A good move for health?: Analyzing urban exposure trajectories of residential relocation and mental health in populations in Bradford

Abstract

Residential relocation can be leveraged as a natural experiment. This study examined the changes in environmental exposures due to residential relocation in two samples of within-city movers in Bradford (UK); 2089 residents (66% women, mean [SD] age, 47.80 [19.88] years) with preexisting common mental disorders-related prescriptions and 12,699 residents (60% women, mean [SD] age, 42.47 [17.40] years) without the same prescriptions at baseline (January-April 2021). Study data were extracted from National Health Service health records. The outcome was the presence of an active prescription for anxiolytics or antidepressants (yes/no) 1 year after relocation (January-April 2022). Change scores were calculated for several exposures, including the normalized difference vegetation index, distance decay to green spaces, coarse (PM₁₀) and fine particulate matter (PM_2.5), and nitrogen dioxide (NO₂) at pre- and postmove addresses. Logistic regression models were used for each change score exposure, adjusting for covariates selected using a direct acyclic graph validated against the data. Participants without prescriptions at baseline were likely to relocate to less green and less polluted areas compared with those with preexisting medication. A total of 15% of participants without prescriptions at baseline had an active prescription at follow-up. For these, increases in normalized difference vegetation index were associated with lower odds of having active prescriptions at follow-up [OR (odds ratio) = 0.93 (95% confidence interval [CI] = 0.88, 0.98), P = 0.007], whereas increases in PM_2.5 [OR = 1.1 (95% CI = 1.04, 1.16), P < 0.001] and PM₁₀ [OR = 1.12 (95% CI = 1.06-1.19), P < 0.001] concentrations were associated with higher odds. Changes in environmental exposures due to residential relocation showed an influence on mental health only for those participants without active prescriptions in the baseline.

Keywords: Air pollution; Antidepressants; Anxiolytics; DAG; Geographic information systems; Green spaces.

Figure 1.

Directed acyclic graph regarding the expected associations between change in environmental exposures and mental health in the follow-up. The arrows indicate causal paths. All the relationships were defined based on previous literature and discussions among the authors. Variables preceded by a “(U)” are unobserved variables.