Schi:ophreniu Research, 7 (1992) 225-231 #Q 1992 Elsevier Science Publishers B.V. All rights
in the clinical expression
1991; revised received
PA, USA 13 April
Gender differences have been reported for a variety of clinical measures in patients with schizophrenia. Clinical characterization may be helpful in identifying symptom clusters which can then be linked to underlying brain function. In this study 74 men and 33 women meeting DSM-IIIR criteria for schizophrenia were studied off medication and rated on measures of symptom type and severity, as well as premorbid and current function. Men were more severely impaired in ratings of negative symptoms, while positive symptoms were not significantly different. There were also differences in premorbid and current functioning, with women manifesting better social functioning than men. Key w,ords: Gender differences;
Gender differences in schizophrenia have been reported for clinical variables which include premorbid adjustment (Klorman et al., 1977; Westermeyer and Harrow, 1984) age of onset (Lewine, 198 1; Loranger, 1984; Goldstein et al., 1989; Hafner et al., 1989) clinical expression of illness (Forrest and Hay, 1971; Gift et al., 1985; Lewine, 1985; Goldstein and Link, 1988; Goldstein et al., 1990; Flor-Henry, 1990; McGlashan and Bardenstein, 1990; Fenton and McGlashan, 199 1), illness course (Goldstein, 1988; Goldstein et al., 1989; Angermeyer et al., 1990; Fenton and McGlashan, 1991) response to pharmacologic and psychosocial interventions (Seeman, 1985; Haas et al., 1990; Test et al., 1990) social and vocational outcome (Nyman and Jonsson, 1983; Salokangas, 1983; Childers and Harding, 1990; Opjordsmoen, 199 l), and health services utilization (Mechanic, 1978). Women are consistently reported to have a more benign form of the illness, typically with Correspondence fo: D.L. Shtasel, MHCRC, Department of Psychiatry, 10th Floor Gates Pavilion, University of Pennsylvania, Philadelphia, PA 19104, USA.
better early development and, at least premenopausally, better long-term outcome. Additionally, the use of different diagnostic criteria lead to varying ratios of males and females with schizophrenia; the more narrow a system the greater proportion of males diagnosed with this disorder (Lewine et al., 1984; Lewine, 1985; Iacono and Beiser, 1989). One strategy to understand the underlying pathophysiology of this disorder has been precise description of clinical subtypes which might then be linked to neural substrates. The development of reliable clinical rating scales specific to schizophrenia has helped quantify observed phenomenology. Lewine (1985) described gender differences in the expression of illness, with men manifesting an ‘amotivational’ syndrome while women had more affective symptomatology. Similar findings are reported by McGlashan and Bardenstein (1990) in their review of initial symptom presentation of the Chestnut Lodge cohort. They described women with more depressive symptoms, interpersonal conflict and self-destructive behavior whereas the men demonstrated more antisocial behavior. In FlorHenry’s (1990) review of gender differences, he further supported the notion that women have
more florid symptoms than men. Goldstein and Link (1988) reported greater affective and paranoid symptoms in women than men; furthermore, they found a positive relationship between psychosis and expression of affective symptoms in women, whereas in men psychosis covaried with withand functional impairment. drawal/isolation Finally, Goldstein et al. (1990) using a latent class analytic approach, examined gender differences in a variety of clinical spheres; despite similar subtype classification there were differences in the relative prevalence of these subtypes between men and women. Men were at greater risk for expressing a predominantly negative subtype whereas women were more likely to manifest predominantly positive symptoms. They pointed out that approaches which look beyond a main effect for gender are helpful in resolving some of the inconsistencies described in these studies. Nasrallah and Wilcox’s (1989) study provides support for a possible theoretical link between this effect and brain function by showing that men with schizophrenia have more evidence for brain insult during childhood associated with negative symptoms, while women have more positive family history. This is in accordance with the hypothesis linking negative symptoms with anatomic abnormalities (Crow, 1980). In contrast, Gift et al. (1985) did not find a relationship between gender and particular psychiatric symptoms. Similarly, in a study of 92 inpatients with DSM-III diagnoses of schizophrenia, Haas et al. (1990) reported gender differences only in character pathology and substance abuse (greater in men). Fenton and McGlashan (1991) in a longitudinal study of the course of schizophrenia reported no sex differences in DSM-III subtypes. It is reasonable to infer that there were no significant gender differences in symptom expression, since this is the basis for subtype classification. The aforementioned studies, which addressed gender differences in the clinical characterization of patients at study entry, for the most paid little systematic attention to the medication status of patients at the time of assessment. Baseline symptoms were extracted from charts in McGlashan and Bardenstein’s (1990) study; medication status in this chronic sample was not described. In Goldstein and Link’s (1988) report of clinical
