DDX // HYPE INDEX V1.3 // METHODOLOGY
DiligenceDx
The HYPE Index/Methodology v1.3
Healthcare AI verification

Separating signal from narrative in healthcare innovation.

Diligence Dx is the applied layer of the HYPE Index, a stage-aware, gap-centered methodology for scoring healthcare AI claims across companies, products, consumer influencers, and research papers.

Healthcare innovation moves faster than evidence. Funding narratives, media velocity, and category enthusiasm often run ahead of clinical validation, regulatory clearance, and real-world performance. The signal is the gap, not the evidence. Diligence Dx applies a structured framework to ask a different question of every claim: what is verified, what is still narrative, and where is the gap?

Built for physicians, investors, journalists, researchers, and the patients and consumers on the receiving end of healthcare AI claims.

The Chain of Hype

Consider how a single biological insight travels through the public information environment. A laboratory publishes a calibrated paper showing that a biomarker is associated with a disease outcome in a specific cohort. A company writes a marketing page that cites the paper, dropping the qualifier. A founder, in a podcast interview, says the biomarker predicts the disease. A clip is shared. The viral caption reads: scientists discovered the marker that determines whether you will get this disease.

The signal does not survive the chain.

By the time the claim reaches a consumer, the cumulative gap is large. The HYPE Index scores each link on the same rubric, which is what makes the propagation of miscalibration measurable.

The four diagnostic axes Companies · Products · Influencers · Research papers
H Hypothesis & History Did this start as a real question? Does it have a paper trail?
Y "Yes, but" & Yields What does the entity say works? What real-world yields can it demonstrate?
P Peer Review & Proof of Work Have qualified outsiders validated this? Has the work been shown?
E Expectation & Evidence The headline axis. The Calibration Gap between public claim and supporting evidence.
How the Index Works
  • Stage-aware: a frontier laboratory and a commercial company are held to different evidentiary expectations.
  • Four entity types on a single rubric: companies, products, consumer-facing influencers, research papers.
  • Sixteen criteria across four axes, each scored on a five-point anchored scale.
  • Two numbers reported: the E score (the headline gap) and the four-axis average (the diagnostic context).
  • Calibration Gap threshold calibrated against retrospective backtest, not stipulated.
Stress-Test Panel

The methodology has been hardened by a twelve-persona stress-test panel. The panel includes a working clinician asking whether the rubric can be applied between admissions, a psychometrician asking whether anchored descriptors hold their measurement-theoretic shape, and an investigative journalist asking whether scored outputs survive in published narrative. It also includes a senior peer reviewer testing whether the research-paper scoring guide captures the gestalt judgment experienced reviewers actually use, a patient advocate asking whether the rubric makes visible the calibration failures that matter to patients on the receiving end of healthcare AI claims, and a regulatory affairs director checking whether the rubric uses language consistent with how regulators evaluate the same calibration questions. Six other personas round out the panel. Each holds a distinct diagnostic question. Full panel synthesis is documented in the v1.3 methodology paper.

Diligence Dx is the editorial layer. The HYPE Index is the scoring layer. Methodology, scoring rubric, and editorial governance are published in full so readers can verify the framework against their own judgment.

What this is not

The HYPE Index is not investment advice. It is not a regulatory instrument. It does not authorize, approve, or clear any entity. The methodology paper defines how to score; the application papers define what was scored.

In development.  Methodology paper v1.3 publishing Summer 2026. · First category evaluations to follow.

The v1.3 methodology paper is available to qualified reviewers on request.

Request the methodology

Send a note below to request the v1.3 methodology paper, ask a question, or help shape the instrument.