ABSTRACT
Objective To construct an air health index (AHI) based on the exposure-response relationships of air pollution and ambient temperature with the years of life lost (YLL) in Tianjin. Methods The time series database of air pollution, meteorological factors, and non-accidental YLL from 2014-2019 in six urban areas of Tianjin were established. The data from 2014 to 2017 were used as the construction set to establish the exposure-response relationships of air pollution and ambient temperature with non-accidental YLL and establish the AHI model. The data from 2018 to 2019 were used as the validation set for verifying AHI. The generalized additive model (GAM) and weighted quantile sum (WQS) model were used to establish the exposure-response relationship between air pollution mixtures and non-accidental YLL. The distributed lag nonlinear model (DLNM) was fitted to assess the exposure-response relationship between ambient temperature and non-accidental YLL. Based on these obtained coefficients, the AHI and air quality health index (AQHI) were built. By comparing the associations between AHI, air quality health index (AQHI), and air quality index (AQI) with daily mortality and YLL and model goodness of fit to evaluate the validity of AHI. Results The formula for AHIt=EYLLt,air pollution+ambient temperature/475.11*10. The validation results showed that each IQR increase in AHI was associated with a higher increase in non-accidental mortality and YLL (10.61% and 353.37 person-year) compared with the corresponding values of AQHI and AQI. In addition, the model goodness of AHI was better than AQHI and AQI model. Conclusion Compared with AQHI and AQI, the AHI based on the integrating health effects of air pollution and ambient temperature has a better health risk prediction ability.