To the best of our knowledge, the present study is the first to use longitudinal data and RLCA to explain the symptomatology schizophrenia in over time. The present findings didn’t suppose the two-dimensional construct of positive and negative symptoms in either the acute or the chronic phase of the illness. A five-/four-class model fit the data relatively better than two- to three-/four- class in the acute/chronic phase by the AIC and BIC criteria. A previous study of Nakaya et al. (1999a) reported that the three-, four- and five-dimensional model on the PANSS fits in 100 admitted patients well in the acute phase and only the five-dimensional model adequately fits the data in the chronic stable phase by using confirmatory factor analysis (CFA). The present finding in the acute phase was analogous to their results of the acute phase, but the result in the chronic phase was a little different. It can be conjectured that the different result was mainly due to different analyzing method
Based on the conclusion from our program, there were five classes labeled: mixed, negative, disorganized thought, delusion and positive under the utilization of RLCA in the acute phase. In the result based on Mplus, the positive class did not emerge, but the a little mixed class replaced the positive class. In addition, the significant demographic variables in the conclusion based on our program were different from that based on Mplus.
These different results may be due to the reason that the initial value in our program was
different from that in Mplus, and the threshold parameter, τ , entered into the mixed model of Mplus. All these factors may lead to different results. However, in the chronic phase, the result based on our program was similar to the result based on Mplus. In the chronic phase, there were four classes labeled: a little mixed, negative, delusion and no-symptoms. The different results between our program and Mplus in the acute and chronic phases could be due to the fact that the latent class model with four-class in the chronic isn’t more complex than the latent class model with five-class in the acute chronic.
The most salient finding for demographic characteristics is that older patients had more mixed symptoms. Men were more likely to develop negative symptoms than women in the acute phase. Patients with fewer years of education were more likely to be in the mixed class or the disorganized thought class in the acute phase, and more likely to be assigned to the classes which had more serious symptoms than no-symptoms class at the chronic phase. Besides, patients without occupation had high probability to be allocated to the a little mixed class or the negative class in the chronic phase, and patients with occupation or older age of onset of psychotic symptom had higher possibility to be assigned to the delusion class, which was the slight class in the acute phase. The analogous results have been reported in some previous studies. For instance, Van Den Oord et al. (2006) has unveiled that negative symptoms were somewhat less severe in females and except for positive and excited, more severe symptomatology was associated with fewer years of education. Reichenberg et al. (2005) also found that the correlation between years of education and negative/cognitive (alike disorganized though) factor is negative. However, according to some previous studies, there are no symptom components correlating significantly with any demographic or clinical variables (Liddle, 1987; Malla et al., 1993; Nakaya et al., 1999b). In addition, patients who had unstable mood or abnormal behavior to interfere with adapting to daily life may have higher tendency to be assigned to the negative class than patients without these characteristics in both the acute and chronic phases. However, none of these previous studies have reported about
the relationship between the environmental factors of present study and symptoms.
In another perspective, we found that in the acute phase, patients with low sustained attention would have high probability to be allocated to the negative class. This finding has demonstrated that the relationship between the undegraded d0 of CPT and negative class was negative and confirmed Liu et al. (1997)’s suggestion that the negative dimension was associated with lower sensitivity index (d0). Based on their report, the positive dimension was not associated with the d0 on the CPT, and this was similar to the result of present study. Furthermore, some previous studies published by Mass et al. (2000) and Good et al. (2004) also mentioned that there was significant correlation between the neuropsychological variables and cognitive/disorganized though. However, in the present study, the disorganized though class only emerged in the acute phase. In the acute phase, the correlation between the neuropsychological variables and symptoms was not investigated because the participants did not assess the neuropsychological variables, except CPT performance. However, in the present study, the sensitivity index (d0) on the CPT was non-significant in the disorganized though class, and this was similar to the result of previous study (Good et al., 2004).
In the field of psychopathology research, both the previous and present studies have examined the symptom structure for two main purposes. First, the recognition of consis-tent patterns of symptom clusters may help identify homogeneous subgroups of patients and provide validation for diagnostic concepts. Second, distinct clusters may hypothet-ically reflect distinct pathophysiologies within the schizophrenic disorder. T Based on exploratory factor analysis(EFA), the structure of symptoms in schizophrenia has been discovered in most previous studies. The structure of PANSS based on RLCA was a little different from the structure of PANSS based on EFA. The RLCA is the categorical ap-proach to posit that schizophrenia may be subdivided in separate and mutually exclusive groups of patients. The dimensional model, such as EFA, proffers that the symptoms of schizophrenia tend to cluster together within different symptom complexes which can
co-exist in individual patients. Since the statistical methodology and heterogeneous clinical characteristics of the disease are different, the symptom structure of PANSS is also dif-ferent. However, the best approach for examining the symptom structure of PANSS still remains unknown. But the best approach to explore it can be considered in the future.
