How to read this source
Start with headings and summary blocks. Use this page to verify citations, claim wording, and chapter-level evidence support before interpreting conclusions.
Internal Research Source
Claim Evidence Master
Structured claim-to-evidence matrix defining claim text, study expectations, thresholds, and pass/fail readiness gates.
internal-csvEvidence status: active sourceUpdated 2026-02-15
Used by chapters
Share
In plain language: this template is a structured evidence table. Use it to map claims, thresholds, risks, and pass/fail criteria consistently.
| claim | risk | reference_standard | dataset_n | study_type | metrics | pass_threshold | abstain_limit | subgroup_gate | labeling_constraints | owner | status |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Early anemia-risk signal from ocular capture | False reassurance | IMDRF N41; AstraCBC §19 | 0 | Prospective paired-label study | Sensitivity; Specificity; NPV; calibration slope | Sensitivity >= 0.90 at predefined threshold | <= 25% overall; <= 35% any subgroup | No subgroup sensitivity drop > 0.07 | Must include confirmatory testing language | Clinical Validation | Draft |
| Hyperbilirubinemia-risk signal from sclera | Missed severe cases | IMDRF N41; ISO 14971 | 0 | Prospective multicenter | AUC; Sensitivity at safety cutpoint; PPV | AUC >= 0.82 and sensitivity >= 0.92 at cutpoint | <= 30% | No subgroup AUC drop > 0.08 | Not for standalone treatment decisions | Clinical Validation | Draft |
| Pupillary-neuro risk signal | False alarm burden | IEC 62366-1; ISO 14971 | 0 | Reader-blinded comparator | Agreement; ICC; classification F1 | ICC >= 0.85 and F1 >= 0.80 | <= 20% in compliant capture mode | No subgroup ICC drop > 0.10 | Output must include urgency and escalation path | ML Lead | Draft |