Wearable health monitoring is the measurement of the body outside the clinic, through sensors worn on the body or used at home, which send their readings to software that alerts a doctor when it matters. The whole arrangement is called remote patient monitoring (RPM). Its buyers are hospitals, clinics and patients themselves.
Every device in this zone does three things. It measures a signal from the body. It sends the readings over the internet, usually through a phone app connected by Bluetooth. And software reads the readings and decides who should see them. Each reading carries the time at which it was taken, so the readings form time series data.
The devices differ in what they measure. A continuous glucose monitor (CGM) on the arm reads blood sugar every few minutes. A personal ECG monitor records the heart's rhythm when the user touches it with their fingers. An ECG, or electrocardiogram, is the tracing of the heart's electrical signals. Remote patient monitoring wearables, such as a patch on the chest, watch the pulse and breathing of a patient sent home from hospital. A home fertility monitor measures hormones in a urine strip. A non-invasive brain-computer interface (BCI) reads the brain's signals through a headset, with no surgery.
Two products bring a clinic's tests to people who have no clinic nearby. A health kiosk is a booth with instruments that checks blood pressure, sugar and more, in a village or a workplace. Remote ECG interpretation lets a small clinic send an ECG to the cloud and receive a heart specialist's reading within minutes.
These devices produce far more readings than any doctor can review. So AI models now read the signal first. A model learns what is normal for a patient and flags anything that departs from it, which is called anomaly detection. It also sorts cases by how urgent they are, which is triage. An AI ECG analysis model can even find heart disease in an ECG that a doctor would read as normal. Regulators now approve such models as medical devices in their own right, as explained in software as a medical device (SaMD).