The problem
Clinical research teams needed a reliable way to compare how multiple coders evaluated the same patient encounters. The existing tool lacked data sharing, collaboration, and meaningful inter-rater reliability analysis. Additionally, there needed to be a way to qualitatively code patient-clinician encounter videos while maintaining the security of patient data.
My role
I developed EnIAC from a small feature request into a full qualitative research platform, designing and building the React/Electron interface, collaboration workflows, agreement tools, and visual interaction.
Methods
UX design, React, Electron, statistical analysis, iterative debugging, information architecture, interaction design, and user-centered problem solving.
-
Build the foundation
I created the clinical coding application with independent export/import capabilities without disrupting researchers' existing process. I built around the frameworks they wanted for the app, not vice versa.
-
Make agreement meaningful
I built and validated multiple calculated measures of agreement among independent qualitative reviews, including ICC, Cohen's kappa, Fleiss' kappa, weighted kappa, and percent agreement. These are incredibly useful in qualitative assessment.
-
Design for real teams
As requirements grew, I added role-based sharing, researcher rosters, progress tracking, transparent agreement views, CSV export, and light/dark theming.
What I learned
Building EnIAC showed me that good research software is more than functionality. In the beginning, I wanted to make something that could be used for any kind of qualititave observation. However, it became more clear that what Mayo Clinic's Knowledge and Evaluation Research unit wanted was an assessment tool that was directly personalized to the work that they were specifically doing. This meant that I had to dig deep and observe their work flow to optimize the performance of the tool. As a plus, now the tool can be used for both the specific assessment frameworks that they requested, and is adaptable to ANY qualitative assessment, clinical or non-clinical research. EnIAC grew into a full-fledged platform used for qualitative assessment and was recognized with an award at the Midwest Endocrinology Society Conference.
