Back

Phenotypic grouping of Catheter-Associated Escherichia coli from COVID-19 isolation wards using Hierarchical clustering in Surabaya

Edbert, D.; Endraswari, P. D.; Mertaniasih, N. M.

2023-07-28 intensive care and critical care medicine
10.1101/2023.07.26.23293190 medRxiv
Show abstract

IntroductionModerate to critical COVID-19 patients may be indicated for urinary catheter use due to the risk of immobility and ventilator or oxygen use. In intensive care units, 18-81.7% of all patients use a urinary catheter. Almost all patients with urinary catheters suffered from bacteriuria within 30 catheter-days. Hospital-associated isolate tracing is mainly performed using complex molecular tests that are not vastly available. This study aims to trace catheter-associated urinary tract infection (CAUTI) isolates using a common hierarchical clustering method that is vastly available MethodsThis is a descriptive study presenting a collection of Escherichia coli culture data performed by dr. Soetomo Public Hospital microbiology laboratory from March, 26th 2020 to March, 31st 2021. Hierarchical clustering was performed using statistical software using Wards clustering method. ResultsThere are 36 E.coli associated with CAUTI. Isolate biochemistry profile and minimum inhibitory concentrations profiles were clustered into 3 clades for each profile. A total of 9 cluster combinations were found. Cluster-ID 1 was melibiose fermenters, Cluster ID 2 was a non-Arginine utilizer, and Cluster ID-3 was an Arginine utilizer. Cluster MIC A consists of third-generation Cephalosporin resistant isolates, Cluster MIC C was multi-susceptible isolates. The Chi-square test between cluster ID and MIC showed no significant differences between the number of isolates per group (X2, p = .430, CI = 95%). ConclusionCAUTI associated E.coli is divided into 9 clusters. This indicates no cluster dominates the isolates, thus CAUTI is not caused by hospital transmission but by normal flora carried by the admitted patient.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.