Medibot

An AI based interactive search interface to search biomedical literature.

Winner of IBM - MobileMonday- API first hackathon, NYC.

Project Demo: Medibot - AI Assisted biomedical Literature Search

Evidence-based medicine (EBM) is an approach to medical practice intended to optimize decision-making by emphasizing the use of evidence from well-designed and well-conducted research. This provides provides with a method to use critically appraised and scientifically proven evidence for delivering quality health care to a specific population.

IMAGE ALT TEXT HERE

one of the most import step in EBM in searching for evidence. I,e. Finding the Biomedical Literature/articles to answer the clinical quentions. But, unfortunately this is not a simple straight forwards search. The following video summerizes the existing practice.

IMAGE ALT TEXT HERE Click on the image for video.

Unfortunately, this is a very time consuming process. According to a 2014 study, although 76% of the nurse practitioner agree that EBM should be part of their practive, more that 50% them agree that they couldnt apply it effectively because it was a bit difficult and time consuming.

So, Addressing this problem we have decided to add a layer of AI(conversational agent) on to the traditional PubMed website, to automate the processes of finding the appropriate medical terms(MeSH) terms, and building an optimized query.

  • The Converstional agent is like siri or cortana with which you can have a conversation in natural language, like you would interact with your assistant.
  • Context based personalized search: Unlike in traditional keyword based search, in our interface each search has an history and context. you would be having a conversation with bot, where it is possible for the bot to ask you additional information to make you search more personalized.
  • Since, the seach is AI assisted, you could find the same literature upto 7X time faster.
  • Voice integrated: It has a built-in speech-to-text commands, so you could you search for literature using voice commands.

This is how our system works. With our system, we could effectively reduce the time by over 70% in most of the cases.

You can try it yourself here: Project Link.