Project developed in the Marche region in collaboration with Vivisol, Almawave, ADiTech and the ASL for the development of an AI system for the identification of serious cases of COVID-19 through the identification of a Clinic Stability Index which alerts general practitioners who are in charge of the patient, who thanks to the developed algorithm can immediately decide whether to hospitalize the patient or treat him at home.
The experimental project was developed in 2021 and used in the Marche region in the municipality of Offagna. The project named RICOVAI_19 observed 129 patients in the 6 months of experimentation. Thanks to the system used, only one patient required admission to intensive care out of the 129 observed.
Telemonitoring and AI against Covid
During the COVID-19 pandemic, in order to reduce the pressure on healthcare systems and in particular to support access to the Emergency Department (ED) in a coordinated way, home telemonitoring of infected patients was strongly recommended. Additionally, numerous papers and articles have demonstrated the potential of using data from wearable medical devices to identify individual user risk lines caused by a viral infection, aggregating the data to determine disease trends and possibly trigger criteria population surveillance.
At the time of the project testing, there was no evidence of using artificial intelligence (AI) applications on digital data collected from patients and doctors in a similar scenario.
With a growing global population of infected users and increased use of connected wearable medical devices, it was thought that this could potentially help improve early diagnosis and management of infected individuals in order to reduce the burden on healthcare systems and in particular to support more appropriate ways of accessing the emergency room.
Artificial intelligence helps decide hospitalizations
During the RICOVAI-19 study (hospital admission with Artificial Intelligence tools in patients with COVid-19) conducted in the Marche Region, 129 subjects monitored at home were evaluated in a six-month period between 22 March 2021 and 22 October 2021 .
During monitoring, on-demand personal health technologies were used to collect clinical and vital data to feed the database and machine learning engine. The output of the artificial intelligence produced a clinical stability index (CSI) which allows the system to provide suggestions to the population and doctors on how to intervene.
The results showed the beneficial influence of the CSI in predicting the clinical classes of the subjects and in identifying which of them needs to be admitted to the emergency room. The same model made it possible to give active support to decision makers for the management of the infected user (in some cases requesting further tests or clinical information or hospitalization of the patient)
In conclusion, the study found how the impact of artificial intelligence tools could contribute to the outcomes of COVID-19 to fight this pandemic by improving the timeliness and precision of epidemic monitoring applicable to other cases of infectious diseases.
Below is attached a copy of the protocol activated during the study.
Vincenzo Gullà
Discover the presentation by Vincenzo Gullà, entitled “Technology in the fight against covid through the use of artificial intelligence”, on May 29th at 12.00 at SPS in the Talk area of the CNI-Fedinger stand.
