Last year, of the 55 million people that died around the world, 11 million died from sepsis. 80% of these people could have been saved, if only we'd had a way to predict sepsis early. The reason we don't, is that almost all sepsis starts at home, and by the time a patient shows up to the hospital, they are too far progressed.
To solve this, Patchd combines proprietary deep learning and existing off the shelf wearables to predict sepsis in at home patients.
When a patient at high risk for sepsis (such as one receiving chemotherapy) is discharged from the hospital. They are provided a wearable for their period of highest risk. This wearable transmits their heart rate, respiration rate, blood pressure, temperature and SpO2 to our deep learning model in the cloud. The deep learning model takes this data along with demographics, medical history and 27 mathematical transformations of the data, processes it, and generates a risk score between 0 and 1. When this risk score crosses a threshold, the system alerts the treating team for that patient with an explanation for why the alert was sent. They can then contact the patient, with next steps.
The algorithm itself is device agnostic, however is currently designed to work with the Biobeat BB-613 smartwatch.
Our latest study has indicated that we can predict sepsis on average 39 hours prior to hospital admission.
The company was started to save one of the co-founders lives after he suffered over 18 episodes of sepsis.