Diagnostic Narrowing Protocol¶
Diagnostic protocol — instantiates Progressive Narrowing
Reduces a differential of possible causes to one working diagnosis by ordering tests to discriminate fastest, keeping can't-miss rare causes alive until ruled out, and reopening the moment the case stops fitting.
A Diagnostic Narrowing Protocol takes a broad differential — the set of explanations that could account for one concrete case — and drives it toward a single working diagnosis using discriminating tests, sequenced so the most decisive and most dangerous possibilities are addressed first. Its distinguishing feature is that it operates on one case, in real time, under consequence asymmetry: a rare but lethal cause must stay on the list even when a common benign one looks likely, and the moment the case stops behaving as the working diagnosis predicts, the protocol reopens rather than doubling down. It narrows in order to act, not in order to publish.
Example¶
A patient arrives at the emergency department with acute chest pain. The differential is wide: musculoskeletal strain, reflux, pericarditis, myocardial infarction, pulmonary embolism, aortic dissection. The protocol does not work through them alphabetically or by base rate; it orders by urgency times likelihood, front-loading the "can't-miss" killers. An ECG and troponin discriminate myocardial infarction quickly. A D-dimer and, if indicated, CT angiography address pulmonary embolism. Asymmetric pulses or a widened mediastinum keep aortic dissection alive until imaging clears it. Musculoskeletal strain, though statistically the most common, is treated as a diagnosis of exclusion, held open until the dangerous causes have been ruled out — the protocol refuses to let the comfortable, likely answer suppress the rare lethal one.
If the workup points to reflux and the pain resolves, that becomes the working diagnosis — but with an explicit reopening trigger: return immediately if the pain recurs or changes character. The protocol's value is that it converges fast enough to treat while keeping the deadly outliers visible until evidence, not reassurance, removes them.
How it works¶
- Order tests by discriminating power and downside. The next test is the one that most shifts the probabilities on the most dangerous candidates, not the cheapest or most convenient.
- Rule out before ruling in. Dangerous causes are actively excluded even while a benign cause is more probable.
- Keep the rare-and-lethal visible. Low-probability, high-consequence explanations stay on the differential past the point base rates would drop them.
- Attach a reopening trigger to the working diagnosis. Convergence is provisional: a defined change — no response to treatment, a new sign — reopens the differential.
Tuning parameters¶
- Urgency-vs-likelihood weighting — how far the ordering favors excluding dangerous causes over probable ones. Heavy urgency-weighting is safe but orders expensive tests early; heavy likelihood-weighting is efficient but risks missing the rare killer.
- Test threshold — how strong a result must be to remove a candidate. Strict thresholds avoid false exclusion but add tests, time, and cost.
- Rare-cause retention depth — how long low-probability lethal causes stay on the list. Longer retention is safer but drives over-testing.
- Reopening sensitivity — what counts as "the case no longer fits." A hair-trigger catches errors but churns; a high bar risks anchoring on a wrong diagnosis.
When it helps, and when it misleads¶
Its strength is that it converges quickly enough to act while structurally guarding against the deadliest error — dismissing a rare lethal cause because a common one is present. The ordering is the safety mechanism.
Its failure mode is that the same speed invites premature closure:[n1] accepting the first plausible diagnosis and stopping the workup, so contradictory evidence gets explained away instead of reopening the list. The classic misuse is anchoring on the triage label — "known reflux" — and ordering tests to confirm it rather than to discriminate. The guarding discipline is to keep at least one alternative explicitly alive to the end and to treat any failure-to-respond as a reopening trigger rather than a dosing problem.
How it implements the components¶
filter_sequence— tests are ordered by discriminating power and consequence (urgency × likelihood), so each step removes the most decisive uncertainty first.evidence_requirement— a candidate is removed only on a discriminating test result adequate to exclude it, not on impression.diversity_retention_rule— rare, high-consequence explanations are held on the differential past the point base rates alone would drop them.reopening_rule— the working diagnosis carries an explicit trigger (non-response, new sign) that reopens the differential.
It resolves one case but frames no formal comparison_frame across rival theories and fires no finalist_commitment_rule on a general question — those are research_hypothesis_elimination, which adjudicates competing explanations of a phenomenon over a research program rather than one patient under time pressure.
Related¶
- Instantiates: Progressive Narrowing — evidence-gated convergence on a single working cause with a live reopening path.
- Sibling mechanisms: Research Hypothesis Elimination · Design Downselection Review · Hiring Shortlist Process · Procurement Shortlisting · Successive Screening · Legal Issue Narrowing · Candidate Disposition Log · Weighted Scoring Matrix · Funnel Process
Editorial Notes¶
Form Classification¶
Form family: Protocol, Workflow & Routine
Rationale: Diagnostic Narrowing Protocol operates as a repeatable ordered procedure or handoff sequence that coordinates action because it reduces a differential of possible causes to one working diagnosis by ordering tests to discriminate fastest, keeping can't-miss rare causes alive until ruled out, and reopening the moment the case stops fitting.
Independent corroboration: The frozen evidence defines Diagnostic Narrowing Protocol as 'Reduces a differential of possible causes to one working diagnosis by ordering tests to discriminate fastest, keeping can't-miss rare causes alive until ruled out, and reopening the moment the case stops fitting', so its operative form is Protocol, Workflow & Routine.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Medicine & Healthcare
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Clinical workup practice cohered ordered discriminating tests, cannot-miss rule-outs, and reopening when a case ceases to fit the working diagnosis.
Related originating lineages:
- Engineering & Design — Fault isolation supplied analogous test ordering and branch reopening in technical systems.
Review resolution: Clinical workup practice cohered ordered discriminating tests, cannot-miss rule-outs, and reopening when a case ceases to fit the working diagnosis. The retained alternate lineages materially shaped the mechanism's form.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Premature closure — in diagnostic-error research (notably Pat Croskerry's work on clinical reasoning), the tendency to stop considering alternatives once an initial diagnosis seems adequate. It is among the most common cognitive contributors to missed diagnoses, which is exactly why an explicit reopening trigger is built into the protocol rather than left to the clinician's discretion. ↩