First Clustering Analysis of COVID in Portugal

Abstract

There is an increasing need to understand the behavior of COVID-19, in this case, what type of medical preconditions can influence the recovery of the infected patient and what age groups are more affected. After the Directorate-General of Health of Portugal (DGS) made available the first records gathered from the infected, it became possible gather some conclusions. In this context, ioCOVID19 project arises, which wants to identify patterns and develop intelligent models able to support the clinical decision.

This article explores which typologies are associated with different outcomes to provide some insights regarding the consequences after the coronavirus infection. To understand which profiles, stand out, a clustering algorithm was used, 65 experiments were carried out, from which 192 clusters were obtained. From this study, the most relevant profiles are the following: the profile associated with death are patients with Diabetes – aged between 44 and 98 years old (19.74%); regrading hospitalized patients who died, the profile achieved was patients with Chronic Kidney Disease – aged between 52 and 102 years old (17.63%); for patients hospitalized in ICU who died the profile obtained was Cardiovascular Diseases – aged between 61 and 88 years old (26.23%); in regards to patients who died after being submitted to ventilatory support the correlated profile are patients with Cardiovascular Diseases – aged between 62 and 99 years old (32.17%). With the completion of this study it was possible to detect a set of profiles that are associated with different clinical conditions.