Source Hierarchy
Every finding in the dataset must be traceable to one of the following source types: PubMed-indexed peer-reviewed publications, NIOSH Health Hazard Evaluations, OSHA inspection records, EPA Superfund or CERCLA site assessments, NARA archival records, or court dockets from federal or state litigation.
Law firm marketing sites, plaintiff attorney websites, and unverified secondary sources are excluded from the source list.
Verification Protocol
Each evidence record undergoes independent verification: the cited source must be accessible at the provided URL or identifiable via the stated accession number. Citation URLs are checked for liveness. DOI and PubMed links are validated against their respective registries.
Research Refresh — September 4, 2026
This dated context update adds authoritative national, veteran, and worker-protection sources without changing any yard-level finding, evidence tier, product record, or the cited v1.5 dataset.
Edwards et al. (2026) used Global Burden of Disease estimates and reported declining U.S. age-standardized mesothelioma incidence and mortality from 1990 to 2023 while absolute cases and deaths increased; the authors estimated occupational asbestos exposure accounted for 95.7% of mesothelioma deaths in 2023. These are modeled population-level estimates, not evidence of exposure at a named shipyard or of individual causation.
VA Public Health lists shipyard work among occupations where Veterans may have been exposed to asbestos and notes 20-to-50-year latency for asbestos-related disease symptoms. OSHA identifies ship repair as a setting where heavy exposure can occur, particularly during removal in renovation, repair, or demolition. Neither overview establishes exposure at a particular yard, workplace, or for a particular person.
Cross-Yard Evidence
Mumma et al. (2019, PMID 31290725): 253,632 U.S. industrial radiographers (1969–2011), 30,724 at shipyards. Shipyard mesothelioma SMR 9.97 (8.50–11.63), asbestosis SMR 18.1. Boice et al. (2026, PMID 41633573): cohort update through 2019, 123,401 radiographers, whole-cohort mesothelioma SMR 6.08 (5.35–6.89), asbestosis SMR 13.4 (11.2–15.9). These cross-yard studies cannot be attributed to any single yard under the no-single-yard-attribution rule.
Regulatory Watch
OSHA Asbestos Standards (RIN 1218-AD58; Docket OSHA-2025-0024) includes 29 CFR Part 1915, Shipyard Employment. Its proposal concerns respirator requirements, not asbestos permissible exposure limits, excursion limits, or monitoring triggers. The limited reopening for ACCSH materials closed August 21, 2026.
EPA Asbestos Part 2 (RIN 2070-AL21) lists statutory deadlines of December 3, 2025 for a proposal and December 3, 2026 for a final rule, while the current Unified Agenda projects February 2027 and March 2028. A projected timetable is not a waiver or a final rule.
Last updated: August 29, 2026. Regulatory items are informational context only and do not constitute legal advice.
OSHA Establishment-Search Reconciliation — September 4, 2026
The v1.6 successor release adds selected official OSHA establishment-inspection and Chemical Exposure Health Data entries only when a result names a canonical yard directly or an official address establishes a specific relationship. The selected entries are administrative records, not a representative census of conditions at a yard or a finding of individual exposure, dose, causation, product source, or liability.
Entity boundaries remain adjacent to the selected evidence: the Brooklyn entry is a G. Marine Diesel tenant record at Building 386, the Newport entry is the named Derecktor establishment at Coddington Cove, and the Todd result is bounded to the San Francisco establishment rather than Alameda. Mare Island and Tampa candidates were excluded for site/entity mismatch. Bulk samples are identified as bulk samples and are not converted into person-level exposure conclusions. The Bath Iron Works entry corrects a prior unsupported exit-route citation by relying instead on the 1987 inspection and linked official sample-search rows.
Current Version
Version 1.6 (2026-09-04) — 37 shipyards, 118 evidence findings, 87 documented products. DOI: 10.5281/zenodo.22313866. v1.5 remains preserved as a historic Zenodo version.