With the rise of telemedicine during the Covid-19 pandemic, it stands to reason that digital diagnostic tools would gain in popularity.
But when I tried the last such tool – a chatbot from an Israeli startup Kahun – I was in for a surprise.
After opening the website, I typed in a symptom that I suffer from: “floaters,” where dislodged lumps of collagen bounce around the vitreous fluid in the eye, casting annoying shadows on the retina.
The chatbot asked me a long list of relevant questions – “Is it in both eyes or in one?” “” Is your vision more blurry at night or during the day? “Are your eyes dry or excessively watery?” “- as well as some unexpected questions -” Do you have a rash? Joint pain? Back pain?
When I reached the end of the full Q&A, Kahun’s chatbot spat out a response: “Sjögren’s syndrome,” an autoimmune disease that can cause all symptoms to which I answered yes.
The only problem: I know from numerous visits to the ophthalmologist that my floaters are not the result of a syndrome that I had never heard of. They started as a result of complications from cataract surgery.
Feeling confused (and a bit combative), I confronted Kahun CEO Eitan Ron. How could the app have gone so wrong? I asked.
Ron was unfazed. He knew exactly where the problem was: Kahun’s chatbot had not yet been programmed to ask questions about other procedures a patient might have undergone, such as cataract surgery.
“It’s a bug,” Ron admitted. “We should have asked. We need to add this to the flow.
Another explanation, Ron said, would be that when the chatbot asked about joint pain and I answered yes, it took that route. “The algorithm is always trying to find the best explanation for the clinical presentation provided,” he told me.
To be fair, I made a query through a competing chatbot – from Israeli startup StuffThatWorks – that did not “diagnose” me incorrectly, but also failed to offer reasonable treatments.
StuffThatWorks is first and foremost a community of patients that draws its information from what patients report themselves.
In contrast, Kahun scours the web for medical information from articles and books written by and for doctors, not patients, and then uses artificial intelligence to order the data in a user-friendly way.
“We created the company to describe medical knowledge not in text form but in a structured way, as part of what Google calls a ‘knowledge graph’,” says Ron.
A knowledge graph is made up of nodes and links that show the relationships between different entities – in this case, between conditions, symptoms, tests, and treatments.
Kahun draws on the 30 million articles and books written on medical topics (with more added daily) that are regularly accessed by specialist search engines such as PubMed. Kahun has already mapped some 20 million associations between clinical features in his database.
To be used in conjunction with a doctor
Kahun’s chatbot has just been launched, so the designation “beta” is appropriate (and makes me a little more forgiving).
The interface is multilingual; I did it in English. There is also a Hebrew version, with other languages, including Arabic, to come.
Ron points out that Kahun’s chatbot is not meant to diagnose, but rather should be used in conjunction with his doctor.
“We view our results as ‘possible or probable causes’ and not as a diagnosis. In most cases, we’ll show a few options.
Kahun’s new chatbot can be deployed in two main ways: before a doctor visit, as a pre-appointment diagnosis, and as part of a telemedicine exam.
Kahun may not be perfect – at least not yet – but he’s doing better than most medical students.
In the “clinical skills” section of the US Medical Licensing Exams (USMLE), Kahun’s app answered 85% of the questions correctly, far exceeding the minimum pass mark. Ron and his team wrote an article on the results, available as a pre-printed version at The Lancet.
Covid tool for hospitals
Kahun was preparing to launch in early 2020 when Covid-19 hit. The new coronavirus had already generated some 4,000 medical papers – and that was only at the start of the crisis.
Ron decided to pivot the business to focus on a single disease and provide doctors with a tool to categorize symptoms and diagnose Covid. The aim was to help hospitals manage their services.
“We had the knowledge and the infrastructure,” says Ron. “So we asked, ‘How can we help? “”
The results of a joint study with the Rabin Medical Center in Israel have been published in the American Journal of Emergency Medicine. Of 18 critically ill patients, 17 were correctly diagnosed. Of 127 patients with non-critical illness, 104 were accurately identified.
