (voice) recorded. Transcripts will be 643 made. Participants will be offered the opportunity to review their transcripts. Multiple 644 readings will be undertaken. Thematic analysis, using an inductive approach, will be 645 undertaken to identify common themes and differences across participants. NVivo will be used 646 to manage the qualitative data. 647 648 649 650 651 652 653 654 *Some participants will have no SAEs, so this will require no time to complete. 655

Research assistants: 656
We will employ a number of research assistants to conduct many of the assessments. In 657 particular the research assistants will conduct the phone assessments (i.e. SDQ, BPM and TPA) 658 and will assist families to complete the DAWBA. The research assistants will have a 659 health/mental health qualification (ideally NZ registered health professionals) and have 660 experience working with children and adolescents. They will be trained in the administering of 661 the weekly assessments and DAWBA. Ongoing supervision and monitoring will be provided to 662 ensure the fidelity of data collection. Double-data entry will be conducted on a random sample 663 of assessments to ensure accuracy. 664

Safety assessments: 665
Generally, an adverse event (AE) is defined as any unintended, unfavourable clinical sign or 666 symptom, any new illness or disease or a deterioration of existing illness or disease whether or 667 not considered treatment related. For this study clinicians will be asked to record whether a 668 child/adolescent participant experienced any serious adverse events and these will be 669 recorded. 670 671  Requires in-patient hospitalisation or prolongs hospitalisation; or, 678

Summary schedule of assessments
 Is another medically significant event that, based upon appropriate medical 679 judgement, may jeopardise the patient and may require medical or surgical 680 intervention to prevent one of the outcomes listed above. 681 682 All serious adverse events (including increased suicide risk) will be managed by the CAMHS 683 clinicians and their team as per their District Health Board's usual risk management protocols 684 and procedures. All serious adverse events (including increased suicide risk) should be reported 685 to the trial steering committee by the clinician within 24 hours of the clinician being aware of 686 the serious adverse event. This notification should be by phone call and fax to Associate 687 Professor Sally Merry (telephone 09 373 7599 ext 86981 or fax 09 373 7013) and should contain 688 all information available at the time. Further information will be communicated as soon as it 689 becomes available. The trial steering committee will review all serious adverse events as soon 690 as possible. However, the clinical responsibility for the care of the clients will remain with the 691 relevant clinicians at the appropriate CAMHS. 692

