Schizophrenia Research Forum - A Catalyst for Creative Thinking
Home Profile Membership/Get Newsletter Log In Contact Us
 For Patients & Families
What's New
Recent Updates
SRF Papers
Current Papers
Search All Papers
Search Comments
Research News
Conference News
Plain English
Current Hypotheses
Idea Lab
Online Discussions
Virtual Conferences
What We Know
Animal Models
Drugs in Trials
Research Tools
Community Calendar
General Information
Member Directory
Researcher Profiles
Institutes and Labs
About the Site
SRF Team
Advisory Board
Support Us
How to Cite
Fan (E)Mail
The Schizophrenia Research Forum web site is sponsored by the Brain and Behavior Research Foundation and was created with funding from the U.S. National Institute of Mental Health.
Online Discussions

Updated 24 May 2007 E-mail discussion
Printable version

Live Discussion: The New Epidemiology of Schizophrenia

John McGrath

View article

View article

Epidemiologists have offered no shortage of leads to the etiology of schizophrenia, but it is no easy task to turn these findings into practical use—either for prevention or treatment. On 11 April, John McGrath of the University of Queensland, Australia, led us in a discussion of the challenges of turning epidemiologic data into new ideas for prevention and treatment. As we await the transcript posting, we invite you to read two of his recent editorials, from Schizophrenia Bulletin and the Archives of General Psychiatry. Please also read and respond to the comments left below.

Our special thanks to the Archives of General Psychiatry for granting open access to this editorial:

McGrath JJ. The surprisingly rich contours of schizophrenia epidemiology. Arch Gen Psychiatry. 2007 Jan;64(1):14-6. View article

John J McGrath. Variations in the Incidence of Schizophrenia: Data Versus Dogma. Schizophr Bull, January 2006; 32: 195-197. View article

View Transcript of Live Discussion — Posted 24 May 2007

View Comments By:
Jim van Os — Posted 4 April 2007
James Kirkbride — Posted 5 April 2007
Jean-Paul Selten — Posted 5 April 2007
Craig Morgan — Posted 9 April 2007
Jonathan Burns — Posted 9 April 2007
Brian Chiko — Posted 10 April 2007
Fuller Torrey — Posted 10 April 2007
Assen Jablensky — Posted 10 April 2007
Jean-Paul Selten — Posted 10 April 2007
Paul Fearon — Posted 10 April 2007
Dana March — Posted 10 April 2007
— Posted 10 April 2007
Sukanta Saha, Saha Sukanta — Posted 11 April 2007
Joy Welham — Posted 11 April 2007
Vera A. Morgan — Posted 11 April 2007
Robert Yolken — Posted 11 April 2007
Preben Bo Mortensen — Posted 11 April 2007
Paul Patterson — Posted 11 April 2007
Patricia Estani — Posted 18 April 2007
Jan van Dijk — Posted 30 April 2007
David Yates — Posted 23 May 2007
Petar Marinov — Posted 29 May 2007
Huda Shalhoub — Posted 13 March 2009

Background Text
By John McGrath

If forgetting that schizophrenia was a brain disease was one of the great aberrations of twentieth-century medicine, ignoring variations in the incidence of schizophrenia must rank as one of the great aberrations of modern epidemiology. However, research has recently provided us with data that cannot be ignored. The incidence of schizophrenia varies widely among sites, and varies according to a range of demographic variables including sex, urbanicity of place of birth/residence, paternal age, season of birth, and migrant status. High-quality prospective cohort studies have also strengthened the case that cannabis is a risk-modifying factor for schizophrenia.

The target articles outline some of the major features of the “new epidemiology” of schizophrenia. The commentary in the Archives of General Psychiatry canvasses various options about how to get the best “bang for our buck” from future epidemiological studies. It is argued that the clues being generated by schizophrenia epidemiology are too important to ignore.

