Study of the Behavior of the pair Klebsiella pneumoniae | Carbapenems in several European Countries by CLUSTERING HEALTH DATA TIME SERIES WITH THE GENERALIZED AFFINITY COEFFICIENT

Antimicrobial resistance (AMR) constitutes a major public health and economic burden in contemporary society. The aim of this work is to identify patterns and trends associated with the resistance behavior of Klebsiella pneumoniae to carbapenems across several European countries in order to better understand its dynamics and public health implications, since it’s a pair considered by the WHO as a critical priority. Variables: The dataset resistance values of K.pneumoniae|Carbapenems for several countries from 2005 to 2021(data available up to 2023) were obtained from the public website of ECDC. Methods: Agglomerative Hierarchical Cluster Analysis (HCA) was based on the generalized Affinity coefficient applied to the estimated ARIMA models. Results:The dendrogram at the cut-off presented four clusters: the first cluster was Slovenia, Luxembourg, Italy, and Estonia; the second cluster was Romania, Malta, Greece, Portugal, and Spain; the third cluster was Denmark, France, and Czechia; and finally, the fourth cluster was Lithuania, Cyprus, Finland, Austria, Belgium, and Slovakia. Although the resistance values show a more visible increasing trend over the years in Clusters II and III, the increase is more pronounced. Data were standardized, and the slopes of the lines obtained between the years 2012 and 2020 were calculated for Clusters II and III. For Clusters I and IV, this was not possible due to the irregularity of the curves. The slopes are higher in the countries belonging to Cluster II. Specifically, the slopes for Cluster II’s countries are m=0.29, m=1.5, m=0.06, m=0.11, and m=0.26 for Portugal, Spain, Greece, Malta, andRomania, respectively, whereas for Cluster III’s countries, the slopes are m=0.02, m=0,08 and m=0.001 for France, Czechia, and Denmark, respectively. It is relevant that within Cluster II, Portugal and Spain show a steeper increase, with Spain showing the most pronounced growth. Conclusions: The application of data analysis, using HCA based on the generalized Affinity coefficient, allows the evaluation of patterns among different regions and gives the possibility to identify alarming groups that need urgent intervention