The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Publications Copernicus
Articles | Volume XLII-4/W14
23 Aug 2019
 | 23 Aug 2019


L. Gianquintieri, E. G. Caiani, P. Brambilla, A. Pagliosa, G. F. Villa, and M. A. Brovelli

Keywords: Public Access Defibrillation, PAD, Automated External Defibrillator, AED, Out-Of-Hospital Cardiac Arrest, OHCA, Catchment Areas Mapping, Health-Geomatics

Abstract. To address the study of the deployment of publicly accessible Automated External Defibrillators (AED), Geomatics allows computing their limited area of effectiveness (i.e. ‘catchment area’, CA), traditionally set as circular surfaces with a 100m-radius. Exploiting open geospatial data related to roads network, also ‘realistic’ CAs, based on the effective walking distance, can be computed. Aim of this study (performed on the territory of Lombardy, Italy, total surface 23,863.65 km2, with open source software as QGIS, PostGIS, pgRouting) was to compare the two approaches, and to evaluate if the territory analysis could support case-by-case decision-making about the preferable mapping technique.

Setting a limit of 200 m, realistic CAs were computed for 7702 known AEDs on the territory (at 28/02/2018). The mean area obtained resulted close to that of the traditional 100m-radius circular area (33,665m2 against 31,415m2), but the spatial coverage of 45043 OHCAs - Out-of-Hospital Cardiac Arrests (Lombardy, 1/1/2015 to 31/12/2018) is very different considering realistic or circular areas (15.35% vs 9.43%). The distribution of the mapping error (realistic-CA – circular-CA) and the computation failures of realistic areas were studied and correlated with the characteristics of the surrounding territory considering attributes related to streets, buildings, and land-use, computing linear correlation coefficients and performing Mann-Whitney U-tests. Results suggest that realistic CAs are not always correctly computable and circular areas are preferable when AEDs are far from the streets in less urbanized and more uniform territories. An automatized decision-making about the best case-by-case mapping technique is therefore feasible with open data and open source software.