Lay Summary

This study investigated if and how cannabis use, deprivation and other setting-level risk factors are associated with new cases of psychotic disorder in 17 settings from 6 countries. This is important because it may help us understand which factors are most relevant to the onset of psychotic disorders. Results will be presented at the congress.

Background

The incidence of psychosis varies by area, social group, environmental risk factors and cannabis use. A previous EU-GEI study investigated psychosis incidence across 17 settings, finding an 8-fold variation between settings, with higher rates in settings with lower levels of owner-occupancy, a putative marker of setting-level deprivation. Another EU-GEI study found an association between higher proportion of daily cannabis use amongst controls in each setting and increased odds of psychosis, especially for high-potency cannabis types. Since setting-level deprivation and cannabis use may be confounded, the aim of the present study was to re-analyse this data to investigate the relative contribution of setting-level cannabis use, deprivation and other setting-level risk factors in explaining variation in the treated incidence of first episode psychosis (FEP).

Methods

The EU-GEI study is an international multisite incidence study. Data of 2774 individuals aged 18-64 years with FEP was collected between 2010-2015 in England, France, Italy, the Netherlands, Spain and Brazil across 17 settings. For the current study, individual-level data on country, setting, age group, sex, ethnicity/minority status and psychotic disorder type was retrieved and used as numerator for the incidence calculation. The denominator was estimated using routine demographic data and stratified by age, sex, setting/country and ethnicity. Next, crude and standardized (for age, sex and ethnicity/minority status) incidence rates per 100,000 person-years and 95% CIs were estimated per setting. Finally, univariable and multivariable random-effects (intercepts) Poisson regression was used to investigate associations between individual- and setting-level risk factors and incidence of FEP across settings. Setting-level exposures were the % of control participants with daily or high potency cannabis use; % of population who owned their own house; % unemployed; % single-person households; latitude; and population density. Cannabis data was retrieved from control participants in the EU-GEI study and other setting-level exposures from routinely available demographic sources, including local censuses.

Results

We hypothesized that after controlling for individual-level age, sex and ethnicity/minority status, deprivation-related variables (owner-occupancy, unemployment) would account for previously observed associations between setting-level cannabis use and FEP incidence. Evidence of these results would suggest that setting-level cannabis reflects exposure to a wider set of social determinants of health which together serve to influence area-level psychosis incidence. We will present the results of this study during the congress.

Conclusion

This study shows the relative contribution of how different social determinants of psychosis risk combine to influence incidence rates and extends our knowledge on environment-environment interactions in psychotic disorders across different contexts. It thereby elucidates potential underlying etiological processes in psychotic disorders.