Investigating 'place effects' on mental health: implications for population-based studies in psychiatry.

Background. Interest in features of our local environments that may promote better mental health and wellbeing continues to rise among decision makers. Our purpose was to highlight a selection of these challenges and some promising avenues for enhancing the quality of evidence. Method. An analysis of approximately 267, 000 people was used to test the local relative deprivation hypothesis, wherein the shortfall of a person's socioeconomic circumstances from their neighbours is said to impact negatively upon mental health. This case was used to anchor further discussion of challenges to identifying and interpreting genuine 'place effects' from spurious correlations. Results. A Median Odds Ratio of 1.29 computed via multilevel logistic regression showed that the odds of experiencing psychological distress (as measured by the Kessler score) varied by geographical area. Approximately 67% of this was attributed to a cross-classified measure of household income and neighbourhood deprivation. Compared to people on high incomes living in affluent neighbourhoods, the odds ratio of psychological distress for people on low incomes in affluent areas was 4.73 (95% confidence interval (95% CI) 4.39, 5.09), whereas that for people on low incomes in deprived areas was significantly higher at 5.83 (95% CI 5.41, 6.28). Conclusions. While no evidence was found to support local relative deprivation hypothesis, the pattern suggests that more affluent areas may contain features that are conducive to better mental health. Selection of bespoke geographical boundaries, use of directed acyclic graphs and more evaluations of natural experiments are likely to be important in taking the field of enquiry onwards.