As nations around the world shift from containing the coronavirus that has wreaked so much havoc towards a strategy of living more alongside the virus, the ability to reduce the health implications and mortality for those infected will be crucial. A new study highlights how artificial intelligence can play a part.
The research, which was part of EIT Health’s Digital Control Centre for Covid-19 project began in April 2020 and has shown promising signs of being able to stratify and personalize treatment of the most serious Covid patients. This results in better treatment responses, and eventually a 50% reduction in mortality.
The main cause of death of Covid patients to date has been respiratory failure, however the team believe that these respiratory problems can be treated if the right care is provided at the right time.
Effective treatment
The researchers demonstrated their tool on a cohort of 786 patients in Hospital Clinic, Barcelona. It was able to predict, with 90% accuracy, the trajectory of the disease in each patient, which subsequently allowed for the right treatment to be administered at the right time.
The team believes that the use of the tool led to such improvement in the condition of each patient that by day five, 93.9% had been given a personalized therapeutic approach.
The AI-based tool crunches through data of all hospitalized Covid patients who are suffering from respiratory problems. It defines three distinct forms of clinical stratification, which reflect the symptom complications experienced by the patient: co-infection, thrombosis, and inflammation.
As with so much in healthcare, the earlier doctors have knowledge of these symptomatic patterns, the earlier they’re able to identify the clinical complications that need to be addressed via a personalized treatment plan.
The AI-based system was trained using over a trillion anonymized data points from Covid patients, and the researchers believe they have provided the hospital with a real-time control centre for all patients admitted.
“The artificial-intelligence system that we have built is capable of supporting clinicians in the early diagnosis of patients more prone to develop complications, thus we have been able to provide timely and personalised treatments. This ‘Central Control System’ can be used for multiple applications beyond COVID-19 and represents a clear example of how AI can improve medicine and health outcomes,” said Carolina García-Vidal, Hospital Clínic de Barcelona.
The next step is to utilize the EIT Health network to expand the solution into other hospitals throughout Europe, including University Hospital UZ Leuven, KU Leuven and Erasmus MC.
