An interesting study, which appeared two days ago in npj Digital Medicine, illustrates how it is possible to diagnose a cardiac arrest through the speech recognition of agonic breath.
Cardiac arrest outside the hospital is one of the leading causes of death worldwide. Rapid diagnosis and initiation of cardiopulmonary resuscitation (CPR) is the cornerstone of therapy for victims of cardiac arrest. Yet a lot of cardiac arrest victims have no chance of survival because they experience an unwitnessed event, often in the privacy of their own homes.
An under-appreciated diagnostic element of cardiac arrest is the presence of agonal breathing, an audible biomarker and brainstem reflex that arises in the setting of severe hypoxia.
The study, conducted by Justin Chan, Thomas Rea, Shyamnath Gollakota and Jacob E. Sunshine, shows how a supporting vector machine (SVM) can classify agonal breathing instances in real time within a bedroom environment, that is where most of these events take place.
Researchers used audio files of emergency 911 calls of confirmed cases of cardiac arrest that may include agonal breathing instances captured during the call.
Researchers used these files to train the SVM and accurately classify agonal breathing instances. They obtained an area under the curve (AUC) of 0.9993 ± 0.0003 and a working point with an overall sensitivity and specificity of 97.24% (95% CI: 96.86-97.61%) and 99.51% (95% CI: 99.35-99.67%).
The researchers then developed a prototype for a proof-of-concept to understand whether a supporting vector machine (SVM) can be trained to detect instances of agonic breathing associated with cardiac arrest in a bedroom and whether the SVM can be used to accurately classify agonic breathing audio in real time using existing smartphones and intelligent speakers.
If you are curious and want to know what conclusions they have reached, I invite you to read the study here.
Digital-health 