features the assessment of symptoms was done retrospectively, and it appears that patients were likely medicated. Haas et al. (1990) conducted symptom ratings at the time of study entry but did not mention medication status of patients. In contrast, Lewine (1985) usually washed patients off medication for at least 1 week. Goldstein et al. (1990) retrospectively rated symptoms based on chart review, but since 82% of this sample were first admissions, and all were identified in the preneuroleptic era, it is reasonable to assume that symptom ratings were uninfluenced by neuroleptic medication. Gift et al. (1985) rated symptoms present in the month prior to the first psychiatric hospitalization; while they did not mention whether or not patients were exposed to medication, it is likely that their sample was unmedicated. Finally, subtype ratings obtained from chart review in Fenton and McGlashan’s (1991) study were conducted with a sample where the majority of patients were either neuroleptic naive or unmedicated at initial assessment. While findings from these studies supported gender differences in clinical features of schizophrenia, methodologic differences and uncertainty of medication status confounded definitive interpretation of their results. The present study attempts to clarify the existence and nature of gender differences in the clinical expression of schizophrenia for unmedicated patients. The following hypotheses will be tested in a prospective sample using standard clinical scales: (1) men have more severe negative symptoms; (2) women have more severe positive and affective symptoms; (3) women have better premorbid adjustment and quality of life.
METHODS Subjects Patients were recruited through the Accrual, Assessment and Care Core of the Mental Health Clinical Research Center for Schizophrenia (MHCRC). Consecutive referrals who met the criteria described below were entered in the protocol; these patients were concurrently entered into brain imaging and neuropsychological protocols.
Inclusion criteriu Patients met DSM-IIIR criteria for schizophreniform disorder or schizophrenia at study entry. Additionally, when reassessed after at least 6 months of illness, their diagnosis was (or remained) schizophrenia. Exclusion criteria (1) History of an Axis I disorder (DSM-IIIR) other than schizophrenia or schizophreniform disorder. This included past or present diagnosis of substance abuse or substance dependence; (2) history of any neurologic disease, including head trauma with loss of consciousness; (3) history of medical illnesses (hypertension, cardiac disease, endocrine disorders, renal disease) which might affect brain function; (4) pregnancy (determined by serum test); (5) English not being a native language; (6) history of electroconvulsive therapy. Of 133 patients initially entered with diagnoses of schizophreniform disorder or schizophrenia, 26 were excluded when follow-up reassessment failed to confirm a diagnosis of schizophrenia (other psychiatric diagnosis, 11, neurologic diagnosis, eight, schizophreniform disorder [ < six mos.], 3, lost to follow-up, 3, unable to cooperate, 1.) Of the 11 excluded for non-schizophrenic psychiatric disorders, four were diagnosed with schizoaffective disorder (three males, one female), three with affective disorders (two males, one female), two with substance abuse disorders (both male), one with brief reactive psychosis (male), and one with atypical psychosis (female). We are thus reporting the phenomenologic characteristics at study entry of 107 subjects whose current diagnosis became or remained schizophrenia. This sample, with an emphasis on comparing first episode to other patients with schizophrenia, is described in Shtasel et al. (1992). Patient characteristics Patient characteristics are presented in Table 1. There were no sex differences in age, t(105) = 1.28, ns; education, duration of illness, and number of previous hospitalizations, all t< 1. As is generally reported, women had a later age of onset, t(44.5) = 1.94, p < 0.025, 1-tailed (df adjusted for unequal variance, F’ (32,73) = 2.36, p < 0.005). Of note, 37 patients (23 males, 14 females) were firsttreatment contact patients (neuroleptic naive). Of
