Clinical Stage Transitions in Persons Aged 12 to 25 Years Presenting to Early Intervention Mental Health Services With Anxiety, Mood, and Psychotic Disorders

Key Points

Question  What demographic and clinical factors are associated with transition from early (subthreshold) to full-threshold major persistent or recurrent psychiatric disorders?

Findings  This longitudinal cohort study of persons aged 12 to 25 years who presented to early intervention services found significant and ongoing risk of transition to major anxiety, mood, psychotic, or comorbid disorders. Poorer social function, psychotic-like experiences, manic-like experiences, and circadian disturbance were associated with illness progression.

Meaning  A clinical staging model for specific youth services may support the efficient allocation of appropriate care to young people and support the evidence-based planning of relevant early intervention and secondary prevention strategies.

Importance  The large contribution of psychiatric disorders to premature death and persistent disability among young people means that earlier identification and enhanced long-term care for those who are most at risk of developing life-threatening or chronic disorders is critical. Clinical staging as an adjunct to diagnosis to address emerging psychiatric disorders has been proposed for young people presenting for care; however, the longer-term utility of this system has not been established.

Objectives  To determine the rates of transition from earlier to later stages of anxiety, mood, psychotic, or comorbid disorders and to identify the demographic and clinical characteristics that are associated with the time course of these transitions.

Design, Setting, and Participants  A longitudinal, observational study of 2254 persons aged 12 to 25 years who obtained mental health care at 2 early intervention mental health services in Sydney, Australia, and were recruited to a research register between June 18, 2008, and July 24, 2018 (the Brain and Mind Centre Optymise Cohort).

Main Outcomes and Measures  The primary outcome of this study was transition from earlier to later clinical stages. A multistate Markov model was used to examine demographic (ie, age, sex, engagement in education, employment, or both) and clinical (ie, social and occupational function, clinical presentation, personal history of mental illness, physical health comorbidities, treatment use, self-harm, suicidal thoughts and behaviors) factors associated with these transitions.

Results  Of the 2254 individuals included in the study, mean (SD) age at baseline was 18.18 (3.33) years and 1330 (59.0%) were female. Data on race/ethnicity were not available. Median (interquartile range) follow-up was 14 (5-33) months. Of 685 participants at stage 1a (nonspecific symptoms), 253 (36.9%) transitioned to stage 1b (attenuated syndromes). Transition was associated with lower social functioning (hazard ratio [HR], 0.77; 95% CI, 0.66-0.90), engagement with education, employment, or both (HR, 0.47; 95% CI, 0.25-0.91), manic-like experiences (HR, 2.12; 95% CI, 1.19-3.78), psychotic-like experiences (HR, 2.13; 95% CI, 1.38-3.28), self-harm (HR, 1.42; 95% CI, 1.01-1.99), and older age (HR, 1.27; 95% CI, 1.11-1.45). Of 1370 stage 1b participants, 176 (12.8%) transitioned to stage 2 (full-threshold) disorders. Transition was associated with psychotic-like experiences (HR, 2.31; 95% CI, 1.65-3.23), circadian disturbance (HR, 1.66; 95% CI, 1.17-2.35), psychiatric medication (HR, 1.43; 95% CI, 1.03-1.99), childhood psychiatric disorder (HR, 1.62; 95% CI, 1.03-2.54), and older age (HR, 1.24; 95% CI, 1.05-1.45).

Conclusions and Relevance  Differential rates of progression from earlier to later stages of anxiety, mood, psychotic, or comorbid disorders were observed in young persons who presented for care at various stages. Understanding the rate and factors associated with transition assists planning of stage-specific clinical interventions and secondary prevention trials.

Quiz Ref IDThe large contribution of psychiatric disorders to premature death and disability is associated with their early age of onset, prevalence, chronicity, and comorbidity.^1,2 Although 75% of major psychiatric disorders begin before the age of 25 years,^3,4 current adult-based thresholds for diagnosis often map weakly onto earlier and nonspecific patterns of illness in young people.^5-8 Common symptoms of psychiatric disorders (eg, depressed mood, sleep disturbance, motor activation) that may seem diagnostically specific are shared across diagnoses, are common among patients with subthreshold and full-threshold disorders, and have variable patterns of association, differentiation, or severity across the illness course.^9-13 In addition, syndromes such as anxiety and neurodevelopmental disorders that are evident before age 12 years often precede the later development of a broad range of syndromes in the same individuals.¹⁴ Consequently, the challenge is to derive classification systems that are consistent with the neurobiological development of young people and patterns of disease development, and that are clinically useful.

