January 17, 2023
The same devices used to take selfies and tweets are being repurposed and commercialized to provide quick access to the information needed to monitor patient health. You can measure your heart rate by pressing your fingertip against the phone’s camera lens. Bedside microphones can screen for sleep apnea. The speakers have also been tapped and use sonar technology to monitor their breathing.
At the best of this new world, data will be sent remotely to medical professionals for patient convenience and comfort, possibly to support clinicians without the need for expensive hardware. increase.
But using smartphones as diagnostic tools is still a work in progress, according to experts. Physicians and their patients have found real-world success in deploying phones as medical devices, but the overall potential remains unfulfilled and uncertain.
Smartphones have sensors that can monitor a patient’s vital signs. Some early application uses, such as concussion assessment, atrial fibrillation monitoring, and performing mental health wellness checks, can be mentioned.
Companies and researchers eager to apply smartphone technology to medicine are taking advantage of the cameras and light sensors built into the latest smartphones. a microphone; an accelerometer that detects body movements; gyroscope; and even speakers. The app then uses artificial intelligence software to analyze the collected sights and sounds to easily connect patients and doctors. Revenue potential and marketability are evidenced by over 350,000 digital health products available on the app store, according to a report by Grand View Research.
Dr. Andrew Gostin, CEO of sensor networking company Artisight, said: Most Americans own a smartphone, according to the Pew Research Center, with more than 60% of people over the age of 65 owning a smartphone, up from just 13% a decade ago. doing. Also, the covid-19 pandemic has made people more comfortable with virtual care.
Some of these products are seeking FDA approval to be marketed as medical devices. That way, if a patient has to pay to use the software, health insurance companies are more likely to cover at least some of the costs. It falls into the same clinical classification as Aid. However, how government agencies treat AI and machine learning-based medical devices is still being adjusted to reflect the adaptability of software.
Ensuring accuracy and clinical validation is essential to obtaining consent from healthcare providers. And many tools still need fine-tuning, says Eugene Yang, PhD, of the University of Washington School of Medicine. Yang is now testing non-contact measurements of blood pressure, heart rate, and oxygen saturation collected remotely via Zoom camera footage of a patient’s face.
These new technologies rely on algorithms built through machine learning and artificial intelligence to collect data rather than the physical tools typically used in hospitals, making decisions difficult. So researchers can’t “compare apples” to medical industry standards, Yang said. Failure to build in such guarantees undermines the technology’s ultimate goal of reducing cost and access. Because doctors still need to validate the results.
“False positives and false negatives lead to more tests and more costs to the healthcare system,” he said.
Big tech companies like Google invest heavily in researching this kind of technology to serve clinicians, home caregivers, and consumers. Currently, the Google Fit app allows users to see their heart rate by placing their finger on the rear camera lens and track their breathing rate using the front camera.
“If you take a sensor out of a phone and out of a clinical device, they’re probably the same thing,” said Shwetak Patel, director of health technology at Google and professor of electrical and computer engineering at the University of Washington.
Our research uses machine learning and computer vision, areas within AI that rely on information from visual inputs such as videos and images. So, instead of using a blood pressure cuff, for example, the algorithm could interpret subtle visual changes in the body to act as proxies and biosignals for the patient’s blood pressure, Patel said.
Google is also investigating the effectiveness of built-in microphones for detecting heartbeats and murmurs, and the effectiveness of using cameras to screen for diabetic eye disease to preserve vision. According to information released by the company last year.
The company recently acquired Sound Life Sciences, a Seattle startup that offers FDA-cleared sonar technology apps. It uses the smart device’s speaker to bounce off inaudible pulses from the patient’s body, identify movement, and monitor breathing.
Israel-based Binah.ai is another company that uses smartphone cameras to calculate vital signs. Its software looks at the area around the eye, where the skin is a little thinner, and analyzes the light returning from the blood vessels to the lens. The company has completed clinical trials in the U.S. and sells its health app directly to insurance companies and other health companies, said Mona Popilian-Yona, a company spokeswoman.
Applications also span areas such as optometry and mental health.
- Using a microphone, Canary Speech uses the same underlying technology as Amazon’s Alexa to analyze a patient’s voice for mental health status. The company’s CEO, Henry O’Connell, said the software would integrate with telemedicine appointments, allowing clinicians to screen for anxiety and depression using voice biomarkers and a library of predictive analytics. says it can.
- Australia-based ResApp Health last year received FDA clearance for an iPhone app that screens for moderate-to-severe obstructive sleep apnea by listening to your breathing and snoring. SleepCheckRx, which requires a prescription, is minimally invasive compared to sleep studies currently used to diagnose sleep apnea. They can cost thousands of dollars and require a battery of tests.
- Brightlamp’s Reflex app is a clinical decision support tool to help manage concussion and vision recovery, among other things. The mobile app uses the iPad or iPhone’s camera to measure how a person’s pupils respond to changes in light. Through machine learning analysis, images provide doctors with data points to evaluate patients. Brightlamp is sold directly to healthcare providers and is used in over 230 clinics. The clinician says that for each account he pays a standard annual fee of $400, which is currently not covered by insurance. The Department of Defense has ongoing clinical trials with Reflex.
For things like the Reflex app, data is processed directly on the phone rather than in the cloud, according to Brightlamp CEO Kurtis Sluss. By handling everything on the device, the app avoids encountering privacy issues. Patient consent is required to stream data elsewhere.
But algorithms need to be trained and tested by collecting large amounts of data, and it’s an ongoing process.
For example, researchers have found that some computer vision applications, such as heart rate and blood pressure monitoring, can be less accurate on darker skin tones. Research is ongoing to find better solutions.
Also, small algorithmic glitches can cause false alarms, scaring patients and deterring widespread adoption. For example, Apple’s new car crash detection feature, available on both the latest iPhones and Apple Watch, automatically initiated his 911 call when people were riding a roller coaster.
“We are not there yet,” Yang said. “That’s the conclusion.”
KHN (Kaiser Health News) is a national newsroom that produces in-depth journalism on health issues. KHN is one of the three main operating programs of KFF (Kaiser Family Foundation), along with policy analysis and polls. KFF is a donated non-profit organization that provides information on health issues to the public.
Subscribe to KHN’s free morning briefing.
