In September 2015, Yonatan Adiri was three months into a clinical trial that was supposed to determine the fate of the company he founded – and the chart that appeared on the screen was not good news.
The first sign was that the values for the protein and blood began to spread out, instead of looking like a continuous diagonal line, the screen filled with dots far from each other. This collection of points belonged to 200 American patients aged 18 to 80 who took two urine tests – one was analyzed in a regular medical lab and the other using a smartphone app and camera.
The tests were conducted in three independent labs in Massachusetts and New York, and they were entered as 10 graphs, one for each component of the urine tested, into the system of Adiri’s company, Healthy.io.
The problem was that the distance between the points meant the results from the two different tests of the samples did not match up. Until that moment, everything looked promising: The product developed over two years successfully met six of the seven strict requirements of the U.S. Food and Drug Administration. The present and last trial was planned for 500 patients, but there was no longer any point in continuing with it.
Adiri refreshed the screen and results during a meeting in Washington, and Ron Zohar, the company’s chief product officer, did the same thing in Tel Aviv – and both of them realized they had failed.
“At that moment the company had $300,000 left out of $3 million we had raised, and I had two fears,” says Adiri. “The first was about the team, we had 11 employees from Weizmann Institute of Science and Tel Aviv University, excellent developers and experts in machine vision. They had where to go, their LinkedIn accounts were filled with good offers, and I feared that when I return to Israel and tell them the trial had failed, they would lose faith in the company and leave me.”
Adiri’s second fear was his board of directors, which included Prof. Dan Aravot, the director of the cardiothoracic surgery department at Beilinson Hospital in Petah Tikva, and Dan Cohen, both of them from Idan Ofer’s Ansonia Holdings investment fund.
“They knew we could very well fail and promised to stand behind us, but at what price?” says Adiri, referring to a major loss of control of ownership of the company, in return for new funding. Adiri and Zohar tried to think about the scenarios that would explain the results in a telephone call.
“We thought that maybe for some of the kits we did not do good enough quality assurance, and that we sent kits with incorrect color cards that messed up the algorithms, or maybe that when the real patients came, they held the components at different angles in front of the light and damaged the kit,” he says.
This is because according to FDA requirements, the final trial uses real patients – and not patients who can be trained, as in previous trials. The failure could have been in the kits, users or algorithm, and Zohar took it on himself to lead the inquiry.
After six weeks he came back with a surprising conclusion: The application doesn’t know how to talk to male patients over 50.
“Our basic assumption was wrong,” says Zohar. “We thought that all it was is dipping a stick in a urine cup. You have a box, you open it and it has a test strip, a urine cup and calibration scale, each of the items numbered. And you have an application through which you need to photograph the stick together with the scale to receive an analysis.”
The first version of our application showed a short clip that was supposed to instruct the users how to carry out the test, but the users did not have to watch it. It led to a lot of mistakes, mostly among men over 50. For example, they would dip the stick in the urine for more than three seconds and ruin it, or wait too long a time before taking the picture,” he says.
Zohar believed that the company might be able to resolve the issue by means of a change in process: in place of a video clip, they would create a chatbot that would carry on a structured conversation with the user, with delineation of the precise details where needed. “Although it is a bit tiresome for people who do have good digital comprehension, for those people who do not have it, it changes the whole story,” he explains.
The investors listened to the findings of the inquiry, and decided to give Healthy an additional $2 million as a convertible note in order to revise the app and then repeat the trial. “The average investor would have given me a down round, but my investors decided that the company value would be determined in the next round, after the trial would succeed, if that. That is what saved my company,” Zohar says.
Four years later, Healthy already has two FDA approvals for two kits, and over 25,500 diagnoses have been carried out with its system. The company has signed contracts to supply approximately 75,000 additional kits in the next six months.
The first approval it received, in July 2018, was for the use of the smartphone camera for the purpose of deciphering a general urinalysis kit with a standard of 10 parameters, which is used primarily for identifying infections of the urinary tract and toxemia of pregnancy. The second approval was received this past July, for use of a smartphone camera for the purpose of deciphering microalbumin (protein) in urine, which aids in the early detection of kidney dysfunction.
The Japanese law
Healthy is valued (on paper) at about $300 million, after having raised $95 million. After Idan Ofer and Ansonia, the company received investments from the Aleph, Samsung Next, Joy Capital and Corner Ventures funds. It now has 135 employees (44% of whom are women).
