The three-syndrome model of schizophrenia: meta-analysis of an artefact.

In two recent studies, Smith et al. (Smith, D.A., Mar, C.M., Turoff, B.K., 1998. The structure of schizophrenic symptoms: a meta-analytic confirmatory factor analysis. Schizophr. Res. 31, 57-70) and Grube et al. (Grube, B.S., Bilder, R.M., Goldman, R.S., 1998. Meta-analysis of symptom factors in schizophrenia. Schizophr. Res. 31, 113-120) used meta-analysis to examine the syndromal structure of schizophrenia. A limitation of both these studies is that the nine subscale scores from Andreasen's Scales for Assessment of Positive and Negative symptoms formed the basis of the analyses. These nine ratings, only four of which represent positive symptoms, do not adequately respresent the diversity of positive symptoms. A review of studies that examined the correlation between the individual items of these scales failed to support the classification of symptoms into these nine subgroups. Studies that indicated low numbers of syndromes suffered from one or more of the following limitations: (1) samples that were restricted to chronic schizophrenia, (2) exclusion of many items from analysis, and (3) a poor fit of the symptom model to the data. Studies not limited in these ways indicated the presence of at least 11 major dimensions of schizophrenic symptomatology, not including affective symptoms. It is concluded that the three-syndrome model of schizophrenia is largely an artefact of inadequate measurement at the symptom level.