Time isn’t the only thing doctors may be up against when diagnosing a child with pneumonia. Noise, like crying, can make it difficult to hear a baby’s congested lungs through a stethoscope. It can be worse in places where soundproofed exam rooms don’t exist. A team of Johns Hopkins researchers is working to change this by developing a smart, low-cost stethoscope that lets doctors zero in on sounds in the lungs.
With acute lower respiratory infections being a leading cause of childhood death worldwide, the stethoscope could make a big difference in diagnosing pneumonia.
“If a child is misdiagnosed, that child goes home and could die,” says pulmonologist William Checkley.
Unlike traditional, nonelectric stethoscopes, five low-frequency microphones inside the programmable stethoscope’s head will make it easier to pick up lung sounds, even if it isn’t ideally placed on the body. A sixth microphone facing outward will collect noise from the environment; an algorithm on the stethoscope’s microprocessor will strip that ambient noise from the lung sounds heard by the physician.
“It’s a very tricky balance with removing the noise,” says Mounya Elhilali, the electrical and computer engineer who developed the algorithm. “If you end up canceling out a sound that actually speaks to what is wrong with their lungs, then you might actually miss the whole diagnosis.”
The team is also developing software to alert users when the stethoscope detects a lung sound—like the crackly noise of pneumonia—that warrants a visit to a health care facility for further evaluation, as well as an option to hear sounds as they would be heard on a traditional stethoscope.
A prototype is being made by Baltimore-based Harbor Designs and Manufacturing. The work is being supported by a $2.5 million, four-year grant from the National Institutes of Health.
Web Extra
Listen to an audio version of this story.
-
00:00:00 Sarah Richards:
Perhaps it comes as no surprise that an interview with Jim West and Mounya Elhilali starts like this.
00:00:07 Jim West:
Right, I notice that you're quieter than any office on this corridor.
00:00:11 Mounya Elhilali:
Well, I used to have noise from here, and I don't anymore.
00:00:15 Sarah Richards:
These 2 researchers are obsessed with sound, how we hear it, how we capture it. In fact, if you use a cell phone today, chances are the microphone inside it is based on Jim West's pioneering work with electret microphones. And Elhilali, she's an expert on what's called signal processing.
00:00:35 Mounya Elhilali:
Part of my interest is to understand how the brain understands sound. I study various aspects of that problem, what we call the cocktail party problem. When you walk into a cocktail party, different people are talking. How does your brain figure out who's talking? Though it sounds trivial, it's a challenge that most audio systems face. When the sound comes in, it comes all combined and it has to entangle it.
00:01:00 Sarah Richards:
Challenge can be problematic for a doctor using a stethoscope to listen to a baby's lungs. The baby could be crying. The clinic could be noisy, especially if the room isn't soundproofed. That's why Elhilali and West are creating a low-cost stethoscope that can zero in on lung sounds. This so-called SMART stethoscope is a $2.5 million, four-year project that's funded by the National Institutes of Health. West and Elhilali want to make it easier for anyone, anywhere in the world, to use a stethoscope to detect pneumonia in a child.
00:01:36 Mounya Elhilali:
This is a true public health problem, and I know that I've seen that pneumonia, or at least lower respiratory diseases, are the number one killer of kids under five years of age.
00:01:46 Sarah Richards:
Now, electronic stethoscopes already exist, but West says they share a common drawback.
00:01:52 Jim West:
With all of the great patents, there's very little on the market that addresses the major problems that stethoscopes face, and that is background noise.
00:02:02 Sarah Richards:
It can also be tricky placing the stethoscope correctly to hear the best lung sound. So, here's how they're getting around these challenges. The team has placed five low-frequency microphones inside the stethoscope's head to make it easier to pick up sound, even if it isn't ideally placed on the body. They've also attached another microphone facing outwards to collect noise from the environment. An algorithm developed by Elhilali on the stethoscope's microprocessor then strips that ambient noise out of the lung sounds. That's actually a lot harder than it sounds.
00:02:37 Mounya Elhilali:
It's challenging for a number of reasons. One of this is an open environment so these other sounds can be coming from different places. The other biggest challenge is you don't want to cancel too much from the sound you're picking from the body because these are... potentially sick kids and you don't know what is wrong with them. So, if you end up canceling something that actually speaks to what is wrong with their lungs, then you might actually miss the whole diagnosis. So, finding that balance between those two is not a trivial thing.
00:03:07 Sarah Richards:
Elhilali and West are also developing software that alerts users when the stethoscope detects a lung sound, say the crackly noise of pneumonia, that warrants a visit to a healthcare facility for further evaluation. Pulmonologists William Checkley and Eric McCollum are helping develop the stethoscope, and a prototype is being made by Harbor Designs and Manufacturing in Baltimore. Testing will start in January, and the team hopes to adapt this stethoscope to work with other illnesses. For Johns Hopkins Medicine, I'm Sarah Richards.