SEARCH

SEARCH BY CITATION

Keywords:

  • Clinical Global Impression;
  • Positive and Negative Syndrome Scale;
  • rating scale;
  • schizophrenia

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Aims:  The aim of the present study was to develop a subscale of the Positive and Negative Syndrome Scale (PANSS) that would be brief and sensitive to changes in the clinical features of schizophrenia (i.e. the Brief PANSS, or bPANSS).

Methods:  The PANSS before and after treatment, and the Clinical Global Impression–Change (CGI-C) was rated for 714 schizophrenia patients. Of these, Clinical Global Impression–Severity (CGI-S) was also evaluated in 30 of these patients. The bPANSS items were extracted from full PANSS items based on the following aims: (i) to develop a brief scale; (ii) to develop a scale sensitive to changes resulting from antipsychotic treatment; and (iii) to reflect the broad spectrum of schizophrenia symptoms.

Results:  The following six items were extracted to serve as the bPANSS: delusion, suspiciousness, emotional withdrawal, passive/apathetic social withdrawal, tension, and unusual thought content. The coefficients of correlation between the bPANSS and full PANSS before and after treatment were 0.86 and 0.92, respectively (both P < 0.001). The coefficient of correlation between the degrees of change in the scores for the bPANSS and the full PANSS was 0.93 (P < 0.001), and that between delta bPANSS and CGI-C was 0.73 (P < 0.001).

Conclusions:  bPANSS is able to capture the overall clinical features of schizophrenia within a short assessment period.

THE POSITIVE AND Negative Syndrome Scale (PANSS) was developed by Kay et al. for comprehensive assessment of the psychopathology of schizophrenia,1 and is one of the most widely used scales for schizophrenia in both psychopharmacological studies and clinical trials. This scale consists of a total of 30 items including seven positive syndrome items, seven negative syndrome items, and 16 comprehensive pathological items, and the severity of each item is rated on a scale from 1 (none) to 7 (most severe). It takes at least 30–40 min to assess the overall schizophrenic symptoms using the PANSS.2

The PANSS is thus an informative rating instrument for capturing the overall psychopathology of patients with schizophrenia. For repeated evaluation, however, it may be too burdensome for both patients and clinicians. In fact, among the 30 items of the PANSS, there are several symptomatic items that do not change even after pharmacological treatment, considering the natural history of schizophrenia. The aim of the present study was therefore to develop a brief version of the PANSS (bPANSS) to complement the full PANSS.

More specifically, in the present study, we tried to extract the least possible number of items from the original PANSS as long as, first, they retained sufficient internal consistency and reliability while reflecting the three domains of the original PANSS; second, they correlated with the overall severity of schizophrenia; and third, they were sensitive to changes in overall severity. We expected that such a scale would be useful and meaningful for the follow up of patients with schizophrenia in routine clinical practice. To our knowledge this study represents the first attempt to extract the core items from the total 30 items of the PANSS.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Ethics consideration

This study was carried out after the approval of the ethics committee of Kochi University and individual attending facilities, in line with the Helsinki Declaration by the World Medical Association and the clinical research guideline defined by the ethics committee at our facilities.

Subjects

A total of 714 patients (409 men, 305 women) were enrolled as subjects in the present study, and were diagnosed as suffering from schizophrenia based on the criteria of the DSM-IV-TR. These included 684 patients who participated in two multi-center clinical trials comparing a new atypical antipsychotic drug against a classical antipsychotic, and 30 patients who had been hospitalized and treated at Kochi Medical School Hospital. The average age of the patients was 43.8 ± 14.0 years.

