Schizophr Bull 1987 Kay 261-76-2
Schizophr Bull 1987 Kay 261-76-2
Schizophr Bull 1987 Kay 261-76-2
2,1967
Abstract
The variable results of positivenegative research with schizophrenics underscore the importance
of well-characterized, standardized
measurement techniques. We report
on the development and initial
standardization of the Positive and
Negative Syndrome Scale (PANSS)
for typological and dimensional assessment. Based on two established
psychiatric rating systems, the 30item PANSS was conceived as an
operationalized, drug-sensitive instrument that provides balanced
representation of positive and negative symptoms and gauges their relationship to one another and to
global psychopathology. It thus
constitutes four scales measuring
positive and negative syndromes,
their differential, and general severity of illness. Study of 101 schizophrenics found the four scales to be
normally distributed and supported
their reliability and stability. Positive and negative scores were inversely correlated once their
common association with general
psychopathology was extracted,
suggesting that they represent mutually exclusive constructs. Review
of five studies involving the
PANSS provided evidence of its criterion-related validity with antecedent, genealogical, and concurrent
measures, its predictive validity, its
drug sensitivity, and its utility for
both typological and dimensional
assessment.
Schizophrenia has long been regarded as a heterogeneous entity,
and over the decades researchers
have sought consistent subpattems
that might explain different aspects
of this complex disorder. Most recently, Crow (1980a, 1980b) and Andreasen (1982; Andreasen and Olsen
1982) have proposed that two dis-
SCHIZOPHRENIA BULLETIN
262
interview (infra) and were then assessed on the PANSS scales plus a
series of measures deriving from
clinical interview, cognitive testing,
motor assessment, and careful review of medical and historical records. These measures are described
in separate articles that chiefly address their relationship to positive
and negative syndromes (Kay,
Opler, and Fiszbein 1986; Opler,
Kay, and Fiszbein 1986).
The assessments were conducted
by two research psychiatrists, one of
whom collected data on 47 patients
and the other on 54. Both psychiatrists first underwent intensive training in the PANSS interview and
rating methods until satisfactory
team concordance was achieved,
and subsequently they rated patients individually. The raters held
no a priori assumptions about the
outcome of data and were unaware
of results on the PANSS, which was
undertaken before other measures
but scored only after the conclusion
of study.
The final sample consisted of 101
subjects of ages 20-68 (mean =
36.81, SD = 11.16), including 70
males, 31 females, 33 whites, 43
blacks, and 25 Hispanics. Twelve
patients were married, 10 divorced,
and the remainder single. Mean
education was 10.09 years (SD =
2.92), with the range extending to 4
years of college in four cases.
Twenty-nine subjects had a first-degree relative who was previously
hospitalized for psychiatric treatment; schizophrenia was specified
in five cases and affective disorder
(depressive, manic, or bipolar) in 10
cases; alcohol abuse was reported in
the nuclear family of 16 patients;
and among 13 subjects there was evidence of family sociopathy, as
judged by record of criminal behavior and prosecution.
On the average, patients were
Full text of the PANSS Rating Manual, which includes the interview procedure, item definitions, anchoring point
descriptions, and rating form, is available on request from the authors.
263
284
SCHIZOPHRENIA BULLETIN
that this scale is statistically or conceptually distinct from the positivenegative assessment (an issue which
also was to be determined by this
study), but only that it may be used
as a yardstick of collective nonspecific symptoms against which to
judge severity of distinct positive
and negative manifestations.
In addition to these three scales, a
bipolar Composite Scale was conceived to express the direction and
magnitude of difference between
positive and negative syndromes.
This score was considered to reflect
the degree of predominance of one
syndrome over the other, and its
valence (positive or negative) may
serve for typological characterization.
The PANSS is scored by summation of ratings across items, such
that the potential ranges are 7-49 for
the Positive and Negative Scales and
16-112 for the General Psychopathology Scale. The Composite
Scale is arrived at by subtracting the
negative from positive score, thus
yielding a bipolar index that ranges
from -42 to +42.