these, 14 entered as schizophreniform disorder ( < 6 months of illness) and 23 entered with diagnoses of schizophrenia, with a mean of 5.3 (i SD. 4.2) and a range of 1.4417.7 years of illness. Medication Status Medication status of patients at study entry is described in Table 2. All patients had been on oral medication with the exception of one female patient, who was on depot medication up to 7 weeks prior to study entry. Since patients also participated in biological studies, we generally excluded those who were treated with depot medication. Procedures Clinical assessment. Potential patients were evaluated by a research psychiatrist of the MHCRC. Comprehensive screening included a detailed medical history and physical examination, complete neurological examination and a semistructured psychiatric interview, followed by the Structured Clinical Interview for DSM-IIIR (SCID-P; Spitzer et al., 1986). Family history was also obtained and the following laboratory data reviewed: SMA-20, CBC, thyroid functions, ANA, RPR, ESR, and toxicology. When consent was obtained, all medical and psychiatric records were reviewed. Assessment scales. Patients were studied while off psychotropic medication. Clinical rating scales were completed within 5 days of study entry. In addition to the SCID, the following instruments were used to provide multidimensional characterization: Brief Psychiatric Rating Scale (BPRS; Overall and Gorham, 1980); Scale for the Assessment of Negative Symptoms (SANS; Andreasen, 1982a,b, 1983); Scale for the Assessment of Positive Symptoms (SAPS; Andreasen, 1982b, 1984); Hamilton Psychiatric Rating Scale for Depression (HAMD; Hamilton, 1960); Premorbid Adjustment Scale (PMA; Harris, 1975); Quality of Life Scale (QOL; Heinrichs et al., 1984). These scales describe phenomenology, premorbid adjustment and social and occupational functioning. Training/reliability. The scales were completed by the research assessment team. This group of raters met weekly to review and complete rating scales on videotaped interviews. The following
All M F
107 74 33
30.0 k 7.6 29.4 k 7.0 31.4k8.7
58 43 15
49 31 18
12.5i2.0 I2.6k2.0 12.5k2.2
22.6k6.2 21.7+5.1 24.6+7.X
Neuroleptic naive >6 months off 5 weeks-6 months off 2 weeks-4 weeks off < 2 weeks off medication Total
23 12 8 27 4 14
14 4 4 9 2 33
37 I6 12 36 6 107
intraclass correlations have been achieved based on 11 separate video sessions by four independent raters for BPRS and 13 sessions for SANS and SAPS. For BPRS total 0.83; SANS item averages 0.98 and for averaged global ratings 0.98; SAPS corresponding correlations were 0.90 and 0.96. In addition, each subject was reviewed at a consensus case conference. Datu Analysis The summary scores for the BPRS, SANS, SAPS, HAMD, PMA and the QOL, were compared between male and female groups using t-tests, with unidirectional tests applied to the SANS, PMA (women were hypothesized to score lower); SAPS, HAMD and QOL (women were hypothesized to score higher). The error term for the t-tests was based on the F’ test for equality of variance (SAS PROC T-Test; SAS Institute, Inc., 1985). Sex differences in subscores of these scales were evaluated by Bonferroni-corrected t-tests, unidirectional for the hypothesized effects, with sex as a grouping factor. Factor scales were used as subscores for BPRS (Guy et al., 1975), and the Global ratings were used for the SANS and SAPS. For the QOL scale. a factor analysis was performed on the 21 items and three factors were identified: (1) social functioning (items: intimate relationships, active acquaintances, social activity, social
7.416.2 7.7k5.7 6.817.3
3.5k6.2 3.5+ 5.2 3.6 k 8.0
network, social initiative, social withdrawal, sociosexual relations, anhedonia); (2) engagement (items: relationships with family, sense of purpose, motivation, curiosity, possession of commonplace objects, engagement in commonplace activities, empathy, engagement with interviewer); (3) vocational functioning (items: occupational role functioning, underemployment, accomplishment. satisfaction with occupational role functioning, time utilization). Since there were sex differences in age of onset (see Table I), the analyses were repeated after removing the effects of onset age using a procedure described by Saykin et al. (1991). Briefly. the following procedure was used: (I) beta weights were generated based on regression of onset age on the clinical and functional ratings; (2) the beta weights were multiplied by each subject’s value for onset age standardized to a mean of zero; (3) the resulting values for onset age were then subtracted from the original rating. Analyses of the corrected values did not alter the effects (available upon request).
