Future directions: what type of research do we need?

There is an urgent need for strategic research designed to generate and rigorously test candidate nongenetic, risk-modifying factors for schizophrenia. A framework for the primary prevention of schizophrenia can be built on this type of research.

While we have some broad clues in the form of risk indicators (urban birth, season of birth, pregnancy and birth complications), we need to 'fine map' these domains.

We should continue to try to identify novel domains from which to generate candidate exposures. One way to look for clues is to identify disease gradients and then search for clues to explain these gradients. These gradients can be across time (season of birth, between-year variability) and across space (neighbourhood variation, urban - rural gradient and latitude gradients in prevalence). Within medical epidemiology, migrant studies have played an important role in generating candidate exposures. They can vary the environmental factors while 'holding' ethnicity constant (and, it is assumed, a degree of genetic variability) (Hennekens and Buring, 1987). The cause of the increased risk of psychosis in second-generation African - Caribbeans born in the UK is still unclear. These studies of schizophrenia in migrant groups may help to generate awareness of novel candidate exposures (Harrison et al., 1997; McDonald and Murray, 2000).

It is essential that the data used to address this problem are derived from epidemiologically informed sampling frames. Sample sizes need to be representative of the underlying population so the PAR values that guide public health planning can be generated. Links between candidate exposures and genetic epidemiology will also be crucial for future research.

{embed="ref/53"}

Just as candidate genes are ranked according to various rules (e.g. are they located in a 'hot spot', how many introns/exons within the gene, biological plausibility, etc.?), those interested in exposures need sorting rules. When rank-ordering candidate exposures for more intense scrutiny, there is a case to prioritise those with universal public health potential (e.g. vaccinations for infectious agents, nutritional supplements). If the exposure is already associated with disorders other than schizophrenia, then this makes it a more attractive candidate from a public health perspective (single interventions could lead to improvements in multiple outcomes).

Epidemiological research can serve to 'sharpen the focus', allowing candidate risk factors to be identified. However, history has shown that risk factor epidemiology can sometimes enter cycles of uninformative replications ('circular epidemiology': Kuller, 1999). Season of birth has been studied as a risk factor for 80 years with little real progress in identifying the underlying risk-modifying exposure. In reaction to this, some research groups are looking to developments in the neurosciences in order to test the impact of candidate exposures on neuronal development, both in vitro and in whole animal studies. Such analyses range from gene expression profiling through to cognitive psychology techniques such as prepulse inhibition.