Ž .Psychiatry Research 97 2000 129]135

Symptom dimensions in recent-onset schizophrenia andmania: a principal components analysis of the 24-item

Brief Psychiatric Rating Scale

Joseph Venturaa,U, Keith H. Nuechterleina,b, Kenneth L. Subotnika,Daniel Gutkinda, Elizabeth A. Gilbert b

aUCLA Department of Psychiatry and Biobeha¨ioral Sciences, Los Angeles, CA 90095-6968, USAbUCLA Department of Psychology, Los Angeles, CA 90095-6968, USA

Received 25 January 1999; received in revised form 21 September 2000; accepted 4 October 2000

Abstract

Ž .Previous four- and five-factor solutions of the 18-item Brief Psychiatric Rating Scale BPRS suggested thepossibility of an affective dimension in psychosis. A principal components analysis was used to analyze psychiatricsymptom data rated on an expanded 24-item version of the BPRS. BPRS data were collected during a period of acutepsychotic and affective illness with 114 young adult, recent-onset schizophrenia and schizoaffective patients and 27bipolar manic patients. Principal components analyses of the 18-item and 24-item BPRS indicated a four-factorsolution was the most interpretable. Principal components analysis of the 24-item BPRS produced a clear maniafactor characterized by high loadings from items added to the 18-item BPRS, which included elevated mood, motorhyperactivity, and distractibility. This factor solution suggests that the 24-item BPRS allows for an expandedassessment of affective symptoms relating to a manic dimension. Potentially important symptoms that were added tothe traditional 18-item version, namely suicidality, bizarre behavior, and self-neglect, also make clear contributions toother factors. Q 2000 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Factor analysis; Schizoaffective disorder; Bipolar disorder; Brief Psychiatric Rating Scale; Schizophrenia

U Corresponding author. UCLA Department of Psychiatry, 300 Medical Plaza, Room 2243, Los Angeles, CA 90095-6968, USA.Tel.: q1-310-206-5225.

Ž .E-mail address: jventura@ucla.edu J. Ventura .

0165-1781r00r$ - see front matter Q 2000 Elsevier Science Ireland Ltd. All rights reserved.Ž .PII: S 0 1 6 5 - 1 7 8 1 0 0 0 0 2 2 8 - 6

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135130

1. Introduction

ŽThe Brief Psychiatric Rating Scale BPRS;.Overall and Gorham, 1962 was initially devel-

oped as a rapid method to assess symptom changein psychiatric inpatients of various diagnoses andhas become one of the most widely used semi-structured instruments in psychiatric research.The original 16-item BPRS was expanded to 18

Ž .items Overall and Klett, 1972 . The 18-item ver-sion was expanded to 24 items through the addi-tion of six symptom items } bizarre behavior,self-neglect, suicidality, elevated mood, dis-tractibility, and motor hyperactivity } to increasesensitivity to a broader range of psychotic and

Ž .affective symptoms Lukoff et al., 1986 . The lat-est version of the expanded 24-item BPRS pro-vides detailed anchor points and probe questions

Ž .for each item Ventura et al., 1993 . Studies haveshown the expanded 24-item BPRS to be a sensi-tive and effective measure of psychiatric symp-toms with good interrater reliability that can be

Žmaintained over time Ventura Green et al., 1993;.Roncone et al., 1999 .

Psychiatric symptom data gathered with theBPRS have been factor analyzed to identify symp-tom dimensions in psychiatric disorders. How-ever, most of the factor analyses have used the16- or 18-item BPRS and were performed onschizophrenia patients who were chronically illŽGuy, 1976; Overall and Beller, 1984; Malla et al.,1993; Mueser et al., 1997; Long and Brekke,

.1999 . To our knowledge, there are only threepublished factor analyses of the 24-item BPRSŽvan der Does et al., 1993; Dingemans et al., 1995;

. Ž .Burger et al., 1997 . Van der Does et al. 1993reported a four-factor solution with a sample ofrecent-onset schizophrenia patients and Dinge-

Ž .mans et al. 1995 a five-factor solution with adiverse sample of psychiatric inpatients. The itemsof Motor Hyperactivity, Distractibility, and Ele-vated Mood, which were added to the 18-itemversion, help to form for van der Does et al. a‘disorganization’ factor and for Dingemans et al.

