LAUSR

Abstract Background We aimed to spatially describe mental illness prevalence in England at small-area geographical level, as measured by prevalence of depression, severe mental illness (SMI) and antidepressant prescription volume in primary care records, and how much of their variation was explained by deprivation, social fragmentation and sociodemographic characteristics. Methods Information on prevalence of depression and SMI was obtained from the Quality and Outcomes Framework (QOF) administrative dataset for 2015/16 and the national dispensing dataset for 2015/16. Linear regression models were fitted to examine ecological associations between deprivation, social fragmentation, other sociodemographic characteristics and mental illness prevalence. Results Mental illness prevalence varied within and between regions, with clusters of high prevalence identified across England. Our models explained 33.4–68.2% of variability in prevalence, but substantial variability between regions remained after adjusting for covariates. People in socially cohesive and socially deprived areas were more likely to be diagnosed with depression, while people in more socially fragmented and more socially deprived areas were more likely to be diagnosed with SMI. Conclusions Our findings suggest that to tackle mental health inequalities, attention needs to be targeted at more socially deprived localities. The role of social fragmentation warrants further investigation, and it is possible that depression remains undiagnosed in more socially fragmented areas. The wealth of routinely collected data can provide robust evidence to aid optimal resource allocation. If comparable data are available in other countries, similar methods could be deployed to identify high prevalence clusters and target funding to areas of greater need.