Assessing the stratification and the prognostic value of SARS-CoV-2 routine data in the context of advanced hospitalization rate 2.0 (HR 2.0)

by Martin Misailovski
Doctoral thesis
Date of Examination:2025-01-21
Date of issue:2025-01-20
Advisor:Prof. Dr. Simone Scheithauer
Referee:Dr. Oliver Bader
Referee:Prof. Dr. Heike Bickeböller
Persistent Address: http://dx.doi.org/10.53846/goediss-11030

 

 

Files in this item

Name:Doktorarbeit_Misailovski.pdf
Size:1.75Mb
Format:PDF

This file will be freely accessible after 2026-01-21.

Abstract

English

Background: Machine learning is becoming increasingly popular in the field of infection prevention and control. This is particularly evident in the context of surveillance, where screening clinical prediction rules (CPR) can be used with routine patient data. Aim: Several machine learning algorithms as CPR were evaluated based on their performance metrics on the example of the so-called Hospitalization Rate 2.0 (HR 2.0), which aims to differentiate cases admitted due to COVID-19 from incidental SARS-CoV-2 positive cases. Materials and methods: A monocentric, retrospective, cross-sectional study design was used for model development and model comparison. Patients with a positive SARS-CoV-2 RT-PCR test at hospital admission were included for model development (01/01/2022 - 06/30/2022). Patients were characterised as primary case (admitted due to COVID-19) versus incidental case (SARS-CoV-2 positive but admitted for another reason) using clinical reasoning and the six-eye principle. 6 models were applied using
Keywords: Surveillance; Machine Learning; Hospitalisation Rate; Clinical Prediction Rule