A response to this challenge has been to apply clinical staging as an adjunct to formal diagnosis.^15-19 Clinical staging recognizes that the boundaries between common psychiatric disorders are often unclear and that an approach that accounts for their comorbidity is needed.^20,21 In other medical specialties (eg, oncology), it is commonly accepted that it is not ideal to choose treatments or plan health care for persons who are likely to experience illness progression or recurrence based solely on a cross-sectional diagnosis.²² Young people experiencing mental ill health vary along a continuum by factors including severity, duration of symptoms, and illness course (eg, first episode vs recurrent illness). Such factors are associated with different patterns of response to psychological or pharmacological interventions,^23-27 and different individuals may benefit from variable types of secondary prevention strategies.^28-32 Consequently, we have proposed a framework for clinical staging among young people presenting with anxiety, mood, or psychotic syndromes.¹⁵

This framework proposes that earlier stages are clinically heterogeneous and more likely to be characterized by lower rates of impairment and different rates of progression to more discrete, persistent, or recurrent disorders. When applied to young people who present for care, the first distinction is between those in early subthreshold phases (stage 1a [nonspecific symptoms] or 1b [attenuated syndromes]) and those who have reached full threshold for major, discrete, and persistent or recurrent disorders (stage 2) (Figure 1). Within stage 1 (subthreshold), we differentiate 2 levels: stage 1b, which we describe as attenuated syndromes and stage 1a, nonspecific anxiety and depressive symptom clusters. Individuals assigned to stage 1b often have the symptoms, duration, and impairment to meet DSM-5 criteria³³ or International Classification of Diseases, Tenth Revision, Clinical Modification criteria³⁴ for specific anxiety or mood disorders; however, compared with stage 2 disorders, these are typically less severe, brief, or not persistent or recurrent. Individuals assigned to stage 1a typically have fewer symptoms and impairment and a shorter duration of illness and, as such, usually do not meet DSM-5 criteria³³ or International Classification of Diseases, Tenth Revision, Clinical Modification criteria.³⁴

Previous longitudinal studies indicate that within 12 months approximately 15% to 20% of patients with disorders rated as stage 1b progress to a later stage and many of these clinical transitions occur within the first 3 months of mental health care.³⁵ The differentiation of these stages is supported by other independent neuropsychological, neurobiological, and circadian markers identified among young people with emerging anxiety, mood, or psychotic syndromes.^36-43

This study describes the characteristics of a cohort of young people presenting with a broad range of anxiety, mood, or psychotic syndromes and examines the rates of transition from stage 1a to 1b, and from stage 1b to 2, and the demographic and clinical characteristics associated with the time course of these transitions.

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. The study was approved by the University of Sydney Human Research Ethics Committee and participants, their guardians, or both gave written informed consent for the use of routinely collected clinical data for research purposes.

Quiz Ref IDParticipants were drawn from a cohort of 6743 individuals aged 12 to 30 years who presented to the Brain and Mind Centre’s youth mental health clinics in Sydney, Australia, and recruited to a research register between June 18, 2008, and July 24, 2018. These clinics include primary care services (ie, headspace, a “1-stop-shop” for young people who need help with mental health, physical health [including sexual health], alcohol and other drugs or work and study support^44-46) as well as more specialized mental health services. Young people were self-referred, referred via a family member or friend, or the community (eg, general practitioner).⁴⁶ All participants received clinician-based case management and psychological, social, medical, or a combination of interventions as part of standard care.

As of December 18, 2018, phase 1 of data entry was completed, and so data were available for 2767 participants, with 78 excluded owing to insufficient data. The inclusion criteria for this study were age 12 to 25 years at baseline and at least 1 month of follow-up.

Data were extracted from clinical files and code inputs according to a standardized form at predetermined time points.⁴⁷ The first available clinical assessment at the mental health service is taken as the baseline time point for each participant, and the date of this assessment is used to determine each of the follow-up time points. If there is no clinical information available for a time point (ie, the participant did not attend the service during that time), that entry is left missing. All clinical notes from the preceding time points, up to and including the current time point are used to inform and complete the current entry.

