This AI Is Using Voice Samples to Detect and Diagnose Type 2 Diabetes

Colin’s Note: Today, we’re talking again about the intersection of artificial intelligence (“AI”) and healthcare.

One company is using an unusual biomarker to detect type 2 diabetes in patients – their voices.

Folks with type 2 diabetes have noticeable changes in their voices, including in the tone and pitch… Something that a human doctor might not be able to detect as easily as an AI analyzing voice samples can.

We’re talking about the power of AI detecting disease with just voice samples. Type 2 diabetes – like many diseases – is already drastically underdiagnosed. Testing can be costly and time-consuming… But imagine if you could get one step closer to diagnosis and treatment with just a smartphone and a few short voice samples.

This is just one of the things we’re talking about when we say AI is going to revolutionize healthcare. I get into all the details in today’s video…

Hello, Bleeding Edge subscribers. Hopefully, you guys are having a wonderful day out there. Colin Tedards here. And today, we’re going to be diving in, once again, to healthcare and AI.

We recently looked at a story about UC San Diego’s new AI that can detect early signs of sepsis in hospital patients. Today, we’re exploring another critical health issue – diabetes.

One concerning aspect of diabetes is its progression is often silent. Here in the United States, out of 37 million adults with type 2 diabetes, only about 28 million of them are diagnosed.

And that leaves a lot of people unaware of their condition. Globally, the situation is even worse. With approximately 462 million people affected by type 2 diabetes, only about half of them realize it.

And despite technological advancements in testing and detection, the primary methods for diagnosing diabetes have required a visit to a doctor.

The first method requires fasting. The second method is an A1C test. Now, these tests are very effective, but they can be inaccessible and unaffordable for many people, especially around the globe.

Enter a groundbreaking development in healthcare: AI-powered voice-based disease detection.

That’s right. Disease detection… just from your voice. This technology offers a revolutionary approach to diagnosing diseases like diabetes using just a smartphone and a brief voice sample.

A company out of Toronto called Klick Health is at the forefront of this innovation with its simple voice test for diabetes. Their method involves recording just a short phrase on a smartphone to determine the likelihood of diabetes.

Their approach leverages biomarkers where AI analyzes your voice pattern to detect subtle changes associated with the disease.

I didn’t know this, but it turns out patients with type 2 diabetes have recognizable changes in their voice, including tone and pitch. The accuracy of this test was surprising. Based on a clinical study published by the Mayo Clinic, the accuracy rate was 86% for men and 89% for women.

Now, you might be wondering how this compares to the traditional diabetes test…

Well, the accuracy is about 92% for the doctor using traditional type 2 diabetes testing. So it’s very close in terms of its accuracy. But that’s not all… and this might not even be the exciting part.

From a technology perspective, the app was remarkably simple to build. The team used voice samples from just 267 people who recorded short clips of their voices about six times per day. After compiling about 18,000 voice samples, the company was able to leverage AI to do the rest of the work.

Now, it’s still the very early days for these types of advancements. But imagine when larger data sets and more patient data are used.

The potential of this technology extends beyond diabetes to other conditions like Alzheimer’s, autism, and maybe Parkinson’s. This is another great example of what we’ve believed about AI from the beginning when it started to take off in notoriety last year.

AI is going to accelerate bringing new technology to the forefront… but it’s not going to replace traditional doctors or testing. Diabetes is a great example of that. With so many people living undiagnosed, a simple app-based test might get more patients the help that they need.

Now, I know what some of you guys are going to say… “There’s got to be some great investment opportunities in healthcare and AI!” And in some cases, I certainly believe that, but this study is an example of the disruption that’s coming in the healthcare space.

If all you need is a few hundred patients, an AI application, and a cell phone to create a test that’s almost as accurate as one that has been used for decades… Well, I tell you what, there’s going to be a lot of disruption to come for healthcare in the coming years and decades ahead.

What is absolutely a guaranteed investment going forward is all the silicon, custom computer chips, networking equipment, data centers, and the backbone of AI… which we’ve had our subscribers invested in since I took over here at Brownstone Research.

Today’s session sheds light on how AI is transforming healthcare, offering hope for early detection and treatment of diabetes. Stay tuned for more insights. We’ll see you again soon.

Like what you’re reading? Send your thoughts to feedback@brownstoneresearch.com.