The discussion will be broad-ranging, and we encourage the readers of Schizophrenia Research Forum to contribute topics for discussion and debate prior to the Live Discussion. For example:

1. What type of epidemiological research needs to be done now?

2. How can we avoid “circular epidemiology” where research perseverates at the ecological level, and fails to move to more refined analytic methods or experimental studies?

3. How can we engage neuroscience in helping to unravel the clues emerging from epidemiology?

Comments on Online Discussion
Comment by:  Jim van Os
Submitted 4 April 2007 Posted 4 April 2007

John makes some excellent points, and his papers help us understand the epidemiology of schizophrenia. I have two points.

1. I was fascinated by the Finnish study published in the Archives of General Psychiatry (Perala et al., 2007); this paper showed the incidence of psychotic disorders was 3.5 percent, of which the diagnosis "schizophrenia" only represented 25 percent! I think we may be missing the point by focusing on a diagnostic category (the criteria of which constantly change) that only represents a minor fraction of the total psychosis morbidity force.

2. The second issue is that not all expression of the psychosis phenotype may be pathological, and that much can be learned by focusing on its subclinical expression.

View all comments by Jim van Os

Comment by:  James Kirkbride
Submitted 5 April 2007 Posted 5 April 2007

The variation in incidence rates from epidemiological studies is now clear to see, not least in terms of ethnicity and urbanicity. I think that we now need to move beyond studies which show such variation to developing models which attempt to elicit the specific environmental "risk factors" which presumably underpin these effects; most likely operating with gene-environment and other interactions.

Future epidemiological studies will need to be more sensitive to the environment and person-environment (cross-level) interactions. It is likely that different environmental exposures have differing "doses" depending on individual susceptibility, particularly given exposure to the ubiquitous risk factors of ethnicity and urbanicity.

Furthermore, I think it will be crucial to elucidate the critical timing of exposure to (socio)environmental factors, be these close to birth, during childhood or adolescence, or close to onset of disorder. I believe all three time points will have a role to play. Clearly, longitudinal multilevel studies will represent the gold standard of...  Read more

View all comments by James Kirkbride

Comment by:  Jean-Paul Selten
Submitted 5 April 2007 Posted 5 April 2007

The points made by John are timely and excellent indeed. I take the opportunity to place some remarks upon the silent epidemic of psychotic disorders among immigrants from non-western countries in Western Europe and about the sociology of science. There are at least three reasons for the concealment of this epidemic:

1. The immigrants are ashamed and ignore it. They threaten psychiatrists and journal editors with accusations of racism.

2. The politicians do not feel confident about psychiatric diagnosis or psychiatric epidemiology and are afraid of being accused of "racism."

3. Many psychiatrists are hopelessly in love with DNA and MRI, cannot change their fixed ideas about equal incidence rates across place and time, or lack the courage to shift their research focus.

The first step forward is to bring this epidemic and the dangers of illicit drug abuse to the attention of the media. Things become "true" when they are on television! We can hope that policy makers at Medical Research Councils, who divide the money, will then encourage researchers to spend their...  Read more

View all comments by Jean-Paul Selten

Comment by:  Craig Morgan
Submitted 9 April 2007 Posted 9 April 2007

The variations in the incidence and prevalence of psychosis point strongly towards the social environment, broadly construed, as being important in etiology.

However, as John McGrath notes, the data offer “clues” only. In the epidemiological research to date, the variables used to represent social environments and experience are usually crude dichotomies—urban vs. rural, migrant vs. non-migrant, abused vs. non-abused. With regard to urbanicity and migration, one consequence is that the precise meaning of observed associations is unclear, and the relevant social conditions or experiences that these variables may index remain unknown.

