Back

Tri-light Warning System for Hospitalized COVID-19 Patients]{A Tri-light Warning System for Hospitalized COVID-19 Patients: Credibility-based Risk Stratification under Data Shift

Xu, C.; Xu, Q.; Liu, L.; Zhou, M.; Xing, Z.; Zhou, Z.; Zhou, C.; Li, X.; Wang, R.; Wu, Y.; Wang, J.; Zhang, L. J.; Zhan, X.; Gevaert, O.; Lu, G.

2022-12-13 radiology and imaging
10.1101/2022.12.11.22283309 medRxiv
Show abstract

OBJECTIVETo develop a tri-light warning system for the early warning of novel coronavirus pneumonia (COVID-19) and stratification of patients. MATERIALS AND METHODSThe system extracts radiomic features from CT images and integrates clinical record information to output a prediction probability and credibility of each prediction. It classifies patients in the general ward into red (high risk), yellow (uncertain risk), and green (low risk) labels. The system was tested using a multi-center cohort of 8,721 patients. RESULTSThe system demonstrated reliability and performance validation under data distribution shifts, and was applicable to both the original strain and variant strains of COVID-19. DISCUSSIONThe tri-light warning system has the potential to improve patient stratification performance and identify epidemiological risks early, thus allowing for timely treatment and optimization of medical resource allocation. CONCLUSIONThe tri-light warning system based on conformal prediction is a reliable and effective method for the early warning and stratification of COVID-19 patients.

Matching journals

The top 8 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.