High Potential Startup #31: Kahun Medical
AI assessment tool that helps in scalable clinical reasoning
At the time of publication: Series A | Total funding raised: ~8mn USD
*****
Human expertise is difficult and expensive to scale. When a lot of expert intuition is required, then the expertise becomes expensive (because intuition requires a lot of training) and the range of outcomes varies. Case in point healthcare.
In many healthcare systems, good doctors are scarce and expensive. Hence the productivity of care needs to improve. If doctors are able to take on more patients than they currently do - everyone wins - health providers can get more revenue and it drives down costs for patients too.
Kahun is an Israel-based AI startup that improves outcomes for patients and increases productivity of doctors.
Kahun currently provides a 'pre-clinical assessment tool' - a chatbot that patients use before doctor appointments (link shared by health providers). The patients answer questions about symptoms and their history.
Kahun then prepares a quick 'case' for the doctor with a summary of the symptoms and other answers provided by the patient and also an initial diagnosis and other diseases to consider.
Kahun says,
“The tool employs the same reasoning expected from trained providers, asking all the right questions and dynamically calculating the best next questions. The questions are tailored to each specific patient in order to take in all the pertinent findings, rule out rare diseases, or urgent conditions. The process takes three to five minutes, leaving no stone unturned, and provides physicians with a clinical summary including recommendations for further evaluation.”
This is a very smart approach. The status quo is to tell the doctor about your symptoms and concerns in person during the appointment. A pre-clinical tool can give the doctor facts and intelligence even before the visit. This is fantastic for both doctor productivity and patient outcomes!
Kahun is able to do this because it has built what it claims is a ‘proprietary map’ of 30 million medical insights from textbooks and journals - the same way a great doctor has to often gain knowledge. Hence Kahun claims it is 'AI that thinks like a doctor'. This is truly 'scalable clinical reasoning'.
I really like the product approach of using a patient chatbot instead of asking doctors to input the symptoms during the appointment. This is great for productivity and also ensures all the symptoms are captured and the patient gets time to explain.
And instead of being a black box AI, this is 'explainable AI' - the reasoning for a diagnosis is also given. This is crucial for trust and adoption by doctors.
Kahun also integrates with Electronic Health Records which ensures that past data about patients is taken into account during assessment and generates a fuller picture.
The pricing model seems unclear but I think there are some very good opportunities. One model is to price per customer chat or a monthly subscription per doctor. Once the product is integrated into the day to day ‘workflow’ of the doctor, there is going to be very low churn making the business very attractive.
Kahun has an interesting ongoing partnership with the New England Journal of Medicine. Kahun powers one of their products called ‘NEJM Healer’ which medical students use to learn ‘clinical reasoning’.
Kahun can seek out more such partnerships for training. Getting the tool in the hands of medical students (even for free) will ensure adoption during professional practice too.
In the future, the chatbot can also request patients for images and voice records (self-serve using a smartphone) for analysis. This will improve diagnosis. These capabilities can be done inhouse or maybe in partnership with other AI companies.
I hope this technology soon reaches the less developed economies where this can really improve patient outcomes and save lives. It just elevates the standard of care by being a co-pilot to the doctor.
There are many reasons why Kahun is a high potential startup - operates in a very large market and has a very good solution approach; has a great value proposition and can become a standard part of a doctor’s ‘workflow’ in the future. Once an integral part of a doctor’s workflow, there is going to be very low churn and very high margins on its product. Its foundational AI engine can potentially be used to build more products (potentially even self-serve products that can offer a second opinion) that can further increase productivity of providers and save costs for patients.
This is definitely useful in saving time. As a former practitioner, I would still not be comfortable with initial diagnosis. It may provide for all possibilities but stay away from provisional diagnosis.