4 Transition of Structure on PANSS
4.1 Background
Most studies demonstrate high agreement that schizophrenia is a heterogeneous disorder with considerable variation in symptoms, premorbid history, clinical course, prognosis, and pathophysiology. Crow (1980) proposed that the structure of schizophrenic symptoms can be discriminated between the positive symptoms, such as hallucinations and delusions, and negative symptoms, such as blunted affect and passive social withdrawal. The researchers began to produce evidence for a syndromic dichotomy in succession (Bilder et al.,1985;
Cornblatte et al., 1985; Andreasen and Grove, 1986; Kay and Sevy, 1990; Mortimer et al., 1990). The dichotomy has been widely accepted and led to the development of reliable scales for measurement of positive and negative symptoms, such as such as the Assessment of Negative Symptoms (SANS; Andreasen, 1983), the Scale for the Assessment of Positive Symptoms (SAPS; Andreasen, 1984). Later, the Positive and Negative Syndrome Scale (PANSS; Kay et al., 1987) was developed in an attempt to provide a comprehensive assessment of all symptoms of schizophrenia. The PANSS is widely used in clinical and research setting and is regarded as a reliable means of symptom assessment (Bell et al., 1992).
Many of these investigations have developed the symptom structures from Crow’s orig-inal two-dimension distinction, and others have found that more than two components are needed to describe the symptoms in Schizophrenia (Liddle, 1987; Arndt et al., 1991; An-dreason et al.,1995; Lindenmayer et al., 1995; Lenzenweger and Dworkin, 1996; Johnstone and Frith, 1996). For instance, Liddle (1987) has proposed the disorganization symptoms.
Later, Cuesta and Peralta (1995) compared seven models by using confirmatory factor analysis, and they found that the three- and four-factor models, which included disor-ganization and/or disorder of a relating syndrome in addition to positive and negative syndromes, obtained higher goodness of fit than one- or two-factor models. According to
a recent study published by Dollfus and Everitt, it is suggested that a four-factor model fit as well as two- and three- factor models (Dollfus and Everitt, 1998). White et al. (1997) also fitted 20 previously proposed models to data from a sample of 1,233 schizophrenics for attempt to reconcile the different research finds. They concluded that none of these models fitted the data adequately, then they derived a new ”pentagonal” model retaining only 25 items of the PANSS, which were labeled: Positive, Negative, Dysphoric mood, Activation, and Autistic preoccupation. Most of the studies that attempt to examine the symptom components in schizophrenia have been limited by the factor that symptoms were measured only cross-sectionally. Therefore, how the composition of the symptom components changes over time remains unknown.
In 1990, Kulhara and Chandiramani (1990) have found 98 schizophrenic inpatients could be divided into three symptom factors (negative symptoms, positive symptoms, thought disorder). However, 18-30 months later, 79 of these patients were reassessed and the composition of these symptom factors had changed, which was that a mixed symptom factor replayed the positive symptom factor. In 1991, Goldman et al. (1991) published the report which indicated that at both time, which were prior to intervention (medication-free baseline) and after 4 weeks of neuroleptic treatment, three symptom factor were evident (negative symptoms, positive symptoms, and unstable behavioral agitation), and the pre- and post-treatment factor loading patterns were similar in 40 schizophrenic inpatients. At the same year, Addington and Addington (1991) also found two symptom factors that possess eigenvalues greater than unity (negative symptoms and thought disorder) in 41 schizophrenic inpatients at the beginning of the study. However, after 6 months, the reality distortion factor appeared in place of the thought disorder.
