Today, I was introduced to a project called Etiologic Classification for Ischemic Stroke (ECIS) through Electronic Health Record-based Natural Language Processing. Lead by Dr. Feifan Liu, a professor at the University of Massachusetts Medical School, this project aims to develop software that automatically classifies the causes of ischemic stroke (IS).
Here is some background information that I learned:
- Stroke is a major issue in today’s world, as it is the 5th leading cause of death in the US and 2nd leading cause of death globally.
- The majority of the burden of stroke is attributed to IS.
- Etiologic subtype classification (identifying the causes) of IS is critical for treatment management and outcome prediction.
- However, classification is currently performed through manual chart review, a time-consuming, error-prone process.
- By developing a deep learning-based NLP system for automatic classification, the project hopes to overcome these limitations.
I was assigned to work on the pre-processing software; my program would use Pandas to manipulate the Electronic Health Record (EHR) data into a NLP system-usable format. I’m eager to contribute to a project that can potentially help improve the treatment of stroke victims, and I’m grateful for the chance to improve my work with Energize Andover using the new Pandas knowledge that I learn.