Epidemiology of Schizophrenia

The epidemiology of schizophrenia

Topics covered:

  • Introduction
  • Epidemiological methods and instruments used in the study of schizophrenia
    • Case-finding
    • Diagnosis
    • Instruments
  • Persons, place, time: completing the description of schizophrenia
    • Prevalence, incidence, and disease expectancy
    • Associations with age and sex
    • Fertility, mortality, and comorbidity
    • Geographical and cultural variation
    • The disease and disability burden of schizophrenia
    • Secular trends: a decreasing incidence of schizophrenia?
    • Factors maintaining the incidence of schizophrenia in populations
  • An epidemiological perspective on antecedents and risk factors
    • Genetic risk: necessary and sufficient?
    • Environmental insults at the early developmental stages
    • The preschizophrenic person
    • The social and family environment
  • Epidemiological issues for the next decade
    • Does the categorical disease concept of schizophrenia constrain aetiological research?
    • Molecular epidemiology of schizophrenia
    • Can schizophrenia be prevented?
  • Summary and conclusions
  • References


Epidemiological research into schizophrenia aims to answer four essential questions.

  1. What is the ‘true' population frequency of the disorder in various populations and how is it distributed across the various groups within populations?
  2. Do the incidence, manifestations, and course of schizophrenia vary in relation to factors of the physical and social environment?
  3. Who is at risk and what forces determine or influence the risk of developing schizophrenia? 
  4. Can the answers to the above questions help explain what causes the disorder and how to prevent it?

The hallmark of the epidemiological method is the referral of any numerical findings about the occurrence of a disorder or its associated characteristics to a population base or denominator, such as person-years. The epidemiological study of a disorder usually proceeds from a description of its frequency and associations (establishing rates of occurrence) to testing hypotheses about risk and causes, using analysis of ratios between rates.

Schizophrenia has been studied extensively from an epidemiological perspective over the past hundred years. (1) Kraepelin, who introduced the concept of dementia praecox in 1896, anticipated the potential of the epidemiological method to ‘throw light on the causes of mental disorder' and proposed comparative population studies of the psychoses and factors predisposing to them across different cultures. (2) In the first half of the twentieth century, epidemiological research into schizophrenia took two divergent paths. While European studies tended to have a strong focus on genetic risks, North American researchers investigated the social ecology of the disorder. A variety of techniques and study designs were explored and applied with considerable success by the pioneers of psychiatric epidemiology, and the contours of the epidemiological map of schizophrenia in Europe and North America were effectively laid down between the two World Wars. The early epidemiological studies were carried out by dedicated researchers who typically spent much of their time collecting data ‘door to door' in small communities. Intimate knowledge of the respondents, access to multigenerational records from the local parish registers, and the co-operation of the community often resulted in studies that remain landmarks of psychiatric epidemiology.

During the last three decades, the scope of epidemiological studies of schizophrenia has expanded to include populations in Asia, Africa, and South America about which little had been known previously. The World Health Organization ( WHO) International Pilot Study of Schizophrenia and the subsequent WHO 10-country epidemiological study (10,11) were the first systematic investigations of the comparative incidence, clinical manifestations, and course of schizophrenia in both developing and developed countries. The WHO programme provided an impetus for similar studies to be undertaken in India, China, the Caribbean, and Australia. Two major studies of psychiatric morbidity in the United States, the Epidemiological Catchment Area project and the National Comorbidity Survey, generated data on the prevalence of DSM-III/IIIR schizophrenia and related disorders in representative population samples. (12,13) Recent studies, initiated in the 1980s and 1990s, tend to make use of large existing databases such as cumulative case registers or birth cohorts to test hypotheses about risk factors in case–control designs, or include methods of genetic epidemiology within population-based studies. There is a tendency towards integrating epidemiological approaches with other types of aetiological research in schizophrenia. (14) This predicts an important role for epidemiology in the coming era of molecular biology of mental disorders.

Epidemiological methods and instruments used in the study of schizophrenia

The measurement of the prevalence, incidence, and disease expectancy of schizophrenia depends critically on the sensitivity of the case-finding method (i.e. its capacity to identify all affected persons in a given population) and the availability of a diagnostic instrument or procedure that selects ‘true' cases (i.e. those corresponding to an established clinical concept).


Case-finding designs fall into three broad groups: case detection in clinical populations (persons in contact with services), door-to-door surveys (including census investigations of entire communities and surveys of population samples), and birth cohort studies. Each method has its advantages and limitations.

While case-finding through the mental health services provides a relatively easy access to a substantial proportion of all persons with schizophrenia, the cases currently in treatment may not be a representative sample of all individuals with the disorder. Different kinds of bias related to gender, marital status, socio-economic factors, culture, or ethnicity are known to affect the probability of being in treatment at any given time in any given setting. Unless such bias is accounted for, generalizations about schizophrenia based on hospital or clinic samples are liable to error. Many of the deficiencies of case-finding through service contacts can be overcome by using cumulative regional or national psychiatric case registers, where such tools exist. Registers cover defined populations and the capacity for linkage to other databases makes them effective research instruments in low-incidence disorders such as schizophrenia.

Surveys essentially involve accounting for every person at risk within a defined community or a population sample in terms of either being or not being a case. Face-to-face interviewing (and following up) every resident in delimited, usually small, communities has been a feature of intensive high-quality research, especially in the Scandinavian countries. However, since the size of the population that can be surveyed in this way is limited, the number of detected cases of schizophrenia is usually too small to generate stable estimates of epidemiological parameters. Single-community studies are also vulnerable to the so-called ‘ecological fallacy' (erroneously inferring that an association observed at the level of the community—say, between the number of pregnant women exposed to influenza epidemics and the number of offspring eventually developing schizophrenia—will also hold at individual level). Therefore surveys of large populations have advantages. There are two basic designs: a single-phase survey of a probability sample drawn from the general population, and a two-phase survey where a screening test is applied to the entire population at risk and only those scoring as screen-positive proceed to a full assessment. In the instance of schizophrenia, mere logistics dictates a choice between assessing large numbers less rigorously and investigating a smaller sample in greater depth. In the absence of a simple and valid screening procedure for schizophrenia, such as a biological or psychological test, the advantages of the two-phase survey may be offset by poor sensitivity or specificity of the screening device which is usually a questionnaire or checklist.

The study of birth cohorts at ages when their members have passed through the greater portion of the period of risk of developing schizophrenia is accomplished either by direct interviewing or by analysing available case register data. If birth cohorts are well characterized, they are the best tools for the study of the incidence of schizophrenia in relation to specific risk factors. This presumes that the population is relatively stable and that an adequate infrastructure is in place for monitoring mortality and morbidity. However, even in settings where such conditions are met, the size of birth cohorts with prospectively collected research data of interest may not be sufficient for conclusive epidemiological inferences.

All this suggests that there is no single gold standard of epidemiological case-finding methods for schizophrenia that could be uniquely applied across all possible settings and situations. Rather, the assets and liabilities of case-finding procedures need to be evaluated in the context of each particular study. This makes the detailed reporting of case-finding methods and procedures a mandatory prerequisite for an ‘evidence-based' epidemiology of schizophrenia.


Variation in diagnostic concepts and practices always explains a certain proportion of the variation in the results of schizophrenia studies, especially if they involve different populations or different periods. Until the late 1960s, the diagnostic rules used in epidemiological research were seldom explicitly stated. The demonstration by the United States–United Kingdom diagnostic study,(15) that American psychiatrists used a broader definition of schizophrenia than their British counterparts, reinforced suspicions that concepts of schizophrenia in different medical cultures could differ to an extent that might invalidate comparisons.

In response to such concerns, the WHO launched the International Pilot Study of Schizophrenia, (10) which examined diagnostic variation across nine countries by comparing the diagnoses made by psychiatrists using a semistructured clinical interview with a standard reference classification by computer algorithm (16) utilizing the same clinical data. The results were reassuring since in the majority of settings psychiatrists were found to use comparable diagnostic concepts corresponding to the definition of schizophrenia in accordance with the Kraepelin–Bleuler tradition. Furthermore, the core of the diagnostic concept of schizophrenia does not seem to have undergone major changes over time. A reanalysis of a sample from Kraepelin's original clinical material demonstrated that descriptions of dementia praecox and manic–depressive psychosis cases of 1908 could be scored and diagnosed using ‘modern' syndrome scales with a resulting agreement of 88.6 per cent between the 1908 diagnosis and the ICD-9 diagnosis assigned by computer. (17) The introduction of explicit diagnostic criteria and rules with the consecutive editions of DSM and the WHO's ICD-10 has resolved some, but not all, diagnostic problems with implications for epidemiology. While ICD-10 (18) and DSM-IV (19) tend to agree well on the core cases of schizophrenia, they agree less well on the classification of atypical or milder cases. Such differences may be less important in clinical practice than in epidemiological and genetic studies. By providing somewhat restrictive criteria for the diagnosis of schizophrenia, both classifications aim to select homogeneous patient groups and to minimize false-positive diagnoses. However, this is not an unequivocal advantage for epidemiology. Applying such criteria at the case-finding stage of a survey may result in the rejection of potential cases which fail to satisfy the full set of criteria at the time of initial assessment. Therefore it is desirable to develop less restrictive screening versions of the DSM-IV and ICD-10 criteria for epidemiological research.

