Web App for prediction of hospitalisation in Intensive Care Unit by covid-19.
Greici Capellari Fabrizzio, Alacoque Lorenzini Erdmann, Lincoln Moura de Oliveira
PMID 38055477WHAT IT FOUND
A new web calculator for covid-19 ICU risk showed a 68% chance in simulated cases, including both low and high risk.
It may help nurses assess severity, but it is not validated for patient care.
Key findings
01The best tested model was a decision tree, with an AUC of 0.668.
02The Web App lets nurses enter 23 patient variables and view a predicted ICU admission probability.
03Efforts were made to reduce false negatives, where patients who needed ICU care were classified as unlikely to need it.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study included only patients with a positive SARS-CoV-2 test and a hospital stay longer than 72 hours. It may not apply to patients who did not meet these criteria. Data were collected by telephone interview from patients who had been discharged to home. This may limit accuracy for patients too unwell to respond and for real-time bedside use. The model used 23 variables but did not include vaccination status. The paper notes this may have affected performance. The best model had an AUC of 0.668, and one example showed model accuracy of 64.7%. The classes were unbalanced before undersampling, and 476 patients remained for model development. This may reduce generalizability. The tool was not validated by users, patients, health professionals, or administrators. The study was done at five Brazilian university hospitals and required fluent Portuguese. Results may not transfer to other settings.
Declared interests
The paper reports that author Fabrizzio GC received a doctoral scholarship from the University Scholarships Program of Santa Catarina (UNIEDU/FUMDES). It also says the study was part of a larger project funded by MCTI/CNPq/CT-Health/MS/SCTIE/Decit call no. 07/2020 and FAPESC call no. 004/2020.
The easy way to misread this
Do not use the app to decide ICU admission. The best model had an AUC of 0.668, the test data included 20 false negatives, and the paper states the tool still needs validation by users, patients, health professionals, and administrators.