The proforma was used to record specific illness course characteristics. More detailed descriptions about the proforma, including the interrater reliability are reported in the supplement and cohort article.⁴⁸ The measures used here include (eAppendix in the Supplement); demographic features, social and occupational functioning (including the Social and Occupational Functioning Assessment Scale [SOFAS],⁴⁹ a clinician-rated measure that assesses functioning on a 0 to 100 scale, with lower scores suggesting functional impairment. The instructions emphasize that the rater should avoid confounding the rating with clinical symptoms,^47-49 and Not in Education, Employment or Training [NEET] status as a measure of participation and engagement with education or work), psychiatric disorder diagnoses, clinical stage, at-risk mental states, self-harm, suicidal thoughts and behaviors, physical health comorbidities, personal mental illness history, and treatment use.

Analyses were performed using R statistical software (R Foundation).⁵⁰ Differences in demographic and clinical characteristics between groups defined by clinical stage at baseline were assessed using 1-way analysis of variance for continuous variables and χ² tests for categorical variables. Pairwise comparisons were carried out for stage 1a vs stage 1b and stage 1b vs stage 2; owing to the number of univariate analyses conducted, an α correction using the Bonferroni method was made. For these analysis, statistical significance was set at 2-sided P < .001.

A multistate Markov model (msm package version 1.6.6; R Foundation)⁵¹ was fitted to determine which demographic and clinical characteristics at baseline were associated with time course of transitions. This analysis modeled 2 transitions: stage 1a to stage 1b and stage 1b to stage 2. The progressive Markov model only allows forward transitions without skipping. The risk of transition for each individual in the cohort is quantified by that individual’s unique transition intensity, which depends on personal demographic features and baseline clinical characteristics; these dependencies form the main parameters of the model. The msm package fits the model to longitudinal panel-observed data, whereby individuals are followed up and classified intermittently, but the exact time of transition to the later stage (eg, stage 1a to stage 1b) is not known. Therefore, these data are assumed to be interval censored, whereby the exact time of transition is only known to lie within an interval (ie, 2 observational time points). Hazard ratios (HRs) and 95% CIs were estimated to determine the change in probability of transitions for each variable at baseline relative to a reference value or absence of that characteristic. Survival probability plots were generated to model the empirical and fitted time-to-transition for transitions using the msm package, which is suited to plotting interval-censored data.

Differences in follow-up time between those who transitioned to stage 2 vs those who did not were assessed using the nonparametric Kruskal-Wallis test. χ² Analyses were undertaken to determine differences in the overall rate of transition from stage 1a and stage 1b at baseline to stage 2 at last follow-up.

The cohort comprised 2254 individuals; 1330 (59.0%) were female with a mean (SD) age of 18.18 (3.33) at baseline and a median (interquartile range [IQR]) follow-up of 14 (5-33) months. Data on race/ethnicity were unavailable. Pairwise comparisons of characteristics according to clinical stage at baseline are in Table 1. Participants who were classified as being at the earlier stages were younger (F_2,2251 = 88.91; P < .001) and more likely to present with less functional impairment, as determined by both lower NEET rates (χ²₂ = 86.67; P < .001) and higher SOFAS (F_2,2231 = 169.17; P < .001).

The pattern of differences between all 3 groups in terms of at-risk mental states, treatment use, and suicidality was positive. Specifically, more-advanced clinical stages were associated with higher rates of manic-like experiences (χ²₂ = 162.32; P < .001); psychotic-like experiences (χ²₂ = 245.54; P < .001), circadian (or sleep-wake) disturbance (χ²₂ = 85.81; P < .001), hospitalization (χ²₂ = 566.16; P < .001), and psychiatric medication (χ²₂) = 455.06; P < .001). There were no differences between participants with stage 1b and stage 2 disorder at baseline in terms of self-harm (χ²₁ = 2.79; P = .08) and suicidal ideation (χ²₁ = 0.15; P = .88); however, participants with stage 2 disorder at baseline were more likely to have a previous suicide attempt (χ²₁ = 24.06; P < .001). Compared with stage 1a, participants with stage 1b disorder at baseline reported higher rates of self-harm (χ²₁ = 135.45; P < .001), suicidal ideation (χ²₁ = 108.27; P < .001), and suicide attempts (χ²₁ = 90.77; P < .001).

