Although the incidence of schizophrenia is comparatively low, its prevalence and lifetime risk of just over 1% indicate that it is by no means a rare disease. In most populations and at any given point in time, at least five out of every 1000 adults would meet the clinical diagnostic criteria for schizophrenia, and the disorder accounts for about one fifth of all chronic and severe disability.
The economic cost of schizophrenia to society is almost half as high as that of myocardial infarction,1 and it is the cause of severe distress to patients and their social networks, as well as to the community.
Few mental disorders have been investigated with such persistence over decades and with such patchy results as schizophrenia. In the broad context of human pathology, schizophrenia exhibits an extraordinary 'behavior'. For instance, no screening or diagnostic test is available to detect behavioral, psychological, or physiological abnormalities specific to schizophrenia.
The diagnostic elusiveness of schizophrenia at the level of community epidemiology is paralleled by a comparable fuzziness at the level of its conceptualization. It is difficult to improve substantially on Kraepelin's original definition:
We designate as dementia praecox the development of a simple, more or less pervasive, state of mental weakness, which manifests itself as an acute or subacute mental disorder. The course of this disease process can exhibit very different patterns.. ..This behaviour indicates, I believe, that in all likelihood we are dealing with an organic change in the brain.2
The current situation in schizophrenia research, which is being dominated increasingly by new technologies, such as molecular biology and brain imaging, evokes two methodological concerns:
* much of the present clinical and biological research in schizophrenia is vulnerable to criticism because of the "law of small numbers" and the insufficient statistical power of research designs to avoid Type II errors, ie, the invalid rejection of hypotheses that may be true, and
* the multi-domain, or "catch-all" approach is not conducive to the crystallization of mutually exclusive, testable concepts and propositions about schizophrenia; this explains why few of the currently coexisting theories and models of schizophrenia can be discarded.
THE ROLE OF EPIDEMIOLOGY
The majority of epidemiologic studies have been descriptive, focusing mainly on:
* persons, place, and time,
* distributions by age, sex, social class, and geographical and cultural location, and
* "natural history."
Few epidemiological inquiries have been designed to test specific hypotheses and, considering the difficulties in case-finding, in determining the needed sample size, and in estimating the statistical power, it is not surprising that the analytical epidemiology of schizophrenia has been less productive than that of cancer or cardiovascular disease.
Much of the descriptive epidemiological research in schizophrenia can be criticized for methodological flaws. As pointed out by Shepherd5:
It is now clear that the design of a study aiming to elucidate the natural history of schizophrenia must take account of at least four factors: (1) the identification of all cases in a defined population during a fixed period of time; (2) the application of standardized diagnostic procedures with criteria of known reliability; (3) prospective follow-up procedures, preferably from the onset of first attack for at least 5 years, with interim as well as endpoint assessments and, preferably, uniform treatment regimes throughout the follow-up period; (4) standardized and independent clinical and social measures of outcome.
THE WHO SCHIZOPHRENIA RESEARCH PROGRAM
The methodological requirements referred to above, as well as the demand for a representative patient population, have been met to a large extent by the World Health Organization's (WHO) multinational epidemiological and clinical studies on schizophrenia, which for nearly two decades have involved psychiatrists and other investigators in 20 centers in 17 countries. In addition to contributing new data, these studies provide a conceptual framework that also allows findings generated outside the WHO network to be evaluated in order to draw an epidemiological profile of schizophrenia.
The WHO program is characterized by:
* simultaneous case finding and data collection, according to a common design and methodology, in geographically-defined areas in different countries and cultures,
* use of standardized instruments,
* clinical assessment of cases by highly-trained psychiatrists whose performance was monitored by means of intra- and Intercenter reliability exercises,
* a system of diagnostic classification that takes into account clinical judgment yet adds a computer-based reference categorization of the cases, and
* multiple follow-up assessments at intervals of 1, 2, and 5 years.
The program comprises three major studies (Table 1). The first (1969 to 1977) was the International Pilot Study of Schizophrenia (IPSS) and included nine centers in Africa, Asia, Europe, and Latin and North America, with a total of 1202 patients selected from consecutive admissions to psychiatric services and meeting specified criteria. Each patient had a detailed standardized clinical assessment and a full reassessment at least twice, at a 2-year and a 5-year follow-up (some of the centers have completed a 10-year follow-up).4,5
The second study, Assessment and Reduction of Psychiatric Disability, was focused primarily on the manifestations and course of behavioral impairments and social disabilities in patients with schizophrenia of a recent onset.6 This study included 520 patients in five European and one African country who were examined initially and then reexamined at 1-year and 2-year follow-up assessments.
