Cardiac arrhythmia is found with software

Can software detect a heart defect better than a doctor? It can and must. It may be because doctors have laid the foundations for a catalog that today distinguishes 24 different types of cardiac arrhythmias; must because technological advancement has produced sensors that are increasingly capable, precise and less invasive. The result is that even a difficult distinction such as that between atrial fibrillation and atrial flutter can become routine when an algorithm analyzes the data. Avoiding this type of confusion means being able to proceed with the best of diagnoses, eliminating initial errors that could compromise the subsequent treatment path.

MODEST: diagnosis and algorithms

MODEST is a program financed by 'European Union aimed at developing mathematical models to support medical diagnosis. Innovation and experience thus become synergistic assets at the service of the patient, strengthened by what machine learning is able to do starting from the previous diagnoses. MODEST immediately returned important results, with a sensitivity in the identification of atrial fibrillation equal to 81%.

So the head Sebastian Sager explains what the experiment moves from:

We tried to to find mathematical models in the literature that accurately describe what is happening, but in the end we had to create our own. The second step was optimization, involving many different disciplines: mathematics, medicine, computer science, biology and machine learning.

Instruments of this type are able to multiply the accuracy of the diagnosis and, consequently, substantially improve the effectiveness of treatments. Much of the future of healthcare passes through projects of the caliber of MODEST (and the related funding), which today can sit on the shoulders of giants with tools and analytical skills that were unthinkable until just a few years ago.