Procedure

These schizophrenia patients were evaluated using the Japanese version of the PANSS,3,4 Clinical Global Impression–Severity (CGI-S) and CGI-Change (CGI-C).5 Six hundred and eighty-four patients participating in clinical trials completed the PANSS at study entry and at the end of the trial, while 30 patients completed the PANSS and the CGI-S upon admission and upon discharge from hospital. Moreover, the CGI-C was also implemented at the end of treatment for all patients. The CGI assesses the overall impression of the clinical status of a patient: for the CGI-S, the overall severity of illness was evaluated in terms of seven stages: 1, normal, not at all ill; 2, borderline mentally ill; 3, mildly ill; 4, moderately ill; 5, markedly ill; 6, severely ill; and 7, most extremely ill, while for the CGI-C the degree of change in amelioration was assessed as: 1, very much improved; 2, much improved; 3, minimally improved; 4, no change; 5, minimally worse; 6, much worse; and 7, very much worse.

Extraction of the items for bPANSS

To extract the items for the bPANSS from the total 30 items of the PANSS, the following criteria were adopted: (i) to develop a scale that would be sensitive to changes resulting from antipsychotic treatment, extracting items with higher Spearman's rank correlation coefficient with the CGI-C as priorities; (ii) to develop a brief scale that can be completed within 10 min (requiring that items be excluded for which the coefficients of correlation with the CGI-S were significantly low, subsequently, no more than eight items were to be extracted from the total 30 items of the PANSS); and (iii) to select at least two items individually from the three domains (positive symptoms, negative symptoms, and comprehensive pathological scales) of the full PANSS to sufficiently reflect the profiles of schizophrenia. The extraction of the bPANSS items was conducted based on the consensus of all authors.

Validation of the bPANSS

Internal consistency of the bPANSS was examined by calculating the Cronbach's alpha. Concurrent validity of the bPANSS was examined by comparing it with the full PANSS according to the whole subjects and the individual subtypes of schizophrenia.

Statistical analysis

Statistical analysis was performed using SPSS version 16.0J (SPSS, Tokyo, Japan). For each item of the PANSS, the Spearman's rank correlation coefficients were calculated between the total PANSS score and CGI-S at two points: at study entry and at the end of treatment. Rank correlation coefficients were also calculated between the change in each item of the PANSS between study entry and the end of treatment and the CGI-C.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient characteristics

Table 1 lists the baseline demographic and clinical characteristics of the patients. The average total PANSS score was 85.1 ± 19.2 at study entry and 74.8 ± 22.8 at the end of treatment. The average CGI-S score was 4.8 ± 1.0 at study entry and 4.0 ± 1.2 at the end of treatment. The average CGI-C score was 2.8 ± 1.4 at the end of treatment.

Table 1.  Baseline patient characteristics (mean ± SD)
  1. CGI-C, Clinical Global Impression–Change; CGI-S, Clinical Global Impression–Severity; PANSS, Positive and Negative Syndrome Scale.

Age (years)43.8 ± 14.0
SexMale 409, Female 305
Total PANSS score at study entry85.1 ± 19.2
Total PANSS score at end of treatment74.8 ± 22.8
CGI-S at study entry4.8 ± 1.0
CGI-S at end of treatment4.0 ± 1.2
CGI-C2.8 ± 1.4
Subtypes of schizophrenia 
 Paranoid type249
 Disorganized type178
 Catatonic type29
 Residual type170
 Unknown88

Item extraction

Table 2 lists Spearman's rank correlation coefficients between individual PANSS items and the CGI-S at study entry and at the end of treatment, as well as those between changes in each PANSS item score and the CGI-C. Based on the aforementioned selection criteria, a number of items were extracted from the 30 items of the PANSS.

Table 2.  Correlation coefficients
ItemWith CGI-S (at study entry)With CGI-S (at end of treatment)Changes with CGI-C
  • *

    P < 0.05;

  • **

    P < 0.01.

  • Bold, extracted item.

  • CGI-C, Clinical Global Impression–Change; CGI-S, Clinical Global Impression–Severity.

