NEW INVALUERS Information Management School (Nova IMS) and the University of Algarve (UALG) have created an AI model capable of predicting the risk of mortality in stroke patients, with a 98.5%hit rate.
In Portugal, a new artificial intelligence tool was created (IA)which anticipates clinical evolution of patients victims of stroke (stroke); and that can transform the way hospital screenings are made.
As a statement from Nova IMS sent to ZAP details, the system was developed based on real clinical data from more than 330 admitted patients in the Algarve University Hospital between 2016 and 2018, collected in the early hours of hospitalization. 24 risk factors were analyzed, from objective clinical indicators to the variable related to the patient’s health history.
Factors with the highest weight in the forecast include: the degree of severity of stroke to the input (NIHSS scale); the respiratory rate; glucose; Hemoglobin levels; Age; and the state of consciousness.
The used model uses an advanced artificial intelligence technique that combines multiple decision trees to identify patterns that are difficult to detect with human analysis only. In addition to high accuracy, the tool allows doctors to understand exactly which factors are influenced each forecast, which makes the clinical decision faster, grounded and transparent.
“This type of system allows health professionals to act earlier in the most severe cases and adapt care To the real needs of each patient, ”explain the authors of the study in the statement.
The investigation is part of the new vacancy of predictive and personalized medicine, which uses algorithms to anticipate risks, reduce complications and support decisions in real time.
The investigation team points out that this innovation can be extended to other conditions: “This model can be adapted to other critical conditions, such as hospital infections or risk of readmission, contributing to more efficient management and more humane medicine.”
