Knowledge Representation Levels by Functional Tier Matrix
Description
One guideline about LDL cholesterol, expressed twelve ways. Read a column downward and you watch one kind of knowledge harden from prose into something an engine can run; read a row across and you see the same level of formality applied to three different jobs. The cell worth pausing on is Structured x Data, where the concept is finally bound to LOINC 2089-1 — that binding is what lets two systems agree on what counts as an LDL result. The Forms/UI column is deliberately muted and badged: it completes the grid so the classification task is well posed, and it is taught in the two chapters that follow. Hovering a row or column header highlights that band and shows the definition of the level or tier.
Lesson Plan
Audience: undergraduate students and healthcare data practitioners. Allow 10-15 minutes. Prerequisites: FHIR resources and value sets, and the CQL-to-ELM compilation step.
Learning objective: Given a description of a clinical knowledge artifact, the learner can classify it by both its knowledge representation level (Narrative, Semi-Structured, Structured, Executable) and its functional tier (Data, Logic, Forms/UI) using the matrix.
- Explore: Hover each of the four row headers in turn and read the level definitions. Then read the Logic Tier column from top to bottom and identify the single step where the artifact stops being interpretable by a person and starts being executable by a machine.
- Explain: Compare Narrative x Logic with Executable x Logic. Explain what is identical about them and what changed, and why nothing in between altered the medicine.
- Transfer: Classify three artifacts of your own — a printed order set, a validated FHIR Questionnaire, and a spreadsheet of screening intervals — by both level and tier. For each, name the feature that fixed the level and the feature that fixed the tier.
Assessment: use the Transfer prompt as an exit ticket. A complete response assigns both coordinates for each artifact and cites a distinct feature for each coordinate rather than one impression covering both.
Embed This MicroSim
Copy this iframe to your website:
1 | |
Quality Checks
Validated September 8, 2026. Completeness rubric: 100/100, scored with the project's calculate-quality-score.py against the standardization checklist. Checked automatically in a browser at 400, 800, and 1200 px: the page loads with no JavaScript errors, no interactive control is clipped outside the viewport, and the document does not scroll horizontally. Controls, selections, and the reset path were exercised by hand during development at desktop width. The page reports its own height to the parent document, so the declared iframe height above serves only as a fallback.
Specification
The full specification below is extracted from Chapter 22: FHIR Resources and Levels of Knowledge Representation.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 | |
Related Resources
References
- Source chapter — supplied the learning objective and the worked example.
- p5.js reference — canvas, drawing, and input handling (accessed September 8, 2026).
- p5.js web editor — paste the JavaScript source to experiment; no hosted sketch has been published.
- HL7 Clinical Guidelines (CPG-on-FHIR) implementation guide — the level-and-tier framing this grid follows (accessed September 8, 2026).