UVEA • ON CALL NOW
North-star On Call Now module. Draft 1 of the written screen content, filled into the On Call Now Module Writer Template. Recognition-and-escalation only; recommendation (non-directive) wording throughout. Layout, imagery, and Canva build come later.
The Writer authors every section of this document. The Faculty Reviewer signs off — they react and approve, they do not co-draft.
There are two review touchpoints, in order: (1) the Faculty Reviewer approves the Section 2 decision tree BEFORE any screen prose is written; (2) after the full draft is submitted, the Faculty Reviewer reads the whole module.
The Section 6 self-check is something the Writer runs on themselves right before submitting. Submitting is what triggers touchpoint (2).
On this module John Gonzales, MD is the senior Writer and Padmamalini Shantha, MD is the proposed Faculty Reviewer (confirmation pending).
The Project Manager uses these fields to route the draft to the right Faculty Reviewer. Confirm the reviewer’s name here before it appears in any assignment email.
| Module title | Hypopyon triage |
| Clinical scenario (one sentence) | A resident is called to a patient with a hypopyon and needs to recognize whether this is time-critical infection (post-procedural or endogenous endophthalmitis, or infectious retinitis) or severe sterile anterior uveitis — and decide who to contact. |
| Writer name | John Gonzales, MD — UCSF Proctor Foundation (senior author) |
| Faculty Reviewer assigned | Padmamalini Shantha, MD (confirmation pending) |
| Draft date | June 11, 2026 |
| Draft 1 target completion | [set ~4 weeks from assignment] |
Plan the tree, then send only this section to your Faculty Reviewer to align on scope. The reviewer confirms it is the right tree; they do not co-author it.
Why this order: a planned tree takes about 15 minutes, while reworking prose built on the wrong tree costs about 2 hours. Lock Section 2 before drafting Sections 3–5.
Status for this module: John drafted the tree below; Dr. Shantha’s sign-off is still pending, so everything downstream (Sections 3–5) is provisional. Self-check item 10 tracks this.
“I have a patient with a hypopyon — what is the most important thing this could be tonight, and who do I need to call?”
Applied to this module, the four features link to the four endpoints as follows:
| Decision feature | Role in the tree | Endpoint reached |
|---|---|---|
| Feature 1 — recent procedure or trauma | Asked as the Screen 1 question; the “Yes” branch. | → Endpoint A |
| Feature 2 — immunocompromised host or posterior findings | Asked as the Screen 2 question; the “Yes” branch. | → Endpoint B |
| Feature 4 — mimic flags (very high IOP + corneal edema, corneal ulcer/infiltrate, known tumor) | Asked as the Screen 3 question; the “Yes” branch. This is the discriminator that gates both endpoints on Screen 3. | → Endpoint D |
| Feature 3 — immunocompetent, anterior-confined inflammation | Not asked as its own question. It is the default reached by answering “No” to Features 1, 2, and 4. | → Endpoint C (default) |
Note the two asymmetries this table makes visible. (1) Endpoint C is a default, not an asked question — a resident only lands there after ruling out everything else, which is the conservative design. (2) Feature 4 has a known leak: herpetic anterior uveitis with very high IOP and corneal edema can trip the Screen 3 question and route to Endpoint D, though it is really an inflammatory case. Endpoint D’s escalation list includes uveitis, so the case still reaches the right service. These five recommendations are written to transfer to the other MVP modules — consider folding them into the blank On Call template’s Section 2 so every writer sees them.
Recent intraocular surgery, intravitreal injection, glaucoma bleb procedure, or open-globe trauma?
Immunocompromised host, or posterior findings — retinitis, retinal whitening, dense vitritis, or no posterior view?
Very high IOP with corneal edema, a corneal infiltrate or ulcer, or a known intraocular tumor?