The third and most recent study, Determinants of Outcome of Severe Mental Disorders, was conducted from 1978 to 1986 and included 1379 patients who were identified and assessed at 12 research centers in 10 countries. The core component of the study was an epidemiological case-finding investigation in which, during 2 consecutive years, all individuals making their first contact with various 'helping agencies' (including not only psychiatric and medical services but also traditional healers or other rarely tapped facilities) in specified geographic areas were screened for symptoms of functional psychosis. Those meeting specified criteria for possible schizophrenic illness were invited for a full clinical and social examination with standardized instruments and were followed for 2 years, with annual reexaminations. The design, methodology, and part of the findings of this study have been published,7,8 and publication of further, detailed data is under way.
EPIDEMIOLOGICAL PROFILE OF SCHIZOPHRENIA
The major findings that can provide a basis for an epidemiological profile of schizophrenia represent a hierarchy. On one hand, there are data that have been adequately replicated in a variety of settings; on the other hand, there are epidemiological data that are less firmly established and, while contributing to the completion of the epidemiological picture, they are less likely to reflect essential features of the condition.
The epidemiological findings that can be regarded as "robust," follow.
* The syndrome of schizophrenia is universal. To date, no population, demographic group, or culture has been shown to be free of schizophrenia. The psychopathological syndromes now generally accepted as characteristic of the disorder occur in many different cultures and geographical areas. Although no single pathognomonic symptom could be identified as invariably present in every patient and in each setting, the overall configuration of the disorder's symptoms is highly consistent across the cultures.
This conclusion about a crosscultural similarity in the clinical expression of the disorder rests on two kinds of evidence. First, the frequency distributions of symptoms elicited in the WHO studies with the Present State Examination indicate that the group profiles of patients in developing and developed countries are very similar (Figure). Second, the CATEGO program, which groups and classifies the cases according to hierarchical rules, assigns high proportions (between 64% and 98%) of the patients diagnosed as schizophrenic in different cultures to classes representing the schizophrenic psychoses in the CATEGO system.
This does not mean that variation in the clinical picture of schizophrenia is absent. Differences occur in the relative frequency of certain symptoms (for example, depressive affect is more common in patients in developed countries while visual hallucinations are more prevalent in patients in developing countries). However, while such variation requires further study, it does not dominate the cross-cultural presentation of schizophrenia. It is important to note that patients in the tropical rain forests of Nigeria and in the fishing villages of Denmark feel and report alike that their thoughts are being stopped, taken away, 'read' by some alien agency, or 'broadcast' at large. The fact that the content of reported subjective experiences of individuals suffering from schizophrenia and the form of their verbal expression are remarkably similar in patients belonging to different cultures and educational backgrounds should arouse the interest of the medical anthropologist at least as much as the variations in symptoms and behavior patterns. Considering the wide differences in cultural beliefs, social norms, and language, the cross-cultural similarity in the subjective experience of 'core' schizophrenic symptoms may be regarded as pointing to a common mechanism of their production at a very basic level of neurobehavioral organization.
The WHO Program of Cross-Cultural Research in Schizophrenia (1967-85)
Profiles on 44 Selected PSE Items of 586 Patients in Developing Countries and 746 Patients in Developed Countries All Meeting 'Broad' Diagnostic Criteria for Schizophrenia and Related Disorders
Stevens and Wyatt9 suggest that patient samples in developing countries may contain many cases of acute transient psychoses that are etiologically different from the rest of the disorders classified as schizophrenia. In the WHO patient sample, considerable heterogeneity was found in regard to course and outcome but very little in regard to presenting symptoms. Disregarding symptom similarities, there are other reasons to think that schizophrenia can be distinguished from the socalled acute schizophreniform psychoses.
Annual Incidence Rates per 10 000 Population, Age 15-54, for IWo Diagnostic Definitions
For example, in all WHO study areas the age- and sex-specific distribution of onsets of disorders meeting the diagnostic criteria for schizophrenia followed a very similar pattern, characterized by an earlier onset in men than in women although the aggregate incidence rates (ie, over the age range of 15 to 54 years) were practically identical in the two sexes. This is different from the pattern of incidence of the acute psychoses, which show a proclivity to an earlier onset in women.