Results
Distribution of Scores. Table 1 summarizes the distribution characteristics of the four scales from the
Positive
Negative
Composite
Mean
Median
SD
Range (potential)
18.20
18
6.08
7 to 49
21.01
20
6.17
7 to 49
2.69
2
7.45
- 4 2 to +42
Range (obtained)
Skewness
Kurtosis
7 to 32
.07
-.97
8 to 38
.48
.06
- 2 5 to +13
.45
.13
General
psychopathology
37.74
36
9.49
16to112
19 to 63
.23
-.30
265
401>
o
z
LJJ
ai
OC
u.
30 --
20 -
Jr\ -
x x
10 -
ir i .
-25
-17
-9
-1
23
31
39
47
55
63
PANSS was examined using coefficient a to analyze its internal consistency and the contribution of the
component items. As detailed in
table 3, each of the items making up
the Positive and Negative Scales correlated very strongly with the scale
total (p < .001), and the mean itemtotal correlations of .62 and .70, respectively, far exceeded the crosscorrelations of .17 (Positive items
with Negative Scale) and .18 (Negative items with Positive Scale). The a
coefficients with single items removed ranged from .64 to .84, and
no perceptible gain on either scale
PANSS SCORE
PANSS, and the full spectrum of
scores is illustrated in figure 1. All
four measures exhibited a roughly
normal distribution pattern, without
substantial skewness or kurtosis.
This observation suggested that the
constructs in question represent typical continua and that their measurement is amenable to parametric
statistical treatment. The obtained
range of scores in all cases was considerably less than the potential
range, suggesting that the scales
were of ample breadth to avoid ceiling restrictions. The medians of the
Positive and Negative Scales were
strikingly close (18 and 20, respectively), and therefore the Composite
Scale, representing their differential,
exhibited a median of -2, which indicated an almost equal contribution
by positive and negative items.
On the basis of the normality of
distribution, it was possible to
convert raw scores for each of the
PANSS scales to percentile ranks
(table 2). This process enables provisional interpretation of individual
scores with reference to a medicated
chronic schizophrenic sample.
Internal Consistency and Test-Retest Reliability. The reliability of the
Positive
Negative
99.9
99
98
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
2
1
0.1
37
33
31
29
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
8
7
40
36
34
32
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
11
8
7
Composite
21
15
13
10
7
5
4
3
2
1
0
- 1
- 2
- 4
- 5
- 6
- 7
- 8
- 9
-11
-13
-15
-18
-20
-25
General
psychopathology
67
60
58
54
50
48
46
44
43
42
40
39
38
36
35
34
33
31
30
28
26
22
18
16
SCHIZOPHRENIA BULLETIN
266
a coefficient
with Item deleted
.78
.48
.66
.55
.64
.61
.59
(a = .73,
p<.001)
<.001
<.001
<.001
<.001
<.001
<.001
<.001
.64
.73
.70
.71
.73
.69
.70
.93
1.08
1.44
1.19
1.34
.45
.30
(5.17
.63
.78
.76
.79
.61
.86
.50
(a = .83,
p<.001)
<.001
<.001
<.OO1
<.001
<.001
<001
<.001
.81
.78
.79
.78
.82
.76
.84
.21
1 .20
1 .06
1 .19
1 .12
.97
1 .10
1 .21
1 .49
1 .14
1 .28
1 .31
1 .30
1 .31
1 .18
1 .18
9.49
.48
.60
.23
.70
.33
.24
.27
.51
.51
.42
.65
.35
.66
.66
.60
.43
(a = .79,
p<.001)
<.001
<.001
<.O2
<.001
<.001
<.O2
<.01
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
<.001
.77
.77
.79
.76
.79
.79
.79
.78
.78
.78
.76
.79
.76
.76
.76
.78
Mean
SD
Positive Scale
Delusions
Conceptual disorganization
Hallucinatory behavior
Excitement
Grandiosity
Suspiciousness
Hostility
Scale total
3.18
3.03
2.50
2.35
2.36
2.70
2.10
18.20
1.52
.42
.70
.24
.56
.24
.14
(5.08
Negative Scale
Blunted affect
Emotional withdrawal
Poor rapport
Passive-apathetic social withdrawal
Difficulty in abstract thinking
Lack of spontaneity & flow of conversation
Stereotyped thinking
Scale total
2.94
3.03
2.58
2.78
3.95
2.87
2.90
21.01
2.39
2.43
1.72
2.35
1.54
1.90
2.09
2.11
3.42
2.09
2.45
3.82
2.10
2.17
2.71
2.48
37.74
Item-total
correlation
(p < .001).
