Vibe Rounds · Clinical Reasoning & Cognitive Frameworks · Draft, untested
[!IMPORTANT] Clinical Disclaimer This module produces learning observations, not a clinical decision, care plan, or monitoring order set. Every item generated requires independent verification against a toxicology/specialty reference or the treating team before it informs any action. Do not enter output from this module into a patient record, handover note, or care plan without review and endorsement by a licensed clinician.
Given a diagnosis or case narrative, generate a structured, source-attributed map of the complications, syndromes, and risks that could unfold ahead — organized by timing and by physiological system — so a learner or clinician can pre-load vigilance rather than being surprised by a late, treatment-related, or system-specific complication.
This module does not produce a deterministic, checkable-against-ground-truth output the way a guideline or interaction check does (see Modules 13, 53). Its reliability comes instead from a fixed output taxonomy, mandatory mechanism justification, and mandatory source-tagging of every item — the goal is grounded synthesis, not free-generation dressed as fact.
Initiation → Execution → Closure/Review
Before generating anything, the case must be entered in a fixed format — this alone removes most of the variance in what the module surfaces.
Prompt:
I am going to give you a diagnosis/case for a complication and trajectory map.
Case:
- Diagnosis (working or confirmed): [X]
- Key exposure/timeline so far: [what happened, when]
- Interventions already given/avoided: [drugs, doses, notable choices — e.g. "succinylcholine avoided for intubation"]
- Current status: [stable/unstable, on what support]
Do not generate the map yet. First, confirm you have enough information, or ask
me for the minimum additional detail you need (e.g. renal function, airway
management specifics) before you can produce a mechanism-justified map.
Application Note: If the model asks a clarifying question here, answer it before moving to Step 58.1. A map built on an underspecified case (e.g. “poisoning, unknown agent, unknown time since exposure”) will default toward generic literature rather than case-specific trajectory.
Prompt:
Using the case above, generate a complication/trajectory map using EXACTLY this taxonomy — do not add or rename categories:
## Timing-based
- Immediate (0–24h): ...
- Early (24–96h): ...
- Delayed (days–weeks): ...
- Late (weeks–months, if applicable): ...
## System-based
- Neurological: ...
- Cardiovascular: ...
- Respiratory: ...
- Renal/metabolic: ...
- Infectious/secondary: ...
- Psychiatric/long-term: ...
(omit any system with nothing relevant — do not pad)
## Treatment-related (iatrogenic)
- Complications caused by the antidote/intervention itself, or that would
have occurred under a different (avoided) intervention choice: ...
## Conditional branches
- If [specific treatment/finding] → watch for [X]
- If [specific treatment/finding was avoided] → this branch does not apply
Rules for every single item:
1. State the mechanism in ≤1 line (e.g. "irreversible enzyme binding," not
just "aging").
2. Tag every item as either [from case] — directly grounded in what was
given above — or [general literature] — standard knowledge for this
diagnosis not confirmed against this specific case.
3. For [general literature] items, search for and cite a source rather than
relying on memory alone if you have that capability. If you cannot
verify a claim, label it [unverified] instead of stating it as fact.
4. Do not include anything you cannot mechanistically justify.
5. Do not repeat the same complication under two categories — cross-reference
instead (e.g. "see Respiratory — Immediate").
Validated Env.: Not yet tested on a live platform against ground truth. Treat as case-report/expert-opinion tier per the repository’s stated evidence base until validated.
Application Note: If the case is high-stakes/rare (e.g. an uncommon toxin, an unusual genetic condition), explicitly ask the model to flag its own confidence as low and to recommend the specific type of specialist/reference to consult — do not accept a confidently flat list for an area where training data is likely thin.
Prompt:
Before I use this map:
1. List every item tagged [general literature] separately from every item
tagged [from case] — I want to review the two groups independently.
2. Of the [general literature] items, which ones did you search/cite a
source for, and which are from memory only?
3. Identify the single complication on this map that, if missed, would be
most dangerous to the patient — and why detecting it early matters more
than the others.
4. Name what this map does NOT cover (e.g. complications outside the
systems/timing categories used, rare presentations, population-specific
risk factors not in the case).
Application Note: Step 58.2 is not optional. This is the step that turns a plausible-sounding list into something a clinician can actually weigh — separating searched claims from memory claims, and forcing the module to name its own blind spots, is what keeps this a reasoning aid rather than a source of quiet overconfidence.
## Timing-based
- Early (24–96h): Intermediate syndrome — sudden return of respiratory,
neck-flexor, and proximal-limb weakness from prolonged AChE inhibition
at the neuromuscular junction. [general literature — search: "intermediate
syndrome organophosphate poisoning"]
- Delayed (1–3 weeks): OPIDN — progressive weakness/sensory loss from
neuropathy target esterase inhibition, distinct mechanism from acute
cholinergic crisis. [general literature]
## Treatment-related
- "Aging" — irreversible phosphorylation of AChE beyond which pralidoxime
cannot reactivate the enzyme; time-critical. [general literature]
- Prolonged paralysis — risk specifically IF succinylcholine used for
intubation (avoided in this case, per input). [from case + general
literature — conditional branch does not apply here]
## Conditional branches
- If succinylcholine avoided (as in this case) → prolonged paralysis risk
from that specific cause does not apply.
- If atropine/pralidoxime infusion titrated down too early → recrudescence
of secretions/wheeze/respiratory distress. [general literature]
Clinical Reasoning & Cognitive Frameworks (companion to Modules 1, 12, 28, 30, 50)
Draft module — not yet peer-reviewed, not yet validated against ground truth. Evidence base for this module’s own design: case-report/expert-opinion tier. Treat every generated map as a learning exercise requiring specialist/reference verification, not a monitoring order set.