Psychiatryand PsychiatricEpidemiology Soc PsychiatryPsychiatrEpidemiol (1993)28:267 274 Social 9 Springer-Verlag 1993 Characteristics of hospital-treated schizophrenia in Paulo, Brazil P. R. Menezes and A. H. Mann Section of Epidemiologyand General Practice, Institute of Psychiatry,London,UK Accepted:23 June 1993 Summary. This paper reports a study on the prevalence of hospital-treated schizophrenia in a large urban centre in Brazil. A sample of 124 individuals from a defined catchment area of Silo Paulo consecutively admitted to psychiatric hospitals due to acute episodes of non-affective psychoses were assessed by standardized instruments for mental state and social adjustment. The sample was predominantly white (72.6%), single (65.2%), Catholic (61.3 % ) and not born in that city (58.9 % ). Eighty-six subjects fulfilled DSM-III-R criteria for schizophrenia (69.3 % ), 15 were classified as schizophreniform psychosis (12.1%), and 7 as schizoaffective (5.6 %). Almost twothirds showed Schneider's 'first rank' symptoms. Social adjustment before admission was poor or very poor in nearly half the sample. Mean age at onset was 3.2 years earlier in men than in women (P = 0.007). No gender differences in mental state or social adjustment were observed. Individuals born outside S~o Paulo State had significantly more florid symptoms than those born in that state. These findings suggest that the features of schizophrenia in a large urban centre of Brazil are consistent with the patterns of schizophrenia described in Western developed countries. though this survey was mainly focused on minor psychiatric disorders, it yielded global prevalence rates for psychoses, as defined by the DSM-III ranging from 0.2% to 2.0 % according to centre. Clinical features of schizophrenia have also been investigated. Sougey et al. (1987) recorded the symptoms of 70 schizophrenics admitted to two psychiatric hospitals in the city of Recife, finding Schneider's 'first-rank' symptoms (FRS) in 46 (65.7 %) patients. Shirakawa (1989) assessed the social adjustment of 44 schizophrenic patients who attended a private practice office in Sgo Paulo for many years. More than 40 % had poor social functioning, and only 6 patients (13.6 % ) received salaries sufficient to support their independence. However, in Brazil there has not yet been any proper description of this disabling condition based upon a systematic sample and standardized assessments. Recent changes in the Brazilian public health system, which lead to regional delivery of care to a larger proportion of the population, has created more favourable conditions for the execution of epidemiological investigations on mental health problems (Santana et al. 1988). Urban development in Brazil Schizophrenia is becoming an important public health problem for developing countries, with a projected increase of 87 % in the number of treated cases between 1975 and the year 2000, based on the effect of population changes (Kramer 1980). In Brazil, it is estimated that from 1985 to the year 2000 there will be an increase of 600,000 cases of schizophrenia, a 50 % growth in the number of cases (Levav et al. 1989). Despite these projections, epidemiological data about the characteristics of schizophrenia occurrence in Brazil are scarce. Studies based on routine clinical diagnoses found that schizophrenia accounted for almost one-third of all psychiatric admissions (Araujo 1955; Souza 1983). Almeida et al. (1992) carried out a community-based survey on the prevalence of psychiatric morbidity in three urban centres of Brazil, one of which was S~o Paulo. A1- Economic growth since 1960 has radically changed the population structure and its distribution in Brazil. In 1965, 55 % of Brazilians lived outside cities, but with rapid industrialization there has been a massive move from country areas to the cities, so that by 1987 only 28 % of the population were living in rural areas (Harvey 1987). The country has developed marked social inequalities, with only 3 % of the population earning US$1500 or more per month and 44.5 % having a monthly income of US$150 or less, according to the 1991 census (VEJA 1993). Illiteracy rates vary according to region, with a national average of 18 %. Although the gross national product per capita is 2.7 times higher and the number of televisions per 1000 people is three times higher than the average for countries of equivalent development, some health indicators are worse: infant and maternal mortality rates, for 268 example, which are 50 % and 35 % higher than average, respectively (Haines 1993). Situated in South East Brazil, Silo Paulo is its largest city and main economic centre, with a population of 11.5 million people (IBGE 1992). During the last decade its average annual population growth rate was 1.15 % and was due equally to natural growth and immigration from other areas of the country, mainly from the North East. Some 8 % of the population live in shantytowns, and another 28 % in slums. The average mortality rate