By Michael Goodman
September 23, 2016 | ResApp Health Limited believes its mobile health app can replace stethoscopes as the go-to technology for diagnosing a variety of respiratory conditions. The Perth, Australia-based company is adequately capitalized, operationally prepared, and armed with competitive data needed to launch ResApp in the United States.
The technology behind the app—based on machine learning—originated in the University of Queensland, funded by a Bill and Melinda Gates Foundation grant. It was spun out in the second quarter of 2015, along with an exclusive license, to ResApp Health, which soon after listed on the Australian Securities Exchange, raising $4 million AUD. In April 2016, it raised an additional $12.5 million AUD in a follow-on offering.
The Technology Behind The App
Each year, approximately 10% of 125 million ambulatory visits to a physician in the US results in a respiratory disease diagnosis. ResApp’s mobile app was developed to diagnose a wide range of conditions, including lower respiratory tract diseases such as pneumonia, bronchitis, COPD, bronchiolitis, and influenza. The rationale for the app was that the stethoscope—the traditional tool for listening to the lungs, a preliminary step in diagnosing a respiratory problem—is neither sensitive nor accurate; the information available in pulmonary sounds has to first pass through the chest musculature, which dampens high-pitched components.
In contrast, the lungs are directly connected to the atmosphere during respiratory events such as a cough. ResApp captures much more information in these audible sounds than is picked up by a stethoscope. Moreover, because contemporary smart phones integrate high-quality microphones, the platform can be delivered without need for additional hardware.
The first iteration of ResApp will be to simply make a diagnosis. Later, it will measure disease severity and, for the consumer version, it will tell patients whether they should be taking medication.
The team at University of Queensland had expertise in medical signal processing, speech recognition, and audio processing. CEO Dr. Tony Keating says they modified techniques used in speech recognition by talking to clinicians about what they listen for when a patient coughs. “We essentially came up with some signatures that we captured from cough sounds, and the app combines those in a unique fashion to make a diagnosis.”
The machine-learning element comes from ResApp’s clinical trials where investigators recorded many cough sounds from patients with different respiratory diseases. The team used machine learning to identify which signatures are the most important for different diseases, and to generate the algorithms that determine which respiratory condition a patient has. Once the tool is on the market, ResApp will turn off the machine learning part. “That just makes it easier for FDA to get comfortable with what we’re doing,” says Keating. “In the future, we plan to turn it back on and take full advantage of all the data we’ve collected.”
Ongoing studies at the Joondalup Health Campus, Wesley Hospital, and Princess Margaret Hospital in Australia have demonstrated the superior sensitivity, specificity, and accuracy of ResApp in a wide variety of upper and lower respiratory tract indications. It correctly identified lower respiratory tract involvement in 97% of cases initially missed by experienced clinicians using a stethoscope. Keating says that “in our clinical studies, we compared our result with the clinical diagnosis made by the treating clinician who typically has access to X-Rays, blood and sputum tests.”
Preliminary data demonstrated separation of bacterial and atypical pneumonia from the viral form with 89% and 90% accuracy. ResApp answers the clinical need in pneumonia for rapid, accurate, and inexpensive diagnosis.
ResApp plans to launch its tool in the US for pediatric and adult use; it is about to begin a registrational trial at Mass General and is lining up a couple of other trial sites. That trial will start at the end of September, and will run for three-four months, with a final read-out in early 2017. The strategy is to launch into several market segments in the US.
First, it hopes launch into the telehealth segment by partnering with major commercial telehealth providers such as Teladoc, American Well, and MDLive. A big concern of the medical community around the use of telehealth is the inability to accurately diagnose certain conditions over a video link. Keating says ResApp is a remote diagnostic tool that will give clinicians more information to make better decisions. And it can quickly be integrated into existing telehealth systems.
While nothing has been announced, Keating is in talks with several telehealth providers. Other telehealth players include consumer chain stores like Walgreens and CVS; integrated care organizations like Kaiser-Permanente; and government agencies like the Veterans Administration. ResApp also plans to directly promote the tool to emergency departments and retail clinics. It is using the $12.5 million funding to recruit a US sales force.
Later on, it will explore a geographic rollout to Europe (where telehealth is starting to gain traction) and Asia. That, along with the big telehealth players, would require a partnering strategy. Finally, it will also explore selling ResApp directly to consumers.