Quiz Ref IDFor the 685 participants initially classified at stage 1a, 253 (36.9%) progressed to stage 1b (Figure 2). Notably, 110 (46.8%) of these transitions occurred within 6 months and 153 (65.1%) within 12 months of baseline. The model identified 6 factors associated with transition from stage 1a to 1b (Table 2): older age (HR, 1.27; 95% CI, 1.11-1.45), lower social functioning (HR, 0.77; 95% CI, 0.66-0.90), engagement in education, employment, or training (HR, 0.47; 95% CI, 0.25-0.91), manic-like experiences (HR, 2.12; 95% CI, 1.19-3.78), psychotic-like experiences (HR, 2.13; 95% CI, 1.38-3.28), and self-harm (HR, 1.42; 95% CI, 1.01-1.99). Childhood attention-deficit/hyperactivity disorder was associated with a lower risk of transition (HR, 0.43; 95% CI, 0.24-0.78).

For the 1370 participants initially classified at stage 1b, 176 (12.8%) progressed to stage 2 (Figure 2). Eighty of the 176 (45.4%) of these transitions occurred within 12 months of baseline. The multi-state Markov model identified 5 factors associated with transition from stage 1b to stage 2 (Table 2): older age (HR, 1.24; 95% CI, 1.05-1.45), psychotic-like experiences (HR, 2.31; 95% CI, 1.65-3.23), circadian disturbance (HR, 1.66; 95% CI, 1.17-2.35), previous use of psychiatric medication (HR, 1.43; 95% CI, 1.03-1.99), and a history of childhood psychiatric disorders (HR, 1.62; 95% CI, 1.03-2.54).

Quiz Ref IDOf the 685 individuals who presented as stage 1a, 18 (2.6%) transitioned to stage 2, compared with 176 of 1370 individuals (12.8%) who presented as stage 1b (χ²₁ = 55.78; P < .001) (Figure 2; Table 3). Those stage 1a participants who did transition to stage 2 were followed up for a median (IQR) of 51 (42) months, compared with 11 (20) months for those who had not transitioned to stage 2 (χ²₁ = 26.32; P < .001). Similarly, those who were stage 1b at baseline and transitioned to stage 2 were followed up for a median of 37 (41) months, compared with 13 (23) months for those who did not transition (χ²₁ = 129.51; P < .001). Among the 194 who transitioned to stage 2, 47 (24.2%) primarily had a psychotic-type syndrome, 86 (44.3%) had a bipolar syndrome, and 61 (31.4%) had a major anxiety or depressive syndrome.

This study reports the longer-term rates of progression from early to later clinical stages and the demographic and clinical factors associated with transitions for a large clinical cohort engaged in active psychological, social, medical, or a combination of interventions as part of standard care. First, this study demonstrates that young people who present at stage 1a are at low risk of early progression to full-threshold disorders; however, a significant subgroup characterized by more complex clinical presentations such as manic-like experiences and psychotic-like experiences are at increased risk of early transition to stage 1b syndromes. Whether this group is also at increased risk of longer-term transition to stage 2 full-threshold disorders is not yet known. Second, a substantial subgroup of stage 1b participants are at risk of early progression to full-threshold, persistent, or recurrent disorders. Psychotic-like experiences and circadian disturbances have some capacity to predict these transitions. Finally, within the transdiagnostic model proposed, the differentiation of stage 1a from stage 1b disorders is justified by the differential rates of progression to stage 2 disorders.

Consistent with previous reports,⁵² 176 of 1370 (12.8%) of participants with stage 1b syndromes at baseline progressed to a more severe stage, with at least 45% of those transitions occurring in the first 12 months after presentation to care. Furthermore, although 34.3% of those at stage 1a progressed to stage 1b, 2.6% progressed to stage 2. The differential rates of progression to stage 2 support the assumptions about longitudinal trajectories that underpin the clinical staging model, namely, that young people at stage 1b at baseline have a higher early risk of developing a major discrete, persisting, or recurrent disorder than young people at stage 1a. Transition to stage 2 disorder does not simply equate with movement from a single subthreshold anxiety, mood, or psychotic-like syndrome to a first episode of mania or single major depressive, bipolar, or psychotic disorder (as described by DSM-5 criteria³³ or International Classification of Diseases, Tenth Revision, Clinical Modification criteria³⁴). In this model,⁵² transition is associated with progression to a more crystallized and enduring syndrome that has distinct features (eg, severely depressed mood associated with psychomotor agitation or retardation) or the development of additional features (eg, psychotic symptoms). The present findings support the health services utility of this model for guiding resource allocation with regard to treatment intensity and strategies toward patients with a greater risk of transition, and avoiding the limitations of premature or arbitrary subclassifications, particularly for young people whose clinical presentations are often mixed or associated with substantial comorbidity.⁵³

