Symptom diagnostics based on clinical records: a tool for scientific research in child psychiatry?

Abstract

Introduction: Child psychiatric diagnoses are generally based on a clinical examination and not on standardized questionnaires. The present study assessed whether symptom diagnostics based on clinical records facilitates the use of non-standardized clinical material for research. Method: Six hundred and eighty-five children, referred to a third level child psychiatric centre in the Netherlands, were, after extensive multidisciplinary examination, classified according to the multi-axial classification scheme for psychiatric disorders in childhood and adolescence (MAC-ICD-9). By two raters 44 behavioural symptoms were scored based on the clinical records of these children. Interrater agreement on symptoms in 50 records was performed. Principal components analysis on symptom scores of all children was performed; factor scores were related with MAC-ICD-9 classifications. Results: Interrater reliability for behavioural symptoms was excellent (κ = 0.88). Many children with psychiatric problems suffer from a large number of behavioural symptoms. Factor scores of the symptoms revealed recognizable and well interpretable entities and indicated overlap in symptomatology and comorbidity. Conclusion: A symptom-based diagnostic approach based on extensive clinical patient files may provide a special dimension to improve the reliability of psychiatric classification.