Wearable Devices and Algorithms for Self-harming Activity Detection

In the United States, there are more than 35,000 reported suicides. Approximately 1,800 of them are psychiatric inpatients. Staff perform intermittent or continuous observation in order to prevent such tragedies, but a study of 98 articles over time showed that 20% to 62% of suicides happened while inpatients were on an observation schedule.
Reducing the instances of suicides of inpatients is a problem of critical importance to both patients and healthcare providers. In this disclosure, we introduce SHARE - A Self-Harm Activity Recognition Engine, which attempts to infer self-harming activities from a sensing accelerometer data using smart devices worn on a subject's wrist. Preliminary classification accuracy of 80% was achieved using data acquired from 4 subjects performing a series of activities (both harmful and not). The results, application, and proposed technology platform are discussed in-depth.