Illustrative project image
Approved by mentor

Medical tags

Clinical need
Remote or self-diagnosis
Public health
Project keywords
Artificial intelligence
Device classification

Project description

In December 2019, a novel coronavirus (now called SARS-CoV-2) was detected in three patients with pneumonia connected to the cluster of acute respiratory illness cases from Wuhan, China. By the end of February 2020, several countries were experiencing sustained local transmission of coronavirus disease (named COVID-19) in people of all ages. Although some people with COVID-19 have mild to moderate symptoms - some people may have no symptoms at all - the disease can cause severe medical complications and lead to death in some people. Complications can include pneumonia in both lungs and/or organ failure in several organs.

It was hypothesized that people who are older or who have existing chronic medical conditions, such as heart disease, lung disease or diabetes, or who have compromised immune systems may be at higher risk of serious illness. However, also young people with no pre-existing medical conditions appeared to become severely ill. Thus, at the moment the conditions that predispose people to develop serious illness are largely unknown and the possibility to understand them is hampered by a general lack of data; when present, data may have a multitude of technical issues, including non-uniform sampling which lead to several missing values. Moreover, such conditions may be aggregated in clusters and such information may be hidden or not interpretable a priori.

Artificial Intelligence (AI) and in particular Machine Learning (ML) algorithms have been widely used in the management of chronic diseases for analyzing electronic health record (EHR) data and have proven useful for dealing with analogous issues. Thus, the aim of the proposed project titled “aiCOVID-19” is focused on developing a decision support system able to prevent, understand, and treat the complications that arise in COVID-19 patients. Specific objectives are:

·         to deal with real-case clinical scenario EHR issues (i.e., incomplete or missing data, registration errors, longitudinal and transversal data sparsity, etc.);

·         to extract EHR clinical features (e.g., specific medications, lab tests) and identify possible clusters related to the onset or development of patient’s complications;

·         to develop innovative solutions and predictions whether a patient belongs to a risk class related to specific COVID-19 complications over a short period of time.

The development of a clinical decision support system integrated in an EHR framework, capable of predicting the risk profile of morbidity and resource consumption of the patients according to their clinical features, would allow more appropriate and intensive care for the patient and a more appropriate use of human and economic resources of the national health system.

The collaboration of governmental organizations and clinical partners will be strategic for the development of the project. To this purpose, already-existing collaborations that can be useful for the project are: DPC - Dipartimento di Protezione Civile;  ASUR Marche - Azienda Sanitaria Unica Regionale Marche; FIMMG - Federazione Italiana Medici di Medicina Generale; INRCA - Istituto Nazionale di Riposo e Cura per Anziani.    

Below some references of the latest published work of our group for the development of innovative ML methods to support the physician in screening and diagnosis tasks using EHR data.

  1. Bernardini, M., Morettini, M., Romeo L., Frontoni, E., Burattini, L., Early temporal prediction of Type 2 Diabetes Risk Condition from a General Practitioner Electronic Health Record: A Multiple Instance Boosting Approach, Artificial Intelligence in Medicine, 2020 (accepted).
  2. Bernardini, M., Morettini, M., Romeo L., Frontoni, E., Burattini, L., TyG-er: An ensemble Regression Forest approach for identification of clinical factors related to insulin resistance condition using Electronic Health Records, Computers in Biology and Medicine, Volume 112, 2019.
  3. Bernardini, M., Romeo, L., Misericordia, P., and Frontoni, E. Discovering the Type 2 Diabetes in Electronic Health Records using the Sparse Balanced Support Vector Machine. IEEE Journal of Biomedical and Health Informatics 2019.