A new U.S. clinical study has begun examining an innovative new diagnostic tool that uses a smartphone app to diagnose respiratory diseases from the sound of a cough.
When it comes to diagnosing respiratory infections and disease, one symptom remains almost universally constant in every patient: a cough. According to The National Center for Health Statistics, more than two million children under the age of 15 visited the emergency room in 2013 due to coughs. Diagnostic measures include listening with a stethoscope (auscultation), body imaging, and lab tests—all of which can be both time consuming and costly.
A new Australian startup known as ResApp aims to change all that through the use of a new diagnostic tool that works to isolate the sounds of a cough to diagnose a wide range of chronic and acute respiratory diseases like pneumonia, asthma, bronchiolitis, and chronic obstructive pulmonary disease (COPD). The novelty of the approach lies within the simplicity of the technology, as the entire process can be achieved through a simple smartphone app.
“Our smartphone app uses machine learning algorithms, originally developed by Dr. Udantha Abeyratne at the University of Queensland, to diagnose respiratory diseases from the sound of a cough,” said Tony Keating, ResApp CEO and managing director. “The basic functionality of the app is for a patient to cough five times with the phone held at about an arms-length. The app recognizes these coughs, extracts them from any background noises, and performs the analysis on the device before the resulting diagnosis is presented.”
Keating said the machine works by matching signatures in a large database of sound recordings with known clinical diagnoses, a process that was created during previous clinical studies carried out at the university. The technology was then equipped with machine learning tools that can find the optimum combination of these signatures to create an accurate diagnostic test that can measure severity along with the type of condition. Their team at the University of Queensland found that these signatures are consistent across all patients, and are not specific to any individual.
The technology was born out of a desire to detect pneumonia in children. With pneumonia one of the leading causes of hospitalization for children under the age of five, researchers were looking to create a technology that could diagnose the condition quickly without the need for lab or imaging tests that can take days to complete.
The result was a technology that has now evolved into a tool that not only diagnoses pneumonia, but can detect a wide range of respiratory diseases in children including bronchiolitis, croup, asthma, and upper respiratory tract infections. Keating added that recent studies have shown that the technology could prove to be even more capable at detecting other diseases as they expand their database.
“We have enrolled over 1800 patients in our Australian studies, both adults and children,” he said. “We have seen sensitivity and specificity of 90% or greater for the differential diagnosis of these diseases. We have been pleased with the capability of the algorithms to even learn new diseases and expand the applicability to adults. The algorithms were initially designed to diagnose pneumonia in children, but now the technology has ‘learned’ to diagnose a wide range of diseases in both children and adults.”
The project was able to get off the ground thanks to initial funding from the Bill and Melinda Gates Foundation, which provided enough financial backing for the group to develop the algorithms that drive the diagnostic forces behind the technology. The company eventually spun out from the University of Queensland in September of 2014 where they were able to raise over $16 million in funding from institutional investors and a public listing on the Australian Stock Exchange.
The company’s focus has now shifted to obtaining FDA clearance here in the United States, where they are currently carrying out a series of clinical studies at Massachusetts General Hospital, The Cleveland Clinic, and Texas Children’s Hospital. So far, clinicians have remained tight lipped while they prepare to publish the results of the U.S. studies by the end of the summer. Following the completion of the studies, the company plans to place a de novo submission for FDA approval.
Keating said that as the group moves forward with their clinical studies, there remains hope that they can not only deliver a new innovative diagnostic tool to healthcare professionals, but that the technology could actually change the process by which patients receive care for respiratory illness.
“Our focus is to deliver an accurate and fast diagnostic test to healthcare professionals,” Keating said. “In in-person settings, this could significantly reduce time to correct treatment, as well as reduce costs. If delivered in a telehealth setting, our tools could increase the accuracy of diagnosis, which could lead to more effective treatment, and could also increase the applicability of telehealth—where patients would not have to have an in-person visit for a chest x-ray. Instead, they could be accurately diagnosed during a telehealth visit.”
Keating said that for now, the technology is designed for use by clinicians only, either in an in-person setting or in a telehealth setting. The key advantage to the technology is that it is designed to use the microphone in the smartphone, so no additional hardware or accessories are required.
Given its simplicity, who knows? It may not be long before a simple app on your phone can help save you or your child a trip to the emergency room.
Kristopher Sturgis is a freelance contributor to MD+DI.
[Images courtesy of RESAPP]