“Our tool could help predict which patients might be discharged home in the next few days and which patients would deteriorate,” Ron notes.
Kahun “has become my go-to outfit during a tense shift in the hospital,” Israeli medical intern Yohai Shraga said last year. “I can use it to make sure I don’t miss anything or consult before going to a senior doctor.”
But over time, doctors and researchers have become less dependent on Kahun’s Covid diagnostic tool.
“Doctors have developed their own feelings about this disease,” says Ron. “If someone comes in and is obese or has high blood pressure or lung disease, they’ll know it’s a high risk. They didn’t need our calculator.
This turned out to be a blessing in disguise, as Kahun could now return to his original mission.
Before the doctor’s visit
A tool that physicians can use in their practices with patients was the first product to be launched. It is a key part of Kahun’s portfolio. Ron says several thousand doctors and medical students around the world use it.
The pre-meeting chatbot diagnosis came next.
It may take up to 10 minutes for the chatbot to run. This saves time during the actual appointment, allowing the doctor to focus on the diagnosis rather than the management. The results of a Kahun conversation can be copied and pasted as a simple text summary.
The tool is more up to date than many doctors who struggle to keep up with the deluge of new medical papers.
Additionally, Kahun’s medic interface asks questions and suggests tests the medic might not have thought of. It’s not just a list of mostly immutable questions, as is the case with other Israeli diagnostic startups. Kahun’s questions change based on the answers provided by the patient, mimicking a doctor’s thought process.
Kahun is transparent about the data he presents: every recommendation or suggestion includes links to the original articles or documents, so doctors understand where the information is coming from.
It should help counter resistance from doctors – you know, the doctor who says, “Please don’t google your symptoms, talk to me instead.” “
Ron adds that compared to searching the web for what ails you, Kahun’s structured question-and-answer approach “is already better than a Google search.” We may discover an important question that the doctor might have missed, such as asking a patient if they are suffering from night sweats.
“Waze for Doctors”
That said, Kahun isn’t really a mainstream app; the chatbot and other tools will be licensed directly to third parties – HMOs, clinics, scheduling services, and telemedicine providers like Teladoc – who will pay a fee depending on the size of the organization or session.
Kahun’s partners will be able to mark it as they wish. The website I used to access Kahun’s chatbot to find my floats is from Kahun’s partner infomed.co.il.
Kahun was founded in 2018 by Ron and Tal Goldberg. The duo had previously created the Human Click customer service application, acquired by LivePerson in 2000. Ron and Goldberg, who served as Managing Director and CTO of Human Click, stayed with LivePerson for seven years, which is almost unheard of. in the rapidly changing tech industry after an acquisition.
A third co-founder, Michal Tzuchman, was a software engineer at Waze before enrolling in medical school to retrain as a pediatrician.
The Waze connection isn’t just made by Tzuchman and Goldberg (who also worked at Waze after Human Click).
Former Waze CEO Uri Levine was the first seed investor in Kahun and is the chairman of the board of the company. It’s no wonder that Kahun is sometimes called “Waze for Doctors”.
Giving meaning to medical knowledge
Kahun has raised $ 5.5 million, including a $ 2.5 million grant from the European Union’s Horizon 2020 program. A team of 26 people work in the company’s offices in Tel Aviv.
The name Kahun has an appropriate backstory: it comes from one of the earliest known medical documents in history – the nearly 4,000-year-old “Kahun Medical Papyrus” found in the Egyptian village of Kahun.
“All medical knowledge in the world today is in text form and the only way to make sense of it is to read it by a doctor or medical staff,” says Levine.
Kahun offered a different approach. “What if we could convert that into data and build apps to use it? This is exactly what we are doing.
As for me and my floats, I will be trying Kahun again in a few weeks. Ron has promised to add “cataract surgery” as one of the options to choose from during my next chatbot session. If I still have “Sjögren’s syndrome”, I can present this information very well to my ophthalmologist.
To learn more about Kahun, Click here