* The incidence rates of schizophrenia are similar across cultures. In the WHO study, annual incidence rates were determined for seven research sites in different cultures on the basis of 2 years of prospective case findings (Table 2). The rates were calculated separately for cases meeting a "broad" clinical definition of schizophrenia and for cases classified according to more restrictive criteria of the "nuclear" schizophrenic syndrome represented by the CATEGO class S + . For the broad definition of schizophrenia, the combined incidence rates for males and females ranged from 1.5 per 10 000 in Aarhus (Denmark) to 4.2 per 10000 in the rural area of Chandigarh (India) . The àifferences across the study areas were highly significant. However, the application of the restrictive definition resulted in lower mean rates, ranging from 0.7 in Aarhus to 1.4 in Nottingham, and to a disappearance of the statistical significance of the differences between the areas (which was not due to the loss of statistical power). Significant geographical variation in the frequency of schizophrenia will be found, of course, if point prevalence and not incidence is considered.
* There is a marked gender effect on the age-specific incidence rate of schizophrenia.The cumulative risk of developing schizophrenia up to the age of 54 is almost identical in males and females. However, the mean age of onset - especially in the period of peak incidence between ages 15 and 34 - is higher in females. This could mean that in the critical age period of reproduction, the rate of "consumption" of the total risk of developing the disorder is either slowed down in females or speeded up in males. Whatever the biological and social significance of this phenomenon, it is cross-culturally consistent and has been replicated in societies that differ markedly with regard to the attributes of the male and female roles. Sex differences in the incidence, manifestations, and prognosis of schizophrenia at present are becoming a focus of special interest.10,11
* Within the schizophrenic syndrome, two major subtypes can be dis tingu ished - sch izoph ren ia of acute onset and schizophrenia of gradual or insidious onset; this distinction explains a high proportion of the variance in course and outcome. In the WHO studies, the mode of onset, defined in terms of the time elapsed since the first appearance of psychotic symptoms and the point at which a recognizable clinical syndrome or symptom-complex is established, turned out to be the best predictor of the pattern of course, ie, of the likelihood that a patient would develop a remitting or a continuous type of illness. The strength of this relationship is of the same order in developing and developed countries, although the proportion of patients with an acute onset and, hence, with a higher probability of a remitting course was higher in non-European settings (Table 3).
Pattern of Course by Type of Onset and Setting (Percentages)
Moreover, the follow-up demonstrated that the different patterns of course, while showing a significant correlation with the mode of onset, tended to cluster at two extremes. At one extreme was the subset of cases of acute onset, in which the cumulative total length of psychotic episodes amounted to less than 15% of the follow-up period. At the other extreme was a subset comprising the majority of cases of an insidious onset, in which the patients tended to remain psychotic for over 75% of the follow-up period. Relatively few cases were of an intermediate type that fell between these two extremes, and the statistical distribution of the variable, "total time in psychotic episodes," was bimodal. It must be emphasized that at the point of the initial psychotic episode, the two groups of different subsequent course and outcome could not be distinguished from one another in terms of their symptoms. Neither the symptoms nor the history of the acute cases distinguished them consistently from the insidious onset cases, and no evidence supported the view that the acute onset/remitting course group might represent a separate class of "acute transient psychoses," which is etiologically different from the rest of the schizophrenic disorders.
A second group of epidemiological findings can be given a second rank in the hierarchy, not because of a lack of consistency or replicability, but as a function of a lower level of specificity. These findings eventually may contribute to a better understanding of the secondary, or pathoplastic, features of the disorder; however, it is unlikely that essential causative or pathogenetic mechanisms would be revealed.
* Culture and the proximal social environment have a significant effect on the course and outcome of schizophrenia. A striking finding of the WHO studies was that schizophrenic patients in traditional cultures (ie, in developing countries) had a significantly better 2-year and 5-year course and outcome than patients in the industrialized world (Table 4). This was true for several different and independent measures of the longitudinal aspect of the disorder, and the results of repeated multivariate statistical analyses demonstrated beyond doubt that the differences could not be explained by sampling bias, diagnostic variation, or unequal proportions of patients with acute and insidious onset in the centers in developing and developed countries.