Diagnostic problems may also arise in studies using ‘historical' databases without direct patient contact, such as information from case registers or birth cohort data. The validity of the original diagnoses in such databases is difficult to ascertain, but several studies in which patient samples have been assessed by research interviews suggest that, in the instance of schizophrenia, serious discrepancies between register and research diagnoses are relatively rare.


The diagnostic instruments used in surveys involving interviewing of respondents fall into two categories: fully structured interviews such as the Diagnostic Interview Schedule (DIS) (12) and the Composite International Diagnostic Interview ( CIDI), (20) written to match exactly the diagnostic criteria of DSM-IIIR/IV and ICD-10, and semistructured interview schedules such as the Present State Examination (PSE) (16) and the Schedules for Clinical Assessment in Neuropsychiatry (SCAN), (21) which cover a broad range of psychopathology and elicit data that can be processed by alternative diagnostic algorithms. While the former can be applied by trained lay interviewers, the latter require clinical judgement and are usually administered by clinicians.

The DIS/CIDI type of instrument is reliable and capable of generating standard diagnoses in a single-phase survey design. However, the range of psychopathology covered is restricted to the diagnostic system to which it is linked, and its clinical validity is open to question because symptoms may not be reported accurately or impairment may be underestimated by the respondent. In contrast, the PSE/SCAN type of interview allows a greater amount of psychopathological data to be elicited but its use in epidemiological studies presupposes the availability of clinically skilled and trained interviewers. While SCAN and other similar interviews are suitable as second-stage diagnostic instruments, there is still a need for a relatively simple and effective screening procedure for case-finding of schizophrenia in field surveys.

Persons, place, time: completing the description of schizophrenia

At present, the epidemiological description of schizophrenia draws on extensive evidence regarding its frequency and age and sex distribution in relatively large populations or population groups. However, less than complete information is available on variations in its epidemiological characteristics that may be found in unusual or isolated populations or on the temporal trends in its occurrence.

Prevalence, incidence, and disease expectancy


Prevalence is defined as the number of cases per 1000 persons at risk present in a population at a given time or over a defined period. Point prevalence refers to the ‘active' (i.e. symptomatic) cases on a given date, or within a brief census period. Since asymptomatic cases (e.g. persons in remission) will be missed in a point prevalence survey, it is useful to supplement the assessment of the present mental state with an enquiry about past episodes of the disorder to obtain a lifetime prevalence index. In schizophrenia, which tends to a chronic course, estimates of point and lifetime prevalence will be closer to each other than in remitting illnesses.

The majority of studies have produced prevalence estimates in the range 1.4 to 4.6 per 1000 population at risk. However, these are usually crude prevalence figures that are difficult to compare because of the demographic differences between populations with regard to factors such as age-specific mortality and migration. Therefore the modal prevalence of 1.4 to 4.6 per 1000 is unlikely to reflect the extent of variation that exists in the prevalence of schizophrenia in different populations.

There are populations and groups which deviate from the central tendency. Unusually high rates (two to three times the national or regional rate) have been reported for several genetic isolates such as an area in northern Sweden and several areas in Finland, and for an area in Croatia characterized by a high level of out-migration during the nineteenth and early twentieth centuries. At the other extreme, a virtual absence of schizophrenia and a relatively high rate of depression has been claimed for the Hutterites of South Dakota, a Protestant sect whose members live in close-knit endogamous communities sheltered from the outside world. (37) Negative selection for schizoid individuals who fail to adjust to the lifestyle of the majority and eventually migrate without leaving progeny has been suggested (but not definitively proven) as an explanation.

Low prevalence rates have also been reported for certain Pacific island populations, (38) but uncertainties about the extent of case-finding makes the interpretation of such reports problematic. Two carefully planned surveys in Taiwan were separated by 15 years during which major social changes had taken place. While the total mental morbidity increased, the prevalence of schizophrenia decreased from 2.1 to 1.4 per 1000. In both surveys, the aboriginal Taiwanese had significantly lower rates than the mainland Chinese who had migrated to the island after the Second World War.

The results of the Epidemiologic Catchment Area study in the United States, (12) which indicate a higher prevalence rate compared with most other studies, should be treated with caution. There are inconsistencies among the study areas that are difficult to interpret (such as a 13-fold difference in the rates for age group 18–24 years across the sites), which suggests that the diagnostic procedure, involving the DIS administered by lay interviewers, may have elicited a number of false-positive diagnoses. In the more recent National Comorbidity Survey, diagnoses of ‘non-affective psychosis' by computer algorithm based on a version of the CIDI were found to agree poorly with clinicians' diagnoses based on telephone reinterviews, resulting in discrepant estimates of the lifetime prevalence of both ‘narrowly' and ‘broadly' defined psychotic illness. (39)

The question of whether major differences exist in the prevalence of schizophrenia in different populations has no simple answer. In the great majority of studies the prevalence rates are similar. On the other hand, there are a small number of populations that clearly deviate from the average. However, the magnitude of these deviations is modest compared with the 10- to 30-fold differences in prevalence observed in other multifactorial diseases (e.g. ischaemic heart disease, diabetes, multiple sclerosis) across populations.


The incidence rate (the annual number of first onset cases in a defined population per 1000 individuals at risk) is of greater interest for the search of risk factors than the prevalence since it is a better estimate of the so-called force of morbidity (the probability of disease occurrence at a point in time) in a given population. Its estimation depends on how reliably the point of onset can be determined. Since it is not possible at present to determine the time of onset of any cerebral dysfunction or biochemical lesion underlying schizophrenia, onset of the disorder is usually defined as the point in time when its clinical manifestations become recognizable and diagnosable according to specified criteria. The first hospital admission, which has been used in many studies, is not a good approximation to the ‘true' onset because of the variable time lag between the earliest appearance of symptoms and the first admission. A better approximation is provided by the first contact, i.e. the point at which any psychiatric or general medical service is accessed by symptomatic individuals for the first time.

The comparison of studies using a ‘broad' definition of schizophrenia (e.g. ICD-8 or ICD-9) suggests that the variation in rates based on first admissions or first contacts is about threefold, in the range from 0.17 to 0.54 per 1000 population per year. Studies using restrictive criteria such as the Research Diagnostic Criteria ( RCD), (51) DSM-III and its successors, and ICD-10 have reported incidence rates that are two to three times lower than those based on ‘broad' criteria.

To date, the only study that has generated directly comparable incidence data for different populations is the WHO 10-country investigation. (11) Incidence counts in the WHO study were based on first-in-lifetime contacts with any ‘helping agency' (including traditional healers in the developing countries) monitored prospectively over a 2-year period. Potential cases and key informants were interviewed by clinicians using standardized instruments, and the timing of onset was ascertained for the majority of the patients. In 86 per cent of the 1022 patients the first appearance of diagnostic symptoms of schizophrenia was within the year preceding the first contact, and therefore the first-contact incidence rate was accepted as a reasonable approximation to the onset rate. Two definitions of ‘caseness', differing in the degree of specificity, were used to determine incidence: a ‘broad' clinical class comprising ICD-9 schizophrenia and paranoid psychoses, and a restrictive definition including only ‘nuclear' schizophrenia manifesting with so-called Schneiderian first-rank symptoms. (52) 

The differences between the rates for ‘broadly' defined schizophrenia (0.16–0.42 per 1000) were significant ( p < 0.001, two-tailed test); however, those for ‘nuclear' schizophrenia were not. Since the mean incidence rates for the ‘nuclear' cases were lower than the mean rates for ‘broadly' defined schizophrenia, the confidence intervals for the lower rates would be expected to be wider than the confidence intervals for the higher rates. In fact, the opposite was true, which suggests that ‘nuclear' schizophrenia comprises a more homogeneous subset of cases than the ‘broad' diagnostic class and occurs with a similar frequency in these different populations. However, no differences were found between cases meeting the ‘broad' criteria and the ‘nuclear' cases with regard to either age or type of onset (acute or insidious), or 2-year course and outcome. Therefore it is unlikely on clinical grounds that ‘nuclear' and ‘broadly' defined cases represent two different disorders.

In recent years, replications of the design of the WHO 10-country study, including its research instruments and procedures, have been carried out with very similar results by investigators in India, the Caribbean, and the United Kingdom.

