Project Background
Medical abstraction involves systematically extracting specific data points or information from medical records, such as diagnoses, procedures, medications, and laboratory results. Medical abstraction is often performed for various purposes, including research studies, quality improvement initiatives, billing and coding, and healthcare analytics.
Historically this is a very manual, tedious process where teams of people with clinical experience comb through charts to find relevant information.
Optum wanted to equip them with a tool to streamline their workflows - using the power of Natural Language Processing.
Methodology
Using in-depth interviews, I performed multiple rounds of research including foundational research to inform “what to build”, evaluative research to guide “how to build it”, and rolling usability testing to make sure we were aligned with user needs.
Insights were gathered from 32 users including abstractors, quality assurance staff, system administrators, and clinical SME’s. Some of this work is ongoing.
Findings
I’m limited in what I can share about the insights, but for more information on the foundational product, visit the link!
