Capturing Investigator Effort from Clinical Notes with AI
Exploring AI interaction patterns to better capture the effort behind public health investigations from clinical and investigative notes.
OVERVIEW
A routine follow-up call may take minutes, while another investigation requires repeated outreach, home visits, and connecting a patient with housing or other services. Yet today, public health teams often rely on a disease priority score to estimate investigator effort.
Tools used: Expert Interviews
Investigators document every action taken in a case as notes
Effort per case is scored by disease rather than the notes
Knowing actual effort is vital to manage and plan staffing
EXPLORATION HIGHLIGHTS
AI-generated effort breakdown
Paste a de-identified SOAP note and receive an overall effort score with a category-by-category explanation of how the score was calculated.
Verify every inference
Every point can be traced back to the exact sentence or passage used by the model, making the reasoning easy to audit instead of treating the output as a black box.
Review inference confidence score
Each assessment includes a confidence indicator, while investigators can dispute, edit, or override individual findings when additional context is needed.
Interested in the details?
This is an ongoing AI design exploration and actively evolving to explore transparent AI experiences for public health. The scoring rubric is representative and continues to evolve alongside backend development. If you'd like to discuss the interaction patterns, implementation approach, or design process, I'd love to chat.