Disease expectancy (morbid risk)

This is the probability (usually expressed as a percentage) that an individual born into a particular population will develop the disease if he or she survives the period of risk for that disease. In the instance of schizophrenia the period of risk is defined as either 15 to 44 or 15 to 54 years. If the age- and sex-specific incidence rates are known, disease expectancy can be estimated directly by a summation of the age-specific rates within the period of risk. Disease expectancy can be estimated indirectly from prevalence data. The estimates of risk produced by a number of studies are fairly consistent across populations and over time. Excluding outliers, such as the northern Swedish isolate, the risk ratio (ratio of highest to lowest disease expectancy) is about 5.0; for the WHO study (11) it is 2.9 (‘broad' diagnostic class) and 2.0 (‘nuclear' schizophrenia). The frequently cited ‘rule of thumb' estimate of disease expectancy for schizophrenia at around 1 per cent (or slightly less) seems to be consistent with the evidence.

Associations with age and sex

There is adequate evidence that schizophrenia may have its onset at any age—in childhood as well as past middle age—although the vast majority of onsets indisputably fall within the interval 15 to 54 years of age. Onsets in men peak steeply in the age group 20 to 24 years; thereafter the rate of inception remains more or less constant at a lower level. In women, a less prominent peak in the age group 20 to 24 years is followed by another increase in incidence in age groups older than 35. While the age-specific incidence up to the mid-thirties is significantly higher in men, the male-to-female ratio becomes inverted with age, reaching 1 : 1.9 for onsets after age 40 and 1 : 4 or even 1 : 6 for onsets after age 60. There seems to be no real ‘point of rarity' between the symptomatology of late-onset schizophrenia and schizophrenia of an early onset.

The sex differences in mean age at onset are unlikely to be an invariant biological characteristic of schizophrenia. For example, within families with two or more affected members, significant differences in age at onset have not been found between male and female siblings with schizophrenia. (53) In some populations (e.g. India) the male–female difference in the frequency of onsets in the younger age groups is attenuated or even inverted. (54)

The question of whether the total lifetime risks for men and women are about the same, or different, has not been answered definitively. In the WHO 10-country study, the cumulated risks for males and females up to age 54 were found to be approximately equal. Some Scandinavian population-based studies which followed up cohorts at risk into a very old age (over 85) reported a higher cumulated lifetime risk in women than in men. (55)

Male–female differences have been described in relation to the premorbid history (better premorbid functioning in women), the occurrence of brain abnormalities (more frequent in men), course (a higher percentage of remitting illness episodes and shorter hospital stay in women), and outcome (higher survival rate in the community, less disability in women). However, there is no unequivocal evidence of consistent sex differences in the symptom profiles of schizophrenia, including the frequency of positive and negative symptoms. Generally, the sex differences described in schizophrenia are more likely to result from the normal sexual dimorphism in brain development, as well as from gender-related social roles, than from sex-specific aetiological factors.

Fertility, mortality, and comorbidity


Earlier studies have described a low fertility in both men and women diagnosed as schizophrenic. The average number of children fathered by schizophrenic men in Sweden was 0.9, and the average number of live births over the entire reproductive period of women treated for schizophrenia in Norway between 1936 and 1975 was 1.8, compared with 2.2 for the general female population. (56) Similar results have been reported from Germany. (57) Yet this phenomenon does not seem to be either universal or consistent over time. For example, according to the WHO 10-country study, (11) the fertility of women with schizophrenia in India did not differ from that of women in the general population within the same age groups and geographic areas. An increase in the fertility of schizophrenic women in more recent decades has been noted and is likely to be sustained as a result of the deinstitutionalization of the mentally ill. Although men with schizophrenia continue to be reproductively disadvantaged, at least one study (58) has found a higher than average fertility among married schizophrenic men. Several studies have examined the fertility of biological relatives of probands with schizophrenia. The results are not entirely consistent, but some of these studies have pointed to a higher than average fertility in parents and siblings of schizophrenic patients. (59)


Excess mortality among schizophrenic patients has been extensively documented by epidemiological studies on large cohorts. National case register data for Norway indicate that, while the total mortality of psychiatric patients decreased between 1926–1941 and 1950–1974, the relative mortality of patients with schizophrenia remained unchanged at a level more than twice that of the general population. (9) Results from cohort and record linkage studies in other European countries and North America are similar, reporting standardized mortality ratios of 2.6 or higher for patients with schizophrenia, which corresponds to a reduction in life expectancy of about 20 per cent compared with the general population. A meta-analysis of 18 studies (60) resulted in a crude mortality rate of 189 deaths per 10 000 population per year and a 10-year survival rate of 81 per cent. Mortality among males was significantly higher than among females, and the difference was almost entirely due to an excess in accidents and suicides in males. Unnatural causes apart, the leading causes of death among schizophrenic patients are similar to those in the general population, with the possible exception of a lower than expected cancer mortality in males with schizophrenia. (61) The latter phenomenon has been replicated by case register and record linkage studies (62) and does not appear to be an artefact. Its causes remain unknown.

Data on successive Danish national cohorts (63) suggest an alarming trend of increasing mortality in first-admission patients with schizophrenia. The 5-year cumulated standardized mortality ratios increased from 5.30 (males) and 2.27 (females) between 1971 and 1973 to 7.79 (males) and 4.52 (females) between 1980 and 1982. Particularly striking was the standardized mortality ratio of 16.4 for male schizophrenics in the first year after the diagnosis is made. Similar data on high mortality after a first episode of schizophrenia have been reported in several other studies. The single most common cause of death among schizophrenic patients at present is suicide (aggregated standardized mortality ratios 9.6 for males and 6.8 for females) which accounts, on the average, for 28 per cent of the excess mortality in schizophrenia. (63) The actual mortality due to suicide is likely to be higher, since a proportion of the deaths classified as accidental or of undetermined cause are probably suicides. Thus the suicide rate in schizophrenic patients is at least equal to, or may be higher than, the suicide rate in major depression. Several risk factors have been suggested as relatively specific to schizophrenic suicide: being young and male, experiencing a chronic disabling illness with multiple relapses and remissions, realistic awareness of the deteriorating course of the condition, comorbid substance use, and loss of faith in treatment. (64) Data from Scotland (65) point to a trend of increasing suicide rate in schizophrenic patients, mostly within the first year after discharge. This trend seems to parallel the significant reductions in the number of psychiatric beds. Whether this new wave of increasing suicide mortality can be attributed to the transition from hospital to community management of schizophrenia remains to be established.


Comorbidity in schizophrenia comprises, above all, relatively common medical problems and diseases that tend to occur among schizophrenic patients more frequently than the expected chance association. Further, certain rare conditions or abnormalities tend to co-occur with schizophrenia.

Substance abuse is by far the most common comorbid problem among schizophrenic patients (66) and may involve alcohol, stimulants, benzodiazepines, hallucinogens, and antiparkinsonian drugs, as well as caffeine and tobacco. In the WHO 10-country study, (11) a history of alcohol use in the year preceding the first contact was elicited in 57 per cent of the male patients, and in three of the study areas drug abuse (mainly cannabis and cocaine) was reported by 24 to 41 per cent of the patients. Cannabis abuse, which exacerbates the symptoms and may precipitate relapse, was a significant predictor of poor 2-year outcome. The prevalence of cigarette smoking among schizophrenia patients is, on the average, two to three times higher than in the general population, taking sex and age group into account. Recent research on the interactions between the nicotinic receptor and the glutamatergic and dopaminergic systems has led to the hypothesis that smoking in schizophrenic patients could be an attempt at self-medication reinforced by the short-term normalizing effect of nicotine on certain neurocognitive deficits such as defective sensory gating. (67)

Physical disease is common but is rarely diagnosed. Between 46 and 80 per cent of inpatients with schizophrenia, and between 20 and 43 per cent of outpatients, have been found in different surveys to have concurrent medical illnesses, including potentially life-threatening conditions in 7 per cent. (68) In addition to increased susceptibility to infection, and especially pulmonary tuberculosis, schizophrenic patients tend to have a higher than expected rate of common diseases, including diabetes, arteriosclerosis disease, and ischaemic heart disease, (69) as well as some rare genetic or idiopathic disorders such as acute intermittent porphyria and coeliac disease. A lower than expected rate of rheumatoid arthritis has repeatedly been found in schizophrenic patients, (70) although one recent record linkage study (71) failed to replicate this. Numerous studies have reported a significantly increased frequency of dysmorphic features and minor physical anomalies, including high-steepled palate, malformed ears, epicanthus, single palmar crease, and finger and toe abnormalities, which may result from deviations in fetal development during the first gestational trimester. (72)

A variety of rare organic brain disorders and anomalies have been described as occurring in association with schizophrenia, including basal ganglia calcification, aqueductus Sylvii stenosis, corpus callosum agenesis, septal cysts, and cerebral hemiatrophy. However, most of these associations have been reported from single case studies; epidemiological evidence of a higher than chance co-occurrence with schizophrenia has so far only been provided for epilepsy (73) and metachromatic leucodystrophy. (74)

