RESUMO
OBJECTIVES: To assess primary impact of selective Licensing (SL), an area-based intervention in the private rented housing market, on individual self-reported anxiety and neighbourhood mental health (MHI-Mental Healthcare Index) and secondary impacts on antisocial behaviour (ASB), population turnover and self-reported well-being. DESIGN: Difference-in-difference (DiD) was used to evaluate effects of SL schemes initiated 2012-2018. 921 intervention areas (lower super output areas) were matched 3:1 using propensity scores derived from sociodemographic and housing variables (N=3.684 including controls). Average treatment effect on treated (ATT) was calculated for multiple time period DiD in area-level analyses. Canonical DiD was used for individual-level analysis by year of treatment initiation while adjusting for age, sex, native birth and occupational class. SETTING: Intervention neighbourhoods and control areas in Greater London, UK, 2011-2019. PARTICIPANTS: We sampled 4474 respondents renting privately in intervention areas (N=17 347 including controls) in Annual Population Survey and obtained area-level MHI population data. INTERVENTIONS: Private landlords in SL areas must obtain a licence from the local authority, allow inspection and maintain minimum housing standards. RESULTS: ATT after 5 years was significantly lower for MHI (-7.5%, 95% CI -5.6% to -8.8%) than controls. Antidepressant treatment days per population reduced by -5.4% (95% CI -3.7% to -7.3), mental health benefit receipt by -9.6% (95% CI -14% to -5.5%) and proportion with depression by -12% (95% CI -7.7% to -16.3%). ASB reduced by -15% (95% CI -21% to -8.2%). Population turnover increased by 26.5% (95% CI 22.1% to 30.8%). Sensitivity analysis suggests overlap with effects of London 2012 Olympic regeneration. No clear patterns were observed for self-reported anxiety. CONCLUSIONS: We found associations between SL and reductions in area-based mental healthcare outcomes and ASB, while population turnover increased. A national evaluation of SL is feasible and necessary.