in children under 5 is 35 per 1000, but varies from 8.1 to 76 per 1000, according to the district of the city (Forattini 1991). Similar wide variations are found regarding education and availability of piped water and sanitation. The costs of living for the poor are the highest in the country. Health care in Brazil The health care system funded by the government has expanded and changed considerably in the last 30 years (Roemer 1986). Until 1987, the administration of medical care deliveries was the responsibility of the social security ministry, whereas preventative activities were the responsibility of the health ministry (Haines 1993). From 1988 the administration of health care has been decentralised to state and municipal control, with regionalised planning for delivery of care. Most of the medical care is provided by private hospitals, which are paid on the basis of fee-forservice by the government. The availability of such care is extremely varied across the country, with the highest concentrations in the South Eastern region. Contrasting with other regions of the country, in the city of Sio Paulo there is a network of public primary health centres covering most of the urban area, and some public hospitals and emergency units have been built in the most deprived areas, as the result of a municipal policy, but their number is still below the actual needs of its population (Iacoponi 1989). In addition to publicly funded services, wealthier people tend to buy into private medical care, whereas the poor purchase non-prescription drugs direct from pharmacies. Although there are no accurate statistics, the role of traditional healing as an alternative to formal medicine has declined, especially in urban centres (Roemer 1986). This trend is exemplified on childbirth statistics: in 1960 the majority of deliveries were supervised by untrained traditional attendants in the mother's home, whereas by 1981 almost 80 % occurred in hospitals. However, alternative sources of help are still sought at the same time as formal medical consultation by many people, commonly when the problem is in the area of mental illness (Iacoponi 1989). Psychiatric services Psychiatric services were almost exclusively located in psychiatric hospitals until the early 1980s (Marl 1983). Since then there have been several government initiatives aimed at changing the orientation towards a community- based approach (Alves et al. 1992). As a consequence, the number of psychiatric beds has fallen by 20 % in the last 8 years, and there has been an increase in the number of non-residential facilities, such as psychiatric out-patient clinics, but these are still far from ideal. The distribution of psychiatric beds is very uneven, ranging from 0.04/1000 of the population in the Northern region to 0.84/1000 in the South East (Alves et al. 1992). The catchment area chosen for this study is a health district located in the Northern region of the city of Sio Paulo, identified as the 'Sistema Descentralizado e Unificado de Safide Regiio 6' (SUDS-R-6). Local psychiatric facilities comprise two psychiatric out-patient clinics, an emergency unit with 24-h on-duty psychiatrists, a 20-bed psychiatric ward in a general hospital, and three psychiatric hospitals with a total of 1500 beds, for both male and female adult admissions. The local population is estimated to be 950,000 inhabitants, indicating a provision of 1.58 psychiatric beds per 1000 of the population. All the psychiatric admissions are centrally recorded in a psychiatric register on a daily basis. The pathway of care for an individual who presents with a psychotic crisis varies. One of the commonest ways of getting into the psychiatric care system is through the emergency unit, because patients will be seen by a psychiatrist on the day of the consultation. They may receive treatment there for a short period (up to 36 h), after which they may be referred to one of the out-patient clinics, be admitted to psychiatric hospital, or be sent back home with a prescription. When discharged from hospital patients may be referred to the out-patient clinics, but many do not seek medical care until the onset of a new crisis. Patients under 18 may have psychiatric care in the out-patient clinics or the ward in the general hospital, or seek attention in specialised centres outside the region. Patients presenting with behavioural disturbances, such as violent or anti-social behaviour, may be taken in by the police and then sent to the emergency unit. The present investigation is the first module of an ongoing prospective outcome study of schizophrenia in Sio Paulo. A sample of individuals who presented with schizophrenic symptoms and were consecutively admitted to psychiatric hospitals in the study catchment area was assessed with standardised measures. This paper presents socio-demographic and socio-economic characteristics, the clinical profile and level of social disability of the patients in this sample, who are to be followed up for 2 years. Method