Comparison of Component of Structure for the PANSS

Till now, a majority of previous studies have performed principal component analysis to explain the structure of the PANSS, and there have been two studies identifying subtypes of the PANSS by cluster analysis (Dollfus et al, 1996) or generalized association plot (GAP, Hwu et al., 2002). The results by carrying out the RLCA without covariates using our program in the present study and these results of previous studies were shown in Table 34. While 12 of the 16 previous studies reported a five-factor solution, the criteria used to select the number of factors differed from study to study, and in fact two of the studies only reported a five-factor model (Marder et al., 1997; Lancon et al., 2000). Other studies, using the conventional method of selecting factors with eigenvalues > 1 actually obtained more than five factors, and then discarded or combined the additional factors for various reasons (Kay and Sevy, 1990; Bell et al., 1994a; Lykouras et al., 2000). Thus, the selecting of the number of factors was arbitrary.

In these studies, we found that all studies had the negative syndrome, whether in the acute/admission phase or in the chronic/discharge phase. The negative syndrome was included blunted affect (N1), emotional withdrawal (N2), poor rapport (N3), pas-sive/apathetic social withdrawal (N4) and lack of spontaneity/flow of conversation (N6) items in these studies. However, in the chronic phase of present study, the negative syn-drome was nested within the negative synsyn-drome of other results. A number of the negative syndrome were added the difficulty in abstract thinking item (N5) (present study; Liu, Yeh and Hwu, 1996; Hwu et al, 2002; Dollfus et al., 1996) or the part of the general psychopathology items.

In addition, a majority of previous studies emerged clearly the positive syndrome, except the studies for the subjects of MPGRP (Liu, Yeh and Hwu, 1996; Liu, Hwu,

Chen, 1997; Hwu et al, 2002) and the study in acute patients of Nakaya et al. (1999b).

However, in these studies, the other syndromes, which were the psychotic factor of the Liu et al. (1996) study at discharge, the factor of delusion/hallucination of the Liu et al. (1997) and of the Nakaya et al. (1999b), and the factor of delusion of the Hwu et al. (2002), were similar to the positive syndrome of other previous studies (such as Kay and Sevy, 1990; Lindenmayer et al. 1994; Dollfus and Petit, 1995, and so on). Dollfus et al. (1996) suggested the positive syndrome was included all positive items (P1-P7) and a part of the general psychopathology items. However, the positive class of our study didn’t included the grandiosity item (P5), and 7 of the previous studies suggested the positive syndrome was only included the delusions (P1), hallucinatory behavior (P3), grandiosity (P5), suspiciousness/persecution (P6) and a part of the general psychopathology items.

In addition, 3 of the previous studies suggested the positive syndrome was only included three positive items, which were delusions (P1), hallucinatory behavior (P3), grandiosity (P5)/suspiciousness (P6), and a part of the general psychopathology items (Kay and Sevy, 1990; Dollfus and Petit, 1995; White et al., 1997). In the study of Mass et al. (2000), the positive syndrome was included the delusions (P1), hallucinatory behavior (P3) and unusual thought content (G9) items. Thus, the components of positive syndrome were different, that possible reason was maybe to use different analysis.

In addition, there were only structure of one study not included the disorganized thought factor by the cluster analysis (Dollfus et al., 1996). However, they obtained the disorganized thought specially by subdividing the positive cluster. All previous studies indicated that the components of disorganized thought (or cognitive) were included the conceptual disorganization (P2) item, except the study of White et al. (1997) and the study in the post-acute patients of Nakaya et al. (1999b). A number of previous studies added the difficulty of abstract thinking (N5) and/or stereotyped thinking (N7) items into the components of disorganized thought. In the present study at the acute phase, the components of disorganized thought were included also the conceptual disorganization

(P2) and difficulty in abstract thinking (N5) items, and added the delusions (P1), hallu-cinatory behavior (P3), unusual thought content (G9) and lack of judgment and insight (G12) items. Thus, the components of disorganized thought of present study seemed to similar to previous studies.

On the other hand, in the present study, we had the delusion class which were in-cluded the delusions (P1) and lack of judgment and insight (G12) items. There were only four previous studies to indicate the delusion/hallucination factor in the structure of the PANSS (Liu et al., 1996; 1997; Hwu et al., 2002; Nakaya et al., 1999b). However, the components of the delusion/hallucination factor in these studies were more similar to the components of the positive syndrome in other previous studies. In addition, the previous study of Dollufs et al. (1996) had the mixed and few symptoms clusters in the structure of PANSS. These were similar to the mixed and no-symptoms class of the present study.

The previous study of Nakaya et al. (1999b) reported that the mixed factor was emerged in the post-acute phase. However, this study used the 14 items of PANSS, which were the positive and negative symptoms, to analyze.

A number of previous studies suggested that the structure of PANSS included the excitement and anxiety/depression factors. Depression and anxiety symptoms loaded as a single factor in the original PANSS analysis of Kay and Sevy (1990), as well as in the majority of subsequent studies. In 1995, Dollfus and Petit also reported the separate anxiety and depression factors and found an anxiety factor at admission and a depression factor at discharge. Besides, Emsley et al. (2003) also found an anxiety factor without a depressive factor.

To draw a conclusion of all the statements listed above, in the present study, the components of the negative, positive and disorganized thought classes of the acute phase were not different from the previous studies. However, the number of components of each structure in the chronic phase was less than the number of components of each structure

in the acute phase because the symptoms of the chronic patients were not obvious. A majority of previous studies hasn’t unveil that the sample was in the acute or chronic phase, and in a number of previous studies, patients in acute and stabilized phases were combined to be analyzed. Thus, it was difficult to discriminate the acute phase from the chronic phase to compare the component of structure for the PANSS. However, it can be discovered that there were much more difference of the results of the chronic phase between the present study and the previous studies than the results of the acute phase.