A systematic review of the prevalence of schizophrenia

A systematic review of the
prevalence of schizophrenia
John McGrath
Sukanta Saha
Joy Welham
David Chant
Schizophrenia - the most common form of psychosis – is
characterised by disorders of cognition (eg paranoia), affect
(mood), communication (thought disorder) and perception
(eg hallucinations) – leading to a loss of contact with reality
along with various forms of impaired behavior. [old]
Because schizophrenia (a) has an early onset and (b) has a
relapsing or chronic course, it is a significant public health
problem. It ranks in the top 10 leading causes of disability,
with a burden of disease comparable to cancer and greater
than heart disease.[old]
Understanding the prevalence of schizophrenia has
important implications for both health service planning and
risk factor epidemiology.[new]
Aims
To systematically identify and collate studies
of the prevalence of schizophrenia
To summarize variation in time, place and
person by examining the distribution of
these estimates of prevalence
To explore factors which may influence
prevalence estimates
Outline
• Research questions
• Methods & Systematic reviews
• Key results
• Caveats and Conclusions
The prevalence of schizophrenia:
Hypotheses
• Sex difference
Males > females
• Migrant status
Migrants > native born
• Urbanicity
Urban born > rural born
Types of prevalence studies
• ‘Core studies’
Sentinel surveys
Register based studies
• Migrant studies
• Cohort studies
• Other special groups
Methods: systematic review
• Electronic data search
– Medline, PsychoInfo, Embase, LILAC
• 1965-2001 inclusive
• (schizo* OR psycho*) AND
incidence OR prevalence)
also
• Review article bibliography
• Wrote to authors
Screen abstract and reviewed papers to cull irrelevant
citations
Rate items and discrete data
• Non-overlapping
Sex
Male, Female
• Overlapping
Age
eg all ages or age 15-54
Diagnosis
eg Catego S+ or Catego SPO + clinical
Diagnostic categories
eg DSMIV Schizophrenia or Schizophrenia + Delusional disorder
Site overlap
eg Denmark or Copenhagen
Epoch overlap eg 1990-92 or 1989-91
Data analysis:
example cumulative distribution
Rate per 100,000
Results
• Electronic search = 834 potential papers
• Manual checking = 249 potential papers
• Letters from 52 authors, who provided an
additional 41 references
Of potentially relevant papers, 74% were
identified from electronic sources
Results (2)
• Rates based on 176,056 potentially
overlapping incident cases
After review
• 158 were included from 32 countries
• Types of studies
– Core studies = 100
– Migrant studies = 24
– Cohort studies = 23
– Other special groups = 14
Core Prevalence Studies
Sex differences
Male : female rate ratio
Migrant status
Urban-rural differences
Cohort studies
– Australia (n=2)
– Denmark (n=3)
– Finland (n=5)
– Israel (n=2)
– Italy (n=1)
– Sweden (n=2)
– The Netherlands (n=2)
– USA (n=3)
– United Kingdom (n=3)
Other special groups
– over age 65
– twins
– various ethnic and/or religious subgroups
– students
– deaf individuals
– workers in a radiation contamination zone
Key findings
• Most of the distributions are ‘data rich’
• Variation
–
–
–
–
Asymmetrical
Long upper tail (>25% rates)
Median 15.2 (10-90% 7.7- 43.0) per 100,000
Five-fold range within the 10-90% quantiles
Key findings (2)
• Males > females
• Migrants > native born
• Urban > mixed urban/rural
• Schizophrenia has a varied and detailed
epidemiological landscape
Caveats
• Comparisons in systematic reviews should
be planned, based on directional
hypotheses & limited to a reasonable
number
• Systematic reviews are best suited to
hypothesis-generation
• Geographical boundaries are administrative
• Mostly ‘treated’ prevalence
Conclusions
There is a wealth of data available on the
prevalance of schizophrenia. Studies come
from many countries, with many different
methodological features, and conducted over
several decades.
The width and skew of the distributions, and the
significant impact of sex, urbanicity and migrant
status on these distributions, indicate
substantial variations in the prevalence of
schizophrenia.
Thus these data may provide leads for further
research into risk factors