Selection of subjects All patients aged between 15 and 44 and resident in the area of the SUDS-R-6 who were admitted to the three psychiatric hospitals in the area or to the psychiatric ward in the general hospital during the period between 1 June to 22 August 1991, and for whom the admission diagnosis was schizophrenia, paranoid psychosis or any other functional psychosis- according to the 9th version of the Inter- 269 national Classification of Diseases ( W H O 1975) - were identified, through weekly visits to the psychiatric register by the main investigator (R R. M.). Medical records were then checked, and exclusion and inclusion criteria were applied. Exclusion criteria were: alcohol or drug dependence in the last year, severe mental retardation, organic psychoses (including drug-induced psychoses), epilepsy, hearing or speech difficulties that could impede the administration of interviews, and intention of leaving S~o Paulo after discharge from the psychiatric hospital (this criterion was set because of the follow-up study). Inclusion criteria were at least one of the following symptoms: mood-incongruent delusions, definitely inappropriate or unusual behaviour (such as talking or laughing on his/her own when there is no-one else about, collecting rubbish), hallucinations, disorders of thinking or language (other than acceleration or retardation); or at least two of the following: psychomotor disturbance (such as excitement, posturing, waxy flexibility, stupor), social withdrawal, overwhelming fear, and self-neglect. Patients eligible for entry into the study were interviewed by the main investigator (R R. M.) within 2 weeks of admission to assess their mental state. An interview with a 'key informant', generally a relative of the patient, was held subsequently but still within 2 weeks of admission. The interview included the scale to assess social adjustment, and was carried out by the investigator (RR.M.) or his research assistant (M.S.). Assessm en t Assessments were carried out in a standardized way for all patients, and information was collected from three main sources: the medical records, the patients, and their key informants. Data gathered included: 1. Socio-demographic data: sex, age, ethnic group, marital status, religion, educational level, occupational status, number of persons per room in the household, and monthly per capita income (total income in the household divided by number of people). 2. Psychiatric history: obtained from medical case records and interviews with key informants. Age at onset and number of previous psychiatric admissions were recorded, as was the clinical diagnosis given by the psychiatrist in the emergency unit at the time of admission, coded according to ICD-9. 3. Present State Examination: mental state was assessed using the 9th version of the Present State Examination (PSE) (Wing et al. 1974). The investigator was trained in the use of the instrument at the M R C Social Psychiatry Unit, in London, and used a Portuguese translation of the instrument (Caetano and Gentil 1983). By addition of a 30-item questionnaire covering longitudinal aspects of the disorder, DSM-III-R (APA 1987) diagnoses were also obtained. 4. Social adjustment: a portuguese translation (Chaves et al. 1990) of the W H O Psychiatric Disability Assessment Schedule ( W H O / D A S ) ( W H O 1988) was administered by one of the two interviewers, a psychiatrist R R. M.) and a psychologist (M. S.). Sections 1 (Overall Behaviour), 2 Table 1. Sodo-demographic characteristics of the study sample by sex (n = 124) Age group (years) 15-24 25-34 35-44 Ethnic group White Black Oriental Mixed Place of birth SP capital SP state South-East North-East Other Religion Catholic Evangelic Other None Marital status Single Married Separated Widowed Education Illiterate Less than 4 years 4-8 years 9 years and more Occupational status Working regularly Some activity No occupation Sickness benefit Retired Housewife Persons per room" 1 or fewer 1.1-2.0 More than 2.0 Monthly per capita income (US$)b 50 or less 51-100 101-150 150-200 201-500 More than 500 Male (%) n = 69 Female (%) n = 55 Total (%) n = 124 13 (18.8) 26 (37.7) 30 (53.5) 7 (12.5) 25 (45.5) 23 (41.8) 20 (16.1) 51 (41.1) 53 (42.7) 57 (82.6) 0 (0.0) 3 (4.3) 9 (13.0) 33 (60.0) 4 (7.3) 2 (3.6) 16 (29.1) 90 (72.6) 4 (3.2) 5 (4.0) 25 (20.2) 27 (39.1) 11 (15.9) 14 (20.3) 16 (23.2) 1 (1.4) 24 (43.6) 8 (14.5) 5 (9.1) 15 (27.3) 3 (5.4) 51 (41.1) 19 (15.3) 19 (15.3) 31 (25.0) 4 (3.2) 47 (68.1) 14 (20.3) 5 (7.2) 4 (5.8) 29 (53.7) 13 (24.1) 9 (16.7) 3 (5.6) 76 (61.3) 27 (21.8) 14 (11.3) 7 (5.6) 52 (75.4) 9 (13.0) 7 (10.1) 1 (1.4) 29 (52.7) 12 (21.8) 12 (21.8) 2 (3.6) 81 (65.3) 21 (16.9) 19 (15.4) 3 (2.4) 4 (5.8) 9 (13.0) 38 (55.0) 18 (26.0) 3 (5.5) 19 (34.6) 18 (32.7) 15 (27.3) 7 (5.6) 28 (22.6) 56 (45.2) 33 (26.6) 18 (26.1) 4 (5.8) 27 (39.1) 9 (13.0) 11(15.9) n = 69 39 (56.5) 20 (29.0) 10 (14.5) 6 (10.9) 5 (9.1) 24 (42.6) 5 (9.1) 2 (3.6) 13 (23.6) n = 54 22 (40.7) 19 (35.2) 13 (24.1) 24 (19.3) 9 (7.3) 51 (41.1) 