Markov-based approaches to support policies makers in environment and healthcare

Perceived need of big cities is focused on policies limiting pollution from cars exhaust, energy production, industrial manufacture, with the aim to reduce the impact on both environment and citizens health/living. Decision support is mandatory, aiming at monitoring the city's status, evaluating policies effectiveness and estimating/predicting impacts of environment on health in order to plan more rationale resources allocation and care provision. We present a multi-period analysis performed through Markov-based approaches and prove how these methods may support decisions in the management of big cities, by integrating environment and health related issues. The study has considered historical pollutants data (2002 to 2011) for assessing air quality in Milan, according to the most updated European Directive; approaches have been used to evaluate the effectiveness of a traffic limitation policy brought into force in the 2007 (the Ecopass). Although several epidemiological studies have provided evidence of a relationship between air quality and mortality/morbidity due to cardiovascular and respiratory diseases, the prediction of emergency hospital admissions in the short term is yet difficult to achieve, even if crucial for a rational healthcare management. In this paper we propose to use Markov-based techniques in order to obtain a model to forecast hospital admissions, in the short term, according to pollution level. This work has been performed 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 (www.lenvis.eu). The Markov-based models, trained and validated on real data, have been deployed into the Lenvis' Health Impact Decision Support System (HIDSS) and made accessible, as services, to decision makers and users accounted. The system has showed its usefulness both for environment authorities and healthcare stakeholders.