Statistical considerations: 693
Study power: Overall, 60 clinicians and 400 patients from a range of CAMHS in New Zealand will 694 provide sufficient power to detect effect sizes of approximately 0.37. This will allow us to detect 695 clinically significant effects sufficient to justify the resources required to introduce MATCH in 696 NZ. We have powered the study so that independent analyses undertaken for Māori and 697 Pacific young people will have sufficient power (80%) to detect moderate effect sizes of 698 between 0.61 to 0.70 as statistically significant (two-tailed α=0.05) (equivalent to the effect 699 sizes found in the USA (32)). This allows for a 15% attrition rate in clinicians and in clients 700 (based on 12.3% attrition rate reported in Weisz et al. (32) and an attrition rate of 10% reported 701 in Merry et al. of a RCT of psychosocial intervention in NZ (51)) and adjustment for the 702 clustering of patients for each clinician (ICC=0.1). We have assumed approximately 8 patients 703 per clinician (approximately 400 patients). If it transpires that fewer clinicians are available 704 within the targeted CAMHS, we will increase the number of participants to ensure that we 705 maintain this level of statistical power. We will recruit approximately 12 clinicians working in 706 Kaupapa Māori services and 12 clinicians working in Pacific services and will investigate effects 707 by service type. The findings will be complemented by data on acceptability, engagement and 708 drop-out rates which will be assessed by ethnicity. We were unable to reach our target recruitment in Kaupapa Māori services (11 clinicians, 722 but only 4 who remained in the study and were able to recruit participants) or Pacific 723 services (5 clinicians). Consequently, we are unable to investigate effects by service type. 724 725 Analysis of costs: We will compare costs between MATCH and UC using cost-per-hour of 726 clinician time, and will discuss these findings in terms of likely ongoing impact of untreated 727 disorder using information from published studies. Although we are collecting the relevant 728 data, we do not have the resources to undertake a formal analysis of costs in this study. If there 729 are positive results from the study, we will apply for funds to do a detailed cost-benefit analysis. 730 731 Analyses: 732 733 Intention-to-treat analyses will be carried out and will include all participants who met inclusion 734 criteria and provided and maintained written informed consent prior to treatment 735 commencing. 736 737 Per-protocol analyses will include participants who completed the study (i.e. not withdrawn), 738 received therapy as per their allocated treatment group (i.e. not treated "off protocol") and had 739 at least four completed therapy sessions. 740 741 We will assess the comparability of baseline of both treatment groups (i.e. MATCH and UC) 742 using descriptive analyses, in terms of age, sex, pre-intervention BPM and SDQ data, diagnosis 743 and ethnicity. Both treatments may vary in content and duration, so that pre-post and follow-744 up (three months after treatment has ended) analyses are potentially confounded by treatment 745 duration and/or dose. To deal with this, we propose to use the trajectories of change across 746 time on the BPM as our primary outcome measure. Intention to treat analyses will be carried 747 out and will include all participants who meet inclusion criteria, and provide and maintain 748 written informed consent prior to treatment commencing. 749 750 Primary outcomes: 751 i. Clinical outcomes: 752 753 Difference between groups on the BPM (internalising and externalising scales) will 754 be compared using differences in trajectories of change across time using a mixed 755 effects regression model with outcome =a 0 (intercept) + a 1 (informant) + a 2 (treatment 756 group) + a 3 (time) +a 4 (treatment x time) and intercept, informant and time (log day) 757 treated as random effects. 758 759 These analyses will be adjusted for: 760  Medication use, 761  Ethnicity, 762  Clinician site, and 763  Adjustment for clinician's previous evidence-based therapy training. 764 765 The addition of the ethnicity and the treatment x time x ethnicity terms to the 766 model above will allow formal statistical testing of the consistency of the treatment 767 effects across ethnicity groups. It is acknowledged that the power for these 768 comparisons will be much reduced compared to the testing of the overall treatment 769 effect and the possibility of type II error in relation to clinically significant 770 differences is somewhat higher. No comparisons between ethnic groups will be 771 made. 772 773 ii. Delivery of EBT. The percentage of EBT content in a random sub-set of sessions 774 assessed by coding audio-taped therapy sessions will be compared between groups 775 using ANOVA following the method used in the USA trial (44). 776 777 iii.
Efficiency of services. Efficiency in delivery of therapy for MATCH and UC will be 778 compared using ANOVA to test for significant differences using the following data 779 which will be collected from the clinical service including: 780  Clinician time (in minutes); 781  Duration of contact with the service (in weeks  Trajectory of change adjusted as above for the SDQ (total problems score; measured 788 monthly) and Top Problems Assessment (measured weekly); 789  Post-treatment diagnoses (number), controlling for pre-treatment diagnoses. A fixed-790 effects analysis of covariance model will be applied to all children and adolescents for 791 whom we have both pre-and post-treatment data on diagnosis using DAWBA. Where a 792 DAWBA assessment is missing it will be assumed that there has been no change. 793 Remission and reductions of DAWBA diagnoses ("++", "+++" or less), total numbers of 794 DAWBA diagnoses pre-to post-intervention, and diagnoses by treatment arm pre-to 795 post-intervention will also be calculated. 796  Changes in the Child Health Utility (CHU9D) assessment will be compared across arms 797 using ANOVA to assess changes between baseline, discharge and at follow-up across 798 treatment groups. 799 Due to the nature of the study intervention, this trial can only be participant and assessor-blind. 819 Clinicians will be aware of treatment allocation and some child/adolescent participants may be 820 unblinded to treatment allocation. Research assistants collecting weekly data and will assist 821 families to complete the DAWBA diagnostic assessments will be blind to group allocation. 822

Interim analysis: 823
An interim analysis is planned after child and adolescent participants recruited in the first six 824 months of the study have completed post-intervention assessments. This analysis will compare 825 treatment arms with respect to the primary outcome variable and serious adverse events. The 826 steering committee will consider the relative efficacy and rates of serious adverse events in the 827 two treatment arms and will determine on this basis, if it is appropriate for the study to 828 continue or not. 829 830 The DMC will consider the relative safety and efficacy and rates of serious adverse events in the 831 two treatment arms and make recommendations to the Steering Committee regarding 832 continuation or termination of the study. 833

Withdrawal criteria: 834
Participants can withdraw from the study at any time. Those participants that remain in the 835 study, but have dropped out of treatment, will still be followed up wherever possible. 836 837 Where a clinician withdraws from the study, a clinician from the same treatment arm will 838 continue to provide the treatment allocated (i.e. either MATCH or usual care). If the participant 839 was randomised to MATCH, and their allocated MATCH clinician withdraws from the study with 840 no one else is available within the team to offer MATCH, then MATCH will be delivered by the 841 MATCH professional support person in that service. 842

Ethics committee approval: 843
Ethics committee approval for this study will be sought from the Health and Disability Ethics 844 Committees and the appropriate locality approval will be obtained from each participating site. 845

Dissemination of results: 846
All study participants (clinicians, children/adolescents and parents/caregivers) who indicated 847 that they would like a brief summary of the results will be sent this summary, along with an 848 outline of their significance and our future research and workforce development plans for 849 MATCH in NZ. Study findings will be discussed at New Zealand meetings, hui and fono as well 850 as at international conferences. Publication of papers in international journals will be sought. 851

Trial registration: 852
The trial has been registered with the Australian New Zealand Clinical Trials Registry 853 ACTRN12614000297628. 854 855