Ž .a ‘mania’ factor. Burger et al. 1997 reported afive-factor solution for the 24-item BPRS with asample of homeless, chronically mentally illpatients, including factors for positive symptoms,

negative symptoms, anxious]depressive symp-toms, hostile]suspiciousness, and ‘activity.’ Thusfar, it appears that the composition of factors forthe 24-item BPRS varies across samples. Further-more, it is unclear whether a meaningful andreplicable fifth factor is present.

In an attempt to replicate previous research,the present study examined the factor structureof symptom ratings on the expanded 24-itemBPRS in a sample of recent-onset schizophreniapatients and bipolar manic patients, most of whomalso represented recent-onset disorders. Thepatients were participating in studies at the re-search center that developed the 24-item versionof the BPRS.

2. Method

2.1. Subjects

The sample comprised 141 psychotic patients,Ž .101 72% of whom were male with a mean age at

Ž .study entry of 23.7 years S.D.s4.8 and meanŽ .education of 12.9 years S.D.s 1.9 . The

ethnicrracial distribution was 70% Caucasian, 8%Latino, 8% African American, 6% Asian, and 6%Other. On the Hollingshead Two-Factor ScaleŽ .Hollingshead, 1957 of social class, these patientsrepresented the full range with a mean in the

Ž .middle class Means3.05, S.D.s1.1 . By DSM-Ž .III-R criteria 1987 , entry diagnoses were

Ž .schizophrenia ns90 , schizoaffective disorderŽ . Ž .ns11 , schizophreniform disorder ns13 or

Ž .bipolar disorder ns27 .The patients with schizophrenia and schizoaf-

fective disorder were recruited for the ‘Develop-mental Processes in the Early Course of Illness’study at the UCLA Department of Psychiatry,part of an ongoing series of longitudinal studiesof the early phases of schizophrenia conducted by

Ž .this research group Nuechterlein et al., 1992 .All of the schizophrenia patients were required tohave a recent onset of psychotic illness, within 2years of study entry. The bipolar patients, who onaverage became ill within 2 years of study entry,were recruited as a comparison sample for the

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135 131

schizophrenia patients. Written informed consentwas obtained from all study participants.

2.2. Procedure

Patients were recruited from successive admis-sions to four large public hospitals in Los AngelesCounty and from referrals to the UCLA Neu-ropsychiatric Institute and Hospital outpatientprogram. DSM-III-R diagnoses were made usingan expanded version of the Present State Exami-

Ž .nation Wing et al., 1974 that included itemsfrom the Schedule for Affective Disorders and

Ž .Schizophrenia Endicott and Spitzer, 1978 . The24-item expanded version of the Brief Psychiatric

ŽRating Scale Lukoff et al., 1986; Ventura Lukoff.et al., 1993 was administered during an acute

psychotic or manic period, typically within 1 weekafter a psychiatric hospitalization. BPRS ratingsreflected a 2-week period of psychiatric symp-toms.

All diagnostic interviewers and BPRS ratersreceived comprehensive training and participated

in an ongoing quality assurance program to re-Žduce rater drift Ventura Green et al., 1993;.Ventura et al., 1998 . Diagnostic interviewers were

trained using criterion ratings and were requiredto achieve a minimum sensitivity of 0.85 for symp-tom presence and minimum specificity of 0.90 forsymptom absence. The BPRS training programparticipants were required to rate taped BPRSinterviews and to conduct live BPRS interviews.All BPRS raters achieved a median intraclasscorrelation coefficient of 0.80 or higher across allBPRS items compared with the criterion ratings.