Positive Syndrome Scale   
P1: Delusion0.290.59**0.53**
P2: Conceptual disorganization0.51**0.70**0.50**
P3: Hallucinatory behavior0.69**0.68**0.48**
P4: Excitement0.62**0.75**0.45**
P5: Grandiosity0.60**0.36*0.24**
P6: Suspiciousness0.40*0.58**0.55**
P7: Hostility0.70**0.55**0.41**
Negative Syndrome Scale   
N1: Blunted affect0.270.65**0.37**
N2: Emotional withdrawal0.290.57**0.46**
N3: Poor rapport0.43*0.71**0.40**
N4: Passive/apathetic social withdrawal0.38*0.56**0.46**
N5: Difficulty in abstract thinking0.180.56**0.34**
N6: Lack of spontaneity and flow of conversation0.48**0.41*0.35**
N7: Stereotyped thinking0.050.320.40**
General Pathological Syndrome Scale   
G1: Somatic concern0.090.110.37**
G2: Anxiety0.130.270.55*
G3: Guilt feeling−0.030.310.32**
G4: Tension0.310.230.56**
G5: Mannerisms and posturing0.360.53**0.43**
G6: Depression−0.020.170.39**
G7: Motor retardation0.250.41*0.37**
G8: Uncooperativeness0.360.62**0.48**
G9: Unusual thought content0.260.59**0.55**
G10: Disorientation0.49**0.38*0.20**
G11: Poor attention0.30.50**0.40**
G12: Lack of judgment and insight0.52**0.53**0.48**
G13: Disturbance of volition0.51**0.48**0.38**
G14: Poor impulse control0.37*0.41*0.40**
G15: Preoccupation0.270.49*0.40**
G16: Active social avoidance0.45*0.56**0.42**

From among the items related to the positive syndrome scale, two items, P1 delusion, and P6 suspiciousness, were selected. These were the top two items with the highest Spearman's rank correlation coefficients relative to the CGI-C, and were considered to be appropriate because their coefficients of rank correlation between the total PANSS score and the CGI-S were maintained to a certain degree. From among the items related to the negative syndrome scale, N2 emotional withdrawal, and N4 passive/apathetic social withdrawal were selected. These were also the top two items having the highest Spearman's rank correlation coefficient relative to the CGI-C, and their coefficients of rank correlation with the CGI-S were also satisfactory. From the comprehensive pathological scale, G4 tension, and G9 unusual thought content, were selected, for reasons similar to those for selection of items from among the positive syndrome scale and the negative syndrome scale. The Spearman's rank correlation coefficient with the CGI-C for G2 anxiety was as high as those for G4 and G9, but the rank correlation coefficient with the CGI-S was only 0.27 at the end of treatment and 0.13 at study entry. Therefore, G2 was not adopted. Based on these considerations, a total of six items (P1, P6, N2, N4, G4, and G9) were therefore extracted as items for the bPANSS.

Internal consistency of the brief PANSS

The Cronbach's alpha for the bPANSS was 0.67 at study entry and 0.80 at the end of treatment. If we excluded any one of the six items, the Cronbach's alpha was always lower (Table 3).

Table 3.  Internal consistency of the brief PANSS (Cronbach's alpha)
ItemsStudy entryEnd of treatment
  1. PANSS, Positive and Negative Syndrome Scale.

P1, P6, N2, N4, G4, G90.670.80
P6, N2, N4, G4, G90.590.76
P1, N2, N4, G4, G90.590.76
P1, P6, N2, G4, G90.650.78
P1, P6, N4, G4, G90.660.78
P1, P6, N2, N4, G90.630.79
P1, P6, N2, N4, G40.610.76

Concurrent validity of the brief PANSS

The coefficient of correlation between the bPANSS and the full PANSS was 0.86 at study entry and 0.92 at the end of treatment (both P < 0.001), and that between the bPANSS and the CGI-S was 0.64 at study entry and 0.84 at the end of treatment (both P < 0.001). Similarly, the coefficient of correlation between the degree of change in the bPANSS score between study entry and the end of treatment (delta bPANSS) and that of the full PANSS score (delta total PANSS) was 0.93 (P < 0.001), while delta bPANSS and CGI-C was 0.73 (P < 0.001).