Geographical and cultural variation

To date, no population or culture has been identified in which schizophrenic illnesses do not occur. Also, there is no strong evidence that the incidence of schizophrenia varies widely across populations, provided that the populations being compared are large enough to allow a low-incidence disorder such as schizophrenia to ‘breed'. The evidence that psychosocial factors or culture play an aetiological role in schizophrenia is also weak. However, there are well-replicated findings of some significant variation in the course and outcome of schizophrenia across populations and cultures which involves, above all, a higher rate of symptomatic recovery and a lower rate of social deterioration in traditional rural communities. Data supporting this conclusion were provided by the WHO studies (10,11) which found a higher proportion of recovering or improving patients in developing countries such as India and Nigeria than in the developed countries. Sampling bias (e.g. a higher percentage of acute-onset schizophreniform illnesses of good prognosis among Third World patients) was not a likely explanation. A better outcome in the developing countries was found in patients with various modes of onset, and the initial symptoms of the disorder did not distinguish good-outcome from poor-outcome cases. What causes such differences in the prognosis of schizophrenia remains largely unknown. The follow-up in the WHO studies demonstrated that outcome of paranoid psychoses and affective disorders was also better in the developing countries. Such a general effect on the outcome of psychiatric disorders may result from psychosocial factors, such as availability of social support networks, non-stigmatizing beliefs about mental illness, and positive expectations during the early stages of psychotic illness, from unknown genetic or ecological (including nutritional) factors influencing brain development, or from an interaction between cultural and biological factors.

The disease and disability burden of schizophrenia

According to World Bank and WHO estimates, (75) no less than 25 per cent of the total ‘burden of disease' in the established market economies is at present attributable to neuropsychiatric conditions. Measured as proportion of the disability-adjusted life-years (DALYs) lost, schizophrenia, bipolar affective disorder, and major depression together account for 10.8 per cent of the total, i.e. they inflict on most communities losses that are comparable to those due to cancer (15 per cent) and higher than the losses due to ischaemic heart disease (9 per cent).

Secular trends: a decreasing incidence of schizophrenia?

A number of studies have suggested that a decrease of 40 per cent or more may have occurred in the first admission rates for schizophrenia over the last 30 years in some of the developed countries. (76) The data are not entirely consistent; the drop in rates is more marked in females or in late-onset cases according to some reports, is more pronounced in young males according to others, or is of the same magnitude in both sexes and in all age groups. Downward trends have been identified mainly by using national or regional admission and discharge statistics. Attempts to replicate such trends on a local level using case register data have not produced consistent results, while studies in which research diagnoses were made after a case review show no decline in incidence rates. Other concurrent changes, such as a reduction in the number of psychiatric beds and in the total number of first admissions, have also been noted in many of these studies. However, increases have been reported in the mortality of schizophrenic patients, in first admissions with diagnoses of borderline states, paranoid psychosis, and reactive psychosis, and in the proportion of patients with schizophrenia who are managed without hospital admission.

Various hypotheses have been proposed to explain the putative decrease in the incidence of schizophrenia, but no study to date has taken into account all confounding factors that may influence first admission or first contact rates. These include changes in the administrative definition of ‘first admission', and changes in diagnostic practices that may have resulted in a greater reluctance among clinicians to diagnose schizophrenia on first admission. Since schizophrenia has a low incidence, its first admission rate would be highly sensitive to such ‘diagnostic drift'. If changes in diagnostic habits play a role, a compensatory increase could be expected to have occurred in diagnoses other than schizophrenia. Such increases have actually been reported for borderline states and affective psychosis. All things considered, the case for a true decline in the incidence rate of schizophrenia is suggestive but not proven.

Factors maintaining the incidence of schizophrenia in populations

Since schizophrenic patients have a reduced rate of reproduction, compared with the general population, the maintenance of a relatively stable rate of schizophrenia in the population requires an explanation. A high mutation rate can practically be ruled out since the rate required for a polygenic disorder would exceed by far the theoretically possible mutation rate. Selective advantages of individuals with schizophrenia (balanced polymorphism) (77) that could offset their low fertility have been suggested, such as resistance to physiological stress or certain diseases, and an increased fertility in their biological relatives has been postulated. None of these hypotheses is at present supported by empirical data, although some evidence of normal or increased fertility among relatives of schizophrenic patients has been reported. However, the whole argument that, in the absence of a reproductive compensation for the low fertility of schizophrenic individuals, the disorder will gradually disappear may be fallacious. (78) The concept of balanced polymorphism, developed to explain the population dynamics of single-gene Mendelian diseases may not be applicable to genetically complex disorders such as schizophrenia where both multiple genetic loci and multiple exogenous factors are likely to be involved. If this is the case, there is no phenomenon requiring an explanation.

An epidemiological perspective on antecedents and risk factors

Studies on clinical samples have proposed a number of possible risk factors in schizophrenia. Considering that clinical samples are vulnerable to bias and may not be representative, epidemiological evidence should be sought to evaluate the significance of such findings. Genetic and environmental risk factors are considered further in here: Genetic and environmental risk factors for schizophrenia

Genetic risk: necessary and sufficient?

Family aggregation of schizophrenia is at present the only epidemiologically well-established risk factor for the disorder, with a relative risk for first-degree relatives of schizophrenics in the range from 9 to 18. Allowing for diagnostic variation, the risk estimates generated by different studies are similar and suggest a general pattern of descending risk as the proportions of shared genes between any two individuals decrease. (79) The weight of the evidence also suggests that genetic vulnerability is necessary but not sufficient to cause schizophrenia and that environmental factors must play a role. However, the evidence for an environmental contribution to the aetiology of schizophrenia remains indirect, stemming primarily from the observation that the concordance for schizophrenia in monozygotic twin pairs is only about 50 per cent. Three general models of the joint effects of the genotype and the environment have been proposed: (80) 1. the effects of predisposing genes and environmental factors are additive and increase the risk of disease in a linear fashion; 2. genes control the sensitivity of the brain to environmental insult; 3. genes influence the likelihood of an individual's exposure to behavioural pathogens, for example by fostering some personality traits. Current epidemiological research into possible environmental causes of schizophrenia focuses on three main areas: pre- and perinatal damage, factors affecting early brain development, and factors operating at the level of the social and family environment. (See also: Genetic and environmental risk factors for schizophrenia.)

Environmental insults at the early developmental stages

Obstetric complications

Maternal obstetric complications are widely cited as an established risk factor in schizophrenia. (81) Several explanatory models have been proposed.

  1. Severe obstetric complication, such as perinatal hypoxia and resulting hippocampal damage, can prepare the ground for adult schizophrenia even if genetic liability is weak or absent.
  2. A genetic predisposition sensitizes the developing brain to lesions resulting from randomly occurring less severe obstetric complications.
  3. A genetic predisposition to schizophrenia leads to abnormal fetal development which in turn causes obstetric complications.
  4. Maternal constitutional factors, partially influenced by genes, such as small physique or proneness to risk behaviour (drug use, smoking during pregnancy) increases the risk of obstetric complications and fetal brain damage.

None of these models has been directly tested and the association between obstetric complications and adult schizophrenia remains tenuous. (82) The majority of studies have been of a case–control design and small to moderate size. Many studies have relied on maternal recall of complications during a pregnancy that had occurred decades earlier (invoking a possible ‘effort after meaning' effect). Two birth cohort studies (83,84) using obstetric data that had been recorded prospectively by midwives reported inconclusive results regarding the association with schizophrenia. However, findings from another birth cohort study suggest that perinatal brain damage might account for as many as 7 per cent of all cases of schizophrenia in the adult population. (85) Such inconsistencies caution against an unqualified acceptance of obstetric complication as a proven risk factor in schizophrenia. Clarification of their role remains an important priority for epidemiological research.

Further information about studies of obstetric complications and hypoxic–ischaemic damage as risk factors for schizophrenia can be found here: Genetic and environmental risk factors for schizophrenia.

Season of birth

A 5 to 8 per cent winter–spring excess of schizophrenic births was first described in 1929 (86) and since then reported by a large number of studies, mainly in the northern hemisphere (southern hemisphere data are inconsistent). On the strength of the number of positive reports, birth seasonality appears to be a robust finding in the epidemiology of schizophrenia. (87) However, seasonality of births has also been reported for various other disorders including bipolar affective illness, major depression, autism, personality disorders, and mental retardation. A large proportion of these studies did not have the sample size needed either to prove or rule out a seasonal effect, or failed to apply the appropriate statistical procedures. As regards the possible explanatory models, few biologically plausible and testable causal hypothesis have been advanced. One of them is the seasonally increased risk of intauterine exposure to viral infection.