2.3. Principal components analysis

Exploratory principal components analysis andvarimax rotation were completed using all 14124-item BPRS symptom assessments. To compareour findings with previous work, separate princi-pal components analyses were conducted for the18-item and 24-item versions of the BPRS. Weextracted a four- and a five-component solution

Table 1aŽ .Varimax-rotated factor loadings of the 18-item BPRS for early course schizophrenia and bipolar patients ns141

Factor Negative Depression- Hostile] Positivesymptoms anxiety uncooperativeness symptoms

Blunted affect 0.83 0.09 0.16 0.13Motor retardation 0.74 0.12 y0.26 0.13Emotional withdrawal 0.70 y0.08 0.15 0.12Depression y0.00 0.69 y0.25 y0.01Guilt y0.12 0.60 y0.03 y0.02Anxiety 0.24 0.60 0.12 0.03Uncooperativeness 0.32 y0.02 0.55 y0.02Tension 0.07 0.29 0.54 0.03Hostility y0.06 y0.01 0.52 0.07Excitement y0.32 y0.09 0.51 y0.17Unusual thought content 0.19 0.26 0.13 0.63Disorientation 0.21 y0.08 y0.03 0.59Hallucinations y0.10 0.09 y0.14 0.49Conceptual disorganization 0.21 y0.11 0.17 0.46Suspiciousness 0.22 0.32 0.14 0.39Somatic concern y0.01 0.30 0.08 0.11Grandiosity y0.18 y0.25 y0.33 0.21Mannerisms and posturing 0.29 0.02 0.19 0.13

Eigenvalue 2.31 1.67 1.56 1.51Variance explained 29% 21% 19% 19%Cumulative 29% 49% 68% 87%

a Factor loadings )0.40 are highlighted in bold-faced type.

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135132

with varimax rotations. To interpret the factors,we focused on BPRS items with factor loadings

Ž0.40 or greater. To derive simple factor or clus-.ter scores from this principal components solu-

tion, we used a unit-weighting procedure thatsummed BPRS items with loadings G0.40.

3. Results

3.1. Principal components analysis of the 18-itemBPRS

First, we conducted a principal componentsanalysis of the 18-item version of the BPRS.Using a principal components analysis, we ex-

tracted and examined the four- and five-factorsolutions. The five-factor solution resulted in afifth factor with loadings G0.40 from only two

Ž .BPRS items, uncooperativeness 0.67 and emo-Ž .tional withdrawal 0.68 . This fifth factor ac-

counted for only 12% of the variance and was notconceptually coherent. The post-rotated eigenval-ues for the first four factors were all above 1.0and were less than 1.0 for the fifth factor andremaining factors. Therefore, the four-factor so-lution for the 18-item version was considered

Ž .most interpretable see Table 1 . Our four-factorsolution of the 18-item version of the BPRS isvery similar to those found in previous four-factorsolutions. The Hostility]Uncooperativeness fac-tor had high factor loadings from two affective

Table 2aŽ .Varimax-rotated factor loadings of the expanded 24-item BPRS for early course schizophrenia and bipolar patients ns141

Factor Manic- Negative Positive Depression-excitement Symptoms symptoms anxiety

Motor hyperactivity 0.87 y0.16 y0.09 0.01Elevated mood 0.77 y0.15 y0.12 y0.07Excitement 0.77 y0.11 y0.20 y0.03Distractibility 0.69 y0.01 0.01 0.09Hostility 0.45 0.10 0.14 y0.03Grandiosity 0.41 y0.08 0.24 y0.21Blunted affect y0.31 0.77 0.05 0.13Motor retardation y0.28 0.70 0.05 0.14Emotional withdrawal y0.13 0.66 0.19 y0.05Self-neglect 0.08 0.41 0.34 0.05Bizarre behavior 0.13 0.28 0.67 y0.11Unusual thought content y0.00 0.21 0.65 0.23Disorientation y0.06 0.22 0.52 y0.04Hallucinations y0.15 y0.14 0.51 0.08Suspiciousness y0.00 0.20 0.48 0.30Depression y0.21 y0.08 y0.06 0.75Anxiety y0.03 0.23 0.06 0.62Suicidality y0.01 y0.02 0.10 0.61Guilt y0.03 y0.13 0.00 0.58Conceptual disorganization 0.11 0.30 0.34 y0.07Tension 0.35 0.22 0.05 0.25Uncooperativeness 0.24 0.38 0.09 y0.01Somatic concern 0.12 0.05 0.03 0.29Mannerisms and posturing 0.04 0.37 0.15 y0.04

