System detects errors in medication when self-administered by patients

“Freedom is the possibility of isolation. You are free if you can withdraw from people or patients, not having to seek them out for the sake of money, company, love, glory, or curiosity, none of which can thrive in silence and solitude. If you can’t live alone, you were born a slave. You may have all the splendors of the mind and the soul, in which case you’re a noble slave or an intelligent servant, but you’re not free. And you can’t hold this up as your tragedy, for your birth is a tragedy of Fate alone. Hapless you are, however, if life itself so oppresses you that you’re forced to become a slave. Hapless you are if, having been born free, with the capacity to be isolated and self-sufficient, poverty should force you to live with others.” 

Fernando Pessoa, The Book of Disquiet

Before saying anything, the question arises, is self-sufficiency good for everything? Especially when it comes to one’s health. Is it okay for a person to self-administer medical drugs? You need to check your slate before going any further into this article.

From swallowing pills to injecting insulin, patients usually take their medication quote-unquote under self-administration. But they are not always correct. Inadequate compliance with doctors’ orders is commonplace, causing thousands of deaths and billions of dollars in medical expenses every year. Researchers at MIT have developed a system to reduce the number of certain types of drugs.

This new technology combines wireless data with Artificial Intelligence to determine if a patient is using an insulin pen or inhaler and markers that can flag a patient’s dosing regimen. “Past research has shown that up to 70% of patients do not take insulin as prescribed and not many patients use the inhaler effectively.” Researchers say the system can be set up at home, alerting patients and caregivers to medication errors and reducing unnecessary hospital visits.

The study is published in the journal Nature Medicine. The study’s lead authors are Mingmin hao, a doctoral student at the MIT Computer Science and Intelligence Laboratory (CSAIL), a former visiting scientist at MIT, and a current faculty member at the University of Prishtina in Kosovo. Other co-authors include former CSAIL Postdoc and current faculty member at Rutgers University, Hao Wang, and CSAIL Ph.D. student Anirudh Raghu.

Some common drugs need precision and have complex administration methodology. “For example, insulin pens need pinning to make sure there are no air bubbles inside. After the injection, you have to hold it for 10 seconds,” Zhao said. “All these steps are necessary for a proper delivery of medicine in the active space.” Each step offers an opportunity for errors, especially when a pharmacist is not present to guide through corrective tips. Patients may not even know when they are making a mistake – so Zhao’s team devised a powerful automated system.

This program can be divided into three broad steps. First, the sensors detect the movement of patients within a radius of 10 meters using radio waves that propagate around each other. Secondly, Artificial Intelligence scours the reflected signals for signs of a patient self-administering an inhaler or insulin pen. Finally, the system will notify the patient or health care provider when a patient self-management error is detected.

The researchers changed their approach from wireless technology they had previously used to monitor people’s sleep patterns. It starts with an object mounted on the wall that emits very low radio waves. When one moves, it changes the signal and shows it on the screen of the device. Each move produces an example of radio waves that the device can decode. “The good thing about this system is that it doesn’t require the patient to challenge the experts,” Zhao said. “It can work through ports, just like you can access your Wi-Fi when you’re in a room other than your router.”

The new AI sensor sits on the wall of the house, like a Wi-Fi router, and intelligently turns the radio waves on. The team developed neural networks to detect keys and signs indicating the use of an inhaler or insulin pen. They train the network to learn the techniques by doing exemplary movements, for example, some movements fit (like using an inhaler) and some do not (like eating) By repeating and encouraging, the network found 96 percent success with insulin with capacity as well as 99% using an inhaler.

“By dividing them into these steps, we can not only see how often a patient uses their device but also evaluate their donor technique to see how well they are doing,” Zhao says. Researchers say a key feature of their system is based on its non-invasiveness. “An alternative way to solve this problem is to install cameras,” says Zhao. “But the use of a wireless signal is much less intrusive. It does not show human appearance.” He adds that their framework could be adapted to medications outside of inhalers and insulin pens – only retraining of the neural network would be necessary to recognize the appropriate sequence of movements. Zhao says that “with this kind of sensor technology at home, we could detect problems early so that a person can visit a doctor before the problem gets worse.”

please click ! for more…


Related Articles


The Role of Health Cloud Platforms in Wellness

Health care, as with any other service operation, …

Read More ...
The Data Confidentiality Law in Healthcare

The Data Confidentiality Law in Healthcare

In the second year of the pandemic, 2021, …

Read More ...

System detects errors in medication when self-administered by patients

“Freedom is the possibility of isolation. You are …

Read More ...

Applications of Artificial Intelligence in Ophthalmology

Artificial Intelligence (AI) has a snowballing effect in …

Read More ...

We are keen to add value to your business.

Please let us have your details and we’ll touch base with you as early as we can.

Subscribe to our newsletter

Join Factspan Community