How This Atlas Grades — Three States, and Instrument is Not Interpretation
A blunt “real vs fake” split would be dishonest. Two distinctions govern every page, and three things are graded separately: plausibility, evidence grade, and tier.
Three epistemic states
Evidence-supported (for specific indications) — a real mechanism with regulatory clearance and grade A–B evidence for at least one indication. The tier attaches to the claim, not the device: off-label extensions of a cleared device sit lower, shown as a dual badge.
Insufficient evidence / under study — the frontier. The mechanism is coherent or at least does not violate physics; the evidence is early or thin; there is no robust refutation. “No evidence yet” is not “evidence of no effect.”
Not supported by evidence-based science — an umbrella that always resolves to one of two sub-states below. Never a flat label.
Instrument is not interpretation
Even where a claim fails, the instrument and the raw measurement are usually real. What fails is the interpretation layered on top. So the strong-rejection tier carries two genuinely different sub-badges:
Tested — not confirmed: the instrument and data are real; the diagnostic or therapeutic interpretation failed testing; a correct reinterpretation remains open. (EAV/Voll, Vega, GDV/Kirlian.)
Contradicts established science: the claim itself is physically incoherent or contradicted by replication — though the hardware may be ordinary electronics. (Rife, scalar, Clark zapper, and others.)
Why two and not one: the tier fails on two different axes — clinical (tested, not confirmed) versus physical (contradicts known science). A single soft phrase would erase the boundary and over-soften a $3 “terahertz” stone into “awaiting a trial.”
The A–F evidence scale
A — multiple positive RCTs / systematic reviews.
B — some controlled trials, mixed.
C — small / low-quality or preclinical only.
D — case reports / announcement-stage.
E — no credible clinical evidence.
F — tested-and-refuted / contradicted.
The grade is consensus on evidence. It is held distinct from physical plausibility (a separate axis) and from the tier — which is why a device can be physically plausible yet grade low, or carry a real substrate yet sit in the strong-rejection tier.
The falsifiable test a claim must pass
If a vendor wants scientific credibility, these are the experiments — and most frequency-medicine marketing avoids exactly this: quantitative dosimetry.
Diagnostic claims (“device X detects condition Y”): a preregistered triple-blind design, a gold-standard comparator, standardized conditions, a sham-device control, independent operators, and a predeclared endpoint — accuracy, sensitivity/specificity, and (for screening) net benefit versus overdiagnosis harm.
Therapeutic claims: first characterise the emitted field under tissue-equivalent loading; show the signal reaching the target is non-trivial and reproducible; then test blinded in culture → animal → human against matched sham plus positive controls. If the signal at target is negligible, or outcomes are no better than sham, the claim fails.