PSYCHIATRIC nosology has been too long the province of speculation and pronouncement, particularly in the area of psychotic disorders, where powerful empirical studies are rare.
Our study1 began with a large epidemiological sample of treated patients with psychotic and affective illness studied an average of 15 years after onset. Based on personal interviews and review of hospital records, we blindly coded symptom and outcome data. We submitted these data to a latent class analysis (LCA): a statistical technique that attempted to explain the observed distribution of symptoms in these probands as resulting from a small number of discrete syndromes.2