European Commission logo
italiano italiano
CORDIS - Risultati della ricerca dell’UE
CORDIS

Kahun - an interactive medical knowledge base, for modeling medical knowledge and managing diagnostic processes, as well medical knowledge related to COVID-19.

Periodic Reporting for period 2 - Kahun (Kahun - an interactive medical knowledge base, for modeling medical knowledge and managing diagnostic processes, as well medical knowledge related to COVID-19.)

Periodo di rendicontazione: 2021-08-01 al 2022-07-31

With 30M+ PubMed articles, the amount of available medical knowledge is mind-boggling and constantly growing. Physicians need help bridging the gap between the clinical reasoning they need to apply when diagnosing their ‎patients and the evidence-based knowledge that is out there. When in doubt, many times physicians rely on access to collogues and specialists they work with, unfortunately this is less relevant and accessible for those working alone in ‎remote locations. Most clinicians are just too overloaded with patients and have less time to consult research to ‎support their intake, diagnosis and decisions.‎
Kahun brings updated, integrated medical knowledge to physicians in an intuitive, actionable ‎way that can’t be achieved through text-based tools. Kahun’s AI engine generates and infers from ‎a huge knowledge graph that links data about diseases, findings, complications, risk factors, etc. The platform is continuously updated to integrate knowledge from ‎across the medical literature and epidemiologic databases.‎

Conclusions of the action
Kahun Medical has successfully met the desired objectives and goals, developing a full, mobile decision assistance tool for medical professionals. The technology’s development was completed, tested and implemented into the Kahun mobile application. Kahun launched its stand-alone mobile app in 2021, and gained thousands of medical professional users from across the world. The platform and app were tested and evaluated as a clinical reasoning tool 2
The original business plan of Kahun was based on offering its mobile app for free, achieving growth and then turning to monetization. As part of Kahun’s business development activity, the company decided to pivot and use its knowledge graph and algorithm to power a chatbot for clinical intake. The product relies on the Kahun’s infrastructure to conduct a gold-standard clinical intake for patients as a self service app, conducted before their visit to a primary care physician. The company is now focused around deploying this solution, especially with telemedicine partners via a dedicated API. The new chatbot was described and presented during the second progress meeting.
Kahun has successfully secured a number of partnerships for the implementation and use of the chatbot:
HelixVM and MediOrobis – Two pure acute and primary care virtual medicine US companies which integrated Kahun into their workflows. Sending a link for pre-visit intake to patients who schedule an appointment. A summary of the intake is reviewed by the provider before the visit starts. The summary includes the history of present illness, review of systems as collected by Kahun, suggested assessment, differential diagnosis and even suggested next steps – all backed up by links to the originating evidence from Kahun’s knowledge graph.
Kahun also recently signed another partner: HealthBridge from South Africa – an HER software developer with more than 3,000 clinics who is currently embedding Kahun into its software.
Another important partnership Kahun made is with the New England Journal of Medicne 1 which chose to work with Kahun’s data to drive its Healer medical simulation tool sold to residency education programs around the world.

Relevant links:
1. https://healer.nejm.org/nejm-healer-collaborates-with-med-tech-start-up-to-enhance-teaching-and-assessing-clinical-reasoning/
2. https://www.kahun.com/news/kahun-at-the-center-of-a-clinical-vignette-study-to-assess-its-performance-as-a-next-gen-ai-driven-diagnostic-support-tool
3. https://www.kahun.com/news/kahun-successfully-identifies-red-flag-covid-19-patients-in-its-pioneering-computer-aided-risk-assessment-calculator
At the beginning of the project Kahun launched a COVID-19 dedicated knowledge map/symptomizer along with a dedicated covid-19 risk calculator3. The results of this work were published in a joint research with Israel’s first Covid-19 turned hospital. Later, based on input from its early pilots, Kahun launched its V2.0 ‎version of the platform with a reworked UI/UX, data input and flow of the Kahun ‎diagnosis assistant to improve ease-of-use, information collection, suggestions accuracy, and ‎input time. The tasks included improving the model and solving algorithm in the backend of ‎V2.0 knowledge graph content building and moderation, and researching compliance ‎requirements. Kahun has also attacked some of the tasks which were part of the “equity work packages” including the next step engine, updating the ‎medical knowledge editor, and communicating and disseminating the project and platform. ‎
Overview of the project’s results
By the end of the project, Kahun have successfully completed all the identified tasks, deliverables and milestones. V2 of the Kahun application, both ‎frontend and backend, including UI/UX, data input methodology, improved model and solving ‎algorithm, and expanded medical knowledge graph, is implemented in the current Kahun mobile application.
The mobile app launched by Kahun in early 2021 was initially offered for free. More than 5,000 medical professional users have used of the product. Running a total of over 50,000 consultation sessions.
Kahun’s knowledge graph is in commercial use via its partnership with the New England Journal of Medicine for medical education worldwide.
As mentioned, Kahun has pivoted its main focus and, based on its core technology, developed a chatbot that improves and streamlines patient intake. State of the art solutions designed for patient use are mostly “symptom checkers” designed to replace a google search and guess a diagnosis based on a short patient intake. Such technologies rely on a mix of patient-record Big Data or expert systems. Kahun is leveraging its knowledge graph, as developed in the project along with the inference engine to offer the first clinical intake chatbot, designed for use by providers. i.e its goal is not to guess a diagnosis, but help the provider by providing a gold standard intake, saving interview and documentation time and improving the overall level of care.
A recent study Kahun performed (pre-publication) has tested Kahun’s intake vs the leading symptom checkers in the market from companies such as Babylon Helath, K and Ada. Kahun came out much better in its ability to uncover pertinent findings in the tested cases with less irrelevant questions.
As noted above, Kahun is in early stages of commercializing this product with several telemedicine partners (see image 1 attached)

The company believes that its explainable AI platform, puts it in a strong starting point to offer its solution to telemedicine players first and later on to other brick and mortar health players. The company intends to focus on offering providers the following values: increased efficiency, reduced variability, and improved overall level of care. Social benefits: 1st class care for billions in developing countries, improve diagnosis and ‎patient care despite reduced expert staff, prevent misdiagnosis, reducing providers burnout and support/guidance to ‎medical students and young physicians. ‎
Kahun’s social benefit was expanded, with its pivot and the introduction of the self-service assessment chatbot. Kahun is now offering the ability to improve quality of care to billions of people in developing countries.
Kahun app
image 1 - publications