14 (11.3) 13 (10.5) 13 (10.5) n = 123 61 (49.6) 39 (31.7) 23 (18.7) n = 66 22 (33.3) 25 (37.9) 6 (9.1) 6 (9.1) 6 (9.1) 1 (1.5) n = 54 20 (37.0) 19 (35.2) 7 (13.0) 5 (9.3) 2 (3.8) 1 (1.9) n = 120 42 (35.0) 44 (36.7) 13 (10.8) 11 (9.2) 8 (6.7) 2 (1.7) a Information missing in one subject b Information missing in four subjects (Social Role Performance) and 5 (Global Evaluation) of the schedule were used. Adjusted scores for sections i and 2 were obtained by summing up all valid scores (different from 9) and dividing by the number of valid items ( W H O 1988). The higher the adjusted score in a section, the worse the functioning in the areas covered by that section. Twenty subjects were each interviewed by the two investigators together for a reliability test. 270 Analysis Frequency distributions of socio-demographic and socio-economic variables, psychopathological status, and social disability levels were examined. Chi-square tests were calculated for categorical variables, and t-tests for comparisons of continuous variables between two groups, with the statistical package Epi Info 5 (Dean et al. 1990). Inter-rater reliability for the DAS was assessed by intra-class correlation coefficients (Bartko 1991). come ranged from US$16.50 to US$ 750; 71.7% of the individuals had a per capita monthly income of US$100 or less. Psychiatric characteristics The mean age in the sample was 32.6 years (SD -- 6.9), and it was predominantly white (72.6 %), single (65.2 %), Catholic (61.3 %) and born outside Sgo Paulo (58.9 %) (Table1). Thirty-five subjects (28.2%) had had fewer than 4 years of education. Fifty-one (41.1% ) had no current source of income and relied on their families to survive; 33 patients had some occupation, either as registered employees or as self-employed; 27 individuals were living on social benefits due to illness, with a majority of males in this situation (Z 2 = 4.75, 1 dr, P = 0.03). Sixty-one subjects (49.6 %) lived in households with up to one person per room, and 21 (18.7 %) lived in very crowded houses, with more than two persons per room. Per capita monthly in- Seventy-three subjects received a diagnosis of schizophrenia on admission to hospital (58.9%), with a predominance of subtype 'unspecified' (n = 22), followed by the subtypes 'paranoid' (n = 19) and 'residual' (n = 15). For most of the remaining subjects (39.5 %) a diagnosis of 'other functional psychoses' (n = 49), subtype 'unspecified' (n = 44) was recorded. Paranoid psychosis was rarely given as a clinical diagnosis (n = 2). Eighty-six subjects (69.3 % ) fulfilled DSM-III-R criteria for schizophrenia, 15 were classified as having schizophreniform psychosis (12.1%), and 7, schizoaffective (5.6 %). Thirteen subjects were classified as having affective disorders, with 7 as bipolar and 6 as depressive; 3 subjects fulfilled criteria for 'other psychoses'. The PSE was applied to all individuals included in the study. Eighty-one subjects (65.3 % ) were classified by the C A T E G O program as class 'S + ', with another subject assigned to class S?. Thirty-three subjects (26.6 %) were assigned to other classes of non-affective functional psychoses (P +, P? and O?); the remaining 8 (6.4 %) were assigned to affective-like classes (M +, M?, D + and R + ), and 1 subject was assigned to class X, which means nonspecific symptoms. The most frequent syndromes were 'Affective Flattening' (77.4 %), 'Nuclear Syndrome' (62.1%), 'Non-specific Psychoses' (63.7 % ), 'Delusions of Reference' (58.9 % ), 'Delusions of Persecution' (57.2 % ) and 'Sexual and Fantastic Delusions' (57.2 %). The average age at onset was 24.3 years (Table 2), ranging from 14 to 44 years. The average lenght of illness was 8.3 years. For 30 subjects (24.4 %) the admission in the study period was the first admission, and another 34 (27.6 % ) had had 10 or more previous psychiatric admissions. Table 2. Psychiatric characteristics of study subjects Social adjustment Results Initially, 217 patients were identified at the psychiatric admissions register, but 93 of these were excluded because they met one or more of the exclusion criteria. Reasons for exclusion included usual residence outside the catchment area (n = 20), alcoholism (n = 19) and drug dependence (n = 10). Ultimately, 124 patients (57.1%), 69 men and 55 women, were included in the study. Socio-demographic characteristics Mean age at onset (years) Mean length of illness (years) Mean DAH subscore Mean BSO subscore S + (%) Total Men (n = 124) (n = 69) 24.3 22.9 8.3 9.5 10.6 9.9 4.9 5.1 65.3 65.2 Women (n = 55) 26.1 6.7 11.5 4.7 65.5 P 0.011 0.021 0.31 0.53 0.98 DAH, Delusions and hallucinations;BSO, behaviour and speech The Disability Assessment Schedule (DAS) was applied to 115 patients (92.7 %). In 3 cases it was not possible to obtain information from the relatives because of lack of contact with the patient in the previous month. The remaining 6 cases had been at home less than 15 days between the index admission and a previous one, which was not sufficient time to