Longest path = 3 taps (Screen 1 → Screen 2 → Screen 3 → endpoint). Highest-consequence sort (surgical history) sits first, per the storyboard’s “red flags before pathophysiology” principle. Red = time-critical (A, B); slate = reconsider (D); amber = default sterile pathway (C).
Everything from here on is downstream of the tree. Treat it as provisional until the Faculty Reviewer has approved Section 2.
Screen title: Red eye with hypopyon
Framing sentence: A layered hypopyon means cells or debris have settled in the anterior chamber. Tonight’s job is to separate time-critical infection from severe sterile inflammation.
First decision question: Has the patient had intraocular surgery, an intravitreal injection, a glaucoma bleb procedure, or open-globe trauma?
Branch buttons
Word count of body text (framing + question) ≈ 40 words — under the 50-word target and well under the 80-word ceiling. Screen 1 is intentionally a binary safety sort, not a diagnosis.
Same status as Screen 1: these screens are fixed by the approved tree, so they stay provisional until Section 2 is locked.
Screen ID: Screen 2
Screen title: Host and posterior segment
Confirmation sentence: With no recent procedure, the next question is whether the host or the posterior segment changes the picture — either can turn a “hypopyon” into a sight- or life-threatening infection.
Next decision question: Is the patient immunocompromised, or are there posterior-segment findings — retinitis, retinal whitening, dense vitritis, or no posterior view?
Branch buttons
Screen ID: Screen 3
Screen title: Anterior picture or mimic
Confirmation sentence: The inflammation looks anterior. Before settling on sterile uveitis, it is worth a moment to confirm this is a true inflammatory hypopyon and not a look-alike.
Next decision question: Does the eye show very high IOP with corneal edema, a corneal infiltrate or ulcer, or a known intraocular tumor?
Branch buttons
Two intermediate screens — at the template’s limit, not over it. Each one narrows the decision rather than rephrasing the last.
Endpoints are where recommendation-not-directive discipline matters most. During review, every red-tier endpoint gets a dedicated wording pass to confirm FDA Non-Device CDS framing (recognition and escalation, never “start / treat as / give”).
Each endpoint follows the same four parts: recognition → rationale → document/escalate → scope disclaimer. No doses, no antibiotic choices, no institution-specific protocols.
The Writer completes this checklist before submitting the draft through the Airtable form. Submitting is what routes the full module to the Faculty Reviewer for touchpoint (2).
Item 10 reaches back to the start of the process: it confirms the Section 2 tree sign-off (touchpoint 1) already happened before the prose was written.
| # | Criterion | Self-assessment / for reviewer |
|---|---|---|
| 1 | Screen 1 is under 80 words. | ≈ 40 words of body text — passes. |
| 2 | Every branch leads to an endpoint (no dead ends). | Traced: all five branch buttons resolve to Endpoint A–D. |
| 3 | Module is 3 taps or fewer to any endpoint. | Longest path Screen 1 → 2 → 3 → Endpoint = 3 taps — passes. |
| 4 | Every endpoint has all four parts. | Endpoints A–D each carry recognition, rationale, document/escalate, disclaimer. |
| 5 | No endpoint uses directive verbs. | No “start / treat as / give / prescribe.” Reviewer: confirm the steroid-caution wording in Endpoint A reads as rationale, not instruction. |
| 6 | No specific drug doses or names. | None included. C acnes named as an organism, not a drug. |
| 7 | Every endpoint includes a rationale. | Yes — Part 2 of each endpoint. |
| 8 | Reading level grade 9–11. | Run full text through Hemingway before submission; not yet verified. |
| 9 | No PHI or identifiable patient detail. | All scenarios generic — passes. |
| 10 | Faculty Reviewer consulted on tree plan (Section 2) before drafting. | Pending Dr. Shantha’s confirmation as Faculty Reviewer. |
The goal is not to write a textbook chapter. The goal is to give a resident on call a 3-tap path from question to a useful answer. — Draft for senior-author and Faculty-Reviewer revision.