RESULTS Symptom prqfile The symptom profiles of males and females are shown in Fig. 1 for BPRS (a), SANS and SAPS (b) subscales. As hypothesized, while the groups did not differ in overall symptom severity. measured by total BPRS (meaniS.D. men: 53.11 k 10.96; women: 50.85 f 9.04), t( 105) = 1.04, n.s., men had more severe SANS scores (2.75 f 1.06) than women (2.27 f 0.97) t( 105) = 2.21, ~~0.025, 1-tailed. Univariate contrasts showed that this difference was significant for affect, t(l05)= 1.85; apathy, t= 2.66; and anhedonia, t= 2.08. However, there was no support for
the hypothesis that women have more severe positive symptoms, measured by the SAPS (men: 2.59 kO.83; women: 2.45 +_0.75), t < 1. The hypothesis of more severe affective symptoms in women was likewise not supported (HAMD men: 17.2555.46; women: 18.71 +6.47), t(105)= - 1.12, n.s. Function The functional ratings for males and females are illustrated in Fig. 2 for premorbid adjustment (a) and the quality of life factor scores (b). As hypothesized, men have worse premorbid adjustment (2.53f 1.12) than women (2.09*0.99), t(l05)= 1.69, p< 0.05, l-tailed, and worse quality of life (men: 1.95+ 1.17; women: 2.58* 1.25), t(105)= 2.34, p < 0.01, l-tailed. Univariate contrasts for the three QOL factors showed that women had better functioning for the social and engagement factors, t=3.03, p=O.O05, and t=1.90, p=O.O5, respectively, but not for the occupational factor, t < 1.
As in most earlier studies, women had a later age of illness onset. Our results also support two of the three hypotheses on differences between men and women in the clinical expression of schizophre1
YEN (*4) IVOYEN(33,
Fig. I, Symptom profile (mean+S.E.M.) for men and women on the BPRS factors (A-D = anxiety-depression; ANE = anergia; THOT = thought disorder; ACTV = activity; HOST = hostility) and the SANS (AF= affect; AL= alogia; AV= avolition; AN = anhedonia; AT = attention) and SAPS (HA = hallucinations; DE = delusions; BI = bizarre behavior: TH = thought disorder) subscales.
Fig. 2. Means (i S.E.M.) of men and women on the premorbid adjustment (PMA) subscales (CHILD = Childhood; ERLAD = early adolescence; LATAD = late adolescence; ADLT = adulthood; GEN =general) and the quality of life (QOL) factors (FACI = social functioning; FAC2 = engagement; FAC3 = vocational functioning). Note that for PMA higher ratings indicate poorer adjustment while for the QOL they indicate better functioning.
nia. The hypotheses of worse negative symptoms and worse level of functioning in men were supported, respectively, by higher severity ratings on the SANS and poorer ratings on the premorbid adjustment and quality of life scales. By contrast, we found no evidence to support the hypothesis of sex differences in positive symptoms. Further, we found no sex differences in depressive symptoms. The higher severity of negative symptoms in men and their lower functioning could not be attributed to the earlier age of onset. The disproportionate numbers of men and women in our sample in part reflect collaborating referral sources: early in the study the Veterans Administration Medical Center provided many of the patients (males). Recognizing the need for a broader base, more recent efforts have been successful in establishing linkages with community agencies and the private sector. Nevertheless, consecutive referrals continue to favor males. Our network of ‘informed referrers’ likely contributes to this composition: patients are referred in an unmedicated state, typically at first presentation or when they present to emergency services having discontinued medication and relapsed. Since women are more likely to seek health services than are men (Mechanic, 1978), it is possible that a subgroup exists which seeks help and is treated before symptoms meet threshold for urgent intervention, thus becoming part of an unreferred uni-
verse of medicated patients. Finally, DSM-IITR, a ‘strict’ system, yields unequal numbers of men and women (more men) diagnosed with schizophrenia (Lewine et al., 1984; Westermeyer and Harrow, 1984; Iacono and Beiser, 1989). Accordingly, a possible explanation for the difference between our study and those which have reported more affective and positive symptoms in women relates to the issue of diagnostic criteria. Conceivably, studies which used broader diagnostic systems included patients more narrowly diagnosed with affective, schizoaffective or other psychiatric disorders. However, the effect of excluding diagnoses which might favor women is likely counterbalanced by the exclusion of substance abuse and dependence disorders, which favor men. Nevertheless, our failure to replicate findings which demonstrated more severe positive symptoms in women is not simply explained and warrants further investigation. Concerning sex differences in level of functioning, our data are consistent with prevailing views that women with schizophrenia have overall better premorbid adjustment and quality of life than do men. In addition, however, we identified a pattern of selective functioning which differed between men and women: a sense of life involvement (as measured by the ‘engagement’ QOL factor) and social role functioning were relatively less impaired in women than in men. Given the selective preservation of social functioning and lack of differences in positive symptoms between men and women, it is appealing to attribute this, at least partly, to the sex differences found in depth of affect. We recognize, however, that these relative differences exist against a backdrop of the severe impairment of this particular sample: only a handful of patients (male or female) work, and none are married. Replication with larger samples of females characterized by a broad range of functional severity is needed before further speculation is warranted. Nevertheless, results of functional analyses are consistent with previous literature (Strauss and Carpenter, 1972) underscoring the need to assess functional spheres independently. The sex differences in negative symptoms and functional measures can reflect social variables or biologic substrates. While men and women differ in the clinical expression of schizophrenia, the nature of these differences and their relationship to underlying brain function remain intriguing.