Eigenvalue 3.28 2.48 2.09 2.07Variance explained 27% 21% 17% 17%Cumulative 27% 48% 64% 82%

a Italicized bold-faced type highlights the additional items from the 24-item BPRS not found on the 18-item BPRS.

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135 133

Ž . Ž .symptoms, hostility 0.52 and excitement 0.51 ,suggesting a tendency toward a manic factor thatwas not well delineated by available BPRS symp-tom items.

3.2. Principal components analysis of the 24-itemBPRS

A principal components analysis with BPRSratings from all 24 items was completed with thesame sample. A four-factor solution was con-sidered the most interpretable in the five-factorsolution; items that loaded on a fifth factor didnot create conceptually meaningful clusters ofsymptoms in that the eigenvalue for the fifthfactor was less than 1.0, and the hostility itemloaded highly on two primary factors.

We found that the addition of elevated mood,motor hyperactivity, and distractibility resulted in

Ž .the formation of a mania factor see Table 2 .Beyond the clear delineation of a mania factor,the other three factors of the 24-item BPRS wereconsistent with the previous four-factor solutions

Žof the 18-item BPRS Guy, 1976; Overall et al.,.1967; Overall and Beller, 1984 . However, the

new item, bizarre behavior, was the most highlyloaded item on the positive symptom factor. Ex-cept for the addition of the new item self-neglect,the negative symptom factor was also consistentwith the 18-item principal components analysis.We did not find any evidence for a disorganiza-tion factor in this combined sample of acutelypsychotic schizophrenia and manic patients or in

Žconsidering the schizophrenia patients alone ns.114 .

4. Discussion

In the present study a mania factor was formed,indicating a clear advantage for the 24-item BPRSas compared to the 18-item BPRS and the Posi-

Ž .tive and Negative Syndrome Scale PANSS . Itemsthat had traditionally loaded on the Hostile]Sus-piciousness factor on the 18-item BPRS loadedon a mania factor in the 24-item BPRS. Thissuggests that affective symptoms which appearedpartially as hostility and uncooperativeness may

actually be associated with a mania dimension.Furthermore, high loadings of elevated mood withother manic symptoms for the 24-item BPRS sug-gest we found a better defined mania factor than

Ž .reported by Burger et al. 1997 and DingemansŽ .et al. 1995 . The 24-item BPRS may also have an

important advantage over the PANSS. Althoughfactor analyses of the PANSS have consistently

Žproduced an excitement or activation factor Lin-.denmayer et al., 1995 , the component items lack

core features of mania found in the 24-item BPRSsuch as elevated mood, motor hyperactivity, dis-tractibility and grandiosity.

In a factor analysis of the 24-item BPRS with arecent-onset sample of schizophrenia patients, van

Ž .der Does et al. 1993 reported a disorganizationfactor. However, a close examination of their‘disorganization’ factor suggests that it would bemore appropriately considered to be a variant ofa mania factor. Several of the items loading highly

Žon their ‘disorganization’ factor excitement, 0.79;motor hyperactivity, 0.76; distractibility, 0.74;

.hostility, 0.65; and elevated mood, 0.65 are clearlypart of our mania factor and do not reflect disor-ganization in the sense used by Andreasen et al.Ž . Ž . Ž .1995 , Arndt et al. 1995 and Liddle 1987 .