An important clinical consideration is whether we can go beyond broad clinical stage and identify individuals who are at greatest risk of progression to full-threshold, more severe, or persistent forms of illness. Herein, we extend earlier illness progression work carried out among ultrahigh-risk groups⁵⁴ by evaluating these transitions among a transdiagnostic sample of young people. For individuals initially classified as having stage 1b syndromes, psychotic-like experiences and circadian disturbance were associated with transition and demonstrate the heterogeneity of clinical characteristics associated with major anxiety, mood, or psychotic disorders. There is ongoing debate as to whether psychotic-like experiences are principally a marker of illness severity or comorbidity, rather than having any specific capacity to predict later psychotic disorders.^55-58 Circadian disturbance is increasingly recognized as a major characteristic of more severe mood disorders, including bipolar disorder, psychotic depression, and atypical depression in younger cohorts.^59-61 Such characteristics may reflect the possible pathophysiological mechanisms (eg, circadian, hypothalamic-pituitary-adrenal axis dysfunction) that differentiate the earliest subthreshold stages from full-threshold syndromes, which is consistent with differential changes in neuropsychological function, structural brain change, and circadian disruption observed previously.^37-43 Importantly, neither suicidal ideation nor previous suicide attempt discriminated between these 2 groups, indicating that the course of suicidality may run somewhat independent of illness progression or specific diagnosis.

With regard to the characteristics of those in stage 1a who transitioned to stage 1b, a broader range of impairment (lower social and occupational functioning at baseline) and clinical (psychotic-like experiences, manic-like experiences, and self-harm) features were associated with transition. Presumably, this reflects the greater underlying heterogeneity of illness trajectories in this subgroup with a mixture of features that is consistent with emerging psychopathology.^52,62 Specifically, despite the absence of substantial severity, specificity, or impairment, these features may indicate a higher degree of complexity, characterized by the presence of major risk factors or underlying pathophysiology. It does suggest that intervention and secondary prevention strategies for this group may need to focus more selectively on those individuals who present with 1 or more of these clinical characteristics.

There were also several clear differences in demographic, clinical, and functional characteristics of participants with early subthreshold (stages 1a or 1b) and later full-threshold (stage 2 or above) disorders at entry to care. The differences in age are notable, with 8.8% of the sample presenting with a discrete disorder (stage 2) and being approximately 3 years older than those presenting at stage 1a. If we are to increase the chance of making clinical contact with young people before they reach stage 2 or higher, we must continue to enhance strategies that attract younger persons to clinical care, potentially at earlier phases of illness.

Quiz Ref IDThis cohort is a selected subset of a larger cohort of young people presenting for care (2254 of a cohort of 6743 [33%]). Given that phase 1 data entry focused on individuals who had also participated in other more detailed neurobiological research studies, this sample may not be representative of all treatment-seeking young people in this region, and this subgroup may be biased toward inclusion of those who already have stage 1b attenuated syndromes, more severe depressive syndromes, more comorbidity and clinical complexity, alcohol or substance misuse, and suicidality. Furthermore, the subgroup used here varied in duration of follow-up, which means that those engaged for longer periods of time tended to be more likely to transition. Consequently, the rates of progression over time and the factors associated with progression may differ if follow-up was more consistent across this sample or among less-severely unwell cohorts including participants with early-phase but less-complex disorders. Yet, given the long periods of untreated illness in the community, the degree of progression among this cohort may actually be in line with comparable cohorts in cities throughout the United States or Europe where early intervention campaigns and services for youth mental health, such as headspace in Australia, are not widespread.

