Even though at-home use of an inhaler or insulin pen can be extremely important, doctors have to pretty much just hope that patients are doing it right. A new system, however, could objectively assess patients’ technique within their homes, and let them know if they’ve got it wrong.
Created by a team at MIT, the artificial intelligence-based technology was adapted from a setup that was previously used to non-invasively monitor changes in sleeping positions. It incorporates a wall-mounted device that emits harmless low-power radio waves within a 10-meter (33-ft) radius.
As long as the patient is located within that area, their body will reflect the waves back to the device. This means that every day when they use their pen or inhaler, they have to go to that approximate location.
Every movement they make modulates the reflected radio waves in a specific manner, which is detected and analyzed by software connected to the device. If those modulations don’t match up with those that the software has been trained to associate with the proper movements, then the system “knows” that the person isn’t using the pen or inhaler correctly. It can then notify them or their physician, via an interface such as an app.
In tests conducted so far, the system has been able to detect 96 percent of insulin pen administrations and 99 percent of inhaler uses. It has additionally identified problems such as the pen being held down for only five seconds instead of the prescribed 10.
And while an optical camera-based system could likely identify such problems visually, the MIT setup has the advantage of not recording any images of users, so there are no privacy concerns.
A paper on the research, which is being led by Prof. Dina Katabi, was recently published in the journal Nature Medicine.