Perhaps we did not find evidence of a cleardisorganization factor because our sample con-sisted mainly of first episode psychotic patients.We believed that the addition of an item whichassesses bizarre behavior might help to form adisorganization factor. However, most of our re-cent-onset schizophrenia patients were rated verylow on the BPRS symptoms that are conceptuallyrelated to disorganization, e.g. mannerisms andposturing, and conceptual disorganization. There-fore, the restricted variance may have preventedthe emergence of a disorganization factor.

As shown in previous studies of the 16- and18-item BPRS, negative symptoms in our recent-onset sample of schizophrenia and manic patientsformed a factor separate from depressive symp-

Žtoms Overall et al., 1967; Overall and Beller,.1984 . In fact, all versions of the BPRS have been

reliably used to differentiate negative symptomsfrom a depression factor. The formation of thesetwo separate symptom clusters left open the ques-

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135134

tion of whether self-neglect was associated withdepression or the negative symptoms of psychoticillness. The loading of self-neglect in our sampleon the negative symptom factor is consistent withan amotivational syndrome and is in agreementwith Burger’s results, but not with Dingemans’finding that self-neglect loaded on a depressionfactor.

Our four-factor solution of the 24-item BPRSdiffers from the five-factor solutions of the 24-item

Ž .BPRS from Burger et al. 1997 and DingemansŽ .et al. 1995 in that a four-factor solution, pro-

vided the most interpretable result. However,there was general symptom dimension consis-

Žtency in all three factor analyses Burger et al.,1997; Dingemans et al., 1995; and the current

.study for factors composed of positive symptoms,negative symptoms, depression]anxiety and ma-nia. A fifth factor has not been consistently iden-tified across studies.

Acknowledgements

The authors wish to thank George Bartzokis,M.D., Craig Childress, M.A., Rosemary Collier,M.S., Rhonda Daily, B.A., David Fogelson, M.D.,Sally Friedlob, M.S.W., Debbie Gioia-Hasick,M.S.W., Michael Gitlin, M.D., Sandy Rappe,M.S.W., Margie Stratton, M.A., and the patientsof the Aftercare Research Program for their con-tributions to this project. This research was sup-

Žported by Research Grants MH37705 P.I.: Keith. ŽH. Nuechterlein, Ph.D. and MH30911 P.I.:

.Robert P. Liberman, M.D. from the NationalInstitute of Mental Health.

References

Andreasen, N.C., Arndt, S., Alliger, R., Miller, D., Flaum, M.,1995. Symptoms of schizophrenia. Archives of GeneralPsychiatry 52, 341]351.

Arndt, S., Andreasen, N.C., Flaum, M., Miller, D., Nepalais,´P., 1995. A longitudinal study of symptom dimensions inschizophrenia. Archives of General Psychiatry 52, 352]360.

Burger, G.K., Calsyn, R.J., Morse, G.A., Klinkenberg, W.D.,Trusty, M.L., 1997. Factor structure of the Expanded Brief

Psychiatric Rating Scale. Journal of Clinical Psychology 53,451]454.

Dingemans, P.M.A.J., Linszen, D.H., Lenior, M.E., Smeets,R.M.W., 1995. Component structure of the expanded Brief

Ž .Psychiatric Rating Scale BPRS-E . Psychopharmacology122, 263]267.

Endicott, J., Spitzer, R.L., 1978. A diagnostic interview: theSchedule for Affective Disorders and Schizophrenia.Archives of General Psychiatry 35, 837]844.

Guy, W., 1976. ECDEU Assessment Manual for Psychophar-macology. U.S. Department of Health, Education and Wel-fare, Rockville, MD.

Hollingshead A.B., 1957. Two-factor index of social position.Available from Augast B. Hollingshead. 1965 Yale Station,New Haven, Connecticut.

