How does a Recurrent Neural Network work?
Recurrent Neural Networks (RNN’s) are best at analyzing time series data. They work by looking at the subtle interactions between different variables and how they change over time. In our case, these variables are things like Diurnal Rhythm, Heart Rate, Temperature etc. By looking at these variables, the RNN develops a “fingerprint” of sorts for sepsis.
Is sepsis really that bad?
According to the Centers for Disease Control (the CDC), Sepsis affects roughly 1.7 Million Americans and kills 270,000. It is currently a national priority for America. Sepsis is also the most expensive hospital admission that exists with per admission estimates at $14,000 to $40,000. If you multiply 1.7M Americans by $14,000 you get a total direct cost of $23.8Bn.
What does the device measure?
The Patchd device measures the same kinds of physiological measurements that are taken in an Emergency Department or ICU, such as Heart Rate, EKG and Respiration Rate.
Why do you target outpatients?
According to the CDC, 80% of sepsis occurs in outpatients. By targeting outpatients we are able to potentially alter the patient journey. Our dream is to keep the patient out of hospital, living an awesome life, and coming in to the outpatient clinic to be seen, or going to a pharmacy to collect antibiotics when required by a physician.