This suggests a need to engage the social (as well as neuro) sciences in helping unravel these new epidemiological "clues." We need to develop appropriate conceptual maps and measurement tools that allow us to study complex social processes, and the contexts in which they occur, if we are to fully understand the nature of the associations between, say, migration and psychosis. These are the very issues that social scientists have...  Read more

View all comments by Craig Morgan

Comment by:  Jonathan Burns
Submitted 5 April 2007 Posted 9 April 2007

Exposing the myth of constant prevalence and incidence rates of schizophrenia is a welcome development for those of us who don't buy into the reductionist gene-centric view of psychiatric disorders. While genetic vulnerability is indisputable, it is ostrich-like behavior to ignore (or pay lip-service) to environmental factors that mediate expression of psychosis in vulnerable individuals.

Jim's point above regarding our diagnostic focus on schizophrenia rather than on psychosis is very important. Studies of first-episode psychosis (FEP) are essential if one is looking at the social environment and its impact on incidence, because one assumes that the likelihood of "social drift" (as an explanation) can be diminished if one is investigating the illness at its onset. In other words, it is easier to argue for an etiological role for environmental factors, if one is looking at FEP.

Despite John's frustration with ecological methodologies, I have just calculated a retrospective treated incidence for FEP for each of seven municipalities here in Durban, South Africa. I...  Read more

View all comments by Jonathan Burns

Comment by:  Brian Chiko
Submitted 10 April 2007 Posted 10 April 2007

At (a non-profit, family oriented schizophrenia education site) we feel that research into schizophrenia epidemiology is vitally important, yet seriously underfunded. We applaud the increased efforts in this important field.

We are also, however, making efforts to educate the public about what recent research is suggesting, and identifying actions that families with a history of schizophrenia may want to consider to lower their risks. As part of this effort we've created a Preventing Schizophrenia section of the web site, which attempts to summarize the state of the research in areas where it can help drive individual and family actions that seem likely to lower risk for future generations. We welcome your comments on this section.

View all comments by Brian Chiko

Comment by:  Fuller Torrey
Submitted 10 April 2007 Posted 10 April 2007

John McGrath has made a major contribution to the epidemiology of schizophrenia, probably the most important contribution since the work of Dr. Edward Hare. By questioning the received wisdom of the textbooks and showing that the incidence of schizopohrneia varies widely, he has opened up new avenues for research. Most such research will have to be done in Europe, especially in Scandanavia, where appropriate databases exist. In the United States, few resources are available that can be used for epidemiological studies, and consequently American interest in such studies is minimal.

In terms of future directions, it would be interesting to link some of the prevalence data to the birth seasonality data, e.g., is the seasonal birth excess greater in those areas where the incidence is higher?

View all comments by Fuller Torrey

Comment by:  Assen Jablensky
Submitted 10 April 2007 Posted 10 April 2007

I have decided to play the role of advocatus diaboli and pose three critical questions which I think are fundamental to defining the role of epidemiology in the present state of schizophrenia research. I'll be greatly interested in the proceedings.

1. What should be the role of epidemiology in the rapidly changing field of research into the etiology and pathogenesis of schizophrenia? Much of what is now driving schizophrenia research and is potentially promising, happens to be within the domains of neuroscience and genetics. Considering that (in contrast to diseases such as cancer) we are at least a decade away from examining population distributions of proven genetic risk polymorphisms for the disorder, can epidemiology at present offer a meaningful complementary approach and disciplinary partnership to neuroscience and genetics?

2. A good deal of the science of epidemiology is about the effects of environmental risk exposures and their interactions with intrinsic genetic vulnerabilities. So far, no single environmental exposure has been definitively...  Read more

View all comments by Assen Jablensky

Comment by:  Jean-Paul Selten
Submitted 10 April 2007 Posted 10 April 2007

We complain that fixed ideas about the epidemiology of psychosis do not disappear from textbooks. I propose that we write to the editors of the 5 or 10 most important textbooks to make this point.

View all comments by Jean-Paul Selten

Comment by:  Paul Fearon
Submitted 10 April 2007 Posted 10 April 2007

I agree with many of the comments already made.

