Early Detection of Severe Apnea-bradycardia in Preterm Infants Using Algorithm Fusion

Apnea-bradycardia is a frequent phenomenon in preterm infants which has been associated with short term complications and alterations in neurodevelopment. Duration and amplitude of the bradycardias depend on the time delay between the beginning of the bradycardia and the intervention of the nurse.

The purpose of the study is to test in clinical units and in real life a new method set for early detection of bradycardia in preterm infants.

Each preterm infant cardio-respiratory traces will be recorded during two 4-hours periods of time (the first one at the time of occurrence of apnea-bradycardias and the second one during the following week). During those periods, the alarms of the monitors usually used in the units will also be recorded through automatic detection and these alarms will be compared with the alarms generated by the algorithm tested, for false positive and precocity of the detection. A false positive alarm is a detection which is not associated with a bradycardia of less than 80 bpm. The system will be considered efficient if it allows an earlier detection of severe bradycardias without increasing the number of false positive detection.