Researchers of the University of Coimbra develop new approach for improving cardiovascular risk assessment

English version: Diana Taborda

A research team of the Centre for Informatics and Systems of the University of Coimbra (CISUC – Department of Informatics Engineering, Faculty of Sciences and Technology), has developed a new approach based on data fusion techniques that improves the assessment of cardiovascular risk following an Acute Coronary Syndrome (ACS) event.

During the research "GRACE PLUS - A data fusion-based approach to improve GRACE score in the risk assessment of Acute Coronary Syndrome", authored by Afonso Neto, PhD student at the Department of Informatics Engineering (DEI), Jorge Henriques and Paulo Gil, CISUC researchers, and José Pedro Sousa, cardiologist at the Coimbra Hospital and University Centre (CHUC), it was possible to conclude that there are risk factors not included in the Global Registry of Acute Coronary Events (GRACE), namely haemoglobin levels at admission.

“We started by looking into ways of adding new risk factors identified by our clinical partner (CHUC) to the GRACE prediction model, with the potential to improve prognostic accuracy, such as haemoglobin at admission and inflammatory markers”, explains Jorge Henriques.

"On the other hand", adds the CISUC researcher, "to preserve the interpretability of the new model and its reliability, perceived by the clinical team, the maintenance of the way GRACE is applied in clinical practice was defined as an additional requirement, albeit subject to a correction factor, due to the input of new risk factors. Hence the name GRACE PLUS".

GRACE can be considered a secondary prevention tool and is usually applied to patients at the time of hospital admission following an event of Acute Coronary Syndrome (myocardial infarction). The goal of this model is to predict the risk of death or of a new ACS event in a specific period of time, usually in the course of the following month or the next six months.

This tool is based on eight independent risk factors recorded at hospital admission, namely age, heart rate, systolic blood pressure, creatinine, Killip classsification, stage of cardiac arrest, elevated cardiac markers and ST-segment deviation on the electrocardiogram (ECG). "By weighting all these factors, GRACE generates a score, which is subsequently discretised in order to provide a risk category, in three distinct classes, namely low risk, intermediate risk and high risk," explains the research team.

Thus, this study focused on the "development of a tool based on data fusion techniques and artificial intelligence techniques, which allows improving the assessment of cardiovascular risk and thus provide a more informed support to clinical decision in a real-life context without, however, changing the way professionals actually make their decisions", says Jorge Henriques.

The authors of the project consider that, "in addition to improving the characterisation of cardiovascular risk, another advantage of this study is to enhance the provision of health care in accordance with the real risk that the patient presents at admission, which may contribute to an effective management of therapies and human resources".

The scientific article, published in the journal Information Fusion, is available at: https://www.sciencedirect.com/science/article/pii/S1566253522001956?via%3Dihub#sec2