I do however respectfully question Prof. Jablensky's third 'Devil's Advocate' point regarding outcome studies. At a time when the oft quoted finding from the WHO studies regarding the supposed better outcome in developing countries is increasingly being questioned (often by the original investigators themselves), surely there is a need for more studies with rigorous follow up. Not only may they provide us with information that may inform aetiology, but they might just allow us to start to answer that most basic of questions (which we still cannot answer!): 'Will my son/daughter get better or remain as they are?'

There does seem to be a window of opportunity at the moment. There have been several first onset studies over the last decade, most of which have been epidemiologically based. Many of these are engaging in follow-ups (of admittedly different time intervals etc). Surely a closer collaboration between these groups worldwide would not only be informative but also, dare one say it, relatively cost-effective??

View all comments by Paul Fearon

Comment by:  Dana March
Submitted 10 April 2007 Posted 10 April 2007

John McGrath asserts that we must be "slaves to the data" in his Schizophrenia Bulletin commentary. Our data doesn’t just drive us, however. Data generated depend on the questions we ask, and the questions we ask shape our approaches and the data we collect, or employ.

In the kernels for our live discussion, there is an emphasis on future studies of schizophrenia, in which we can tailor our search for environmental risk factors. Let us not forget, however, the trove of available longitudinal data, which with careful consideration and creativity, can be plumbed for clues to etiology, particularly the timing of relevant exposures. Travel across the epidemiologic landscape is far from complete, and our studies, especially those begun many decades ago, are far from exhausted.

Like extant epidemiologic data, information from specialized studies conducted within the bounds of other disciplines has not been completely mined. To understand the full range and sequence of causes of schizophrenia, research from genetics, neuroscience, psychology, sociology, and even history,...  Read more

View all comments by Dana March

Comment by:  Sukanta SahaSaha Sukanta
Submitted 11 April 2007 Posted 11 April 2007

The variation in the incidence of schizophrenia is now firmly established. There is also large body of evidence to suggest that prevalence, and mortality (Arch Gen Psychiatry 2007 in press) in schizophrenia varies widely among sites. I agree with Jean-Paul Selten that it is timely to write to the editors of the important text books to correct it. The ‘silent epidemic’ of psychotic disorders among immigrants in many parts of the world also needs to be documented. As an immigrant I can see the problem of psychotic illness among various groups.

I am an advocate of more epidemiological studies from different populations to answer what’s and why’s? Firstly, as Craig Morgan wrote, the precise meaning of associations between observed dichotomised variations (male/female, urban/rural, and migrant/non-migrants) is not clear. Secondly, in our systematic reviews of incidence, prevalence, and mortality, we found very few studies from developing countries. We do not know whether these dichotomised variations are the same across populations. For example, an influential study from...  Read more

View all comments by Sukanta Saha
View all comments by Saha Sukanta

Comment by:  Joy Welham
Submitted 11 April 2007 Posted 11 April 2007

There have been some varied and interesting comments about the “New Epidemiology of Schizophrenia”

Similar debates about the future of epidemiology have taken place in other areas, but this seems to be resolving into a healthy mix of studies. A personal brief survey of topics covered in contemporary general epidemiological journals, such as the current edition of the International Journal of Epidemiology, shows a range of studies conducted at multiple levels—from the social and geographical, through to the genetic, with some studies modelling genetic and environmental interactions. Can this be the future of the epidemiology of schizophrenia, at least the foreseeable future?

I’d like to add a comment and a question. The comment is: Although socio-demographic factors may be associated with variation in incidence/prevalence over time and place, other factors may also be operating: infectious agents and climate come to mind. There may well be an interaction between these and the genetic and socio-demographic factors already discussed.

The question is: There is...  Read more

View all comments by Joy Welham

Comment by:  Vera A. Morgan
Submitted 11 April 2007 Posted 11 April 2007

An issue of critical importance, and relevant to comments raised by Assen Jablensky and Paul Fearon, is the funding of longitudinal incidence studies. While grant funding bodies steer away from these, most governments are reticent to put money into studies that are so costly and time-consuming. The government timeline for funding is, more often than not, related to that government's cycle in office: they want results that they can use with political expedience. How do schizophrenia researchers persuade governments to fund long-term research programs that will help build up the bank of epidemiological evidence?

