A spatial multilevel analysis of the impacts of housing conditions on county-level life expectancy at birth in China.

Housing and health studies rarely concentrate on aggregate health indicators. There is a need for more research to disentangle the complexity within housing types and between housing and socioeconomic development as they influence health. To enhance the evidence base for the understanding of the association between housing and health, from a geographic perspective, we applied spatial multilevel analyses to census and statistical yearbook data of China to investigate what and how housing conditions were associated with life expectancy at birth (LEB). The results found significant spatial auto-correlation of LEB across nearby counties and significant spatial clustering of LEB across provinces. Spatial multilevel models fit the data better than regular regression models. Housing conditions, including housing tenure and internal housing quality, were significantly associated with LEB independent of socioeconomic factors. Compared to access to clean energy and tap water, where the associations with LEB were conditional on housing tenure, having a washroom inside the house, crowding and housing tenure were robustly related to LEB. The associations between internal housing quality and LEB were confounded by housing tenure. The coefficients of housing facilities decreased or became insignificant and the coefficients of crowding increased after housing tenure was considered. Housing tenure might be one of the mediators of the economy to indirectly impact LEB, and an independent social indicator which might affect LEB directly. The associations between housing conditions and LEB were stronger in counties situated in less-developed provinces. These findings indicate that, to narrow regional gaps in life expectancy, improving the housing facilities in the least developed regions should be a priority. With rapid urbanization and the vast migration, ensuring urban dwellers have access to houses and reducing housing congestion should be the primary policy concerns all over China, but especially in the less-developed regions. • Spatial autocorrelation and clustering issue of life expectancy are considered. • Housing tenure and housing quality are independently associated with life expectancy. • Housing tenure is one of the mediators of the economy to influence life expectancy. • The impact of housing quality is confounded by housing tenure. • The impacts of housing condition are stronger in counties in less-developed provinces. [ABSTRACT FROM AUTHOR]