Concurrent validity by schizophrenia subtype

A total of 626 of the 714 subjects for whom the subtypes of schizophrenia had been diagnosed were analyzed. These subjects consisted of 249 with the paranoid type, 178 with the disorganized type, 29 with the catatonic type, and 170 with the residual type. The coefficients of correlation between the bPANSS and the full PANSS, delta bPANSS and delta total PANSS, and delta bPANSS and CGI-C versus subtypes are shown in Table 4. All of these correlation coefficients had satisfactorily high values and were highly significant (P < 0.01).

Table 4.  Concurrent validity vs schizophrenia subtype
SubtypeSubjects (n)Correlation coefficient
bPANSS and total PANSS at study entrybPANSS and total PANSS at end of treatmentdelta bPANSS and delta total PANSSdelta bPANSS and CGI-C
  • **

    P < 0.01.

  • delta bPANSS, degree of change in the brief PANSS score between study entry and end of treatment; PANSS, Positive and Negative Syndrome Scale.

Paranoid2490.86**0.93**0.93**0.78**
Disorganized1780.88**0.93**0.92**0.70**
Catatonic290.94**0.96**0.92**0.72**
Residual1700.86**0.90**0.93**0.64**

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The bPANSS, consisting of six items selected from the 30 items of the full PANSS, showed a high correlation with the full PANSS score, being 0.86 at study entry and 0.92 at the end of treatment. Also, the correlation between delta bPANSS and delta total PANSS was 0.93 (P < 0.001), while that between delta bPANSS and CGI-C was 0.73 (P < 0.001). Furthermore, the correlation between bPANSS and CGI-S was 0.64 at study entry and 0.84 at the end of treatment (both P < 0.001). Considering that it takes a short time to complete the six items, it is believed that the requirements of the bPANSS, which is sensitive to changes in psychotic manifestations and also reflects the profile of schizophrenia to a certain extent, have been satisfied.

Attempts to select several items from among the existing clinical rating scales for certain purposes have been reported for a number of scales. For example, Montgomery and Asberg developed the Montgomery and Asberg Depression Rating Scale,6 which is a representative rating scale for depression consisting of 10 items selected from the Comprehensive Psychopathologic Rating Scale (CPRS), which contains 67 items in all: 40 items for subjective psychopathological symptoms, 25 items for objective psychopathological symptoms, and two items for global rating and assumed reliability.7 For schizophrenia they also selected 12 items that were the most sensitive to changes resulting from treatment, and they subsequently proposed an acute schizophrenia rating scale (brief CPRS) that was more sensitive to therapeutic response than the Brief Psychiatric Rating Scale.8

The six items in the bPANSS and the 12 items in the brief CPRS are shown in Table 5. Ten of the 12 items are related to subjective pathological symptoms in the brief CPRS. This appears to be because the items that respond readily to treatment, such as hallucination and delusion, tend to be measured mainly as subjective pathological symptoms, and the balance of the number of items representing subjective pathological symptoms and objective pathological symptoms is not taken into consideration. In contrast, the items in the bPANSS have been selected from the three psychopathology domains of the original PANSS in a balanced manner. Considering that the correlations of the two selected items within the individual three domains are relatively high (r = 0.517–0.685), it may be possible to simply select three items instead of six.

Table 5.  Brief PANSS vs brief CPRS
Brief PANSSBrief CPRS
  1. CPRS, Comprehensive Psychopathological Rating Scale; PANSS, Positive and Negative Syndrome Scale.