Further investigation of gender differences in schizophrenia, which links such differences to measures of brain function, will be helpful in clarifying these issues.
We thank Margaret M. Taleff, and Diane Sandefur for assistance in phases of this project. This research is supported in part by NIMH grants MH43880, MH42191, lT32MH19112, Research Scientist Development Award MH00586 (REG) and the Clinical Research Center at the University of Pennsylvania Medical Center 5MO l-RR00040.
Andreasen, N.C. (1982a) Negative symptoms in schizophrenia. Arch. Gen. Psychiatry 39. 784-788. Andreasen. N.C. (1982b) Negative vs. positive schizophrenia. Arch. Gen. Psychiatry 39, 7899794. Andreasen, N.C. (1983) The scale for the assessment of negative symptoms (SANS). The University of Iowa, Iowa City, IA. Andreasen, N.C. (1984) The scale for the assessment of positive symptoms (SAPS). The University of Iowa, Iowa City, IA. Angermeyer. M.C., Kuhn, L. and Goldstein, J.M. (1990) Gender and the course of schizophrenia: differences in treated outcomes. Schizophr. Bull. 16, 293-307. Childers, S.E. and Harding, C.M. (1990) Gender. premorbid social functioning, and long-term outcome in DSM-III schizophrenia. Schizophn. Bull. 16, 3099318. Crow, T.J. (1980) Molecular pathology of schizophrenia: More than one disease process? Br. Med. J. 280, 66668. Fenton, W.S. and McGlashan, T.H. (1991) Natural history of schizophrenia subtypes I. Longitudinal study of paranoid. hebephrenic, and undifferentiated schizophrenia. Arch. Gen. Psychiatry 48. 969-977. Flor-Henry, P. (1990) Influence of gender in schizophrenia as related to other psychopathological syndromes. Schizophn. Bull. 16, 21 l-227. Folstein, M.F., Folstein, S. and McHugh, P.R. (1975) Minimental state: a practical method for grading the cognitive state of patients for the clinician. J. Psychiatry Res. 12. 1899198. Forrest, A.D. and Hay, A.J. (1971) Sex differences and the Acta Psychiatr. Stand. 47. schizophrenic experience. 137-149. Gift. T.E., Harder, D.W., Ritzier, B.A. and Kokes, R.F. (1985) Sex and race of patients admitted for their first psychiatric hospitalization: Correlates and prognostic power. Am. J. Psychiatry, 1447-1449.
231 Goldstein, J.M. and Link, B.C. (1988) Gender and the expression of schizophrenia. J. Psychiatry Res. 22, 141-155. Goldstein, J.M. (1988) Gender differences in the course of schizophrenia. Am. J. Psychiatry 145, 684-689. Goldstein. J.M., Tsuang, M.T. and Faraone, S.V. (1989) Gender and schizophrenia: Implications for understanding the heterogeneity of the illness. Psychiatry Res. 28, 243-253. Goldstein, J.M., Santangelo, S.L., Simpson, J.C. and Tsuang, M.T. (1990) The role of gender in identifying subtypes of schizophrenia: A latent class analytic approach. Schizophn. Bull. 16, 2633275. Guy, W., Cleary, P. and Bonato, R.R. (1975) Methodological implications of a large central data system. In: Proceedings of IXth Congress, CINP, Excerpta Medica. Amsterdam. Haas, G.L.. Click, I.D., Clarkin, J.F.. Spencer. J.H. and Lewis, A.B. (1990) Gender and schizophrenia outcome: A clinical trial of an inpatient family intervention. Schizophn. Bull. 16, 2777292. Hafner, H., Riecher, A., Maurer, K., Loffler, W., MunkJorgensen, P. and Stromgren, E. (1989) How does gender influence age at first hospitalization for schizophrenia? Psychol. Med. 19, 903-918. Hamilton, M. (1960) A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 23. 56-62. Harris, J.G. Jr. (1975) An abbreviated form of the Phillips Rating Scale of Premorbid Adjustment in schizophrenia. J. Abnorm. Psychol. 