Bridging Perspectives: Clinician-Adolescent Agreement on Psychopathological Severity in the European MILESTONE Cohort

Abstract

Objectives: Adolescents transitioning from child and adolescent mental health services (CAMHS) to adult mental health services (AMHS) may face challenges in accurately identifying and reporting their mental health symptoms, often leading to discrepancies between clinician and patient evaluations. Using data from the MILESTONE project, this study aims to assess clinician-adolescent concordance over 24 months and identify domains of psychopathology with the highest disparities.

Methods: Participants were assessed at baseline, 9, 15, and 24 months using the Health of the Nation Outcome Scales for Children and Adolescents (HoNOSCA) scale and were categorized in four diagnostic groups. Hierarchical cluster analysis identified symptom-based subgroups of patients based on clinician and patient-rated HoNOSCA scores. Concordance was evaluated through multilevel linear regression models, while Bland-Altman plots examined agreement between scores across time points.

Results: Two clusters of patients were identified: one characterized by lower severity and greater prevalence, the other by higher complexity and fewer individuals. Clinician-patient concordance increased over time, rising from 77% to 83% by the second time point and stabilizing. Concordance varied across diagnostic categories, with anxiety showing the highest agreement and ADHD the lowest.

Conclusions: Improved communication, psychoeducation, and tailored interventions may facilitate greater patient-clinician alignment, thereby supporting more favorable outcomes during this critical developmental period.

Trial registration: ISRCTN83240263; NCT03013595.

Keywords: CAMHS to AMHS transition; HoNOSCA; clinician–adolescent agreement; longitudinal assessment; psychopathological severity.

Figure 1

Average item scores of patients and clinicians' evaluation in the two clusters at Time Points 1 and 4. Descriptive for complexity group extracted from HCA procedures. Patients evaluations: Time Point 1: N patients in Cluster (i) = 334; N patients in Cluster (ii) = 131; Time Point 4: N patients in Cluster (i) = 462; N patients in Cluster (ii) = 11; Clinicians evaluations: Time Point 1: N patients in Cluster (i) = 562; N patients in Cluster (ii) = 53; Time Point 4: N patients in Cluster (i) = 611; N patients in Cluster (ii) = 14.

Figure 2

Percentage of concordance at the four time points: (a) between clinician and patient evaluations; (b) across time points; (c) stratified by diagnosis. Concordance defined as assignment to the same severity cluster in different HCA procedures: (a) comparison of the cluster assigned by HCA based on clinician evaluation and HCA based on patient evaluation; (b) comparison of the clusters assigned at the four time points, separately for clinician and patient; (c) as (a), stratified by diagnosis. The overall Cochran's Q test was not significant across all time points, but a significant increase in agreement was observed between T1 and T2 only for patient evaluation (a). In stratified analyses (c) significant increases were also observed between T1 and T2 for depressive and anxiety disorders, and between T1 and T3 for OCD/eating and ADHD groups based on patient ratings.

Figure 3

Bland–Altman plot for evaluating differences between patients and clinicians' ratings. Bland–Altman plots for assessing agreement between clinicians and patients across four time points.