Spatial variation of heat-related morbidity: A hierarchical Bayesian analysis in multiple districts of the Mekong Delta Region.

This study examined spatial variability of heat-related morbidity in multiple districts of the Mekong Delta Region (MDR), Vietnam. It was conducted in 132 district/cities of the MDR. We used a series of hierarchical Bayesian models to examine the region-wide and district-specific association between temperatures and hospitalizations during the period of 2010-2013. The potential effects of seasonality, long-term trends, day of the week and holidays were controlled in the models. We also examined influences of socio-demographic factors on the temperature-hospitalization relationship. The results indicate that an increase of 5 °C in average temperature was associated with a 6.1% increase (95%CI: 5.9, 6.2) in region-wide hospital admissions. However, the district-level risks ranged from a 55.2% decrease {95%CI: (-54), (-56)} to a 24.4% increase (24.3-24.6) in admissions per 5 °C increase in average temperature. This reflects the heterogeneous magnitudes of temperature-hospitalization risk across districts. The results also indicate that temperature-hospitalization risk increased by 1.3% (95%CI: 1.2-1.4), for each increase of 1000 persons/km in population density, 2.1% (95%CI: 2.04-2.11) for each 1% increase in percent of females, and 2.7% (95%CI: 2.6-2.8) for each 1% increase in percent of pre-school students. In contrast, the temperature-related hospitalization risk decreased up to 6.8% {(95%CI: (-6.6)-(-6.9)} for each 1% increase in rural population. Public health intervention measures for both short-term and long-term effects of heat-related health risk should be developed with consideration of the use of city/district scale for the factors rather than the province scale. The province scale of factors does not accurately represent the variability of health risk due to exposure to high temperatures.