complete schedule sections 1, 2 and Table 3. Distribution of DAS adjusted scores for sections 1 and 2 Overall behaviour" Social role Performancea 0.0043.99 45 (39.1) 30 (26.3) 1.00-1.99 21 (18.3) 16 (14.1) DAS, World Health Organization PsychiatricDisability Assessment Schedule Percentages in brackets 2.00-2.99 22 (19.1) 26 (22.8) 3.00-3.99 19 (16.5) 22 (19.3) 4.00-4.99 8 (7.0) 20 (17.5) 271 100 . -~_~ -_~-_- ~ _ ~ .Q 90 80 ,o e~ 70 60 50 "5 40 0 30 E .e e "~ o-e / 9 o" males ---o-- females 20 10 0 " ~ " " " ~" . . . . . . . . . . . l I I 1 I t ~ t 1 I 1 I I 1 I a g e a t onset 80 e 9 SP :2 70 e GO e ---~-- Other '' /T 50 8 40 ~ O. 30 20 10 0 I I I I I I I I I I I I I I ; ; ; ; I I I I Fig. 1. Cumulative percentage of age at onset by sex (n = 124) Fig.2. PSE syndrome profiles according to place of birth (n = 124). NS, Nuclear syndrome; CS, catatonic syndrome; IS, incoherent speech; RS, residual syndrome; DD, depressive delusions and hallucinations; SD, simple depression; ON, obessional neurosis; GA, general anxiety; SA, situational anxiety; HT, hysteria; AE affect flattening; HM, hypomania; AH, auditory hallucinations; PE, delusions of persecution; RE, delusions of reference; GR, grandiose and religious delusions; SF, sexual and fantastic delusions; VH, visual hallucinations; OH, olfactory hallucinations; OV, overactivity; SL, slowness; NP, non-specific psychosis; DE, depersonalization; ED, special features of depression;AG, agitation; NG, self-neglect; IR, ideas of reference; TE, tension; LE, lack of energy; WO, worrying, etc.; IT, irritability; SU, social unease; IC, loss of interest and concentration; HZ hypochondriasis; OD, other symptoms of depression; OR, organic impairment; SC, "subcultural" delusions of hallucinations; D/, doubtful interview syndrome 5. The inter-rater reliability of the DAS was good, with intra-class correlation coefficients (ICC) for each item ranging from 0.88 (95% CI = 0.72-0.95) to 1.00 (95% CI = 0.99-1.00). The ICC for global evaluation was 0.93 (95 % C.I. = 0.83-0.97). The average adjusted scores for sections 1 (Overall Behaviour) and 2 (Social Role Performance) were 1.67 and 2.25, respectively. Subjects tended to have higher scores in section 2 than in section 1 (Table 3). In section 5 of the schedule, global evaluation of social disability, 57 subjects were rated as 3 or 4 ('poor adjustment' and 'very poor adjustment', respectively), with the remaining 57 subjects evenly distributed between scores 0, 1 and 2 ('excellent or very good', 'good' and 'fair' adjustments). For 1 subject it was not possible to give a global evaluation score. Gender differences A significantly greater proportion of white subjects were male (Z z = 7.86, 1 dr, P = 0.005) (Table 1). The proportion of single subjects was also greater in male than in female patients (Z 2 = 6.92, 1 df P = 0.008). The proportion with 4 or more years of education was higher for men than for women; the difference was statistically significant (Z 2 = 6.76, 1 dr, P = 0.009). The main difference between genders concerned age at onset (Fig. 1). Men became ill at younger ages than women (Table 2); the relative risk of age at onset less than 26 years for men over women was 1.74 (95 % C.I. = 1.16 to 2.62, P--0.002). Men had more years of illness than women (Table 2), with averages of 9.5 years and 6.7, respectively. Women were significantly more represented in the first admission group (19vs 11men, X 2= 5.78, P = 0.016). PSE syndrome profiles for men and women were similar. Of the 38 syndromes, only 3 differed between sexes at a statistically significant level. These were 'depressive delusions and hallucinations' (DD) ( P = 0.012), 'olfactory hallucinations' (OH) (P=0.012), and 'lack of energy' (LE) ( P = 0.048). Concerning social adjustment, gender differences for adjusted scores of sections 1 and 2 were small and non-significant. There was no significant difference for 'global evaluation' scores either. Place of birth Those born in the state of Silo Paulo were compared with those born in other states of Brazil (Table 4, Fig. 2). There were no significant differences between these two groups regarding age, sex, mean age at onset of illness, and mean 272 Table 4. Psychiatric characteristics according to place of birth length of illness. Those not born in Sgo Paulo State were significantly more often classified as class S + by the CATEGO program than those born in that state. Their average subscore for delusions and hallucinations (DAH) was also significantly higher, and this difference remained after controlling for length of illness, years of education and per capita monthly income. The average subscore for observed problems with behaviour and speech (BSO) was also higher, but not at a statistically significant level. There were no cases excluded due to intention to leave Silo Paulo, nor did any patients refuse to participate in the study. Inclusion criteria were based on a very broad range of symptoms, and might have allowed the inclusion of a reasonable number of other