As expected, both scales correlated strongly with the Composite
Scale, and they yielded coefficients
VOL13.NO 2,1987
267
268
SCHIZOPHRENIA BULLETIN
Variable
Demographic/historical
Number of hospital admissions
Years of education
Male gender
Family history of illness
Sociopathy
Unspecified psychosis
Major affective disorder
Total psychiatric illness
Positive Negative
.20
-.29
.21
Clinical (BPRS)
Unusual thought content
Anxiety
Preoccupation
Disorientation
Motor retardation
Somatic concern
.33
.21
.29
-.21
.28
.20
Cognitive/psychometric
Egocentricity of Thought Test
(CDB)
Random number fluency
Color Form Preference Test
(CDB)
Affective (MARS)
Angry affective tone
Affective lability
Total affective impairment
Dull facial expression
Impoverished thought content
Global unrelatedness
Lack of vocal emphasis
Slow response latency
Global immobility
Lack of expressive gestures
Soft voice level
Poor eye contact
Increased noncommunicative
movements
General
Composite psychopathology
-.34
-.33
.24
.39
.27
.46
.31
.23
.64
.54
.52
.50
.49
.47
.43
.41
.32
-.41
-.40
-.49
-.42
-.40
-.36
-.43
-.37
-.30
.47
.24
.27
.30
.39
.46
.41
.73
.38
.28
.26
.50
.33
.22
.20
-.28
NoteBased on study of 101 chronic schizophrenics (Kay, Opter & Rszbein 1986). Shown are the
significant (p<.05) nonovertapping covartates of the Positive and Negative Scales and the correlates
of the Composite and General Psychopathology Scales, excluding those clinical items that enter Into
the latter scale. Abbreviations.COB - Cognitive Diagnostic Battery (Kay1982); MARS - Manifest
Affect Rating Scale (Alpert and Rush 1983); BPRS - Brief Psychiatric Rating Scale (Overall and
Gorham 1962); PANSS - Positive and Negative Syndrome Scale.
VOL13.NO. 2,1887
270
SCHIZOPHRENIA BULLETIN
Discussion
We have described the development
and initial standardization of the 30item PANSS as an instrument for
measuring the prevalence of positive
and negative syndromes in schizophrenia. A major impetus of its development was the need tor a psychometrically sound procedure to
serve typological and dimensional
assessment. Perhaps its most important contributions are the provision
of specified interview guidelines and
assessment criteria, and the inclusion of two additional scales that
consider positive-negative syndromes relative to one another and
relative to general severity of
psychopathology.
The PANSS method derives from
two established psychiatric rating
scales for which interrater agreement and treatment sensitivity have
been demonstrated. As such, it proceeds from reliable techniques that
are familiar to clinicians and researchers, requiring relatively little
additional training. For the purpose
of the PANSS, however, precise op-
271
272
SCHIZOPHRENIA BULLETIN
273
SCHIZOPHRENIA BULLETIN
275
Appendix
Mohan Singh, whose conceptualization of schizophrenic phenomena influenced the definitions of many
scale items.
The Authors
Stanley R. Kay, Ph.D., is Assistant
Clinical Professor, Department of
Psychiatry, Albert Einstein College
of Medicine/Montefiore Medical
Center, and Co-Director, Research
Unit, Bronx Psychiatric Center,
Bronx, NY. Abraham Fiszbein,
M.D., is Resident in Psychiatry,
PI. Delusions. Beliefs which are unfounded, unrealistic, and idiosyncratic. Basis for rating: thought
content expressed in the inteview
and its influence on behavior.
1. AbsentDefinition does not apply.
2. MinimalQuestionable pathology; may be at the upper extreme of
normal limits.
3. MildPresence of one or two delusions that are vague, uncrystallized, and not tenaciously
held. Delusions do not interfere
with thinking, social relations, or behavior.
4. ModeratePresence of either a kaleidoscopic array of poorly formed,
unstable delusions or of a few wellformed delusions that occasionally
SCHIZOPHRENIA BULLETIN