Neighbourhood ethnic density and psychosis - Is there a difference according to generation?

Background: For different migrant groups living in an area with few people from the same ethnic background is associated with increased psychosis incidence (the ethnic density effect). We set out to answer the question: are there generational differences in this effect? Methods: Analysis of a population based cohort (2.2 million) comprising all those born 1st January 1965, or later, living in Denmark on their 15th birthday. This included 90,476 migrants from Africa, Europe (excluding Scandinavia) and the Middle East, with 55% first generation and the rest second-generation migrants. Neighbourhood co-ethnic density was determined at age 15 and we adjusted for age, gender, calendar period, parental psychiatric history and parental income. Results: For first-generation migrants from Africa, there was no statistically significant difference (p = 0.30) in psychosis rates when comparing lowest with highest ethnic density quintiles, whereas the second generation showed a 3.87-fold (95% CI 1.77-8.48) increase. Similarly, for migrants from the Middle East, the first generation showed no evidence of an ethnic density effect (p = 0.94) while the second showed a clear increase in psychosis when comparing lowest with highest quintiles, incidence rate ratio (IRR) 2.43 (95% CI, 1.18-5.00). For European migrants, there was some limited evidence of an effect in the first generation, (IRR) 1.69 (95% CI, 1.19-2.40), with this slightly raised in the second: IRR 1.80 (95% CI, 1.27-2.56). Conclusions: We found strong evidence for an ethnic density effect on psychosis incidence for second-generation migrants but this was either weak or absent for the first generation.