psychiatric conditions than schizophrenia. However, the use of standardised instruments and definitions limited the extent to which this may have happened. More than 80 % of the subjects fulfilled DSMIII-R criteria either for schizophrenia or for schizophreniform psychosis, and more than 90 % were classified in classes S, P or O by the CATEGO. Information on age at onset was obtained from the key informants. The observed gender difference cannot be attributed to differential misclassification due to recall bias, and agrees with other studies. Observer bias might have occurred in the assessment of mental state and social disability, but the practised use of standardized instruments makes this unlikely. The assessment of social adjustment was not always blind to the assessment of psychopathology, but good inter-rater reliability observed for the DAS suggests that there was no significant observer bias. Discussion Socio-demographic characwristics About two-thirds of this consecutive sample of psychotic patients admitted to psychiatric hospitals in a region of Silo Paulo presented with Schneider's FRS, and almost 70 % fulfilled DSM-III-R criteria for schizophrenia. More than 40 % had no income of any sort, depending entirely on their relatives. Approximately half of the sample had a poor social adjustment in the month prior to admission to hospital. Men became ill at younger ages than females, but there were no other gender differences in clinical features or social adjustment. A specific finding was that subjects not born in Sgo Paulo State showed more positive schizophrenic symptoms than those born in Silo Paulo state. A recent survey on the prevalence of psychiatric disorders in the community, which examined a representative sample of the population of Silo Paulo, provided data on socio-demographic and socio-economic characteristics of that population (Mari and Andreoli 1990). Subjects included in the present study differed from the general population in many ways. The proportion of single individuals was much higher in the study sample (65 % vs 33.3% in the general population). Low proportions of married schizophrenics have been found in several studies (Gmur 1991; Isele and Angst 1985; McGlashan 1984). The study sample also had fewer years of education than the general population, suggesting that people suffering from schizophrenia have their educational development impaired by their illness. The disadvantages for subjects included in this study are also evident from occupational status. Whereas more than half of the general population, as assessed by the prevalence survey, were working regularly as employees or self-employed, less than 20 % of the subjects in this study were doing so. In the general population 9 % were unemployed, contrasting with almost 30 % of the study subjects. These findings are in accordance with the results of several studies on schizophrenia (Campbell et al. 1990; Isele and Angst 1985; McCreadie 1982; Mignolli et al. 1991; Tsuang and Fleming 1987). Age mean (years) Proportion male (%) Mean age at onset (years) Mean length of illness (years) % S+ DAH mean BSO mean Place of birth Silo Paulo State (n : 69) 32.6 53.6 23.8 8.9 56.5 9.0 4.5 Other states (n : 55) 32.5 46.4 25.0 7.5 76.4 12.6 5.5 P 0.90 0.61 0.65 0.27 0.021 0.017 0.31 Methodological limitations This is a study of hospital-treated schizophrenia and excluded those not admitted to hospital; it is known that in some countries some 20 % of all schizophrenic patients do not receive hospital care (Westermeyer and Harrow 1988). Those aged 45 or over were also excluded from the study, because it was designed to allow comparisons with the W H O studies on schizophrenia. In fact, patients aged 45 or over accounted for about 20 % of all admissions due to non-affective functional psychoses in the catchment area during the study period. There were not cases under the age of 18. One possible explanation is that local psychiatric hospitals admit only those aged 18 or over, with the exception of the psychiatric ward in the general hospital, and patients with psychotic illnesses arising when they are still under 18 years of age may seek treatment in services outside the catchment area. The sample is not representative of those schizophrenic patients with concomitant severe alcohol- or drug-related problems. Symptomatology The percentage of C A T E G O class S + observed in this study is similar to those found in the three studies organized by the World Health Organization, the International Pilot Study of Schizophrenia (WHO 1973), the W H O Collaborative Study on the Assessment and Re- 273 duction of Psychiatric Disability ( W H O 1988), and the Determinants of Outcome of Severe Mental Disorders (Jablensky et al. 1992). Almost two-thirds of the patients presented with Schneider's 'first rank' symptoms, a figure in agreement with those found in several studies in different settings (Carpenter and Strauss 1974; Ndetei 1988; Wing et al. 1974). underlying concepts and methodological