View all comments by Vera A. Morgan

Comment by:  Robert Yolken
Submitted 11 April 2007 Posted 11 April 2007

The question of whether one can develop animal models for environmental exposures is an interesting one. Models can be developed for relatively high risk exposures, such as smoking and lung cancer. However, they are more difficult to develop for more subtle exposures where only some of the exposed animals will develop the consequences of the exposures and where some of the risk may be genetically determined. The development of relevant animal models for these types of exposures remains an important goal for research into complex disorders.

View all comments by Robert Yolken

Comment by:  Preben Bo Mortensen
Submitted 11 April 2007 Posted 11 April 2007

I think Assen Jablensky's concerns regarding the potential for epidemiology to contribute to the identification of causes are very relevant. However, I do not agree that "we are at least a decade away from examining population distributions of proven genetic risk polymorphisms for the disorder," and I think that this is certainly a necessary next step of the highest priority. From an epidemiological point of view it is of course the most important set of confounders and/or interacting agents for any environmental factor we could suspect. However I think it is less well realized at present that this is equally important for the progress of genetic studies. I think one of the real benefits epidemiology can bring to the field is a more rigorous description and handling of biases and confounders that are mostly ignored in genetic studies (let alone neuroimaging), and bringing this perspective into the field will be necessary. So, rigorous design with a strong sense of the population base of the study, and more exact measurement of the exposures/confounders in question seem to me...  Read more

View all comments by Preben Bo Mortensen

Comment by:  Paul Patterson
Submitted 11 April 2007 Posted 11 April 2007

Regarding the question of whether there are useful animal models based on epidemiological findings, the answer is certainly yes.

For schizophrenia, there are now quite a few papers based on the findings (most definitively from Allan Brown and colleagues) that maternal respiratory infection increases the risk in the offspring. They calculate that this risk factor can account for 14-21 percent of schizophrenia cases. Mice given a respiratory infection at mid-gestation yield offspring with striking abnormalities in behaviors related to schizophrenia (anxiety, social interaction, latent inhibition, prepulse inhibition), some of which can be corrected by antipsychotic drugs. The effect of maternal infection on fetal brain development can be mimicked by activating the maternal immune response by injection of dsRNA. Thus, the virus is not required, and other mechanisms of activating the maternal immune response can presumably cause similar changes in the offspring. We have recently found that the cytokine IL-6 mediates the effects of maternal immune activation on the...  Read more

View all comments by Paul Patterson

Comment by:  Patricia Estani
Submitted 12 April 2007 Posted 18 April 2007

I would like to agree with two points from Dr. Jablensy's excellent commentary.

In the field of the new epidemidemiology of schizophrenia, I think that the temporal projection of the sistematic gradients of several variables, such as the variables of urbanicity and sex, for example, could help researchers to design some strategies of prevention with the collaboration of other professionals of the medical services.

In this sense, research on the risk factors could make the field of epidemiology a more dynamic and a more contributive field in relation to the pathogenesis and the causes of the disease. Thus, the epidemiology could tell us substantial information about the relationships between genetic polymorphisms and the environmment.

View all comments by Patricia Estani

Comment by:  Jan van Dijk
Submitted 27 April 2007 Posted 30 April 2007

The papers of McGrath reminded us that the incidence of schizophrenia is significantly higher in males than in females with a male:female risk ratio of 1.4:1. Interestingly the first whole-genome association study of Todd Lencz et al. (Lencz et al., 2007) revealed a novel schizophrenia locus that is present on both chromosomes X and Y. Since there is only a clear difference in the two sex chromosomes, I wonder if this schizophrenia locus can explain this variation in incidence by sex?


Lencz T, Morgan TV, Athanasiou M, Dain B, Reed CR, Kane JM, Kucherlapati R, Malhotra AK. Converging evidence for a pseudoautosomal cytokine receptor gene locus in schizophrenia. Mol Psych. 2007, March 20. Advanced online publication.

