A markov-based model to forecast emergency hospital admissions due to air pollution: the lenvis project approach

Antonio Candelieri1, Ilaria Giordani2,3, Paolo Testa2,3, Gaia Arosio1, Francesco Archetti1,2,3

1 Department of Information, Systems and Communications, University of Milano-Bicocca, Italy, viale Sarca 336, 20126, Milano
2 Consorzio Milano Ricerche, Italy, via Roberto Cozzi, 53, 20125 Milano, Italy
3 Consorzio Milano Ricerche, Italy, via Roberto Cozzi, 53, 20125 Milano, Italy


Several epidemiological studies proved pollutant levels and exposure may increase risk of morbidity and mortality, in particular for respiratory and cardiovascular diseases. However, performing a short term estimation of the hospital admissions due to air quality remains difficult even if crucial for a rational healthcare management. In this paper we present a Markov based approach aimed at estimating short term emergency hospital admission trends. This predictive model has been developed within the European project Lenvis (Local ENVIronmental Services), a collaborative network of services able to retrieve and analyze heterogeneous and geographically dispersed data sources in order to deliver environment and health information. One of the services of Lenvis is the Health Impact Decision Support System (HIDSS) whose inferential engine is provided by a Markov-based model trained on real data related to pollutant levels and emergency hospital admissions in Milan, Italy. HIDSS has shown, in several use cases, its usefulness both for environment authorities and healthcare stakeholders.


air quality; emergency hospital admissions; environmental health; Markov models; short term forecasting

Full Text:

 Subscribers Only