shortcomings, and an alternative explanation has been proposed, i.e. that the association of migration with higher psychiatric morbidity is confounded by the condition of being displaced from the labour forces (Almeida 1980). Further research specifically designed to investigate this issue is needed. Social disability Conclusion The higher adjusted scores in section 2 than section 1 of the DAS indicate that subjects tended to have problems in specific areas of behaviour, rather than in overall behaviour. In a study carried out in South Verona, Italy, dysfunctions in the sexual and occupational roles, items in section 2, were more frequent than dysfunctions in the items in section 1 (Mignolli et al. 1991). In the ' W H O Collaborative Study on the Assessment and Reduction of Psychiatric Disability', 'sexual role' and 'work role' appeared to be altered first, whereas deterioration in 'selfcare' happened later ( W H O 1988). The same pattern was also observed in a cohort of Dutch schizophrenics (de Jong et al. 1985). This is the first study carried out in Brazil in which a consecutive sample of hospital-treated schizophrenic patients was assessed by means of standardised instruments. Results suggest that schizophrenic patients in a large urban centre of Brazil share similar clinical features with patients from Western developed countries. The high proportion of patients without any kind of social security benefits highlights the need of developing a proper social security policy to guarantee even minimal living conditions for them. Acknowledgements. Dr. R R. Menezes is supported by a grant from the 'Conselho Nacional de Desenvolvimento Cientffico e Tecnoldgico-CNPq', Brazil. The authors thank M. Scazufca for her help with DAS interviews, and Dr. K. Lloyd, Dr. H. Cope, and two anonymous referees for their comments on earlier drafts of this paper. Gender differences The difference in educational level (men had more years of formal education than women) follows the pattern found for the general population (Mari and Andreoli 1990). Age at onset differed significantly between the sexes. On average, men became ill 3.2 years earlier than women, comparable with the 3.2-4.1 years observed by Hafner et al. (1993), and within the range of differences observed in the W H O Collaborative Study on Psychiatric Disabilities (Hambrecht et al. 1992). This difference in age at onset may be one of the reasons why there are more men than women in this sample, since the maximum age for entry in the study was set as 44 years, which might have excluded women with later onset. Regarding symptomatology, there were only minor differences in accessory symptoms, whereas for symptoms of nuclear schizophrenia there was no difference. This is in line with the findings from the W H O multicentre study (Hambrecht et al. 1992) in Western E u r o p e a n centres. Place of birth The present study was not originally designed to investigate the relationship between place of birth and schizophrenia, so that the finding that those not born in S~o Paulo State have more positive schizophrenic symptoms must be interpreted with caution. Some authors have suggested that immigration to large urban centres might be a risk factor for the development of mental illness, especially in Latin American coutries, due to the stress of acculturation and cultural marginalisation (see Almeida 1985, 1987). However, studies carried out in Latin America focusing on this issue have been criticised for their References Almeida NF (1980) Urbanizacao e doenca mental. Rev Assoc Psiq Bahia 1:61-77 Almeida NF (1985) Cultura e psicopatologia. Revisao da literatura epidemiologica latino-americana. J Bras Psiq 34:357-364 Almeida NF (1987) Social epidemiology of mental disorders. Acta Psychiatr Scand 75:1-10 Almeida NF, Mari J J, Coutinho ESF, Franca JF, Fernandes JG, Andreoli SB, Busnello E (1992) Estudo multicentrico de morbidade psiquiatrica em areas urbanas brasileiras (Brasilia, S~o Paulo, Porto Alegre). Rev ABP-APAL 14:93-104 Alves DSN, Seidl EMF, Schechtman A, Silva RC (1992) Elementos para uma analise da assistencia em saude mental no Brasil. J Bras Psiq 41:423426 American Psychiatric Association (1987) Diagnostic and statistic manual of mental disorders, 3rd edn, revised (DSM-III-R). APA, Washington Araujo D (1955) Atividades da Secao de Cooperacao do SNDM: Dados estatisticos. Arq Serv Nac Doenca Mental 4:24%251 Bartko JJ (1991) Measurement and reliability: statistical thinking considerations. Schizophren Bull 17:483-489 Caetano R, Gentil V (1983) Glossario das definicoes dos sintomas incluidos na 9a. edicao do PSE mimeo, Silo Paulo Campbell PG, Taylor J, Pantelis C, et al (1990) Studies of schizophrenia in a large mental hospital proposed for closure, and in two halves of an inner London borough served by the hospital. In: Weller M (ed) International perspectives in schizophrenia: Biological, epidemiological and social findings. Libbey, London Paris, pp 185-202 Carpenter WT Jr, Strauss JS (1974) Cross-cultural evaluation