Sources · provenance · limitations
Our evidence approach
An interpretation should be inspectable: what question it addresses, which sources it uses and where its limits begin.
Preliminary and unreviewed.
This is an intended approach, not an approved clinical methodology. No named clinical reviewer or reviewed public dossiers are established by this preview.
It contains no clinical evidence, treatment recommendations or patient-specific advice. Clinical workspace in development.
01 / Sources
Start with a question and a defined scope.
We intend to identify the question, study design, population, intervention and outcomes relevant to each interpretation. Source records would identify the title, authors, publication year and persistent identifier where available, and disclose whether the basis is an abstract or full text.
Human findings would be separated from animal, laboratory and other preclinical work. Preclinical observations do not establish clinical effects in people. A limited search would be described as limited; missing information would remain explicit.
02 / Interpretation
Put limitations beside the claim.
The planned presentation would show study limitations, conflicting findings, gaps and applicability alongside the interpretation they qualify. A successfully opened citation or a source identity check does not establish claim support.
This preview assigns no clinical certainty grades or evidence scores. Existing synthetic demonstrations test software mechanics, not clinical outcomes. Their draft safety rules are unvalidated and for synthetic use only; they are not a basis for patient care.
03 / Provenance
Preserve the exact source and its rights.
The intended provenance record would retain the specific source revision used, its capture date and a locator for any permitted supporting passage. A later source version should not silently replace the basis of a historical interpretation.
Access to a source does not establish permission to reproduce it. Unknown rights do not permit excerpts. If a passage cannot be located or displayed under its rights, that limitation should be stated rather than filled in. No real source passages or dossiers are published in this preview.
04 / Review
Keep review status explicit.
Source identity, claim support and clinical review are different questions. We intend to record their status separately. An identity check would not label a claim clinically reviewed, and an unchecked draft would remain unreviewed.
No clinical reviewer or staffed review process is represented as available here. The draft policy pages describe plans and limitations; they do not establish completed review or clinical readiness.
05 / Change history
Use manual, dated releases and corrections.
The intended process uses manual, dated evidence releases with a record of material changes and corrections. Source last checked, evidence last reviewed and page materially updated would be separate dates. Checking that a source is reachable would not renew a clinical review date.
No automatic refresh schedule or response-time commitment is offered. The draft corrections policy explains the proposed correction path and current limits; it does not promise a staffed corrections service. This preview is not a dated clinical evidence release.
06 / Today
Read the approach with its limits.
No clinical service, public evidence library or paid plan is offered here. No sign-ups or email collection are available. Do not submit patient information.
The draft methodology and draft editorial policy provide further context on the intended scope. They remain preparation drafts.