USP’s app reaches hospitals to calculate chances of death – Health – Tecnoblog

The University of São Paulo (USP) has developed an application that calculates the chances of worsening among patients who contracted COVID-19. It will start being tested in the coming weeks by 85 doctors from 14 hospitals across Brazil. The tool has an algorithm that uses data from the patient’s tests for SARS-CoV-2, providing information that is able to assist physicians in medical care.


Application will use PCR test data for COVID-19 (Image: Itapevi Prefecture/ Flickr)

RandomIA, from USP, predicts worsening in COVID-19

The platform is called RandomIA. The test with the app that predicts the evolution of the infectious picture of patients with COVID-19 will be carried out in a coordinated survey carried out by the Faculty of Public Health (FSP), at USP. The app is able to anticipate possible cases of infection, hospitalization and death in Brazilian hospitals.

At the moment, researchers are evaluating the adjustment of RandomIA according to the needs of each healthcare professional, so that the tests use each patient’s clinical information.

To USP JournalAlexandre Chieavegatto Filho, the FSP professor who leads the app’s testing research, says that the app allows the inclusion of any artificial intelligence system. The first tests will be with software that gather data from COVID-19.

The researcher gave details about how the app that monitors patients and is able to predict even deaths works:

“The algorithm is first trained with real data about the disease, from the hospital where it is tested, and then incorporated into the application. We are currently in the validation phase of the application structure according to the physicians’ preferences.”

In the testing phase, a draw will be made to determine which doctors should receive the results of the application, while another group should not. Then, the prognoses of the two groups will be compared. In this way, the research is similar to the clinical study of vaccines and other drugs, where one group is medicated and the other receives only a placebo.


Only data from patients who undergo RT-PCR will be used in the RandomIA testing phase (Image: Federal Senate/ Flickr)

The app uses data from patients who took the RT-PCR exam, regardless of whether it was negative, positive, or even inconclusive. Roberta Wichmann, who is a postdoctoral fellow at FSP and oversees the project’s operations, highlighted:

“It is important to mention that no patient data is shared, nor with the coordinating center of the research, being only used locally by the physician to test the application.”

Once the physician fills in all fields with the patient’s information, it is possible to estimate both the patient’s diagnosis and prognosis by COVID-19. They are presented through graphs that tell how the patient may (or may not) develop the disease in a more severe form.

“Usability in e-health apps is little explored”

Roberta says that the use of apps in the e-health sector in Brazil is not very common, which motivated researchers to create the tool. “It is noted that the concern with the usability of e-health applications is still little explored and, in relation to applications that use artificial intelligence to support clinical decisions, the lack of scientific information is even greater”, explains the supervisor of the Randomia.

Also according to the researcher at the head of the project, the AI ​​tool also works as a guide to guide physicians’ decisions about COVID-19. In this sense, the data collected can be used to combat the pandemic, in general, within emergencies and emergency rooms.

RandomIA’s research is led by FSP’s LABDAPS (Big Data and Predictive Health Analysis Laboratory), under Alexandre’s command. Already under the leadership of Roberta, the operational part is supported by the CNPq (National Council for Scientific and Technological Development) and by Microsoft.

With information: Jornal da USP

Leave a Comment