Doctors’ time is very valuable, so any tool that can help them shorten appointment times without harming the quality of treatment can lead to more people getting the help and treatment they need. So, a paediatric specialist and two entrepreneurs who have already had exits in the past created a chatbot to do so. Their startup is called Kahun .
Let the chatbot start the diagnosis
Kahun develops an engine that is based on statistical and mathematical algorithms that enable the mapping of medical studies and medical books. This is different from what is normally done, as oftentimes past data, patient files, and expert opinions from the past are considered which can lead to challenging biases that are not helpful to the diagnosis. The result of their engine is a kind of medical knowledge map that can find connections between risk factors, medication use, side effects and more. And why is all this good? Because now you don’t need to google your symptoms and be frightened by the list of scary possibilities it could be, but rather you have a tool that is actually able to understand the whole picture and which diseases the symptoms you report indicate; there are quite a few companies that are trying to crack this concept.
"Kahun’s artificial intelligence engine questions the patient until it reaches a differential diagnosis, that is, a list of all the diseases that can explain the patient's complaints. The questioning is carried out dynamically with each new piece of information that is received (and not according to a static decision tree); the system recalculates the differential diagnosis as a doctor would do and based on that it calculates what is the most correct question to ask next in order to strengthen or reject optional diagnoses”, Eitan Ron, CEO of Kahun, explains. This questioning phase takes place, for example, through a chatbot in telemedicine services, before the remote meeting with the doctor, thus making it possible to shorten the initial questioning process that the doctor does with you at each meeting. At the end of the process, which should take several minutes, the doctor receives the initial diagnosis. Moreover, it allows for documentation to already begin, as the doctor does not have to then type all the symptoms the patient complains about– it is all already there.
Ron immediately clarifies that the system is not intended to replace doctors, because the patients are the ones who convey subjective information that cannot be used to make a final diagnosis. "As a result of the questions asked by the bot, the doctor receives recommendations for further action: what tests should be done, what the system thinks is necessary to examine and with what probability. That is, it provides a differential diagnosis that will help reach a final diagnosis as accurately and as quickly as possible. In other words, the system does not replace doctors but helps them perform their work in a thorough, high-quality, and faster way. By saving time for questioning and documentation, the doctor can devote more quality time to the patient. By using AI, it is possible to ensure a thorough and high-quality questioning of all patients, regardless of the doctor's level of experience, the time available to them or other distractions that might get in the way," Ron stated. Kahun is already working in emergency departments to question patients who come for treatment before the final diagnosis is carried out by the on-call medical team.
Waze’s first employee recruits from his former CEO
Kahun was founded in 2018 by an unusual group: Dr. Michal Tzuchman Katz, a paediatrics specialist and retired software engineer, along with Tal Goldberg and Eitan Ron. The two founded HumanClick in Israel, which was sold to LivePerson for $9 million in 2004. Goldberg has another interesting fact under his belt: he was the first employee to join Waze. This isn't just a trivial fact because it is what got him his foot in the door for Kahun to raise their first round of funding. Just before the weekend, the startup announced the completion of a seed round of funding for $8 million led by LocalGlobe, the EIC fund of the European Union, and the TFK fund of – you guessed it – Uri Levine, the founder of Waze and Goldberg's former boss. The company also completed a pre-seed round of $5 million from Levine’s fund in 2018, so it seems that the relationship between the two was quite beneficial to the startup at the beginning of its journey.
Before (and during) she started studying medicine, Dr. Tzuchman Katz said that she already developed quite a bit of familiarity and experience with software development. She claims that the very fact that she stayed in these two worlds– though very different– simultaneously allowed her to examine her own thinking process for diagnosis, and try to see how it can be imitated and improved: "I identified the different patterns in medical science – those based on statistical and numerical data that require memory and knowledge, and those which are based on the recognition of pathophysiological processes of cause and effect and thus a comprehensive understanding of the natural course of diseases. In the practical work in medicine and depending on the level of complexity of the patient, the brain simultaneously uses two systems, one of acquired knowledge and the other of patterned perception.” According to her, the model developed at Kahun presents complex medical knowledge as a graph and logically illustrates relationships between different elements, objects, and connections.
"The ability to describe medicine in this way required an understanding of both worlds. One world of software languages and databases– technologies that allow the use of graphs and the ability to describe clinical thinking in mathematical terms. The second world is an understanding of theoretical and practical medical science. Throughout the development of Kahun’s system, challenges arose in adapting the model to different areas of medicine, different health systems and different clinical environments. These challenges required and still require the ability to bridge the development teams with the medical teams at every stage. My experience and familiarity with both worlds serve as a bridge that helps in dealing with these challenges," says Dr. Tzuchman Katz.