Spatiotemporal relationship analysis of the 2019-nCoV patients hospitalized in Istanbul: A retrospective database analysis

Abstract

Aim: The COVID-19 epidemic has reached every country in the world. Control strategies require effective tracing and isolation activities. Electronic mapping techniques are used in the visualization of spreading characteristics of COVID-19. The Geospatial Information System became an exceedingly popular open web tool to inform professionals and the public. These systems allow public health authorities to monitor the spreading characteristics and plan effective control strategies. The objective of this study was to identify the spatiotemporal mutual relationship of COVID19 patients living in two of the biggest districts of Istanbul (Kadikoy and Uskudar) who were admitted to the hospital. Methods: A total of 672 adult patients who were diagnosed with possible or confirmed COVID19 infection were included in the analysis. COVID19 diagnosis was confirmed either with positive RT-PCR test or radiographic chest imaging plus the presence of symptoms of the infection. Pearson correlation analysis and Moran’s correlation analysis were applied to the data set. Small pieces of regions [100,000 x 100,000] were set for the districts, and each event origin was fitted into the proper region using cartesian coordinate information. Getis-Ord hot spot analysis was performed to pinpoint the infections with higher concentration over time. Results: Pearson’s correlation revealed no significant results, while Moran’s analysis showed a significant correlation between distance and admission date [I: 0.64]. We identified at least 10 relevant hot spots in 3 districts. Conclusion: Determining the spatiotemporal relationship among cases of a central hospital may inform local authorities about dissemination patterns and help improve control measures against epidemics.