Van der Does et al. (1995) rated 65 schizophrenic patients at the acute phase, 3 months later, and 1 year after the second assessment. They found that there was a different factor structure at each assessment, but a four-dimensional structure (disorganization, negative symptoms, positive symptoms, and depression) was stable over time. According to a
study which observed 86 newly admitted schizophrenic patients and was conducted by Nakaya et al. (1999), four symptom factors were investigated in the acute phase (negative symptoms, excited, delusion/hallucinatory, and thought disorder). However, in the post-acute phase, three symptom factors were evident (negative symptoms, mixed symptoms, and though disorder). Therefore, they suggested that the negative symptom component is stable while the difference in the phase of illness has some effects on the symptom structure of schizophrenia. In a word, each previous study led to different findings about the composition of symptom components over time, and the sampling and assessment methods differed among the previous studies, making any comparison difficult.
Although a part of previous studies has explored the symptomatology of schizophrenia in different phase, how the patients change between the acute phase and the chronic phase is still unknown. In addition, Nakaya et al. (1999b) reported that the difference in symptomatology between the acute and post-acute phase of schizophrenia. Therefore, the present study mainly focuses on the changes in latent class of the PANSS over time, and the study reported in this article aims to examine the changes in the structure of the PANSS items in both the acute phase and the chronic phase under latent transition analysis (LTA). Furthermore, LTA with demographic variables, environmental factors or neuropsychological variables are all applied to explore the changes of the structure of the PANSS after the adjustment of demographic variables, environmental factors or neuropsychological variables.
4.2 Method
4.2.1 Subjects
The subjects were composed of three projects, the Multidimensional Psychopathology Group Research Projects (MPGRP), the Multidimensional Psychopathological Study on Schizophrenia (MPSS) and the Study on Etiological Factors of Schizophrenia (SEFOS).
The initial project started as the MPGRP from July 1993 till June 1998. The subsequent
project following the initial MPGRP, was the MPSS started in July 1998 till June 2001.
Both MPGRP and MPSS were successfully carried out from July 1993 to March 2001, and up to the time of sending this SEFOS proposal as the subsequent study on the pathogenesis of schizophrenia, a further step of psychopathological study on schizophrenia.
The focus of the MPGRP was to study the clinical manifestations of schizophrenia and the family situation in a cohort of schizophrenia patients. The MPGRP also con-centrated on the phenotype definition of schizophrenia using CPT manifestation in the schizophrenia family. In the MPSS project, the focus was on the follow-up neuropsycho-logical evaluation of the schizophrenia cohort collected in the MPGRP, other than the descriptive follow-up clinical data collection. The Program Project Grant (PPG) entitled SEFOS from January 2002 till December 2005, which aimed to search for the separate etiological factors under the understanding that schizophrenia is a complex disorder. The PPG of SEFOS formulated a dynamic etiological hypothesis of schizophrenia and was a retrospective/prospective study. The PPG of SEFOS designs 3 projects of: (1) A Study on Neurobiology of Schizophrenia; (2) A Study on Environmental insults/stress of schizophrenia; and (3) Molecular Genetics Study of Schizophrenia. The main purpose of these projects is to find different levels of neurobiological and anatomical abnormalities, to discover different levels of environmental insults/stress, and to locate vulnerability genes in different chromosome regions respectively.
The recruitment procedures have been described in detail in earlier reports of MPGRP project (Liu et al., 1997; Chen et al., 1998b; Chang et al., 2001). Briefly, from August 1, 1993 to June 30, 1998, all patients consecutively admitted to the acute inpatient wards of three hospitals, National Taiwan University Hospital, Taipei City Psychiatric Center, and Taoyuan Psychiatric Center, were included in MPGRP if they met DSM-IV (Ameri-can Psychiatric Association, 1994) criteria for schizophrenia and consented to participate.
The diagnoses were re-evaluated at discharge by consensus among three senior psychia-trists using all information available from clinical observations, medical records, and key
informants. Up to 1998, the final year of MPGRP and the starting point for MPSS study, the MPGRP cohort would have been in their 2-5 years’ of follow-up period. On this ground, further follow-up of the MPGRP cohort into the long term course, supplemented by neuropsychological evaluations, would provide unusual opportunities for an integrated clinical and neuropsychological approach. The MPSS project thus recruit MPGRP pa-tients who agree to receive further follow-ups. Averagely, papa-tients in the MPSS project were also included in the MPGRP for three follow-up years. In addition, the family which had two schizophrenia sib-paired children - one schizophrenia parent and the other one should be normal - was the inclusion criteria for SEFOS.