Another limitation is that the data are extracted from clinical records, rather than via prospective structured assessments. However, the data collection is structured, conducted by trained staff, completed independently of treating clinicians, and has acceptable interrater reliability. A final limitation, the use of a prespecified list of factors that did not include other potentially relevant factors such as temperament or other social factors, means that there may be other relevant factors that were not considered by the current analyses but may be important.

These findings, when considered alongside the concurrent neurobiological data we have presented previously,^37-43 provide the basis for the implementation of the clinical staging model in daily practice with young people; the design of specific youth clinical service models to support the efficient allocation of appropriate care^63-65; and the evidence-based planning of stage-based early intervention and secondary prevention studies. In association with those clinical studies, there is a need to investigate other potentially differentiating neurobiological, psychosocial, or pathophysiological markers within those young people who present for care at early stages of illness.

Accepted for Publication: June 3, 2019.

Corresponding Author: Frank Iorfino, PhD, Youth Mental Health Team, Brain and Mind Centre, University of Sydney, 94 Mallet St Camperdown, Sydney, New South Wales, Australia 2050 (frank.iorfino@sydney.edu.au).

Published Online: August 28, 2019. doi:10.1001/jamapsychiatry.2019.2360

Author Contributions: Drs Hickie and Iorfino had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs E. M. Scott and Iorfino are joint first authors.

Concept and design: Iorfino, E. M. Scott, Carpenter, Cross, Hermens, J. Scott, McGorry, Hickie.

Acquisition, analysis, or interpretation of data: Iorfino, Carpenter, Cross, Hermens, Killedar, Nichles, Zmicerevska, White, Guastella, J. Scott, Hickie.

Drafting of the manuscript: Iorfino, Hermens, Killedar, J. Scott.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Iorfino, Killedar, J. Scott.

Obtained funding: Hickie.

Administrative, technical, or material support: E. M. Scott, Carpenter, Zmicerevska, White, Guastella, Hickie.

Supervision: E. M. Scott, Cross, Hermens, Guastella, J. Scott, Hickie.

Conflict of Interest Disclosures: Dr E. M. Scott is medical director, Young Adult Mental Health Unit, St Vincent’s Hospital Darlinghurst; discipline leader of Adult Mental Health, School of Medicine, University of Notre Dame; Research Affiliate, The University of Sydney; and Consultant Psychiatrist. She has received honoraria from Laboratoires Servier and Eli Lilly and Company for presenting educational seminars related to the clinical management of depressive disorders and has participated in a national advisory board for the antidepressant compound Pristiq, manufactured by Pfizer. She was the national coordinator of an antidepressant trial sponsored by Laboratoires Servier. Dr J. Scott is a visiting professor at Diderot University, the Norwegian University of Science and Technology, Swinburne University of Technology, and The University of Sydney and a Science Without Borders fellow (Brazil). She has received grant funding from the UK Medical Research Council and from the UK Research for Patient Benefit program. Dr Hickie was an inaugural Commissioner on Australia’s National Mental Health Commission (2012-18). He is the co-director, Health and Policy at the Brain and Mind Centre (BMC), University of Sydney. The BMC operates an early-intervention youth service at Camperdown under contract to headspace, National Youth Mental Health Foundation Ltd. Dr Hickie has previously led community-based and pharmaceutical industry–supported (Wyeth, Eli Lily, Servier, Pfizer, AstraZeneca) projects focused on the identification and better management of anxiety and depression. He was a member of the Medical Advisory Panel for Medibank Private until October 2017, a board member of Psychosis Australia Trust, and a member of Veterans Mental Health Clinical Reference group. He is the Chief Scientific Advisor to, and a 5% equity shareholder in, InnoWell Pty Ltd. InnoWell was formed by the University of Sydney (45% equity) and PricewaterhouseCoopers ([PwC] Australia; 45% equity) to deliver the $30 million Australian Government-funded Project Synergy (2017-2020; a 3-year program for the transformation of mental health services) and to lead transformation of mental health services internationally through the use of innovative technologies. No other conflicts were reported.

Funding/Support: This work was partially supported by grants from the National Health & Medical Research Council Centre of Research Excellence grant 1061043, and Australia Fellowship 511921 (both, Dr Hickie).

Role of the Funder/Sponsor: The funders had no input into the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We would like to thank all the young people who have participated in this study and all the staff in the Youth Mental Health Team at the Brain and Mind Centre, past and present, who have contributed to this work.