It should be noted that in spite of the increasing refinement of the research design and the instruments used in consecutive WHO studies, the nature of the "culture effect" could not be explained fully. A modest but consistent impact of stressful life events was demonstrated in relation to the acute onset on psychotic episodes in different cultures, and differential levels of expressed emotion within the family were found to be associated with differences in the relapse rates in schizophrenic patients in India and Denmark. However, a more general effect of culture seems to exist that cannot be reduced to single explanatory variables. Together with the observation that schizophrenic disorders exhibit a great variety of patterns of course, the effects of culture reinforce the evidence that the course of schizophrenia is highly malleable by internalized or extrinsic environmental influences (or both). The reason for attributing a second rank in the epidemiological hierarchy to this finding is that it does not seem to have etiological implications and may not be specific to schizophrenia. In all likelihood, the social environment of traditional cultures favors symptom recovery and restoration of function in other mental disorders as well.
* Schizophrenia can occur as a symptomatic disorder in association with a variety of cerebral and physical diseases. Disorders with schizophrenic features occur in association with no less than 12 major groups of neurological, endocrine, and infectious diseases; intoxications; and intracranial lesions.12 Epilepsy is the most notable example,13 but recent additions to the list include idiopathic basal ganglia calcification14 and aqueduct stenosis.15 Many chromosomal or genetic disorders, ranging from XXX karyotype to acute intermittent porphyria, are associated with a significantly increased risk of schizophreniform psychosis.16 Rare conditions, such as metachromatic leukodystrophy, sometimes may be discovered by chance or through screening in patients diagnosed as schizophrenic.
Unfortunately, the search for cerebral or somatic pathologies associated with schizophrenic symptoms has not been systematic. Few of the observations collected so far lend themselves to epidemiological analysis. However, if the clinical characteristics of psychotic illness secondary to cerebral or physical disease eventually prove to be indistinguishable from "true" schizophrenia, then a number of discrete entities linked to specific genetic or exogenous causes eventually may be identified, resulting in a fragmentation of the nosological group and a progressive shrinkage of the share of idiopathic forms.
* Schizophrenia shows both positive and negative associations with other conditions and abnormalities. The Oxford Record Linkage Study17 has demonstrated an increased relative risk in schizophrenics for arteriosclerotic heart disease, for the malabsortion syndrome associated with celiac disease, and for myxedema. More interesting, however, are the negative associations. Among the diseases reported to occur in schizophrenic patients with a significantly lower incidence than expected, the best documented evidence pertains to rheumatoid arthritis.18 In contrast, ankylosing spondylitis and uroarthritis, both associated with the HLA B-27 antigen, are frequent among schizophrenics. These data point to an area for investigation of possible disease markers, which should be aided by modern record linkage techniques and molecular studies, to explore the relations among schizophrenia, joint disease, and the major histocompatibility complex.
In a recent WHO study, based on a record linkage between the Danish national psychiatric case register and the cancer registry (comprising a total of nearly 100 000 person-years), a significant reduction of the relative risk for cancer of different localization was found for schizophrenic patients.19 Subsequent data analyses20 indicated that cancer risk reduction was positively correlated with the length of neuroleptic treatment, which suggests a protective effect of phenothiazines with regard to malignant neoplasia.
* There is a consistent seasonality effect on the distribution of births of persons who subsequently develop schizophrenia. There seems to be a significant but modest excess of births in the season with the lowest average temperatures, regardless of latitude. It should be noted that seasonality of births also has been described in other disorders - for example in diabetes mellitus, malformations of the cardiovascular system, anencephaly, congenital hip dislocation, Down's syndrome, and bipolar affective disorder. The seasonality effect, therefore, is not specific to schizophrenia and regardless of the different causal explanations proposed (eg, viral infection, perinatal complications), is unlikely to be of great importance for understanding the nature of schizophrenia.
What conclusions can be drawn from the epidemiological evidence, if we assume that the proposed hierarchy of the conceptual significance of the findings is valid?
IMPLICATIONS OF THE EPIDEMIOLOGICAL EVIDENCE
First, the evidence will rule out, or at least render less plausible, certain models and theories of schizophrenia. For example, the evidence that the schizophrenic syndrome may be a manifestation of a horizontally transmissible viral or other infection, or of an autoimmune disease, is scant. At the other end of the etiological spectrum, theories of a pure psychogenic or sociogenic causation probably can be rejected.