This study included the 219 acute patients who had complete information from the PANSS at admission in the MPGRP project. The 122 chronic patients were assessed the PANSS in the first year of MPSS project and the 103 chronic patients had complete assessment of PANSS in the SEFOS project. Thus this study included the 225 chronic patients who participated in the MPSS or SEFOS project. On the other hand, the 115 subjects among these patients included were both assessed the PANSS in the MPGRP and MPSS projects. Thus, the patients in the MPGRP project was divided two groups, which one was follow-up into the MPSS project and the other was loss to follow-up into the MPSS project. Table 1 shows that the characteristics of two groups of patients. In the Table 1, it seems that the characteristics of the dropout patients were non-different from the non-dropout patients.
4.2.2 Instruments
The main applied instrument in this study is the PANSS, which is an assessment of the clinical symptoms of the patients. It has 33 items rated from 1 to 7 based on a semi-structured interview with detailed descriptions for symptom ratings, and it consists of four subscales: positive (seven symptoms: P1-P7), negative (seven symptoms: N1-N7), general psychopathology (sixteen symptoms: G1-G16), and supplementary excitability
(three symptoms: S1-S3). Each item on the PANSS is accompanied by a complete def-inition as well as detailed anchoring criteria for all seven rating points, which represent increasing levels of psychopathology: 1 = absent, 2 = minimal, 3 = mild, 4 = moderate, 5 = moderate-severe, 6 = severe, 7 = extreme. The subscales of positive and negative syndromes are assumed to cover the core symptoms in these two dimensions (Kay et al., 1991). The subscales of general psychopathology and supplement items for the aggression risk profiles are considered to be the separated index of severity of illness (Kay et al., 1986).
The Chinese version of the PANSS, the PANSS-CH, was translated from the English ver-sion specifically for the MPGRP. The details of development of the PANSS-CH and the reliability test were published in earlier literature (Cheng et al., 1996). Psychopathology was further evaluated by a semi-structured interview using the PANSS-CH within 1 week after admission by attending psychiatrists who had completed the PANSS-CH reliability training. In an inter-rater reliability study, the coefficients of agreement (Kay, 1991) were satisfactory: 12 items were above 0.80, 17 items between 0.70 and 0.79, and the remaining four items between 0.66 and 0.69 (Cheng et al., 1996).
All subjects on admission of the MPGRP project have received psychiatrists’ clinical assessments with the PANSS. After their condition stabilized during the index hospital-ization, subjects were tested with the Continuous Performance Test (CPT; Rosvold et al., 1956). All subjects in each follow-up projects (MPSS and SEFOS) were assessed the PANSS ratings and the CPT performance. However, a part subjects didn’t complete the CPT at each projects.
4.2.3 Study Variables Demographic Variables
Demographic variables include variables of age, gender, years of education, marital status (single versus married), occupation (with versus without occupation), and age of onset of psychotic symptom. Note that the married marital status consists of people living
together and people getting married; housewives, students, people who never worked, who are unemployed or who already retired are included in people without occupation.
Environmental Factors
In this study, the environmental factors are related to obstetric complications, prena-tal growth retardation, special personal behavior, the psychological problem, and so on.
There are three environmental factors, described as follows separately.
(1) The patient has brain injury in the growth, such as prenatal growth retardation, brain damage, retarded intelligence and so on.
(2) Before getting disease, the patient had the unstable mood or abnormal behavior to interfere with adapting to the daily life, including angry, timid, depressed, inactive, having behavior problems, and so on.
(3) Before getting disease, the patient had the psychological problems to interfere with adapting to life in their infancy, including bad relation between parents, getting along badly with sibling or parents, getting disease about body, unforeseen happen-ings of family, and so on.
The first environmental factor was rated by a 3-point scale with 0 as no circumstance, 1 as slight (have not obviously heart body obstacle) and 2 as obvious (have obviously heart body obstacle). Due to the ratio of obvious subjects with the first environmental factor was too low, we combined the slight subjects with the obvious subjects in the first environmental factor. The others were rated by a 3-point scale with 0 as no circumstance, 1 as slight (have not obviously influenced routine life) and 2 as obvious (have obviously influenced routine life). There were one dummy variable for the first environmental factor, two dummy variables for the others.
Continuous Performance Task (CPT; Rosvold et al., 1956)
We used a CPT machine from Sunrise Systems, version 2.20 (Pembroke, MA, USA).
We used a CPT machine from Sunrise Systems, version 2.20 (Pembroke, MA, USA).