Authors:
Rute Almeida
1
;
Ana Cláudia Teodoro
2
;
Hernâni Gonçalves
3
;
Alberto Freitas
3
;
Ana Sa-Sousa
3
;
Cristina Jácome
3
and
João Fonseca
4
Affiliations:
1
University of Porto and CMUP - Centre of Mathematics of the University of Porto, Portugal
;
2
Faculty of Sciences and University of Porto, Portugal
;
3
University of Porto, Portugal
;
4
University of Porto, University of Porto and CUF Porto Institute & Hospital Estrada da Circunvalac¸ao, Portugal
Keyword(s):
Asthma Exacerbation, NDV I, Temperature, NO2, Air Pollution, Meteorological Parameters, Forecasting System.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Business Analytics
;
Cardiovascular Technologies
;
Computing and Telecommunications in Cardiology
;
Data Engineering
;
Decision Support Systems
;
Decision Support Systems, Remote Data Analysis
;
Health Engineering and Technology Applications
;
Knowledge-Based Systems
;
Symbolic Systems
Abstract:
Asthma has a major social impact and is prone to exacerbations. It is known that environmental factors, such as meteorological conditions and air pollutants, have a role over their occurrence. In a previous work, positive associations were found between hospital admissions due to asthma exacerbation at highly urbanized regions of Portugal and higher atmospheric NO2 levels, lower vegetation density and higher air temperatures, estimated using remote sensing. In this study we propose the use of georeferenced environmental factors to forecast the risk of hospital admissions due to asthma exacerbation. We applied linear discriminant analysis using monthly averages based in 2003–2007 environmental data to forecast positive monthly admission rates in municipalities of Lisboa district (Portugal) during 2008. Space-time estimates of nitrogen dioxide (NO2), vegetation density from MODIS Normalized Difference Vegetation Index (NDV I) and near-surface air temperature (Ta) were considered as ind
ependent variables. We identified over 65% of the combinations months/municipalities having hospital admissions in the testing set, with less than 10% of false positives. These results confirm that NO2, NDV I and Ta levels obtained from remotely sensed data can be used to predict hospital admissions due to asthma exacerbation, and may be helpful if applied in warning systems for patients in the future.
(More)