84, 129-137. Heinrichs. D.W.. Hanlon, T.E. and Carpenter, W.T. (1984) The quality of life scale: An instrument for rating the schizophrenic deficit syndrome. Schizophn. Bull. IO, 388-398. Iacono, W.G. and Beiser, M. (1989) Age of onset, temporal stability, and eighteen-month course of first-episode psychosis. In: D. Cicchetti (Ed.), The Emergence of a New Discipline: Rochester Symposium on Developmental Psychopathology, Vol. I : Lawrence Erlbaum Associates. Hillsdale, N.J., pp. 221-260. Klorman, R., Strauss, J.S. and Kokes, R.F. (1977) Premorbid adjustment in schizophrenia, part III: The relationship of demographic and diagnostic factors to measures of premorbid adjustment in schizophrenia. Schizophn. Bull. 3, 214-225. Lewine, R.R.J. (1981) Sex differences in schizophrenia: Timing or subtype? Psychol. Bull. 90, 4233444. Lewine, R.R.J., Burbach, D. and Meltzer H.Y. (1984) Effect of diagnostic criteria on the ratio of male to female schizophrenic patients. Am. J. Psychiatry 141, 84487. Lewine, R. (1985) Schizophrenia: An amotivational syndrome in men. Can. J. Psychiatry 30, 316-318.
Loranger, A.W. (1984) Sex difference in age at onset of schizophrenia. Arch. Gen. Psychiatry 41, 1577161. McGlashan, T.H. and Bardenstein, K.K. (1990) Gender differences in affective, schizoaffective, and schizophrenic disorders. Schizophn. Bull. 16, 319-329. Mechanic, D. (1978) Sex, illness, illness behavior, and the use of health services. Sot. Sci. Med. 128. 2077214. Nasrallah, H.A. and Wilcox, J.A. (1989) Gender differences in the etiology and symptoms of schizophrenia. Ann. Clin. Psychiatry I, 51-53. Nyman, A.K. and Jonsson, H. (1983) Differential evaluation of outcome in schizophrenia. Acta. Psychiatr. Stand. 68, 4588475. Opjordsmoen, S. (1991) Long-term clinical outcome of schizophrenia with special reference to gender differences. Acta Psychiatr Stand. 83, 3077313. Overall, J.R. and Gorham, D.R. (1980) The brief psychiatric rating scale. J. Oper. Psychiatry 11, 48864. Salokangds, R.K.R. (1983) Prognostic implications of the sex of schizophrenic patients. Br. J. Psychiatry 142, 1455151. SAS User’s Guide, Version 5. (1985) SAS Institute, Inc., Cary, NC. Saykin, A.J., Cur, R.C., Gur, R.E., Mozley, P.D., Mozley, L.H., Resnick, S.M., Kester, D.B. and Stafiniak, P. (1991) Neuropsychological function in schizophrenia: Selective impairment in memory and learning. Arch. Gen. Psychiatry 48, 618-624. Seeman, M.V. (1985) Interaction of sex, age, and neuroleptic dose. Compr. Psychiatry 24, 124-128. Shtasel, D.L., Cur, R.E., Gallacher, F., Heimberg, C., Cannon, T. and Cur, R.C. (I 992) Phenomenology and functioning in first episode schizophrenia. Schizoph. Bull., (in press). Simpson, G.M., Lee, J.H., Zoubek, B. and Gardos, G. (1979) A rating scale for tardive dyskinesia. Psychopharmacology 64, 171I179. Spitzer, R.L., Williams, J.B.W., and Gibbon, M. (1986) Structured Clinical Interview for DSM-IIIR-Patient Version @CID-P, 10/l/86). New York State Psychiatric Institute, New York, NY. Strauss, J.S. and Carpenter, W.T. (1972) The prediction of outcome in schizophrenia. I. Characteristics of outcome. Arch. Gen. Psychiatry 27, 7399746. Test, M.A., Burke, S.S. and Wallisch, L.S. (1990) Gender differences of young adults with schizophrenic disorders in community care. Schizophn. Bull. 16, 331-344. Westermeyer, J.F. and Harrow, M. (1984) Prognosis and outcome using broad (DSM-II) and narrow (DSM-III) concepts of schizophrenia. Schizophn. Bull. 10, 624-637.