The AI voice app that detects COPD flare-ups earlier

Category: (Self-Study) Health

Storyline

Hide Storyline

Researchers are working on a project to develop an AI-powered app for people with chronic obstructive pulmonary disease (COPD). COPD is a lung condition that can cause breathing problems, including shortness of breath, coughing, wheezing, chest tightness, and frequent chest infections.

The idea is that an algorithm can analyze what a patient’s voice sounds like on any given day and detect dangerous flare-ups before symptoms become severe. The technology is based on the idea that indirect changes in breathing and lung function can affect the sound of a person’s voice before symptoms worsen.

The project, led by pulmonologist Dr. Sami Simons at Maastricht University Medical Centre, asks patients to record short voice samples. The app, which is still under research, then uses artificial intelligence to analyze the recordings and identify patterns that may signal declining respiratory health.

“A flare-up of asthma COPD means that people suddenly become worse. So, they get more dyspnea, more phlegm, more cough. So, what usually happens is that they are at home. Or they are at work. So, they need to contact the healthcare provider. And so, there’s a delay in a way of getting the right care. And that delay can last up to a couple of days,” says Dr. Simons.

The team hopes the technology will reduce hospital visits and provide patients with a simple way to monitor their condition from home.

As artificial intelligence becomes increasingly common in healthcare, privacy concerns remain a major issue for patients. Because the app analyzes a person’s voice, researchers say protecting that sensitive data has been a priority from the start.

Dr. Simons says the app was developed with the help of privacy experts to ensure voice recordings remain in a secure environment.

This article and video were provided by The Associated Press.

Script

Hide Script

[Laptops, one playing a recording of COPD patient’s voice recorded via app, and one showing a sound frequency graph]

[COPD patient’s voice recording frequency graph on laptop screen]

[Dr. Sami Simons monitoring the voice recording of a COPD patient]

Dr. Sami Simons (interview): “So, a flare-up of asthma COPD means that people suddenly become worse. So, they get more dyspnea, more phlegm, more cough. So, what usually happens is that they are at home. Or they are at work. So, they need to contact the healthcare provider. And so there’s a delay in a way of getting the right care. And that delay can last up to a couple of days. So, people are worse sitting at home and the idea is that the app enables the people to detect and treat apps more early in the home environment. I do acknowledge that, there is the AI component, so there needs to be some human oversight. But the potential of getting the care from the hospital into the homes of people. I think that’s very promising.”

[Dr. Simons working with the app on his laptop]

Dr. Sami Simons (interview): “Voice indeed is something very private. You recognize people with their own voice. What we develop the app is we develop it also together with people who are experts in privacy. So this app is developed to ensure that the voice does not travel anywhere else besides the app and the environment, which is protected. So we really ensure that the voice stays in a protected environment and it’s not possible to access this remotely, or by anyone else who doesn’t have access to the app.”

[Dr. Simons working with the app on his laptop]

[COPD patient’s voice recording frequency graph on laptop screen]

Dr. Sami Simons (interview): “I think the app has its pros and cons. I think one of the real pros is that it’s really easy. I had a patient of mine saying, well, it’s as easy as brushing my teeth. That’s the pro. I think the buzz of AI and voice is also a pro, a lot of people really want this app on the market tomorrow. I think the cons is that we really need to ensure the app is safe. We really need to ensure that the signal it’s giving the patient is a true signal. So, it’s really an exacerbation. You don’t want any misdiagnosis in patients, you don’t want any over-diagnoses. So, this is the technology and the clinical challenge we still face.”

[Dr. Simons walking through Maastricht hospital corridor to his office]

[Dr. Simons opening the COPD voice recording app on his smartphone]

Dr. Sami Simons (interview): “So, this is the latest version of the app. This is the home screen, and patients just tick on the voice task, and they perform three voice tasks daily. The first was just to tell them about your day. And well, I’m a bit nervous, so maybe you can see it in my voice. That’s a very simple task. And then the next task is to saying ah, a very loud ah, and E and O. This is necessary to see the balance in your voice. So I will just say, aaaaaaaaaaaaaaa, eeeee, ooooo. And the last task is reading out a sentence, it’s a read text, which is a balanced text. So, you can see here, “they agree that whoever forced the traveller to remove his cloak would be considered the stronger.” These were the three tasks, and then you push the button, finished. The algorithm runs in the background, and then here you see the results. This was not shown in the study, but currently people can see how smooth their voice is, how breathiness they are, and pitch. And here you see my voice was more breathy, indicating I was a bit nervous.”

[Exterior of Maastricht University Medical Center]

[Visitors and patient entering the hospital]

[Hospital signboard reading “Maastricht UMC (University Medical Center)”]

This script was provided by The Associated Press.