# Data audit (P1, 2026-10-30)
| file | SHA-256 | wc -l | notes |
|---|---|---|---|
| data/assertions.json | 2965ab3acc6f326653169b83a5345698b51fac88a556de65faf89210aea0cad4 | (JSON) 6 top-level items | ids A1–A6, unique; A1–A5 cite arxiv-2609.27732v1, A6 cites arxiv-2606.14621v1; no missing fields |
| data/README.md | 5348bd49912099735108dbbe19ae09fa72bd600cb3d12b878658e6b9806198a7 | 14 | provenance; listed extract hashes |
| data/sources/arxiv-2609.27732v1.txt | ee2d444d2f78f2af2ca2cf69b59a61fcd7ac5e10ea29e8ddcd67917bf07d6a8b | 134 | matches README hash |
| data/sources/arxiv-2606.14621v1.txt | 27c463c0a144d9d5a3916dc641f15b6029f6c68b5e574ad9074e7df1aeb0a238 | 58 | matches README hash |

Schema: DCR extract — Table 6 (ATT, SE, % change by 1/3/5 km), Table 7 (ATT, SE by outcome × buffer), Section 6.1, Table A19 (5 km by estimator). Units: log points of hospitalisation rate per 100,000; SE clustered by MedStat region. RESEA extract — Table 2 (group means, T–C diff, ATE with p and 95% CI), text, Table 3. Units: weeks, $, pp.

Anomalies / internal inconsistencies:
1. DCR Table A19 CS-DR column prints ATT -0.431** with SE (0.218), whereas Table 6 5 km (same estimator/spec) prints SE (0.210).
2. DCR Table A19 note says 640 never-treated comparison regions; the table body and Tables 6/7 say 608.
3. DCR abstract says fatal outcomes "fall by 35%"; Table 7 5 km fatal ATT -0.354 implies 100(exp(-0.354)-1) = -29.8%, matching Section 6.1's "approximately 30%".
4. DCR Figure 7 note says "four" sub-categories while the text and Table 7 have five; Table 7 note gives ICD codes for four only. Not decision-relevant.
5. RESEA text prints CIs for reductions as positive magnitudes ([1.75, 2.24]) while Table 2 prints signed values ([-2.24, -1.75]); equivalent. Rendering defect "p ¡ 0.001" for "p < 0.001".
6. Neither extract reports any sex-specific estimate.
No assertion cites a missing source; no availability-rule violation in the inputs. No rows excluded.