Study management: 857
The co-principal investigators are primarily responsible for the conduct of the study. The 858 principal investigators will facilitate the trial steering committee and study management 859 committee. 860

Trial steering committee: 861
The trial steering committee will consist of the study investigators and at least two New 862 Zealand experts in the field of child and adolescent mental health. It is responsible for 863 providing strategic guidance for the trial, including developing and maintaining the study 864 design, approval of protocol changes, statistical analyses, presentation and publication of 865 results. The committee will meet at least quarterly (or more frequently if required) and will be 866 conducted using video conference at a time suitable for our American investigators (to allow for 867 those outside of Auckland) to easily access meetings. The committee will review information 868 relating to recruitment and serious adverse events, plus any problems and issues raised by the 869 study management committee. The trial steering committee will review data from the planned 870 interim analysis and recommend whether the study continues or is terminated. 871 872 All serious adverse events will be reviewed by at least two or three of the clinicians on the trial 873 steering committee ( Jenny Herren, Dr Monique Faleafa and Dr Mathijs Lucassen) to establish whether these could 876 be related to the treatment received. If they have any concerns they will forward the 877 information to the full trial steering committee. In addition this information (even if no 878 concerns are raised during the initial review) will be reviewed at the regular trial steering 879 committee meetings. 880

Study management committee: 881
The study management committee consists of Associate Professor Sally Merry, Dr Sue Crengle, 882 Dr Karolina Stasiak, Dr Mathijs Lucassen (1 Oct 2013 -1 June 2015), and Dr Sarah Hopkins (1  883 June 2015 -30 September 2016) who are involved in the daily operation of the study, and will 884 review study materials, deal with study problems, recruitment and logistical issues. 885

MATCH training group: 886
MATCH face-to-face training will be provided by the MATCH trainers from Professor Weisz's 887 team (i.e. Assistant Professor Sarah Kate Bearman, Dr Ana Ugueto and Dr Jenny Herren) in New 888 Zealand. The MATCH trainers will provide weekly group consultation sessions (via telephone or 889 Skype) on the use of MATCH for the duration of study recruitment and the active treatment of 890 child and adolescent participants (estimated to be 15 months). Professor John Weisz will 891 provide oversight to ensure that the MATCH training and consultation provided to New Zealand 892 clinicians meets the necessary standards. Dr Ainsleigh Cribb-Su'a (or a delegate) will provide 893 support and supervision to the MATCH consultants to help ensure MATCH training and 894 consultation sessions are responsive to the needs to Māori. Dr Monique Faleafa (or a delegate) 895 will provide support and supervision to the MATCH consultants to help ensure MATCH training 896 and consultation sessions are responsive to the needs to Pacific participants in the study. 897

Protocol, participant information sheets and other revisions: 898
Any protocol deviations will be documented. All revisions to the protocol will be discussed and 899 approved by the study management committee. If the revision is an "administrative letter", 900 one of the co-principal investigators will sign it and submit it to the ethics committee for their 901 information. If the revision is an "amendment", one of the co-principal investigators will sign it 902 and submit it to the ethics committee for review and approval prior to implementation. 903 Documentation of approval signed from the chairperson or designee of the ethics committee 904 will be sent to the co-principal investigators. 905 906 If an amendment alters the study such that the participant information sheet (PIS) needs to be 907 updated: 908  The PIS/consent form will be revised and submitted to the ethics committee for review 909 and approval; 910  The revised PIS form will be posted to participants; and, 911  The new PIS/consent form will be used for new participants as soon as it is approved by 912 the ethics committee. 913

Reporting: 914
The co-principal investigators will provide annual reports to the ethics committee and the 915 funder (the Health Research Council). 916

Record retention and security: 917
The co-principal investigators will not dispose of any records relevant to this study until 10 918 years after the study completion. The investigators shall take responsibility for maintaining 919 adequate and accurate hard copy source documents of all data generated during this study and 920 will be securely stored in locked premises at the University of Auckland. Electronic data will be 921 securely stored, using participant codes (except for the form pertaining to demographics) and 922 will be password protected. Participating CAMHS clinicians will have access to clinically 923 relevant data (e.g. a weekly summary of how their participating child/adolescent client is 924 progressing), however they will not have access to the raw data. 925 926

Ownership of data and publication policy: 927
The data from this study remain the property of the investigators. All publications from this 928 study should be approved by the trial steering committee. It is intended that the results of this 929 study will be presented at appropriate conference/s and published in peer-reviewed journals, 930 regardless of the outcome of the study. 931

Confidentiality: 932
All information generated in this study is considered highly confidential and will not be 933 disclosed to any persons not directly concerned with the study. Documents and data will be 934 stored securely with controlled access to prevent loss, tampering or unauthorised access. 935 Documentation for the study will be managed according to the University of Auckland's 936 research policy. Data will be destroyed after 10 years according to standard research practice. 937

Quality control and assurance: 938
Each study site/CAMHS will be monitored on a regular basis to ensure integrity of the data 939 supplied. 940