Diagnostic Atlas¶
Diagnostic reference — instantiates Archetype Pattern Indexing
Maps presenting symptoms to candidate patterns with the look-alikes to rule out and a stated confidence in the fit.
A Diagnostic Atlas indexes patterns for identification under uncertainty: you enter with a set of presenting signs and it returns the candidate patterns that fit, the look-alikes you must rule out, and how sure you can be. Its defining move is the differential — it never returns a single answer but a ranked short-list of possibilities plus the discriminating tests that separate them, because the whole hazard in diagnosis is prematurely committing to the first plausible match. Retrieval runs from symptom to candidate rather than from name to definition, and every match arrives with an explicit confidence so a resemblance is treated as a hypothesis to confirm, not a conclusion.
Example¶
An emergency clinician meets a patient with acute chest pain, shortness of breath, and a raised D-dimer. She consults a diagnostic atlas organized by presenting symptom. Entering "pleuritic chest pain + dyspnea + elevated D-dimer," the atlas surfaces a differential: pulmonary embolism at the top, but with myocardial infarction, pneumonia, and pericarditis listed as the look-alikes that share those signs. Each candidate carries its distinguishing findings — the ECG changes that would favor MI, the fever and focal consolidation that would favor pneumonia — and a confidence note that D-dimer is sensitive but not specific, so a positive result raises suspicion without confirming. The atlas does not diagnose the patient; it hands her a structured differential with rule-out tests and honest confidence, which is exactly what keeps her from anchoring on the first name that fit. She orders the CT angiogram the atlas flags as the confirmatory test, precisely because the index refused to let the top candidate masquerade as certainty.[n1]
How it works¶
The atlas is indexed on findings, not on pattern names: entries are reachable by the symptoms and signs a case presents. Each candidate pattern carries a signature stated as a finding profile — the constellation that must be present for it to be in play — and, crucially, a differential set of look-alikes that share surface signs but are ruled in or out by specific discriminating tests. Every candidate also carries a confidence qualifier: how specific each finding is, and therefore how much a match should be trusted before confirmation. Retrieval returns a ranked short-list, not a verdict; the atlas's job ends at "here are the possibilities, here is how to tell them apart, here is how sure you can be," handing the actual next-step decision to the clinician.
Tuning parameters¶
- Symptom-index breadth — how many findings each entry is reachable by. Broad indexing catches atypical presentations but returns longer, noisier differentials.
- Differential depth — how many look-alikes each candidate lists. Deeper differentials guard against premature closure but slow the reader and can induce over-testing.
- Confidence calibration — how conservatively the fit qualifiers are set. Well-calibrated confidence prevents false certainty; miscalibration either alarms or falsely reassures.
- Rule-out test coupling — how tightly each differential is tied to a discriminating test versus left to judgment. Tight coupling makes the atlas actionable; loose coupling keeps it lightweight but less directive.
When it helps, and when it misleads¶
Its strength is disciplined breadth: by always returning a differential with rule-out tests and calibrated confidence, it directly counters the two great errors of diagnosis — anchoring on the first plausible match and mistaking a suggestive sign for a confirmed one.
Its failure mode is a poorly calibrated confidence layer — an atlas that presents a suggestive match as authoritative manufactures exactly the premature closure it exists to prevent, and a retrieved pattern feels far more certain than the evidence warrants. The classic misuse is stopping at the top candidate and skipping the differential because it "obviously" fits. The guarding discipline is to require that the discriminating tests be run before a candidate is accepted, and to keep confidence qualifiers honest about the specificity of each finding, so the atlas keeps surfacing the look-alikes rather than quietly collapsing to one answer.
How it implements the components¶
pattern_signature— each candidate is defined by a finding profile (the constellation of signs that puts it in play), stated as diagnostic criteria rather than an adoption recipe.retrieval_index— the atlas is indexed on presenting symptoms, so a case is matched from its findings to candidate patterns rather than from a remembered name.counterexample_set— the differential of look-alikes, each with its discriminating test, is the core guard against committing to the first plausible match.confidence_label— every candidate carries a calibrated fit qualifier, keeping a match a hypothesis to confirm rather than a conclusion.
It issues no response_guidance — the atlas identifies and stops, leaving the treatment or leverage move to System Archetype Index and Anti-Pattern Catalog; and it draws no neighbor_distinction map, because its lookalikes are handled as case-level differentials, not as a stable inter-archetype boundary.
Related¶
- Instantiates: Archetype Pattern Indexing — it supplies the symptom-driven, differential side of the index, built for identification under uncertainty.
- Consumes: Tagging Schema provides the symptom-and-finding vocabulary its entries are indexed on.
- Sibling mechanisms: Design Pattern Catalog · Anti-Pattern Catalog · Case Library · System Archetype Index · Solution Archetype Archive · Pattern Card Template · Tagging Schema · Pattern Library.
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: Diagnostic Atlas operates as a non-executable information artifact that externalizes static or prospective structure because it maps presenting symptoms to candidate patterns with the look-alikes to rule out and a stated confidence in the fit.
Independent corroboration: The frozen evidence defines Diagnostic Atlas as 'Maps presenting symptoms to candidate patterns with the look-alikes to rule out and a stated confidence in the fit', so its operative form is Representation, Specification & Plan.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Medicine & Healthcare
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Clinical medicine cohered symptom-indexed diagnostic atlases that present candidate conditions, look-alikes, discriminating tests, and confidence.
Related originating lineages:
- Biology & Ecology — Taxonomic field guides supplied parallel pattern atlases with look-alike discrimination.
Review resolution: Clinical medicine cohered symptom-indexed diagnostic atlases that present candidate conditions, look-alikes, discriminating tests, and confidence. The retained alternate lineages materially shaped the mechanism's form.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Differential diagnosis — the standard clinical discipline of holding a ranked list of candidate conditions consistent with the findings and using discriminating tests to narrow it, rather than committing to the first that fits. The atlas's differential set and confidence qualifiers are the indexing form of this discipline. ↩