Working notes on health information & teaching
This page collects in-progress essays and small reflections on clinical information quality, multilingual care, and the quiet, structural problems I keep seeing in both healthcare and math education.
These are drafts meant to stay honest and specific. They help me think more clearly about what I want to study in a PhD and how my lived experience shapes my questions.
-
1 · From unreadable paper charts to structured EHR: what changed, what didn’t
Growing up in China, I watched doctors write notes that looked more like drawings than language. In the U.S., I discovered EHRs and patient portals. This piece contrasts those two worlds and asks: when information looks more structured, how much is actually gained?
-
2 · Missed windows and delayed diagnoses: what never made it into the record
My mother’s cancer was not a sudden event; there were earlier warnings that never became part of a coherent story. Here I trace how those moments were documented, where they broke, and how EHR timelines can hide as much as they reveal.
-
3 · Documentation variability: same patient, different note, different reality
Across nurses, physicians, and clinics, my mother was described in very different ways. I write about how small phrasing choices and shortcut templates reshape the “truth” of a patient in the database, and why this matters for downstream models.
-
4 · Semantic drift in clinical language: when shorthand quietly changes the meaning
Over time, the same phrase in the chart can mean different things to different people. This essay explores how shorthand, copy-forward behavior, and time pressure cause concepts to drift—creating subtle misalignments that are hard for models to see.
-
5 · Interpreter versus family member: whose words make it into the EHR?
For six years I accompanied my mother to appointments as an informal interpreter, even when a formal interpreter was present. I compare what was actually said in the room with what ended up written down, and how translation choices reshape clinical reality.
-
6 · Where information breaks: triage, nursing notes, and the physician’s summary
In theory, information flows smoothly from triage to nurse to physician to plan. In practice, each handoff is a place where details are compressed, reworded, or dropped. I outline the specific breakpoints I’ve seen and how they might be studied systematically.
-
7 · Healthcare as a noisy, biased network: a math student’s view
I think about healthcare as a noisy network where each node rewrites the signal a little. This piece sketches how ideas from probability, networks, and information theory might help describe what I watched happen to my mother’s data over time.
-
8 · Cross-cultural care and “data inequity” for multilingual patients
Multilingual patients often arrive with rich histories that do not map cleanly into English clinical language. I explore how this mismatch produces thinner, more fragile data trails, and what that implies for models trained on “real-world” EHR data.
-
9 · MyChart, agency, and quietly reading ahead of the doctor
While my mother waited anxiously for appointments, I read every test result the moment it appeared in the portal. Here I write about what it meant to be “ahead” of the doctor, and how portals redistribute both power and anxiety in care.
-
10 · When a student multiplies the wrong sides: what that reveals about structure
A tenth-grade student once chose two random sides of a triangle, multiplied them, and divided by two, calling it “area.” I unpack what this moment reveals about how we teach structure, and how similar misunderstandings appear in clinical workflows.
-
11 · Designing visual math flashcards for neurodivergent learners
I began designing math flashcards for neurodivergent students who struggled with dense, symbolic explanations. This essay connects that work to broader questions about layout, visual hierarchy, and why information structure matters in any complex system.
-
12 · Why information structure matters more than the model
A closing note on my current belief: before we reach for more complex models, we need to understand how the underlying information is shaped, constrained, and made legible. This piece sketches how that belief guides my PhD goals.