Maternal influenza

In utero exposure to influenza has been implicated as a risk factor since a report that an increased proportion of adult schizophrenia in Helsinki was associated with presumed second-trimester in utero exposure to the 1957 A2 influenza epidemic. (88) Over 30 studies have subsequently attempted to replicate the putative link between maternal influenza and schizophrenia. While several studies have replicated the original Finnish findings, negative results have more recently been reported from an increasing number of studies based on large epidemiological samples in different parts of the world. (89,90) Two studies (91,92) in which access to data on actually infected pregnant women was available found no increased risk of schizophrenia among their offspring.

The preschizophrenic person

Premorbid social impairment

Individuals who develop schizophrenia as adults are more likely to manifest difficulties in social interaction during childhood and adolescence than individuals who do not develop schizophrenia. Among children at increased genetic risk (having a parent with schizophrenia), poor social competence at age 7 to 12, and passivity and social isolation in adolescence, have been found to be common in those who go on to develop the disorder as adults. (93) The association between such ‘schizoid' traits and the risk of adult schizophrenia is not restricted to such high-risk populations. Evidence of early developmental peculiarities in children who develop schizophrenia as adults has been provided by prospectively collected data on a national birth cohort in the United Kingdom. (94) Preschizophrenic children had an excess (odds ratios, 2.1–5.8) of speech and educational problems, social anxiety, and preference for solitary play.

A cohort study from Sweden, (95) involving a 15-year follow-up of 50 087 men conscripted into the army at age 18 to 20, found that poor social adjustment during childhood and adolescence was significantly more common among the 195 individuals who subsequently developed schizophrenia than among the rest of the cohort. However, the early behavioural traits that are associated with schizophrenia in adult life are of such low specificity that their presence in a child or adolescent is of negligible predictive value.

Further information about studies of premorbid social impairment can be found here: Genetic and environmental risk factors for schizophrenia.

Premorbid intelligence (IQ)

In the same Swedish cohort, the subjects who subsequently developed schizophrenia were compared with the rest of the cohort on the performance of IQ-related tests and tasks at the time of conscription into the army. After controlling for confounding effects, the risk of schizophrenia increased linearly with the decrement of IQ. The effect was mainly attributable to poor performance on verbal tasks and tests of reasoning. (96)

Neurocognitive and neurophysiological markers

Several specific deficits in sustained attention, (97) event-related brain potentials, (98) and saccadic eye-movement control (99) have been found in clinical and laboratory research to be common in schizophrenic patients and also in a proportion of their clinically normal biological relatives, but to be rare in control subjects drawn from the general population (see: The neurobiology of schizophrenia). Their specificity to schizophrenia needs to be investigated in larger population samples. Should such variables be validated by epidemiological studies as biological markers of schizophrenia, the power of risk prediction at the level of the individual may increase substantially.

The social and family environment

Social class

Since the 1930s, numerous studies in North America and Europe have consistently found that the economically disadvantaged social groups contribute disproportionately to the first admission rate for schizophrenia. Two explanatory hypotheses, of social causation (‘breeder') and of social selection (‘drift'), were originally proposed. (100) According to the social causation theory, the greater socio-economic adversity characteristic of lower-class living conditions could precipitate psychosis in genetically vulnerable individuals who have a constricted capacity to cope with complex or stressful situations. In the 1960s this theory was considered to be refuted by a single study (101) which found that the social class distribution of the fathers of schizophrenic patients did not deviate from that of the general population, and that the excess of low socio-economic status among schizophrenic patients was mainly attributable to individuals who had drifted down the occupational and social scale prior to the onset of psychosis. As a result, aetiological research in schizophrenia in recent decades has tended to ignore such ‘macrosocial' variables. However, the possibility remains that social stratification, socio-economic status, and acculturation stress are important factors in the causation of schizophrenia.

Migrants and ethnic minorities

An exceptionally high incidence rate of schizophrenia (about 6.0 per 1000) has been found in the African-Caribbean population in the United Kingdom. (102,103) This excess morbidity is not restricted to recent immigrants and is, in fact, higher in the British-born second generation of migrants. Similar findings of nearly fourfold excess over the general population rate have been reported for the Dutch Antillean and Surinamese immigrants in Holland. (104)

The causes of the phenomenon remain obscure. Incidence studies in the Caribbean do not indicate any excess morbidity in the indigenous populations from which migrants are recruited. Explanations in terms of biological risk factors have found little support. (105,106) A finding in need of replication is the significant increase of schizophrenia among the siblings of second-generation African-Caribbean schizophrenic probands compared with the incidence of schizophrenia in the siblings of white patients. (107) Such a ‘horizontal' increase in the morbid risk suggests that an environmental factor may be modifying (increasing) the penetrance of the genetic predisposition to schizophrenia carried by a proportion of the African-Caribbean population. Psychosocial hypotheses involving acculturation stress, demoralization due to racial discrimination, and blocked opportunities for upward social mobility have been suggested but not yet properly tested.


Earlier hypotheses that urban environments increase the risk of psychosis, either by contributing to causation (the ‘breeder' effect) or by attracting vulnerable individuals (the ‘drift' effect), have been revived in the light of recent epidemiological findings suggesting that urban birth is associated with a moderate but statistically significant increase in the incidence of schizophrenia, affective psychoses, and other psychoses. (108) It remains unclear whether the effect is linked to a factor operating pre- or perinatally, or a factor influencing postnatal development (see also: Genetic and environmental risk factors for schizophrenia).

Marital status

There is evidence that marital status is significantly associated with first admission rates, the age at onset, and course and outcome of schizophrenia. Single males appear to be over-represented in schizophrenia samples, including epidemiological studies such as the WHO 10-country study (11) and a large German study. (109) Since both overt schizophrenia and preschizophrenic social impairments reduce the chances of marriage, married schizophrenics may represent a selected group with a milder form of the disease. Alternatively, marriage itself (or living with a partner) may have an effect of delaying the onset of schizophrenia or cushioning its impact. Neither of these two hypotheses can be definitively rejected on the basis of available descriptive epidemiological data. However, a statistical analysis of the WHO data, in which confounding factors such as age, premorbid personality traits, and family history were controlled, found that married men experienced a statistically significant delay (1–2 years) in the onset of psychotic symptoms compared with single men. (110)

Early rearing environment

Support for an effect of the early rearing family environment on the risk of developing schizophrenia is provided by a study of a Finnish sample of adopted children of schizophrenic parents (a high-risk group) and a control sample of adoptees at no increased genetic risk. (111) While the rates of adult psychosis or severe personality disorder were significantly higher in the high-risk group compared with the control group, the difference was entirely attributable to the subset of high-risk children who grew up in dysfunctional or otherwise disturbed adoptive families—a result consistent with the model of genetic control of sensitivity to the environment.

Epidemiological issues for the next decade

The rapid growth in basic knowledge about the human genome and the brain suggests that novel approaches to the study of complex disorders such as schizophrenia, integrating molecular genetics, neuroscience, and epidemiology, are likely to arise within the next decade.

Does the categorical disease concept of schizophrenia constrain aetiological research?

It has been suggested that the categorical disease concept of schizophrenia is no longer tenable and may be an obstacle to further progress in aetiological research. (112,113) The reasons advanced include the variation in the clinical phenotype, the likely genetic heterogeneity, and the absence, following several complete genome scans in large samples of families, of clear evidence for genetic linkage of the diagnostic entity. Reasons for revising the original formulation of the problem of the psychoses were given by Kraepelin himself, who concluded in 1920 that schizophrenia and manic–depressive illness do not represent particular pathological processes but rather indicate which ‘areas of our personality' are affected by them. (114)

Whether schizophrenia is a single disease or a syndrome arising as a ‘final common pathway' for a variety of pathological processes, the validity of the concept is supported by the epidemiological evidence. This makes it unlikely that the concept will be abandoned unless a clearly superior alternative is proposed. Existing alternatives include dimensional models derived from factor-analytical clinical studies (e.g. specifying a ‘psychotic', a ‘disorganized', and a ‘deficit' dimension), (115) subtyping schizophrenia according to the presence or absence of particular neurophysiological or neurocognitive abnormalities as ‘correlated phenotypes', (116) and subdividing the psychoses into multiple, presumably homogeneous, clusters along the lines proposed by Leonhard. (117). While experimenting with such concepts may generate new aetiological hypotheses, the resulting fragmentation of schizophrenia will reduce the statistical power of epidemiological and genetic studies. It is more likely that the diagnostic concept of schizophrenia will be retained, but will be refined by using some of the above models as complementary, rather than alternative, approaches.

Molecular epidemiology of schizophrenia

Notwithstanding the difficulties accompanying the genetic dissection of complex disorders, novel methods of genetic analysis will eventually identify genomic regions and loci predisposing to schizophrenia. The majority are likely to be of small effect, although one cannot exclude the possibility that genes of moderate or even major effects will also be found, especially in relation to the neurophysiological abnormalities associated with schizophrenia. Clarifying the function of such genes will be a complex task. Part of the solution is likely to be found in the domain of epidemiology, since establishing their population frequency and associations with a variety of phenotypic expressions, including personality traits and environmental risk factors, is a prerequisite for understanding their causal role. Thus a molecular epidemiology of schizophrenia is likely to be the next major chapter in the search for causes and cures.