Healthy’s UTI diagnostic kit is sold over the counter for 10 pounds ($12.90) at Boots, the British pharmacy chain and a subsidiary of Walgreens Boots Alliance, which also enables its customers to receive antibiotics on the spot on the basis of the exam result, without requiring a physician visit. According to estimates, Healthy enjoys the profits shared with Boots from income on the entire service, which includes the dispensing of antibiotics when a positive diagnosis is received. Company sources say that 75% of the examinations produce positive results.
In Israel, the company is collaborating with the Clalit and Maccabi health maintenance organizations, both of which issue a pack of urinalysis kits to pregnant women for the purpose of monthly monitoring, without making them come into the clinic. Thanks to the high quality of community medicine in Israel, individuals at risk of kidney disease are meticulous about conducting periodic urine tests, such that Healthy’s microalbumen examination kit is less relevant to the local market.
In the United States, where Healthy is making moves to enter the market, the situation is altogether different. The low response rate to preventive examinations is a primary cause of the situation in which the costs of kidney disease - the only organ in the body that is fully insured by government health insurance in the U.S. - amounts to some $100 billion.
Adiri is proud of the fact that Healthy is the first company that has made the smartphone camera into an approved diagnostic tool (Class 2) by the FDA. Making the urine tests an accessible, off the shelf product, with the actual testing conducted independently, may effect a total change in the quantity of urine tests that are carried out, and improve the capabilities of preventive medicine. And yet, the company did not have to reinvent the science. The urine test is evidently the most ancient medical examination, which signaled the onset of laboratory medicine. It now serves for the assessment of pathological conditions such as diabetes, infections, kidney diseases and reproductive cancers.
The urine test is conducted through the use of an examination strip - a five millimeter-wide plastic or paper strip that includes 10 pads saturated with chemicals (reaction agents). The pads change color when soaked in urine, and reading the results is carried out by means of a comparison between colors of the reaction agents and the color ladder provided by the manufacturer.
In the laboratory, the dipstick is deciphered by an automatic reader that includes a darkened chamber; whereas with the Healthy system, the reading is conducted by an algorithm that analyzes the photo of the dipstick and the color ladder as they are sent from the user’s smartphone.
The company’s innovation is its capability of reaching a clinical level of precision in reading the color of the dipstick, which is photographed in differing lighting conditions, with a variety of cameras, by untrained individuals. In addition, the camera’s interaction with the device’s operating system can also produce a variety of sensitivities.
The European regulator has already approved the company’s use of a methodology with which it independently approves future smartphones, but Healthy still has to persuade the FDA that it can work not only on the existing versions of smartphones based on which it has received the agency’s approval.
In February 2018, Adiri met with Simon Stevens, the deputy director of the National Health Service in Britain, to speak with him about the urinalysis kits; in this meeting, Adiri discovered the company’s future direction of development.
Through Stevens, Adiri was exposed to the world of chronic ulcers, a widespread phenomenon among elderly persons, primarily those suffering from additional health conditions such as diabetes or high cholesterol. In 2018, NHS spent nearly 5 billion pounds sterling on the treatment of chronic ulcer wounds, an expenditure that doubled in the past decade due to increased life expectancy and aging of the population.
“He told me that the treatment is carried out by a nurse who specializes in ulcer sores. Each week, she measures the wound – color, depth, size – and gives an ointment relevant to the progression graph of the wound,” Adiri explains. “I was in shock when I heard about the process. The nurse carries out measurement by hand with the use of a paper ruler, like you find in IKEA, and there are studies showing a 20-30% deviation between nurses measuring the same wound. From one week to the next, a different nurse measures the wound and then relies on the notations of the nurse who came before her. I thought to myself: this is astounding, there is a color and a size that we can teach our algorithm to identify, there are insane costs of 4,200 pounds sterling per wound, and this is a market without data. We can produce this data.”
Two companies are already active in this sector – the American company Tissue Analytics, and the Canadian firm Swift Medical, but neither of them is active in Britain.
After the interview, Adiri sent a message saying, “Please don’t frame the article as ‘They’re killing it.’ Medicine will not look different before or after us and we’re not offering a cure for cancer. We are part of a wave of companies that deal with a specific segment, and all of us together will have a cumulative effect on the prevention and deterrence of disease. We’re a camel herd. At this point in time, the rationale of the entrepreneur is that you have to build a camel, because the entrepreneurial world is a world of deserts, not of legends and unicorns.”
Want to enjoy 'Zen' reading - with no ads and just the article? Subscribe todaySubscribe now