Pneumonia kills more children than any other cause, according to the World Health Organization. While treatable with antibiotics, each year the lung infection claims roughly 1.4 million kids before their fifth birthday, 18 percent of under-five deaths. Most of the deaths occur in South Asia and sub-Saharan Africa, where access to doctors and drugs is limited.
Hon Weng Chong, a 24-year-old student at Melbourne Medical School, has developed a quick way to diagnose children in areas where health care is scarce. He’s built a stethoscope that can plug into a smartphone and doesn’t require medical training to use. “It seems simple that you just have to diagnose someone and give them antibiotics,” Chong says. “The problem is, how do you diagnose them?”
The Australian began developing his device on the advice of a professor while training in pediatrics earlier this year. Long a hobbyist programmer, he’d recently spent a year studying health informatics at Johns Hopkins University and was looking for a project to enter in the Imagine Cup, Microsoft’s (MSFT) annual student technology competition.
In March he built the first hardware prototype in his garage for less than $20. He connected the round part of the stethoscope, the chest piece, to a small digital microphone encased in rubber tubing, which plugs into the phone’s audio jack.
Designing StethoCloud, the software to analyze the data, proved tougher. “I was way out of my league,” Chong says. He recruited two programmers and another med student to help build the algorithms.
Say a boy in a remote African village suffers from a persistent wheezing cough. A health worker with minimal training need only grab her smartphone and plug in the stethoscope attachment. The app asks how long the child has felt sick, then prompts her to check for two types of abnormal breathing, using video clips to demonstrate. The worker then records the boy’s breathing and uploads the sound files. An algorithm counts the breath rate and returns a diagnosis with treatment instructions.
While counting breaths is the most basic way to diagnose pneumonia, “It is surprisingly difficult to count a wriggling, unwell child’s respiratory rate,” says Dr. Jim Black, one of Chong’s mentors at Melbourne.
StethoCloud hasn’t had a clinical trial yet, but Chong is beginning to recruit patients in Melbourne for a study and is discussing similar test runs with doctors in Mozambique and Malaysia. He’s seeking funding to continue development and is considering pursuing the venture full time instead of practicing medicine.
The lack of speedy networks and relative paucity of smartphones in many countries limits StethoCloud’s reach for now, Chong concedes. More audio samples are needed to improve its diagnosis of subtler breathing cues. And doctor-free diagnoses won’t get medicine to sick kids. Still, the need is great enough, and the device simple enough, that Chong envisions it becoming “as ubiquitous as a thermometer at home.”