Study to Develop and Validate a Digital Tool to Detect Normal and Abnormal Heart Sounds in Newborns Under 30 Days Old

What this Trial Is About

This research aims to improve the early detection of congenital heart diseases CHD in newborns, focusing on rare but serious ductal-dependent cardiovascular conditions. Current newborn screening methods, including physical examination and oxygen saturation measurement, have limitations in sensitivity and specificity, which can lead to missed diagnoses or unnecessary follow-up tests. The study is conducted by IRCCS Azienda Ospedaliero-Universitaria di Bologna and other centers to develop a new digital tool that analyzes heart sounds using machine learning to better distinguish healthy from abnormal cases. The study has two main phases a derivation training phase and a validation phase. In the derivation phase, heart sounds from newborns with known heart conditions will be recorded using a digital stethoscope and analyzed to create a binary classification algorithm distinguishing normal from abnormal sounds. In the validation phase, the algorithm will be tested on a separate group of newborns undergoing standard cardiovascular screening. All babies will have echocardiograms for confirmation. The study will also compare the costs and benefits of digital screening versus current methods. Participants will be newborns under 30 days old, with heart sounds recorded in a calm setting without sedation. Clinical exams, oxygen saturation measurements, and echocardiograms will be performed. The study will measure the accuracy of the digital classifier in detecting heart abnormalities, including sensitivity and specificity, and evaluate economic impacts. Data will be securely stored and analyzed, with a total study duration of about one year for outcome measurement.

Digital diagnoSis of Cardiac sOUNd in peDiatric Patients [DI-SOUND Study]