Expert Case Recall Checklist¶
Elicitation protocol — instantiates Nearest-Exemplar Response Reuse
Forces an expert's tacit "this reminds me of a case" into an explicit, auditable comparison — which case, why it's close, and where the resemblance breaks.
An Expert Case Recall Checklist targets the moment an experienced practitioner thinks "I've seen this before" — a recognition that is fast, powerful, and almost entirely tacit[1]. The checklist does not replace that recognition; it makes it say its reasons out loud. Through a short, fixed set of prompts it forces the expert to encode the present case explicitly, to name the specific prior case being recalled and its outcome, to state why the two are close, and — the load-bearing prompt — to say where the match breaks and how confident they are. Its whole value is auditability: it converts a private hunch reasoned directly from a concrete remembered case into a legible, checkable comparison that a colleague or the expert's own slower judgment can catch. It elicits and interrogates recall; it does not score similarity on a fixed rubric and it does not adapt a stored solution.
Example¶
An emergency physician sees a 19-year-old athlete with sharp chest pain that worsens when he presses on the sternum. Her trained pattern-recognition fires instantly: benign costochondritis — looks just like the kid I saw last month. Left there, that recognition is fast and usually right and occasionally catastrophically wrong. The recall checklist intervenes before she moves on.
Prompt one: profile the present case — age, pain quality, exertional or not, ECG findings, family history. Filling it out, she notes she has not yet asked about family history. Prompt two: name the case you are recalling and its outcome — "last month's costochondritis, discharged, fine." Prompt three: why is it close — same age band, same reproducible-on-palpation pain, reasoning straight from that concrete patient rather than from a rule. Prompt four, the one that bites: where does this case differ, and what is your confidence? Asking the family-history question now, she learns his uncle died suddenly at 30. That is exactly where the resemblance breaks. Her confidence drops, she orders further workup instead of discharging, and the checklist has turned a silent near-miss into a documented decision.
How it works¶
- Encode the present case. A structured prompt captures the decisive features before any prior case is named, so the recall is checked against the actual situation rather than a remembered gestalt.
- Name the exemplar, not a category. The expert states which concrete case is being recalled and its outcome — reasoning from the specific remembered instance, which the checklist records as such.
- State the fit, then force the break. Why the cases are close, then the mandatory counter-prompt: where they differ and how confident the expert is. The differ-prompt is what makes the tool debiasing rather than confirming.
- Route on low confidence. A stated break-point plus low confidence is an instruction to slow down, gather more, or get a second reader — not to proceed.
Tuning parameters¶
- Prompt granularity — a four-line card versus a detailed structured form. More structure catches more silent gaps but invites checklist fatigue and rote filling.
- Mandatory differ-prompt — whether "where does this break?" is required or optional. Required is the whole point, but forcing it on routine cases breeds resentment and box-ticking.
- Confidence scale — coarse (high/medium/low) versus numeric. Coarse is honest about how soft the judgment is; numeric invites false precision.
- Trigger scope — every case versus only high-stakes or atypical ones. Narrow triggers preserve the tool's bite; universal use dulls it.
- Independent review — whether a second reader checks the filled checklist on high-stakes calls.
When it helps, and when it misleads¶
Its strength is that it surfaces availability bias — the vivid recent case that comes to mind is not always the nearest one — and it makes tacit expertise teachable and auditable at almost no cost, catching the silent near-miss where a confident expert would otherwise sail past the one detail that breaks the pattern.
Its failure mode is post-hoc rationalization: an expert who has already decided can fill the checklist to justify the gut call rather than test it, writing a break-point they have privately dismissed. Checklist fatigue erodes it further, turning the differ-prompt into a reflexive "none." And it cannot fix a genuinely wrong sense of similarity — it only exposes the reasoning, it does not recompute it. The guarding discipline is to keep the differ-and-confidence field mandatory and specific ("name the single feature most likely to break this match"), to reserve the tool for cases where recognition is most dangerous, and — on high stakes — to have a second reader check the filled comparison, so the audit is not graded only by the person being audited.
How it implements the components¶
query_case_profile— the structured prompts that elicit the present case's decisive features before any prior case is named.abstraction_bypass_note— records that the expert is reasoning directly from a specific recalled case rather than an abstract rule, and makes that reliance explicit.confidence_and_coverage_label— the forced statement of confidence and of exactly where the recalled match breaks down.
It does not compute a similarity_metric (Similarity Search over Case Embeddings) or grade the match into a case_quality_label on fixed weighted criteria — that's Case Similarity Rubric; the checklist elicits and audits the expert's own recall rather than scoring an assigned candidate. It also does not adaptation_delta_note-adapt a stored solution (Case-Based Reasoning System) or write outcome_feedback_update_record (Exemplar Feedback Registry).
Related¶
- Instantiates: Nearest-Exemplar Response Reuse — this is the archetype's human-judgment variant, forcing tacit recognition into an auditable explanation.
- Sibling mechanisms: Case-Based Reasoning System · K-Nearest-Neighbor Case Matcher · Similarity Search over Case Embeddings · Precedent Matching Workflow · Incident Playbook Lookup · Case Similarity Rubric · Exemplar Feedback Registry
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Expert Case Recall Checklist operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it forces an expert's tacit 'this reminds me of a case' into an explicit, auditable comparison — which case, why it's close, and where the resemblance breaks.
Independent corroboration: The frozen evidence defines Expert Case Recall Checklist as 'Forces an expert's tacit 'this reminds me of a case' into an explicit, auditable comparison — which case, why it's close, and where the resemblance breaks', so its operative form is Assessment, Review & Assurance.
Nearest alternative: Interface, Display & Cue — The checklist evaluates the expert's recalled comparator and where resemblance breaks, while prompts are its point-of-use surface.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Cognitive Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Externalizing analogical case retrieval and testing where resemblance breaks follows expertise and case-based reasoning research.
Related originating lineages:
- Law & Governance — Precedent reasoning materially shaped the requirement to state both similarities and distinguishing breaks.
- Organizational & Management Science — Knowledge-management practice materially contributes auditable reuse of practitioner precedent. Knowledge-management practice materially shaped the checklist that externalizes tacit case memory.
Review resolution: Both reviewers agree that cognitive_science is primary. I retain organizational_management, law_governance only as formative origin lineages; cross_disciplinary_synthesis is appropriate because the final form materially combines the agreed primary with the retained formative lineages. Reach is multi_domain because the structure transfers across several fields but is not a near-universal human pattern, an applicability judgment kept separate from provenance. Encyclopedia synthesis is true because the exact generalized artifact is an encyclopedia-authored combination or refinement. No unresolved historical ambiguity remains after reconciling the secondary fields.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
References¶
[1] Klein, G. Sources of Power: How People Make Decisions. MIT Press (1998). Supports expert recognition as a fast and powerful decision resource grounded in prior experience. registry ↩