Can schizophrenia be prevented?

The increasing interest in the early diagnosis and treatment of first episodes of schizophrenic and affective psychoses has raised the questions of whether people likely to develop schizophrenia can be reliably recognized prior to the onset of symptoms, and whether early pharmacological, cognitive, or social intervention can prevent the development of the disorder. While early diagnosis and treatment of symptomatic cases are feasible and may have the potential of improving the short- or medium-term outcome, the pre-onset detection of likely cases with a view to preventative intervention is problematic. It has been proposed that screening young age groups in the population by using predictors of high risk (such as a family history of psychosis, obstetric complications, or abnormal eye tracking) could identify potential patients long before onset. (118) However, the poor specificity of such putative risk factors is likely to result in low positive predictive values. Other candidate risk factors have not been evaluated at all epidemiologically. Problems of reliability of measurement apart, using combinations of such variables would decrease further the positive predictive value of the screening procedure and pose the practical and ethical problem of having to treat a large number of individuals who do not have the disorder. From an epidemiological point of view, presymptomatic detection and preventative intervention in schizophrenia do not appear to be feasible at present.

Summary and conclusions

After nearly a century of epidemiological research, many essential questions about the nature and causes of schizophrenia still await answers. Two major conclusions stand out.

First, the clinical syndrome of schizophrenia is robust and can be identified reliably in diverse populations, regardless of wide-ranging demographic, ecological, and cultural differences among them. This suggests that a common pathophysiology is likely to underlie the characteristic symptoms of schizophrenia. On balance, the evidence suggests that no major differences in incidence and disease risk can be found across populations at the level of large population aggregates. However, the study of ‘atypical' populations such as genetic isolates or minority groups may be capable of detecting unusual variations in the incidence of schizophrenia that could provide novel clues to the aetiology and pathogenesis of disorder.

The second conclusion is that no single environmental risk factor of major effect on the incidence of schizophrenia has yet been discovered. Further studies using large samples are required to evaluate potential risk factors, antecedents, and predictors for which the present evidence is inconclusive. Assuming that the methodological pitfalls of risk-factor epidemiology (such as the ‘ecological fallacy') can be avoided and that a number of variables will eventually be identified as risk factors of small to moderate effect, the results will complement those of genetic research which also implicate multiple genes. All this suggests that the key to understanding schizophrenia is likely to be found in gene–environment interactions.


1. Jablensky, A. (1995) Schizophrenia: the epidemiological horizon. In Schizophrenia (ed. S.R. Hirsch and D.R. Weinberger), pp. 206–48. Blackwell Science, Oxford.

2. Kraepelin, E. (1904) Vergleichende Psychiatrie. Zentralblatt für Nervenheilkunde und Psychiatrie, 27, 433–7. English translation: Comparative psychiatry. In Themes and variations in European psychiatry (ed. S.R. Hirsch and M. Shepherd), pp. 3–6. John Wright, Bristol, 1974.

3. Koller, J. (1895), Beitrag zur Erblichkeitsstatistik der Geisteskranken im Canton Zürich. Vergleichung derselben mit der erblichen Belastung gesunder Menschen durch Geistesstörungen u. dergl. Archiv für Psychiatrie, 27, 269–94.

4. Luxenburger, H. (1928). Vorläufiger Bericht über psychiatrische Serienuntersuchungen an Zwillingen. Zeitschrift für die gesamte Neurologie und Psychiatrie, 116, 297–326.

5. Klemperer, J. (1933). Zur Belastungsstatistik der Durchschnittsbevölkerung. Psychosehäufigkeit unter 1000 stichprobemässig ausgelesenen Probanden. Zeitschrift für die gesamte Neurologie und Psychiatrie, 146, 277–316.

6. Brugger, C. (1931). Versuch einer Geisteskrankenzählung in Thüringen. Zeitschrift für die gesamte Neurologie und Psychiatrie, 133, 252–390.

7. Essen-Möller, E., Larsson, H., Uddenberg, C.E., and White, G. (1956). Individual traits and morbidity in a Swedish rural population. Acta Psychiatrica et Neurologica Scandinavica, Supplement 100.

8. Hagnell, O. (1966). A prospective study of the incidence of mental disorder. Svenska Bokforlaget, Lund.

9. Ödegaard, Ö. (1946). A statistical investigation of the incidence of mental disorder in Norway. Psychiatric Quarterly, 20, 381–401.

10. World Health Organization. (1979). Schizophrenia. An international follow-up study. Wiley, Chichester.

11. Jablensky, A., Sartorius, N., Ernberg, G., et al. (1992). Schizophrenia: manifestations and course in different cultures. A World Health Organization ten-country study. Psychological Medicine Monograph Supplement 20. Cambridge University Press.

12. Robins, L.N. and Regier, D.A. (ed.) (1991). Psychiatric disorders in America. The Epidemiologic Catchment Area Study. Free Press, New York.

13. Kessler, R.C., McGonagle, K.A, Zhao, S., et al. (1994). Lifetime and 12-month prevalence of DSM-IIIR psychiatric disorders in the United States. Results from the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19.

14. Jones, P. and Cannon, M. (1998). The new epidemiology of schizophrenia. Psychiatric Clinics of North America, 21, 1–25.

15. Cooper, J.E., Kendell, R.E., Gurland, B.J., Sharpe, L., Copeland, J.R.M., and Simon, R. (1972). Psychiatric diagnosis in New York and London. Oxford University Press, London.

16. Wing, J.K., Cooper, J.E., and Sartorius N. (1974). The measurement and classification of psychiatric symptoms. Cambridge University Press.

17. Jablensky, A., Hugler, H., von Cranach, M., and Kalinov, K. (1993). Kraepelin revisited: a reassessment and statistical analysis of dementia praecox and manic–depressive insanity in 1908. Psychological Medicine, 23, 843–58.

18. World Health Organization (1992). International statistical classification of diseases and related health problems, 10th revision. WHO, Geneva.

19. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th edn). American Psychiatric Association, Washington, DC.

20. Robins L.N., Wing, J., Wittchen, H.U., et al. (1988). The Composite International Diagnostic Interview: an epidemiologic instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Archives of General Psychiatry, 45, 1069–77.

21. Wing J.K., Babor, T., Brugha, T., et al. (1990). SCAN. Schedules for Clinical Assessment in Neuropsychiatry. Archives of General Psychiatry, 47, 589–93.

22. Strömgren, E. (1938). Beiträge zur psychiatrischen Erblehre, auf Grund von Untersuchungen an einer Inselbevölkerung. Acta Psychiatrica et Neurologica Scandinavica, Supplement 19.

23. Bøjholm, S. and Strömgren, E. (1989). Prevalence of schizophrenia on the island of Bornholm in 1935 and in 1983. Acta Psychiatrica Scandinavica, 79 (Supplement 348), 157–66.

24. Lemkau, P., Tietze, C., and Cooper, M. (1943) A survey of statistical studies on the prevalence and incidence of mental disorder in sample populations. Public Health Reports, 58, 1909–27.

25. Sjögren, T. (1948). Genetic-statistical and psychiatric investigations of a West Swedish population. Acta Psychiatrica et Neurologica Scandinavica, Supplement 52.

26. Böök, J.A. (1953). A genetic and neuropsychiatric investigation of a North Swedish population (with special regard to schizophrenia and mental deficiency). Acta Genetica, 4, 1–100.

27. Böök, J.A., Wetterberg, L., and Modrzewska, K. (1978). Schizophrenia in a North Swedish geographical isolate, 1900–1977: epidemiology, genetics and biochemistry. Clinical Genetics, 14, 373–94.

28. Rin, H. and Lin, T.Y. (1962). Mental illness among Formosan aborigines as compared with the Chinese in Taiwan. Journal of Mental Science, 198, 134–46.

29. Lin, T.Y., Chu, H.M., Rin, H., et al. (1989). Effects of social change on mental disorders in Taiwan: observations based on a 15-year follow-up survey of general populations in three communities. Acta Psychiatrica Scandinavica, 79 (Supplement 348), 11–34.

30. Bash, K.W. (1967). Untersuchungen über die Epidemiologie neuropsychiatrischer Erkrankungen unter der handesbevölkerung der Provinz Fars, Iran. In Beiträge zur vergleichenden Psychiatrie (ed. N. Petrilowitsch), pp. 162–78. Karger, Basel.

31. Crocetti, G.J., Lemkau, P.V., Kulcar, Z., and Kesic, B. (1971). Selected aspects of the epidemiology of psychoses in Croatia, Yugoslavia, II. The cluster sample and the results of the pilot survey. American Journal of Epidemiology, 94, 126–34.

