Can AI Improve Stethoscopes?
Written: March 2023
AI is becoming more and more prevalent in our business and personal lives. It is changing the world we live in at a rapid pace. I have observed over the past 50 years every new technology brings bad actors along with it. They immediately look for ways to commit fraud, steal, engage in illicit activities, spread mis-information, and disrupt the stability of societies around the world. This has been true for cable TV, CD ROMs, the Internet, the World Wide Web, e-commerce, social media, and now artificial intelligence (AI). AI is not going to go away. In fact, big tech, small tech, and thousands of startup companies are moving as fast as they can to exploit AI in their products and services.
Major media will alert us to the bad things about AI. In my writings, I will focus on the good things about AI. I will start with the stethoscopes.
Connor Landsgraf, CEO of San Francisco startup Eko Devices, believes it is time to upgrade the stethoscope, which he pointed out has not changed since the 1880s. He claims many physicians do not get adequate training on how to interpret the sounds they hear with a classic stethoscope. He says the result is “rampant misdiagnosis”.
Eko has developed CORE, an FDA cleared, complete electronic stethoscope with analog and digital capabilities. It can amplify heart and lung sounds and reduce ambient noise. CORE includes a computerized insert for stethoscopes which provides data from the stethoscope to a smartphone app. The app can record, visualize, and save the data with robust data management. The app offers live streaming and HIPAA-compliant sharing to make it easy to share your condition with your health team. In addition, you can sync with your electronic health record (EHR) to help provide a more comprehensive picture of overall well-being.
Adding AI to the digital stethoscope data could have a profound impact. Every time a practitioner listens to a patient’s chest or stomach, the sound can be digitized and sent to the cloud. Over time there will be millions of data sets. The data can be enhanced by adding the confirmed patient diagnosis. The reason diagnoses are not always accurate is because the listener may not have enough experience to recognize the diagnosis. With the application of neural networks and machine learning, both subsets of AI, the patient’s stethoscope data can be compared to the millions of datasets and produce an accurate diagnosis. Physicians who make data-driven decisions will achieve improved patient outcomes. Some believe a medical resident with a digital stethoscope can produce a more accurate diagnosis than a pulmonologist with decades of experience.
Researchers at Johns Hopkins University have taken the digital stethoscope to a higher level as they focused on pneumonia. In 2015, about 5.9 million children died before the age of five worldwide. Pneumonia accounted for 15.5 percent of those deaths around the world and was the top cause of death in sub-Saharan Africa. Early detection is key to curing the disease, but diagnoses based on a hunch leads to costly treatments which may not be necessary, and doing a full work up of tests is unaffordable.
The Johns Hopkins team concluded the limitations of stethoscopes is preventing affordable and effective diagnoses. The limitations have been lived with for more than 100 years. The first limitation is noise. A quiet environment is key. The sounds of pneumonia coming from the lungs can be subtle and background noise can make it difficult to differentiate the disease from the background noise. Unfortunately, there is a big difference between a quiet doctor’s office in the United States and an African village where there may be a lot of noise from children and animals. A second limitation is training. It takes years of experience for a clinician to know exactly where to position the stethoscope on the chest and being able to interpret the sounds, which is more of an art than a science. Experts don’t always agree on a diagnosis and resort to X-rays which may not be available in remote locations.
Eko describes its mission, “We believe the ability to listen is the ability to care. Which is why we’re dedicated to providing healthcare professionals with industry leading digital tools to hear and understand their patients’ hearts and lungs, allowing them to detect with higher accuracy, diagnose with more confidence, manage treatment effectively, and ultimately give their patients the best care possible.” A short video from Eko explains what they are doing. The video is here.