View all comments by Jan van Dijk

Comment by:  David Yates
Submitted 23 May 2007 Posted 23 May 2007

A few comments from a clinician and a carer.

Is their some merit in dividing schizophrenia outcome into silent and florid? The condition fluctuates. Sometimes there is not enough visible behavior to make for a diagnosis. The DSM, which lists delusional illness, schizoaffective disorder, schizophreniform disorder, even psychosis accompanying physical illness, allows for sharpness of definition, appreciation, and overlapping such as to negate comparability between populations. Yet the condition can be there in the very restricted and limited lifestyle.

The illness expression, and therefore the diagnostic listing, will anyhow vary as the proportion of people in any population divide into different personality types: the introvert, anancastic, obsessional people may develop negative schizophrenia; the extrovert, sociable, novelty-seeking people, more florid—‘positive'— illness.

The social and domestic life of women in growing up, and in the balance of relationships is different from that of men. The age at which schizophrenia arrives—after or...  Read more

View all comments by David Yates

Comment by:  Petar Marinov
Submitted 27 May 2007 Posted 29 May 2007

The discussion is very interesting. From a forensic point of view we found 8.3/1 male/female ratio in Bulgaria with respect to homicidal schizophrenic offenders (Marinov, 2006). I wonder if this huge difference in aggressivity is related to the gender distribution or whether it can be explained mainly by hormonal differences.

Marinov, P. (2006). Proximity of the victim to the perpetrator of schizophrenic and nonschizophrenic homicides ­ an attempt for comparative analysis. Receptor. 3, 49-53.

View all comments by Petar Marinov

Comment by:  Huda Shalhoub
Submitted 4 March 2009 Posted 13 March 2009

The discussions set forth have been well expressed in terms of their explanations that the incidence and prevalence of schizophrenia are not uniform across different variables (migrants vs. non-migrants, males/females, and urban/rural) and that we need to do something about disseminating such knowledge. In terms of exploring the current situation, I would say that for those who are willing to dig deeper into the breadth of research on the topic, they will not be able to miss this fact as there have been well-established meta-analyses and well-grounded research findings up to this date that unambiguously share that knowledge (such as Cantor-Graae and Selten, 2005).

The question to ask at this point in time in terms of advancing our repertoires to improve treatment and the etiological factors behind schizophrenia's variations would be to start using a microscopic view, in contrast to large-scale studies (i.e., epidemiological and longitudinal) that now are only statistically quantifying what we already know.

One way of...  Read more

View all comments by Huda Shalhoub
Submit a Comment on This Online Discussion
Make a comment on this live discussion. 

If you already are a member, please login.
Not sure if you are a member? Search our member database.

*First Name  
*Last Name  
Country or Territory  
*Login Email Address  
*Confirm Email Address  
*Confirm Password  
Remember my Login and Password?  
Get SRF newsletter with recent commentary?  
Enter the code as it is shown below:
This code helps prevent automated registrations.

Please note: A member needs to be both registered and logged in to submit a comment.


(If coauthors exist for this comment, please enter their names and email addresses at the end of the comment.)


SRF News
SRF Comments
Text Size
Reset Text Size
Email this pageEmail this page

Live Discussion FAQs

Webinar: A Webinar is a seminar conducted remotely over the Web. Attendees view the slides through their Web browser and hear the presentations over their own telephones.

Registration: All participants are to register by clicking on the "Register for the Webinar" link.

Access: After you register, you will receive an e-mail with a link to the Webinar and a phone number.
New Schizophrenia Fact Sheet for Patients and Families

Latest BBRF Research Breakthroughs

Visit our Facebook page and our Blog.

Support the Brain & Behavior Research Foundation Today.

Research Participants
Copyright © 2005- 2014 Schizophrenia Research Forum Privacy Policy Disclaimer Disclosure Copyright