32. Dube, K.C. and Kumar, N. (1972). An epidemiological study of schizophrenia. Journal of Biosocial Science, 4, 187–95.

33. Rotstein, V.G. (1977). Material from a psychiatric survey of sample groups from the adult population in several areas of the USSR. Zhurnal Nevropatologii i Psikhiatrii, 77, 569–74 (in Russian).

34. Keith, S.J., Regier, D.A., and Rae, D.S. (1991). Schizophrenic disorders. In Psychiatric disorders in America. The Epidemiologic Catchment Area Study (ed. L.N. Robins and D.A. Regier), pp. 33–52. Free Press, New York.

35. Jeffreys, S.E., Harvey, C.A., McNaught, A.S., et al. (1997). The Hampstead Schizophrenia Survey 1991. I: Prevalence and service use comparisons in an inner London health authority, 1986–1991. British Journal of Psychiatry, 170, 301–6.

36. Jablensky, A., McGrath, J. Herrman, H., et al. (1999). People living with psychotic illness. National Survey of Mental Health and Wellbeing, Report 4. Commonwealth Department of Health and Aged Care, Canberra.

37. Eaton, J.W. and Weil, R.J. (1955). Culture and mental disorder. A comparative study of the Hutterites and other populations. Free Press, Glencoe, IL.

38. Torrey, E.F., Torrey, B.B., and Burton-Bradley, B.G. (1974). The epidemiology of schizophrenia in Papua New Guinea. American Journal of Psychiatry, 131, 567–73.

39. Kendler, K.S., Gallagher, T.J., Abelson, J.M., and Kessler, R.C. (1996). Lifetime prevalence, demographic risk factors, and diagnostic validity of nonaffective psychosis as assessed in a US community sample. Archives of General Psychiatry, 53, 1022–31.

40. Walsh, D. (1969). Mental illness in Dublin—first admissions. British Journal of Psychiatry, 115, 449–56.

41. Häfner, H. and Reimann, H. (1970). Spatial distribution of mental disorders in Mannheim, 1965. In Psychiatric epidemiology (ed. E.H. Hare and J.K. Wing), pp. 341–54. Oxford University Press, London.

42. Murphy, H.B.M. and Raman, A.C. (1971). The chronicity of schizophrenia in indigenous tropical peoples. Results of a twelve-year follow-up survey in Mauritius. British Journal of Psychiatry, 118, 489–97.

43. Lieberman, Y.I. (1974). The problem of incidence of schizophrenia: material from a clinical and epidemiological study. Zhurnal Nevropatologii i Psikhiatrii, 74, 1224–32 (in Russian).

44. Helgason, L. (1977). Psychiatric services and mental illness in Iceland. Acta Psychiatrica Scandinavica, 53 (Supplement 268), 1–140.

45. Castle, D., Wessely, S., Der, G., and Murray, R.M. (1991). The incidence of operationally defined schizophrenia in Camberwell, 1965–84. British Journal of Psychiatry, 159, 790–4.

46. Nicole, L., Lesage, A., and Lalonde, P. (1992). Lower incidence and increased male:female ratio in schizophrenia. British Journal of Psychiatry, 161, 556–7.

47. Rajkumar, S., Padmavati, R., Thara, R., et al. (1993). Incidence of schizophrenia in an urban community in Madras. Indian Journal of Psychiatry, 35, 18–21.

48. Hickling, F.W. (1991). Psychiatric hospital admission rates in Jamaica, 1971 and 1988. British Journal of Psychiatry, 159, 817–21.

49. McNaught, A., Jeffreys, S.E., Harvey, C.A., et al. (1997). The Hampstead Schizophrenia Survey 1991. II. Incidence and migration in inner London. British Journal of Psychiatry, 170, 307–11.

50. Brewin, J., Cantwell, R., Dalkin, T., et al. (1997). Incidence of schizophrenia in Nottingham. British Journal of Psychiatry, 171, 140–4.

51. Spitzer, R.L., Endicott, J., and Robins, E. (1978). Research Diagnostic Criteria. Rationale and reliability. Archives of General Psychiatry, 35, 773–82.

52. Koehler, K. (1979). First rank symptoms of schizophrenia: questions concerning clinical boundaries. British Journal of Psychiatry, 134, 236–48.

53. DeLisi, L.E., Goldin, L.R., Maxwell, M.E., Kazuba, D.M., and Gershon E.S. (1987). Clinical features of illness in siblings with schizophrenia or schizoaffective disorder. Archives of General Psychiatry, 44, 891–6.

54. Murthy, G.V.S., Janakiramaiah, N., Gangadhar, B.N., and Subbakrishna, D.K. (1998). Sex difference in age at onset of schizophrenia: discrepant findings from India. Acta Psychiatrica Scandinavica, 97, 321–5.

55. Helgason, T. and Magnusson, H. (1989). The first 80 years of life. A psychiatric epidemiological study. Acta Psychiatrica Scandinavica, 79 (Supplement 348), 85–94.

56. Ödegaard Ö. (1980). Fertility of psychiatric first admissions in Norway, 1936–75. Acta Psychiatrica Scandinavica, 62, 212–20.

57. Hilger, T., Propping, P. and Haverkamp, F. (1983). Is there an increase of reproductive rates in schizophrenics? Archiv für Psychiatrie und Nervenkrankheiten, 233, 177–86.

58. Lane, A., Byrne, M., Mulvany, F., et al. (1995). Reproductive behaviour in schizophrenia relative to other mental disorders: evidence for increased fertility in men despite decreased marital rate. Acta Psychiatrica Scandinavica, 91, 222–8.

59. Srinivasan, T.N. and Padmavati, R. (1997). Fertility and schizophrenia: evidence for increased fertility in the relatives of schizophrenic patients. Acta Psychiatrica Scandinavica, 96, 260–4.

60. Brown, S. (1997). Excess mortality of schizophrenia. A meta-analysis. British Journal of Psychiatry, 171, 502–8.

61. Dupont, A., Jensen, O.M., Strömgren, E., and Jablensky A. (1986). Incidence of cancer in patients diagnosed as schizophrenic in Denmark. In Psychiatric case registers in public health (ed. S.H. ten Horn, R. Giel, and W. Gulbinat), pp. 229–39. Elsevier, Amsterdam.

62. Mortensen, P.B. (1994). The occurrence of cancer in first admitted schizophrenic patients. Schizophrenia Research, 12, 185–94.

63. Mortensen, P.B. and Juel, K. (1993). Mortality and causes of death in first admitted schizophrenic patients. British Journal of Psychiatry, 163, 183–9.

64. Caldwell, C.B. and Gottesman, I.I. (1990). Schizophrenics kill themselves too: a review of risk factors for suicide. Schizophrenia Bulletin, 16, 571–89

65. Geddes, J.R. and Juszczak, E. (1995). Period trends in rate of suicide in first 28 days after discharge from psychiatric hospital in Scotland, 1968–92. British Medical Journal, 311, 357–60.

66. Rosenthal, R.N. (1998). Is schizophrenia addiction prone? Current Opinion in Psychiatry, 11, 45–8.

67. Dalack, G.W. and Meador-Woodruff, J.H. (1996). Smoking, smoking withdrawal and schizophrenia: case reports and a review of the literature. Schizophrenia Research, 22, 133–41.

68. Jeste, D.V., Gladsjo, J.A., Lindamer, L.A., and Lacro, J.P. (1996). Medical comorbidity in schizophrenia. Schizophrenia Bulletin, 22, 413–30.

69. Herrman, H.E., Baldwin, J.A., and Christie, D. (1983). A record-linkage study of mortality and general hospital discharge in patients diagnosed as schizophrenic. Psychological Medicine, 13, 581–93.

70. Eaton, W.M., Hayward, C., and Ram, R. (1992). Schizophrenia and rheumatoid arthritis: a review. Schizophrenia Research, 6, 181–92.

71. Lauerma, H., Lehtinen, V., Joukamaa, M.R., Helenius, H., and Isohanni, M. (1998). Schizophrenia among patients treated for rheumatoid arthritis and appendicitis. Schizophenia Research, 29, 255–61.

72. Murphy, K.C. and Owen, M.J. (1996). Minor physical anomalies and their relationship to the aetiology of schizophrenia. British Journal of Psychiatry, 168, 139–42.

73. Bredkjær, S.R., Mortensen, P.B., and Parnas, J. (1998). Epilepsy and non-organic non-affective psychosis. National epidemiologic study. British Journal of Psychiatry, 172, 235–8.

74. Hyde, T.M., Ziegler, J.C., and Weinberger, D. (1992). Psychiatric disturbances in metachromatic leucodystrophy. Archives of Neurology, 49, 401–6.

75. Murray, C.J.L. and Lopez, A.D. (ed.) (1996). The global burden of disease. A comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Harvard School of Public Health, Cambridge, MA.

