In an age when artificial intelligence and computer vision can determine the specific contents of virtually every image captured, you can easily find out what type of flower you’re looking at or filter all the cat photos in your library. That’s not exactly life-changing stuff, of course, but changing lives is just what Healthy.io, wants to do.
The company transforms your smartphone camera into a clinical-grade medical device through urinalysis — one of the most common methods of medical diagnosis.
Healthy.io uses color recognition, computer vision, and AI — in conjunction with a urine sample kit and testing strips — to allow what could be termed a “medical selfie.”
“One of the challenges of at-home clinical-grade urine testing is ensuring that regular users, who may not be tech savvy, can successfully conduct the test in any environment with an arbitrary phone,” CEO and founder Yonatan Adiri told me. “Through rapid prototyping and repeated user testing, we have developed a test kit and smartphone app that has achieved more than 99 percent usability across age groups ranging from 18-80.”
So how does a user conduct a test at home with nothing more than the sample kit and a smartphone?
“The process for the user is very simple. We like to say ‘If you can text, you can test’,” Adiri said. “The user simply opens the app and is walked through the process step by step — aided by our chatbot nurse named Emily. The user opens the kit, fills the cup, dips the stick and places it on our patented colorboard. After waiting for 60 seconds (timed within the app) both the colorboard and dipstick are scanned, similar to how a QR code is scanned. The image is normalized and data points are sent to our cloud, where they are then classified into the correct clinical result.”
Healthy.io’s solution is CE and ISO 13485-certified. The dipstick measures 10 parameters and indicates a range of infections, chronic illnesses, and pregnancy-related complications.
Of course, smartphone cameras and sensors vary wildly, as do lighting conditions. How is Healthy.io ensuring a clinical-grade result under such variable circumstances?
“The image normalization and classification both pose serious technical challenges,” Adiri said. “We knew we needed to develop a highly accurate calibration method and sophisticated computer vision algorithms and to understand the phone/image pipeline (how different smartphones manipulate the image) and generate more than 1 million correctly tagged images across the color spectrum to train our algorithms.”
That data, combined with AI, is how Healthy.io manages to deal with differing conditions and hardware.
“It’s the combination of our user-centric design and cutting edge computer vision, paired with a rigorous clinical approach, that has enabled us to turn the camera into a urinalysis device, equivalent to current devices used at the doctor’s office,” Adiri said.
When it comes to the list of medical problems the system can detect, Healthy.io isn’t trying to reinvent the wheel. Urinalysis has been around for a long time, and it offers accurate results for a range of diseases.
“It is important to first note that we haven’t tampered with the standard urine dipstick,” Adiri said. “It has been a tried and tested diagnostic tool for decades and is well established across many clinical pathways. We test for ketones, leukocytes, nitrites, glucose, protein, blood, specific gravity, bilirubin, urobilinogen, and pH. These indicators span a wide range of pathologies, from urinary tract infection to ketosis, kidney disease, health in pregnancy, and bladder cancer.”
The company has plans to add additional tests soon, and it has an interesting take on the future of smartphones and their use in at-home health care.
“We are bringing an albumin:creatinine ratio test to the market soon, which is critical for the 76 million Americans (people with diabetes, hypertension etc.) that need to get their urine tested for signs of chronic kidney disease every year,” Adiri said. “We see a leap forward in smartphone hardware that happens roughly every four years. Our urinalysis product would not have been possible five years ago because the image quality of smartphone cameras simply wasn’t good enough.”
The latest innovation — 3D imaging — has been driven by the augmented reality industry, which will impact mainstream consumers via smartphones long before it converts the masses to 3D glasses.
“We are using these 3D capabilities to bring standardization and accuracy to chronic wound care,” Adiri said. “Diabetic ulcers, pressure wounds, and similar issues cost about $50 billion every year in the U.S. alone. Today, they are still largely assessed by eye or with a simple paper ruler, which is not only inaccurate but makes it very difficult to coordinate treatment between nurses and doctors. Our next product is tackling this exact problem.”
Healthy.io is not currently available as an off-the-shelf product, but it does save those in need of regular testing several trips to a medical facility.
“We think the greatest value lies in making testing more convenient and accessible to the patient while enabling doctors to receive the results of the urinalysis using existing medical records and CPT codes in a reliable fashion,” Adiri said. For example, women with at-risk pregnancies who need to get their urine tested several times a week can have kits sent to their homes and automatically share results through their electronic medical record. This saves the patient a trip while maintaining the diagnostic flow required and trusted by doctors.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more