Liddle, P.F., 1987. The symptoms of chronic schizophrenia: are-examination of the positive]negative symptom dimen-sion. British Journal of Psychiatry 151, 145]151.

Lindenmayer, J.P., Bernstein-Hyman, R., Grochowski, S., Bark,N., 1995. Psychopathology of schizophrenia: initial valida-tion of a 5-factor model. Psychopathology 28, 22]31.

Long, J.D., Brekke, J.S., 1999. Longitudinal factor structure ofthe Brief Psychiatric Rating Scale in schizophrenia. Psycho-logical Assessment 11, 498]506.

Lukoff, D., Nuechterlein, K.H., Ventura, J., 1986. Manual forthe Expanded Brief Psychiatric Rating Scale. SchizophreniaBulletin 12, 594]602.

Malla, A.K., Norman, R.M.G., Williamson, P., Cortese, L.,Diaz, F., 1993. Three syndrome concept of schizophrenia: afactor analytic study. Schizophrenia Research 10, 143]150.

Mueser, K.T., Curran, P.J., McHugo, G.J., 1997. Factor struc-ture of the Brief Psychiatric Rating Scale in schizophrenia.Psychological Assessment 9, 196]204.

Nuechterlein, K.H., Dawson, M.E., Gitlin, M., Ventura, J.,Goldstein, M.J., Snyder, K.S., Yee, C.M., Mintz, J., 1992.Developmental processes in schizophrenic disorders: longi-tudinal studies of vulnerability and stress. SchizophreniaBulletin 18, 387]425.

Overall, J.E., Beller, S.A., 1984. The Brief Psychiatric RatingŽ .Scale BPRS in geropsychiatric research: I. factor struc-

ture on an inpatient unit. Journal of Gerontology 2,187]193.

Overall, J.E., Gorham, D.R., 1962. The Brief Psychiatric Rat-ing Scale. Psychological Reports 10, 799]812.

Overall, J.E., Klett, C.J., 1972. Applied Multivariate Analysis.McGraw Hill, New York.

Overall, J.E., Hollister, L.E., Pichot, P., 1967. Major psychi-atric disorders: a four-dimensional model. Archives of Gen-eral Psychiatry 16, 146]151.

Roncone, R., Ventura, J., Impallomeni, M., Falloon, I.R.H.,Morosini, P., Chiaravalle, E., Casacchia, M., 1999. Reliabil-ity of an Italian standardized and expanded Brief Psychi-

Ž .atric Rating Scale BPRS 4.0 in raters with high vs. lowclinical experience. Acta Psychiatrica Scandinavica 100,229]236.

van der Does, A.J.W., Linszen, D.H., Dingemans, P.M., Nugter,M.A., Scholte, W.F., 1993. A dimensional and categorical

( )J. Ventura et al. r Psychiatry Research 97 2000 129]135 135

approach to the symptomatology of recent-onsetschizophrenia. Journal of Nervous and Mental Disease 181,744]749.

Ventura, J., Green, M., Shaner, A., Liberman, R.P., 1993.Training and quality assurance on the BPRS: ‘The DriftBusters’. International Journal of Methods in PsychiatricResearch 3, 221]244.

Ventura, J., Lukoff, D., Nuechterlein, K.H., Liberman, R.P.,Green, M., Shaner, A., 1993. Appendix 1: Brief Psychiatric

Ž . Ž .Rating Scale BPRS Expanded Version 4.0 scales, anchor

points and administration manual. International Journal ofMethods in Psychiatric Research 3, 227]243.

Ventura, J., Liberman, R.P., Green, M.F., Shaner, A., Mintz,J., 1998. Training and quality assurance with the Structured

Ž .Clinical Interview for DSM-IV IrP SCID . Psychiatry Re-search 79, 163]173.

Wing, J.K., Cooper, J.E., Sartorius, N., 1974. The Measure-ment and Classification of Psychiatric Symptoms: An In-struction Manual for the PSE and Catego Program. Cam-bridge University Press, London.