Direct assessment of health impacts on hospital admission from traffic intensity in Madrid

Abstract

In this paper we establish the attributable risk on respiratory and cardiovascular disorders related to traffic intensity in Madrid. In contrast to previous related studies, the proposed approach directly associates road traffic counts to patient emergency admission rates instead of using primary air pollutants. By applying Shapley values over gradient boosting machines, a first selection step is performed among all traffic observation points based on their influence on patient emergency admissions at Gregorio Marañon hospital. A subsequent quantification of the relative risk associated to traffic intensity of the selected point is calculated via ARIMA and log-linear Poisson regression models. The results obtained show that 13% of respiratory cases are related to traffic intensity while, in the case of cardiovascular disorders, the percentage increases to 39%.

Keywords: ARIMA; Attributable risk; Hospital admissions; Poisson regression; Traffic intensity.