In regard to genetics, current work on pedigrees and DNA recombinant techniques almost certainly will result in the identification of linkage polymorphisms and, eventually, gene systems that are significantly associated with an increased risk of clinical schizophrenia and related disorders. However, such findings will only signal the beginning, not the end, of the road. The unraveling of the intermediate links that fill the gap between the DNA locus and the clinical symptom is likely to be a lengthy and laborious process. Besides, there is the strong possibility that schizophrenia is both genetically and pathophysiologically heterogeneous. Facing a situation of such labyrinthine complexity, it would be natural to look for shortcuts that might be suggested by epidemiological approaches.
One such approach would be the search for explanatory models and possible analogues of schizophrenia in disorders that are at present better understood and "behave" epidemiologically in ways similar to schizophrenia. One such disorder that offers a certain analogy to schizophrenia is diabetes mellitus. Shepherd21 pointed out that until recently:
Little was known about its pathogenesis apart from the facts that the syndrome was hereditaiy, the genetics was probably multifactorial, and more than one disease process was represented by the clinical picture... The recent work on auto-immunity, HLA systems, and virology has led to a more searching mode of classification which, though still speculative, maps the complexity of the condition. If the promise of current biological research in psychiatry is eventually fulfilled, a model of this type may point the way to a more fundamental approach to the classification of the schizophrenias...
However, even if the recent progress in understanding diabetes can serve as a methodological model for research in schizophrenia, it is unlikely that the nosology of diabetes as such could serve as a prototype for the nosology of schizophrenia. Epidemioiogically, the incidence and prevalence distribution of diabetes across populations is quite different from that of schizophrenia. For example, while the differences in the incidence rate of schizophrenia between populations as distant from one another as Denmark and India are relatively minor, the difference in the incidence rate of diabetes between these two countries may be as large as 500-fold.
A closer analogy may be found in conditions that exhibit comparable levels of incidence, are ubiquitous in human populations, and as far as we know, are of a multifactorial causation. The epidemiological list of such conditions must include, for example, moderate and severe mental retardation, and in particular, epilepsy. Epilepsy, apart from its wellknown pathophysiological affinity to psychosis of a schizophreniform type, is one of the relatively small number of disorders occurring at comparable rates in very different populations.22 Although a major proportion of the variance of the total risk of developing epileptic seizures can be explained by a genetic predisposition, there are subgroups of "symptomatic" epilepsies due to exogenous insults and lesions. The ictal discharge seems to be a response modality appearing at a certain level of neurophysiological organization that can be activated by a variety of lesions and stimuli.
If schizophrenia is not a single disease but a "common final pathway" for a variety of pathological processes and neurodevelopmental anomalies, similar rates of its incidence in different populations could be seen to be a manifestation of a more or less uniformly distributed liability for a schizophrenic type of reaction to different causes. This liability must have a genetic basis and its expression is likely to be mediated by a dysregulation of the dopaminergic system operating under threshold conditions.
Such a model would be in agreement with the epidemiological data referred to above. It also would take us back to Kraepelin's idea that:
The affective and schizophrenic forms of mental disorder do not present the expression of particular pathological processes but rather indicate the areas of our personality in which these processes unfold.. ..It must remain an open question whether hereditary factors make certain areas more susceptible and accessible to pathological stimuli... .The various syndromes of illness may be compared to the different registers of an organ, any of which may be brought into play according to the severity and extent of the pathological changes involved. They impart a characteristic tone to the illness quite irrespective of the mechanism which has brought them into play.23
A last point concerns the much debated issue of the diagnostic definition of schizophrenia. Should this definition exclude, by introducing an arbitrary cut-off point, cases of recent onset that otherwise meet widely accepted descriptive criteria of the syndrome (as required in the DSM-IH system and, to a lesser degree, in the RDC)? While such exclusion might be justified in selecting homogeneous patient samples for genetic or other biological studies, it would be totally inappropriate for epidemiological research, which must take into account the entire range of variation, temporal and descriptive, of the condition in a given population.
In summing up, it is very likely that the epidemiology of schizophrenia eventually will find a new role in guiding neurobiological research to more clearly defined targets and in making sense of the multitude of data emerging from such research. However, the question of whether a neurodevelopmental, a "common final pathway," or any other reductionist model of pathogenesis is capable of providing us with the ultimate understanding of the phenomenological reality of schizophrenia will remain an open one for a long time. The subjective world of schizophrenia, which is the cornerstone of the diagnosis, may remain refractory to current explanatory models until the nature of the interface between the conscious mind and cerebral mechanisms becomes transparent to neuroscience.
