About Swanology

• Reference ranges are transcribed from the chapters and papers in the project library (Kern in Braunwald, Naderi in Practical Cardiology, Bootsma, Foucha, Nishimura & Carabello) and from the ESC/ERS 2022 pulmonary hypertension guideline. Where authors disagree, every value is kept and the disagreement is shown rather than averaged away.
• Rules are a declarative table (54 rules, version 2026.10-r1-draft). Each has a condition over the derived metrics, teaching text, a tone, the population the threshold came from, and citations. Bands tile each metric’s axis, and each band edge says which side owns the exact edge value; a test enforces that the bands leave no gaps. The criteria shown on every finding are generated from the same condition the engine evaluates. All rules are drafts pending review by a board-certified cardiologist.
• Threshold kinds. Every rule and every reference-range entry is tagged as a reference range, a diagnostic definition, a prognostic cutoff, a population-specific finding, a teaching framework, or a descriptive pattern. The kinds are kept apart in wording and layout: for example, a cardiac power output of 0.6 W is a prognostic cutoff from cardiogenic shock, not a normal range, and the two different wedge lines (15 for pulmonary hypertension classification, 18 for congestion) are not interchangeable. The Sources & rules page shows the kind of every threshold and has a list of what was and was not verified against the literature.
• Randomized cases are 12 physiologic archetypes with ranges for every input. The generator samples numbers from a seeded random stream, derives BSA-consistent output and a Fick-consistent mixed venous saturation, and then verifies with the same interpretation engine that the archetype’s defining findings fire. Because the engine both interprets and grades these cases, they test internal consistency, not clinical truth: if a rule is wrong, the answer key is wrong in the same way. A case link reproduces the same case.
• Independently authored reference cases are written separately from the engine so that its output can be checked against an expected reading. They are awaiting clinician review and should not be treated as validated until that is done.
• Waveforms come from a parametric model of one cardiac cycle (timed a, c and v waves; x and y descents; respiratory and rhythm effects). They are model-generated teaching tracings, not patient recordings. Real tracings will be shown only when they are properly licensed and deidentified.

Read every range, rule, and citation on the Sources & rules page →

1. v1 · 2024An R/Shiny application with a hemodynamic calculator, a rule-based interpretation, and the first CVP waveform exercises.
2. v2 · 2026Rebuilt in TypeScript as a static web app: a versioned, citable rule table, live parametric waveforms, randomized cases, and no server.