Modeling the prevalence of respiratory chronic diseases risk using satellite images and environmental data

Abstract

Several studies have demonstrated that air quality and weather changes have influence in the prevalence of chronic respiratory diseases. Considering this context, the spatial risk modeling along the cities can help public health programs in finding solutions to reduce the frequency of respiratory diseases. With the aim to have a regional coverage and not only data in specific (point) locations, an effective alternative is the use of remote sensing data combined with field air quality data and meteorological data. During the last years, the use of remote sensing data allowed the construction of models to determine air quality data with satisfactory results. Some models using remote sensing based air quality data presented good levels of correlation (R2 > 0.5), proving that it is possible to establish a relationship between remote sensing data and air quality data. In order to establish a spatial health respiratory risk model for Quito, Ecuador, an empirical model was computed considering data between 2013 and 2017, using the median data values in each parish of the city. The variables are: i) 46 Landsat-8 satellite images with less than 10% of cloud cover and some indexes (normalized difference vegetation index NDVI, Soil-adjusted Vegetation Index SAVI, etc.); ii) air quality data (nitrogen dioxide - NO2, Ozone - O3, particulate matter less than 2.5μ - PM2.5 and sulfur dioxide - SO2) obtained from local air quality network stations and; iii) the hospital discharge rates from chronic respiratory diseases (CRD). In order to establish a probability model to get a CRD, a logistic regression was used. The empirical model is expressed as the probability of occurrence during the studied time. All the procedures were implemented in R Studio. The methodology proposed in this work can be used by health and governmental entities to access the risk of getting a respiratory disease, considering an application of remote sensing in the environmental and health management programs.