Understanding the prevalence of disease in different regions of the world is important for physicians and researchers to understand disease drivers. In addition to looking at large-scale regions we can home in on risk factors that make individuals susceptible to disease by exploring regional and population level factors.
What it does
Our pipeline allows physicians to dictate clinical notes. These notes are then converted into text and parsed with NLP, which extracts relevant clinical information and updates a database about the patient's location, disease and time of disease. A website of different regions of the world draws information from this database and updates the distributions of different diseases all around the world.
How we built it
We built our backend in python and our front end is built in D3.
Challenges we ran into
NLP on clinical notes is very different from regular text. Using ontologies and training our own system to recognize the patterns of clinical notes will greatly improve our ability to extract meaningful information.
Accomplishments that we're proud of
We built an awesome proof of concept in one weekend.
What we learned
There will be lots of nuances we have to take care of when we implement the final product. For example, we'll have to handle different languages, different ways of naming the same disease, and remaining HIPAA compliant.
What's next for UDDeR
We will work on smarter NLP tools and easier ways to integrate our product into the clinical workflow.