76. Jablensky, A. (1995). Schizophrenia: recent epidemiologic issues. Epidemiologic Reviews, 17, 10–20.

77. Huxley, J., Mayr, E., Osmond, H., and Hoffer, A. (1964). Schizophrenia as a genetic morphism. Nature, 204, 220–1.

78. Haverkamp, F., Propping, P., and Hilger, T. (1982). Is there an increase of reproductive rates in schizophrenia? I. Critical review of the literature. Archiv für Psychiatrie und Nervenkrankheiten, 232, 439–50.

79. Jablensky, A. and Eaton, W.W. (1995). Schizophrenia. In Baillière's clinical psychiatry, Vol. 1, No. 2, Epidemiological psychiatry (ed. A. Jablensky), pp. 283–306. Baillière Tindall, London.

80. Kendler, K.S. and Eaves, L.J. (1986). Models for the joint effect of genotype and environment on liability to psychiatric illness. American Journal of Psychiatry, 143, 279–89.

81. Cannon, T.D. (1997). On the nature and mechanisms of obstetric influences in schizophrenia: a review and synthesis of epidemiological studies. International Review of Psychiatry, 9, 387–97.

82. Geddes, J.R. and Lawrie, S.M. (1995). Obstetric complications and schizophrenia: a meta-analysis. British Journal of Psychiatry, 167, 786–93.

83. Done, D.J., Johnstone, E.C., Frith, C.D., et al. (1991). Complications of pregnancy and delivery in relation to psychosis in adult life: data from the British perinatal mortality survey sample. British Medical Journal, 202, 1576–80.

84. Buka, S.L., Tsuang, M.T., and Lipsitt, L.P. (1993). Pregnancy/delivery complications and psychiatric diagnosis: a prospective study. Archives of General Psychiatry, 50, 151–6.

85. Jones, P.B., Rantakallio, P., Hartikainen, A.L., Isohanni, M., and Sipila, P. (1998). Schizophrenia as a long-term outcome of pregnancy, delivery, and perinatal complications: a 28-year follow-up of the 1966 North Finland general population birth cohort. American Journal of Psychiatry, 155, 355–64.

86. Tramer, M. (1929). Über die biologische Bedeutung des Geburtsmonates, insbesondere für die Psychoseerkrankung. Schweizerischer Archiv für Neurologie und Psychiatrie, 24, 17–24.

87. Torrey, E.F., Miller, J., Rawlings, R., and Yolken, R.H. (1997). Seasonality of births in schizophrenia and bipolar disorder: a review of the literature. Schizophrenia Research, 28, 1–38.

88. Mednick, S.A., Machon, R.A., Huttunen, M.O., and Bonett, D. (1988). Adult schizophrenia following prenatal exposure to an influenza epidemic. Archives of General Psychiatry, 45, 189–92.

89. Morgan, V., Castle, D., Page, A., et al. (1997). Influenza epidemics and incidence of schizophrenia, affective disorders and mental retardation in Western Australia: no evidence of a major effect. Schizophrenia Research, 26, 25–39.

90. Grech, A., Takei, N., and Murray, R.M. (1997). Maternal exposure to influenza and paranoid schizophrenia. Schizophrenia Research, 26, 121–5.

91. Crow, T.J. and Done, D.J. (1992). Prenatal influenza does not cause schizophrenia. British Journal of Psychiatry, 161, 390–3.

92. Cannon, M., Cotter, D., Coffey, V.P., et al. (1996). Prenatal exposure to the 1957 influenza epidemics and adult schizophrenia: a follow-up study. British Journal of Psychiatry, 168, 368–71.

93. Cannon, T.D., Mednick, S.A., and Parnas, J. (1990). Antecedents of predominantly negative- and predominantly positive-symptom schizophrenia in a high-risk population. Archives of General Psychiatry, 47, 622–32.

94. Jones, P., Rodgers, B., Murray, R., and Marmot, M. (1994). Child developmental risk factors for adult schizophrenia in the British 1946 birth cohort. Lancet, 344, 1398–402.

95. Malmberg, A., Lewis, G., David, A., and Allebeck, P. (1998). Premorbid adjustment and personality in people with schizophrenia. British Journal of Psychiatry, 172, 308–13.

96. David, A.S., Malmberg, A., Brandt, L., Allebeck, P., and Lewis, G. (1997). IQ and risk for schizophrenia: a population-based cohort study. Psychological Medicine, 27, 1311–23.

97. Cornblatt, B.A. and Keilp, J. (1994). Impaired attention, genetics, and the pathophysiology of schizophrenia. Schizophrenia Bulletin, 20, 31–46.

98. Freedman, R., Adler, L.E., Myles-Worsley, M., et al. (1996). Inhibitory gating of an evoked response to repeated auditory stimuli in schizophrenic and normal subjects. Archives of General Psychiatry, 53, 1114–21.

99. Clementz, B.A., McDowell, J.E., and Zisook, S. (1994). Saccadic system functioning among schizophrenia patients and their first-degree biological relatives. Journal of Abnormal Psychology, 103, 277–87.

100. Mischler, E.G. and Scotch, N.A. (1983). Sociocultural factors in the epidemiology of schizophrenia: a review. Psychiatry, 26, 315–51.

101. Goldberg, E.M. and Morrison, S.L. (1963). Schizophrenia and social class. British Journal of Psychiatry, 109, 785–802.

102. Bhugra, D., Leff, J., Mallett, R., Der, G., Corridan, B., and Rudge, S. (1997). Incidence and outcome of schizophrenia in Whites, African-Caribbeans and Asians in London. Psychological Medicine, 27, 791–8.

103. Harrison, G., Glazebrook, C., Brewin, J., et al. (1997). Increased incidence of psychotic disorders in migrants from the Caribbean to the United Kingdom. Psychological Medicine, 27, 799–806.

104. Selten, J.P., Slaets, J.P.J., and Kahn, R.S. (1997). Schizophrenia in Surinamese and Dutch Antillean immigrants to The Netherlands: evidence of an increased incidence. Psychological Medicine, 27, 807–11.

105. Hutchinson, G., Takei, N., Bhugra, D., et al. (1997). Increased rate of psychosis among African-Caribbeans in Britain is not due to an excess of pregnancy and birth complications. British Journal of Psychiatry, 171, 145–7.

106. Selten, J.P., Slaets, J., and Kahn, R. (1998). Prenatal exposure to influenza and schizophrenia in Surinamese and Dutch Antillean immigrants to The Netherlands. Schizophrenia Research, 30, 101–3.

107. Hutchinson, G., Takei, N., Fahy, T.A., et al. (1996). Morbid risk of schizophrenia in first-degree relatives of White and African-Caribbean patients with psychosis. British Journal of Psychiatry, 169, 776–80.

108. Marcelis, M., Navarro-Mateu, F., Murray, R., Selten, J.P., and van Os, J. (1998). Urbanization and psychosis: a study of 1942–1978 birth cohorts in The Netherlands. Psychological Medicine, 28, 871–9.

109. Häfner, H., Riecher-Rössler, A., van der Heiden, W., et al. (1993). Generating and testing a causal explanation of the gender difference in age at first onset of schizophrenia. Psychological Medicine, 23, 925–40.

110. Jablensky, A. and Cole, S.W. (1997). Is the earlier age at onset of schizophrenia in males a confounded finding? Results from a cross-cultural investigation. British Journal of Psychiatry, 170, 234–40.

111. Wahlberg, K.E., Wynne, L.C., Oja, H., et al. (1997). Gene–environment interaction in vulnerability to schizophrenia: findings from the Finnish adoptive family study of schizophrenia. American Journal of Psychiatry, 154, 355–62.

112. Kringlen, E. (1994). Is the concept of schizophrenia useful from an aetiological point of view? A selective review of findings and paradoxes. Acta Psychiatrica Scandinavica, 90 (Supplement 384), 17–25.

113. van Praag, H.M. (1993). ‘Make-believes' in psychiatry or the perils of progress. Brunner–Mazel, New York.

114. Kraepelin, E. (1920). Die Erscheinungsformen des Irreseins. Zeitschrift für die gesamte Neurologie und Psychiatrie, 62, 1–29. English translation: Comparative psychiatry. In Themes and variations in European psychiatry (ed. S.R. Hirsch and M. Shepherd), pp. 7–30. John Wright, Bristol, 1974.

115. Liddle, P.F. (1987). The symptoms of chronic schizophrenia: a re-examination of the positive–negative dichotomy. British Journal of Psychiatry, 151, 145–51.

116. Tsuang, M.T. and Faraone, S.V. (1995). The case for heterogeneity in the etiology of schizophrenia. Schizophrenia Research, 17, 161–75.

117. Leonhard, K. (1979). The classification of endogenous psychoses (5th edn). Halsted Press, New York.

118. McGlashan, T.H. and Johannessen, J.O. (1996). Early detection and intervention with schizophrenia: rationale. Schizophrenia Bulletin, 22, 201–22.