Positive Syndrome ScaleSubjective items
P1. Delusion1. Sadness
P6. Suspiciousness5. Inability to feel
 6. Pessimistic thoughts
Negative syndrome scale28. Depersonalization
N2. Emotional withdrawal29. Feeling controlled
N4. Passive/apathetic social withdrawal30. Disrupted thoughts
 31. Ideas of persecution
General Pathological Syndrome Scale33. Delusional mood
G4. Tension36. Other delusions
G9. Unusual thought content37. Commenting voices
 Objective items
 45. Lack of appropriate emotion
 50. Perplexity

Andreasen et al. extracted eight of the 30 items of the whole PANSS in order to define remission from schizophrenia.9 A total of three items, delusion (P1), passive/apathetic social withdrawal (N4), and unusual thought (G9), overlapped with those currently selected for the bPANSS. These three items are considered to be essentially important for evaluating both remission from schizophrenia and the response to antipsychotic drugs.

Possible limitations of the present study may include the following. First, most of the study subjects were patients enrolled in two multi-center active-drug-controlled clinical trials of a new atypical antipsychotic drug, for whom CGI-S was not evaluated. In other words, we placed more emphasis on the bPANSS to reflect longitudinal changes in psychopathology, rather than reflecting the cross-sectional severity of schizophrenia. Second, all the subjects were receiving active treatment, and therefore changes due to treatment were unable to be differentiated from those due to the natural course of the disorder.

In contrast, the strengths of the bPANSS may be summarized as follows. First, it is brief and therefore exceedingly less burdensome for both patients and clinicians than the full PANSS. Second, it still covers the three broad domains of psychopathology represented by the original scale in a balanced manner. Third, it has high internal consistency reliability. Fourth, the bPANSS shows satisfactory concurrent validity with the full PANSS and the CGI-S for any of the schizophrenia subtypes. Fifth, the bPANSS is sensitive to changes resulting from treatment.

We therefore conclude that the bPANSS is a useful rating instrument that is easy to administer within a short time and is sensitive to changes in the overall clinical features of schizophrenia. One recommended practical application of the scales would be to administer the full PANSS at baseline and then at extended intervals to gain a comprehensive picture of the psychopathology, while in the meantime repeatedly monitoring the changes in the patients using the bPANSS.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Kay SR, Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 1987; 13: 261276.
  • 2
    Kay SR. Positive and Negative Syndromes in Schizophrenia: Assessment and Research (Clinical and Experimental Psychiatry). Brunner-Routledge, Hove, 1991.
  • 3
    Kay SR, Opler LA, Fiszbein A. Positive and Negative Syndrome Scale (PANSS) Rating Manual. Multi-Health System, Toronto, 1991 (Translated into Japanese by Yamada H, Masui K, Kikumoto K. Seiwa Shoten Publishers, Tokyo, 1991).
  • 4
    Igarashi Y, Hayashi N, Yamashina M et al. Interrater reliability of the Japanese version of the Positive and Negative Syndrome Scale and the appraisal of its training effect. Psychiatry Clin. Neurosci. 1998; 52: 467470.
  • 5
    Guy W. Clinical global impression (CGI). In: NIMH Psychopharmacology Research Branch (ed.). ECDEU Assessment Manual for Psychopharmacology. US Department of Health and Human Services, Public Health Service, Alcohol Drug Abuse and Mental Health Administration, National Institute of Mental Health, Rockville, MD, 1976; 218222.
  • 6
    Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br. J. Psychiatry 1979; 134: 382389.
  • 7
    Asberg M, Montgomery SA, Periss C, Schalling D, Sedvall G. A comprehensive psychopathological rating scale. Acta Psychiatr. Scand. 1978; 271: 527.
  • 8
    Montgomery SA, Taylor P, Montgomery D. Development of a schizophrenia scale sensitive to change. Neuropharmacology 1978; 17: 10611063.
  • 9
    Andreasen NC, Carpenter WT Jr, Kane JM, Lasser RA, Marder SR, Weinberger DR. Remission in schizophrenia: Proposed criteria and rationale for consensus. Am. J. Psychiatry 2005; 162: 441449.