Understanding which patients admitted to intensive care for Covid-19 are most at risk of fatal events thanks to Machine Learning techniques. This is what the aim is to make the work of health professionals more and more efficient. A recent study coordinated by the Covid-19 Veneto Icu Network and the Biostatistics Epidemiology and Public Health Unit of the University of Padua opens our eyes to the importance of new techniques for analyzing clinical data through Machine Learning.
The role of Artificial Intelligence
Machine Learning (or machine learning) is one of the main branches of artificial intelligence and is a fundamental technology for managing and understanding the enormous amount of data, health and non-health, that we produce on a daily basis. The use of Machine Learning in medicine allows the faster and more precise identification of the mechanisms underlying a disease or its degeneration, but also to define a therapy based on the personal characteristics of the patient.
Blood cancer, Artificial Intelligence and machine learning for tailor-made treatments
by Irma D’Aria
Research
The study, published in the Journal of Anesthesia, Analgesia and Critical Care and coordinated by Paolo Navalesi, director of the Covid-19 Veneto Icu Network, and by Dario Gregori, director of the Epidemiology and Public Health Biostatistics Unit of the Department of Cardio-Thoraco-Vascular Sciences and Public Health of the University of Padua, highlights the advanced age of patients as one of the predictors of greater mortality for Covid-19 in intensive care.
A research and innovation center for aging well
by Letizia Gabaglio
Networking with biostatisticians
The research is the result of a joint effort between a network of anesthesiologists, resuscitators and biostatisticians. “It is important to underline that the awareness of the need to network is becoming the new modus operandi to face the challenges of epidemiological emergencies”, he explains Paolo Navalesi, director of the Covid-19 Veneto Icu Network. Even the intensive care units are no longer an island within the hospitals. Never before has their function and operation been at the center of public attention as in recent months. The 25 Operational Units of the Veneto are already working online to collect the data of each hospitalized patient. For this study we asked for the involvement of our Milan colleagues to add the vision of a non-regional reality but equally committed on the Covid front. But above all, the network that has seen us involved in an innovative way and that allows us to have an important interpretation, is the one created within our University between the Anesthesia and Intensive Care Institute and the Biostatistics Unit “.
SLA, Italian study: with Artificial Intelligence the diagnosis comes first
by Irma D’Aria
The development of the algorithm
The work aimed to develop a tool, through a Machine Learning approach, capable of identifying those hospitalized for Covid-19 most at risk of fatal events. In the 25 ICUs in the Veneto region, the researchers studied the relationship between patient characteristics and mortality cases, examining various elements including age, gender, organ failure assessment score, need invasive mechanical ventilation, tracheostomy or re-intubation, prone position during ICU stay and ICU readmission. Work yes
is based on the data of 1616 patients admitted to the intensive care units of the Covid-19 Veneto Icu Network and the Irccs Ca ‘Granda Ospedale Maggiore Policlinico di Milano from February 28, 2020 to April 4, 2021. The models highlighted age as the predictive parameter more important on the mortality of these patients.
Person-based clinical research
Since the onset of the pandemic, the development of predictive models has attracted great interest due to the initial lack of knowledge about Covid-19 diagnosis, treatment and prognosis. “These techniques are the methodological tool at the basis of person-centered clinical research, that is, clinical research based on the person and personalized medicine, based on the analysis of clinical data, integrated with algorithms and with the big data available”, he underlines Dario Gregori, director of the Epidemiology and Public Health Biostatistics Unit of the Department of Cardio-Thoraco-Vascular Sciences and Public Health of the University of Padua. The clinical variables investigated represent only a small number of parameters potentially relevant and capable of influencing the outcomes of critically ill patients. Additionally, several patients had incomplete records, due to the enormous workload for ICU doctors during the Covid-19 pandemic. Nevertheless, the results of this research represent the umpteenth demonstration of the great importance of the collaboration between biostatisticians and doctors to better understand the dynamics of the progress of each disease ”.
.