Not surprisingly, when patients are responsible for their own medication administration – whether pills, inhalers or insulin shots – adherence levels are often less than ideal, and potentially dangerous errors often occur.  

But according to a study published recently in Nature Medicine, researchers at MIT have enlisted AI to try to reduce the chance of errors, at least for some types of medications. The new tool uses wireless sensing and AI to determine when a patient is using an insulin pen or inhaler, and flags potential errors in the patient’s administration method.

“Some past work reports that up to 70 percent of patients do not take their insulin as prescribed, and many patients do not use inhalers properly,” said Dina Katabi, the Andrew and Erna Viteri Professor at MIT.

The system works by using a sensor, adapted from a technology previously used to monitor people’s sleeping positions, to track a patient’s movements within a ten-meter radius, using radio waves that reflect off their body. Then, AI analyzes the reflected signals for signs of a patient self-administering an inhaler or insulin pen. Finally, the system alerts the patient or their healthcare provider when it detects an error in the patient’s self-administration.

The new sensor sits in the background at home, like a Wi-Fi router, and the team developed a neural network to key in on patterns indicating the use of an inhaler or insulin pen. They trained the network to learn those patterns by performing example movements, some relevant (e.g. using an inhaler) and some not (e.g. eating). Through repetition and reinforcement, the network successfully detected 96 percent of insulin pen administrations and 99 percent of inhaler uses.

“One nice thing about this system is that it doesn’t require the patient to wear any sensors,” said Mingmin Zhao, a PhD student in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). “It can even work through occlusions, similar to how you can access your Wi-Fi when you’re in a different room from your router.”

After successfully detecting relevant movements, the system showed that it could detect errors as well. Because every proper medication administration follows a similar sequence, the system can flag anomalies in any particular step. For example, the system can recognize if a patient holds down their insulin pen for five seconds instead of the prescribed ten seconds. The system can then relay that information directly to the patient’s doctor so they can fix their technique.

“By breaking it down into these steps, we can not only see how frequently the patient is using their device, but also assess their administration technique to see how well they’re doing,” said Zhao.

According to the team, a key feature of the system is its noninvasiveness, which could encourage patients to participate more proactively in their own healthcare.