[ad_1]
Through Writing La Rochelle
Published on
Numa Health, one to boot from La Rochelle, developed and is testing his own at the city hospital Software computer that can predict potential risks complications for the ill of coronavirus within five days.
Here is one she studies which could revolutionize the treatment of patients of COVID-19 in the next months.
In collaboration with the La Rochelle-Ré-Aunis hospital group, which performs the test and study this application is undergoing scientific validation.
A simple blood test
Imagine that you can predict, as soon as you arrive at the hospital, whether your health will deteriorate or not. It seems unlikely, yet this is what this software does thanks to algorithms and blood tests.
The principle is to use a simple blood test to analyze which patients might see their condition worsen.
He says that when every Covid-19 patient enters the hospital, all of his markers aren’t necessarily red, but the software can predict that.
The algorithms could therefore allow better identification of patients showing signs of deterioration and thus save lives through faster treatment.
Encouraging results
If the study is only in the first phase, also validated last Thursday, the results are still very promising.
Indeed, the hospitals of La Rochelle-Ré-Aunis carried out retrospective tests, “that is, observing the results a posteriori”, corresponding to the first part of the study.
In other words, the software was trained on 50 confirmed cases previously passed to the La Rochelle emergency room. “We compared the predictions of the algorithms with the real future of the patients,” explains David Chalvet.
And the results are compelling according to Numa Health. As for patients with a low risk of degradation, “the application was correct in 70% of cases”, reveals the director.
For high-risk patients, “we increase to 95%, which means that patients have died from the virus or entered intensive care within the next five days, as announced by the software.”
Place in stage 2
After a first part of a validated retrospective study, the time has come for phase 2.
“It consists of including other patients in order to increase the power of the study, but also to introduce additional factors (history, weight, radiology)”. It should finish in early January.
This phase is expected to be extended to 90 patients before moving to a prospective study in 2021. “ER doctors will be able to use this score live as soon as patients arrive, and then immediately judge cases in real time.”
This application may be available in late 2021, early 2022 at the latest.
By Corentin Cousin.
Source link
