Changes in stability of first-admission psychiatric diagnoses over 14 years, based on cross-sectional data at three time points

Background: Stability of diagnoses over time is an important criterion of reliability of any diagnostic system. Aims: To compare the stability of diagnoses among hospitalized psychiatric patients over 14 years, based on cross-sectional data at three time points. Method: Diagnoses extracted from the National Psychiatric Case Register concerning psychiatric patients first admitted during 1989 (n=2,996) were compared to those who were admitted during 1996 (n=3,021) and 2003 (n=4,041). Stability of diagnostic categories was measured by positive predictive value of admission diagnosis at discharge from hospital. Results: There was no significant difference in diagnostic stability for most diagnostic categories between 1989 and 1996 patient cohorts. However, over the seven following years (cohort 2003) the diagnostic stability had been substantially increased: by 19% for affective disorders, by 18% for childhood disorders, by 17% for organic conditions, by 14% for neurotic disorders, by 12% for both schizophrenia and drug and alcohol dependence. Conclusions: In long-term perspective, reliability of most diagnostic categories of mental disorders has clearly improved.