Written by ERIC WAGNER / April 20, 2016
Mayank Goel had run into a challenge. A PhD student of Professor Shwetak Patel’s in the CSE Department at the University of Washington, Goel is part of a team working on SpiroSmart, a mobile app that will let physicians use phones to test a patient’s lung capacity. (SpiroSmart comes from spirometry, the name of the standard pulmonary function test.)
For SpiroSmart, a patient holds their phone at arm’s length, takes a deep breath, and exhales. The microphone picks up the sound of the breath, and the program translates those sounds into a flow curve. It then analyzes the curve to show whether the patient’s lungs are functioning normally or are obstructed, and if so, by how much. A physician can then start to diagnose whether the obstruction is due to asthma, cystic fibrosis, or some other chronic lung condition.
Goel was developing SpiroSmart’s algorithm, and for that he needed to know what a range of obstructed lungs sounded like. The problem was that, while he had access to plenty of data for both normal and mildly obstructed lungs from Seattle-area hospitals, he didn’t have much from lungs that were severely obstructed. “It was a big hole,” Goel says. “We needed data from lungs that were functioning at only twelve or fifteen percent of capacity.”
The solution came from Jim Stout MD, one Goel and Patel’s collaborators. Dr. Stout is a professor of pediatrics at the UW School of Medicine who specializes in asthma care. He is also ther creator and of a suite of UW web-training programs called Spirometry360®, offering on-line training and feedback to help primary-care physicians and their staff interpret spirometry tests.
“Spirometry is one of those things that’s easy to get wrong, but not too hard to do well,” Stout says. For a test to be valid, a patient must hold the device at the correct distance, breathe in a particular way, and so on. A lot of little things can go awry—the patient might not take a deep enough breath, or might cough—and even the smallest deviations can screw up the results. Spirometry360’s training module can prevent that. It is now used at thousands sites all over the world. (It is even translated into Dutch and part of the standard training regimen in the Netherlands.)
With a grant from CoMotion, Spirometry 360 had expanded its offerings to clinics in developing countries, and Stout told Goel about one such clinic in Khulna, Bangladesh, where he had developed an excellent relationship with the primary physician, Dr. Monsur Habib. “Dr. Habib’s practice just blew us all away,” Stout says. “The staff there just really embraced the spirometry training, and they were doing really good work.”
Dr. Habib’s clinic was the only one to specialize in pulmonary ailments for two hundred square miles. Patients with lung conditions came from all over to be tested and get treatment. As such, where some clinics in Seattle might see fifty patients in a week, Dr. Habib’s clinic saw that many in a day, and as many as a thousand a month. For Goel, it was a goldmine of data, and last December, he went to visit Dr. Habib in Bangladesh and see the operation for himself. He was able to spend more than two days testing SpiroSmart in the field, comparing its results with those of the standard spirometry test. “It was tremendously valuable,” he says.
Dr. Habib has since provided more than two thousand head-to-head tests, and on a variety of ailments. From this, the SpiroSmart team will be able to improve their algorithm. He, Patel and Stout are hoping to launch SpiroSmart later in 2016, and are working to get FDA approval. The hope is that, with SpiroSmart, patients will eventually be able to use their own phones to track their lung function at home. And that, says Goel, will give them more control over their health, and their lives.