of Schneider's first-rank symptoms of schizophrenia: A report from the International Pilot Study of Schizophrenia. Am J Psychiatry 131:682-687 Chaves AC, Sarin LM, Mari JJ (1990) Escala de avaliacao de incapacitacao psiquiatrica (DAS). Escola Paulista de Medicina (mimeo), S~o Paulo De Jong A, Giel R, Slooff CJ, Wiersma D (1985) Social disability and outcome in schizophrenic patients. Br J Psychiatry 147:631-636 274 Dean AG, Dean JA, Burton AH, et al (1990) Epi Info, Version 5.01: A word processing, database and statistics program for epidemiology on microcomputers. USD, Georgia Forattini OP (1991) Qualidade de vida e meio urbano. A cidade de S~o Paulo, Brasil. Rev Saude Publica 25:75-86 Gmur M (1991) The 12-year clinical course of schizophrenia. Soc Psychiatry Psychiatr Epidemio126: 202-211 Hafner H, Maurer K, Loftier W, Riecher-Rossler A (1993) The influence of age and sex on the onset and early course of schizophrenia. Br J Psychiatry 162:80-86 Haines A (1993) Health care in Brazil. BMJ 306:503-506 Hambrecht M, Maurer K, Harrier H (1992) Gender differences in schizophrenia in three cultures. Results of the WHO Collaborative Study on Psychiatric Disability. Soc Psychiatry Psychiatr Epidemio127:117-121 Harvey R (1987) Clumsy giant. A survey of Brazil. Economist, special supplement, 26 April Iacoponi E (1989) The detection of emotional disorders by primary care physicians. A study in S~toPaulo. Brazil. PhD thesis, University of London IBGE - Fundacao Instituto Brasileiro de Geografia e Estatistica (1992) Censo demografico 1991. IBGE, Silo Paulo Isele R, Angst J (1985) Social disabilityin schizophrenia: the controlled prospective Burgholzili study. I. Case-finding, research design, and characteristics of the sample. Eur Arch Psychiatry Neurol Sci 234:341-347 Jablensky A, Sartorius N, Ernberg G, Anker M, Korten A, Cooper JE, Day R, Bertelsen A (1992) Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study. Psychol Med Monogr Suppl 20 [whole issue] Kramer M (1980) The rising pandemic of mental disorders and associated chronic diseases and disabilities. In: Stromgren E, Dupont A, Nielsen JA (eds) Epidemiological research as basis for the organization of extra-mural psychiatry. Acta Psychiatr Scand Suppl 62:382-397 Levav I, Lima BR, Lennon MS, Kramer M, Gonzalez R (1989) Salud Mental para todos en America Latina y el Caribe. Bases epidemiologicas para la accion. Bol Of Sanit Panam 107:196-219 Marl JJ (1983) Psychiatric care in Brazil. In: Brown S (ed) Psychiatry in developing countries. Gaskell, London, pp 7-9 Marl JJ, Andreoli SB (1990) Estudo multicentrico de transtornos mentais: relatorio preliminar- setor Silo Paulo. Escola Paulista de Medicina, Silo Paulo McCreadie RG (1982) The Nithsdale Schizophrenia Survey. I. Psychiatric and social handicaps. Br J Psychiatry 140:582-586 McGlashan TH (1984) The Chestnut Lodge Follow-up Study. I. Follow-up methodology and study sample. Arch Gen Psychiatry 41: 573-585 Mignolli G, Fanccincani C, Platt S (1991) Psychopathology and social performance in a cohort of patients with schizophrenic psychoses. A seven-year follow-up study. Psychol Med Monogr Suppl 19:17-54 Ndetei DM (1988) Psychiatric phenomenology across countries: constitutional, cultural, or environmental? Acta Psychiatr Scand Supp1344:33-44 Roemer MI (1986) The changeability of health care systems - Latin American experience. Med Care 24:24--29 Santana VS, Almeida N, Fo., Marl JJ (1988) Revisao dos estudos epidemiologicos em saude mental no Brasil, part 1. J Bras Psiq 37: 22%231 Shirakawa I (1989) O ajustamento social na esqnizofrenia. PhD thesis, Escola Paulista de Medicina, Silo Paulo Sougey EB, Carvalho TFR, Bandin JM, Albuquerque JOM (1987) Frequencia de sintomas de primeira ordem de Kurt Schneider numa populacao de esquizofrenicos do Nordeste brasileiro. Rev ABP-APAL 9:151-156 Souza CM (1983) Epidemiologia em medicina clinica. Educ Med Salud 17:83-87 Tsuang MT, Fleming JA (1987) Long-term outcome of schizophrenia and other psychoses. In: Hafner H, Gattaz WF, Janzarik W (eds) Search for the causes of schizophrenia. Springer, Berlin Heidelberg New York, pp 88-97 VEJA (1993) Urea piramide feita de desigualdade. VEJA 1281: 7881 Westermeyer JF, Harrow M (1988) Course and outcome in schizophrenia. In: Tsuang MT, Simpson JC (eds) Handbook of schizophrenia, vol 3: Nosology, epidemiology and genetics. Elsevier Science, Amsterdam, pp 205-244 Wing JK, Cooper JE, Sartorius N (1974) Measurement and classification of psychiatric symptoms: an instruction manual for the PSE and Catego Program. Cambridge University Press, London World Health Organization (1973) The international pilot study of schizophrenia, vol 1. World Health Organization, Geneva World Health Organization (1975) International classification of diseases, 9th version. WHO, Geneva World Health Organization (1988) Psychiatric Disability Assessment Schedule (WHO/DAS). WHO, Geneva Dr. P. Menezes Section of Epidemiology and General Practice Institute of Psychiatry De Crespigny Park London SE5 8AF UK