Smartphone App Detects Sleep Apnea

Mobile technology could fill a gap in medical diagnosis

23 January 2015
Photo: Michael Bodmann/Getty Images

Loud snoring, restless sleep, morning headaches, and changes in mood are signs that someone may have sleep apnea, a disorder in which people stop breathing repeatedly throughout the night. Most people with the disorder, which increases the risk of a heart attack or stroke, are often unaware and go undiagnosed. To help solve this problem, one Ph.D. student has been working on an app for mobile devices to help detect sleep apnea. Eventually, he hopes it will be accurate enough to be used as a screening tool for the disorder.

Smartphone apps are already available to measure sleep activity such as tossing and turning, waking during the night, heavy breathing, and snoring, which could all be signs of a sleep disorder. Joachim Behar thought he might use these apps to detect sleep apnea, but after spending two months reviewing some 40 of these apps, he and his team at Oxford found them all lacking. They concluded that most were scientifically unsound and did not have any clinical evidence that they are accurate. Taking matters into his own hands, Behar designed SleepAp to help detect sleep apnea with the help of colleagues in Oxford’s department of engineering science. He also received support from the Oxford Centre for Affordable Technology, which is headed by his professor Gari Clifford, and the sleep unit at the Oxford Centre for Respiratory Medicine at Churchill Hospital.

His research, “SleepAp: An Automated Obstructive Sleep Apnoea Screening Application for Smartphones,” was published in the IEEE Xplore Digital Library. Behar recently cofounded the start-up SmartCare, which is focused on using mobile technology to find ways to identify specific medical conditions, including sleep disorders.

The usual way to determine whether someone has a sleep disorder is for the person to be studied overnight at a sleep clinic. Such a study involves a polysomnography test, in which sensors are placed on the patient’s body to record brain activity, eye movement, heart rate, blood pressure, and the amount of oxygen in the blood.

“Using a mobile app first can be a useful way to identify if someone is at risk and determine whether further testing in a clinic is needed,” Behar says.

SLEEP ON IT

Just as in clinical settings, the SleepAp first requires users to fill out a questionnaire that covers predictors such as gender, age, daytime sleepiness, and even neck size. A thick neck, for example, can narrow the airway and be a potential cause of sleep apnea.

The original version of Behar’s app used the smartphone’s built-in sensors to monitor movement using the device’s accelerometer and snoring with the phone’s audio recorder. He enabled the app to read an external medical device, a pulse oximeter, which can be plugged in through the device’s headphone jack. The oximeter’s fingertip clip measures oxygen levels in the blood, which indicates how well the user is breathing. The app also suggested that users connect an external microphone to capture better-quality audio.

However, the design was complicated because people had to accurately set up these external devices. Moreover, smartphones each have their own internal sensors and software, which would make it difficult to regulate as a medical device without doing so for each type of phone. For these reasons, the app has been upgraded to what is now called the SmartCare Sleep app, which is able to read wireless medical devices, including a wireless pulse oximeter, through Bluetooth. Like many wireless fitness wearables, the upgraded sleep app can sync to any device and display comparable results.

The app’s software then analyzes the data and provides a score to indicate whether a patient is at risk of having sleep apnea.

To evaluate the effectiveness of the algorithm, the researchers randomly selected 856 patients from the respiratory medicine unit at Churchill Hospital who had been referred to the clinic with suspected sleep apnea, and they found that the results were about 90 percent accurate. Their goal is to make the app one that eventually can mimic a polysomnography test. 

HOME CLINIC

The SmartCare Sleep app is first and foremost a way to identify a potential problem, Behar points out. Users can bring the results to their doctor to determine whether further tests are needed. It could also cut health care costs by avoiding unnecessary overnight stays in a sleep clinic.

Behar is currently working to increase the app’s accuracy to further improve its reliability as a screening tool. This year, the SmartCare team will be working with Diego Mazzotti, who received his Ph.D. in psychobiology at the Federal University of São Paulo, in Brazil, to improve the app to match those of clinical tests and validate its results. If effective, the SmartCare Sleep app would be the first smartphone app to accurately screen for sleep apnea. The next step will be to receive approval from government regulators, such as the U.K. Medicines and Healthcare Products Regulatory Agency. 

Learn More