1. Andrews G, Hall W, Goldstein G. Lapsley H, Bartels R, Silove D. The economic costs of schizophrenia. Arch Gen Psychiatry. 1985; 42:537-543.
2. Kraepelin E. Psychiatrie: Ein Lehrbuch für Stutherende und Aerate. Leipzig, Germany: Abel; 1899.
3. Shepherd M. Formulation of new research strategies on schizophrenia. In: Gattaz WF, Janzarik W, eds. Search for the Causes of Schizophrenia. Heidelberg, West Germany: Springer; 1987:29-38.
4. World Health Organization. Report of the International Pilot Study of Schizophrenia, Vol. I. Geneva, Switzerland: World Health Organization; 1973.
5. World Health Organization. Schizophrenia. An International Follow-up Study. Chichester, United Kingdom, John Wiley & Sons; 1979.
6. Jabiensky A, Schwarz R, Tomov T. WHO collaborative study on impairments and disabilities associated with schizophrenic disorders. Acta Psychiatr Scand. 1980; 62:(suppl 285): 152-163.
7. Sartorius N, Jablensky A, Korten A, et al. Early manifestations and first-contact incidence of schizophrenia in different cultures. Psychol Med. 1986; 16:909928.
8. Jablensky A. Multicultural studies and the nature of schizophrenia: a review. J R Soc Med. 1987; 80:162-167.
9. Stevens IR, Wyatt RJ. Similar incidence worldwide of schizophrenia: case not proven. Br J Psychiatry. 1987; 151:131-132. Letter to the Editor.
10. Seeman MV. Gender differences in schizophrenia. Can J Psychiatry. 1982; 27:107-112.
11. Loranger AW. Sex difference in age at onset of schizophrenia. Arch Gen Psychiatry. 1984; 41:157-161.
12. Davison K, Bagley CR. Schizophrenialike psychoses associated with organic disorders of the central nervous system : a review of Ae literature. In: Herrington RN, ed. Current Problems in Neuropsychiatry. British Journal of Psychiatry Speical Publication No. 4. Ashford, United Kingdom: Headley Brothers; 1969:113184.
13. Flor-Henry P. Cerebral Basis of Psychopathology. Boston, Mass: John Wright; 1983.
14. Francis A, Freeman H. Psychiatric abnormality and brain calcification over four generations. J Nerv Ment Dis. 1984; 172:166-170.
15. Roberts JKA, Trimble MR, Robertson M. Schizophrenic psychosis associated with aqueduct stenosis in adults. J Neurol Neurosurg Psychiatry. 1983; 46:892-893.
16. Propping P. Genetic disorders presenting as "schizophrenia." Karl Bonhoeffer's early view of the psychoses in the light of medical genetics. Hum Genet. 1983; 65:1-10.
17. Baldwin JA. Schizophrenia and physical disease. Psychol Med. 1979; 9:611-618. Editorial.
18. Osterberg E. Schizophrenia and rheumatic disease. Acta Psychiatr Scand. 1978; 58:339-359.
19. Dupont A, Moeller-Jensen O, Stromgren E, Jablensky A. Incidence of cancer in patients diagnosed as schizophrenic in Denmark. In: ten Horn GHMM, Giel R, Gulbinat W, Henderson JH, eds. Psychiatric Case Registers in Public Health. Amsterdam, Netherlands: Elsevier; 1986:229-239.
20. Mortensen PB. Environmental factors modifying cancer risk in schizophrenia. Presented at the World Psychiatric Association Symposium; August 1986; Copenhagen, Denmark.
21. Shepherd M. Contributions of epidemiological research to the classification and diagnosis of mental disorders. In: Mental Disorders, Alchohol- and Drug-related Problems. International Perspectives in Their Diagnosis and Classification. Amsterdam. Netherlands: Elsevier; 1985:337-341. Excerpta Medica International Congress Series No. 669.
22. Neugebauer R, Susser M. Epilepsy: some epidemiological aspects. Psychol Med. 1979; 9:207-215.
23. Kraepelin E. Die Erscheinungsformen des Irreseins. Zeitschrift fur die gesamte Neurologie und Psychiatrie. 1920; 62:1-29.
The WHO Program of Cross-Cultural Research in Schizophrenia (1967-85)
Annual Incidence Rates per 10 000 Population, Age 15-54, for IWo Diagnostic Definitions
Pattern of Course by Type of Onset and Setting (Percentages)