MED01-P1 — Correction-round prompts (2026-09-24)

Paste each into the same chat the researcher used for its pass (the pass is frozen; the correction is filed beside it as …-correction-1.md). Corrections inherit full verification (base Rule 9): every returned identifier is re-resolved before anything changes in the ledger. Neither prompt reveals the other researcher’s findings.

To ChatGPT

Correction round for your MED01-P1 pass. Your original submission is frozen as received; reply as a separate, numbered correction note. Do not revise or restate the rest of the pass. Verification confirmed 38 of your 39 references exact and every quantitative clinical claim I could check at abstract level. Two items need resolution, plus one request:

1. PMID 35817852 — cited in Q1 for "a later period review indicates dozens of cannabinoids were already known by 1980." That PMID resolves to Busquets-García et al., Nature Metabolism 2022, "Metabolic Messengers: endocannabinoids." State whether you intended a different source; if so give it with a resolvable identifier, or withdraw the citation. Do not complete an identifier from memory.

2. Ng & Chang 2022 — you gave the co-author as "Chang YK"; Crossref lists Nathan Chang. Confirm, and state whether the decade figures you quoted (US 33 records in the 1960s; 979 in 1970–79; 29,802 total 1829–2021) were read from the full text or from the abstract/PMC landing page.

3. Your PubMed count query, re-run by the verifier on 2026-09-23 through NCBI ESearch, returns 8,663 records for 1960–1999 (252 / 2,492 / 2,153 / 3,766 by decade). No action needed; this is for your record and will be ledgered as verifier-found.

Format: item number · resolution · identifier(s) · "verified live" or "unverified". Nothing else.

To Gemini

Correction round for your MED01-P1 pass. Your original submission is frozen as received; reply as a separate, numbered correction note. Do not revise or restate the rest of the pass. Verification confirmed the substance of most of your citations, but several identifiers, page numbers, and attributed magnitudes did not resolve to the sources you cited. For each item below: give the corrected detail with a resolvable identifier (DOI or PMID), or state that you cannot verify it and withdraw it. Do NOT complete an identifier from memory; a wrong identifier is worse than none. Do not consult or cite the other researcher.

IDENTIFIERS
1. Beal 1995 — your DOI 10.1016/0885-3924(94)00111-7 returns 404 at Crossref. Provide the correct DOI or PMID.
2. Tashkin 1973 — your DOI 10.1056/NEJM197308162890701 resolves to a different NEJM paper (Gordis et al., rheumatic fever), and "289:333–336" is not the Tashkin paper's location. Provide the correct identifier and pages.
3. Gill, Paton & Pertwee 1970, Nature 228 — you gave pages 1193–1194. PubMed indexes the Gill/Paton Nature 1970 paper at 228:134–136 ("Preliminary experiments on the chemistry and pharmacology of cannabis"). Confirm which paper you meant and whether it reports THCV isolation.
4. "Mechoulam et al. 1967, JACS 89:4552 — total synthesis." The indexed 1967 paper is a stereospecific synthesis of (−)-Δ1 and (−)-Δ1(6)-THC. State which paper you regard as the first total synthesis of Δ9-THC and its year, with identifier.
5. Tashkin 1999, Am J Respir Crit Care Med 160(5):1658–1666 — no PubMed record found at that location. Provide PMID or withdraw.
6. Hall, Room & Bondy 1999, Addiction (the excised WHO comparative chapter) — not found in PubMed by author/year/journal. Provide DOI/PMID or withdraw.
7. Claussen, von Spulak & Korte 1966, Tetrahedron 22:1477 (CBC) — not indexed in PubMed; provide DOI if one exists, otherwise mark "unverified, older chemistry."

ATTRIBUTED MAGNITUDES — each must come with the source that contains the number, or be withdrawn
8. Hepler & Frank 1971: "25–30% decrease in intraocular pressure" and "11 healthy volunteers." The cited JAMA item is a one-page letter with no abstract in PubMed. Which document contains these figures?
9. ">75% of federal research budgets" allocated to harms (P3). Source with page.
10. "over 90% accuracy" with which patients/assessors identified the active arm (§E). Source.
11. Washout periods "as short as 24 to 48 hours" (§E). Source.
12. Oral THC 15–20 mg "roughly equivalent to 60–120 mg oral codeine" (§C5, Tier 1). The Noyes 1975 abstracts describe the effect as "mild" and do not give an equivalence. Source, or withdraw.
13. Compassionate IND: 1978 start; June 1991 halt; March 1992 closure; 27 approved; 13–15 grandfathered (§F2, CAND-11). You cited a press release, a memoir, and IOM 1999 p. 44. Give the primary document (PHS/FDA record, Federal Register, or the exact IOM page text) that supports each number.
14. Young ruling 6 Sept 1988 and 54 FR 53767 (29 Dec 1989): confirm you opened the Federal Register page, and quote the sentence that states the Administrator's rejection.
15. Nabilone approvals Canada 1982 / UK 1983: source document for each.

METHOD
16. CAND-24 cites "EMBASE scans" — EMBASE does not appear in your search log. State whether EMBASE was actually searched; if not, restate CAND-24's basis.
17. Your PubMed query, re-run verbatim by the verifier on 2026-09-23 via NCBI ESearch, returns 10,630 (414 / 3,679 / 2,516 / 4,021 by decade), not 11,241 (342 / 4,118 / 2,789 / 3,992). State how your figures were obtained (interface, date, any filters), or accept the verifier's count.
18. Chang 1979 (§D3, Q10): the abstract reports oral and smoked THC within one randomized trial but gives no route-versus-route contrast. State whether the full text reports one; if not, the "smoked vs oral" framing will be reworded.

Format: item number · resolution · identifier(s) · "verified live" or "unverified". Nothing else.

Source file in the research repository: medical/MED01/correction-prompts.md. Private links, some names and practitioner-record passages are removed from the public copy, and broken character encoding from file conversion is repaired; nothing else is changed.