Documentation Gap Analysis: Independent Audit of TGA COVID-19 Vaccine Safety Monitoring Plan
Description
Research hypothesis: The TGA cannot produce verifiable documentation demonstrating systematic implementation of its COVID-19 Vaccine Safety Monitoring Plan (February 2021), which specified 17 enhanced-pharmacovigilance outputs across five objectives defining "enhanced monitoring" required for provisional approval, plus two additional governance outputs, totalling 19 assessed outputs. Data: Documentary evidence gathered 2022-2026 via FOI requests, OAIC decisions, Senate testimony, public TGA reports, and peer-reviewed literature. Assessed against ISO 19011:2018 and ANAO standards using the Plan's 19 outputs as assessment criteria. Findings: 3 of 19 outputs (16%) fully documented. 16 (84%) partially or not documented. Signal detection achieved 0% full implementation across 8 outputs; governance achieved 0% documentation. OAIC Decision [2025] AICmr 54 confirmed requested records "cannot be found or do not exist." Senate testimony (9 October 2025) confirmed Plan "never systematically tracked" and monitoring occurred through "day-to-day processes." No documentation verifies enhanced monitoring conditions were met before 68.4 million provisional doses transitioned to full registration. Interpretation: Absence of proof that enhanced monitoring was systematically implemented, not proof that no monitoring occurred. Findings establish evidence base for formal audit by oversight bodies, parliamentary scrutiny, FOI appeals, and further research. The classification criteria are documented within the dataset, and the codebook defines the variables, classification codes, evidence tiers, and rating scales used in the evidence matrix underlying the TGA COVID-19 Vaccine Safety Monitoring Documentation Dataset. Reproducibility: The dataset is structured to support direct verification of headline findings. All summary statistics — status distribution, status-by-objective cross-tabulation, severity aggregation — are reproducible from the underlying matrix using standard analytical tools (spreadsheet software, R, Python). Companion data paper (SSRN): doi.org/10.2139/ssrn.6610438 Zenodo: doi.org/10.5281/zenodo.20175557 The complete audit repository with all primary source documents, version history, and detailed methodology is available on GitHub: https://github.com/paulrekaris/TGA-COVID19-Vaccine-Safety-Monitoring-Audit
Files
Steps to reproduce
Audit Methodology: Reproducible Framework for Regulatory Accountability This methodology outlines a systematic process for an independent documentation audit of TGA's COVID-19 Vaccine Safety Monitoring Plan. It is transparent, falsifiable, and adaptable for scrutinising other government commitments. 1. Guiding Standards ISO 19011:2018, ANAO, and GAO Yellow Book for evidence-based assessment and performance evaluation. ICH E2E, WHO/EMA GVP Module IX, and CIOMS VIII for pharmacovigilance. ISO 15489-1:2016 for records management. PGPA Act 2013 for Commonwealth accountability. OSINT principles for systematic collection. Open science via CC BY 4.0, Git, blockchain timestamping, and FAIR data principles. 2. Core Steps Step 1: Identify Target Framework. Select a public plan with clearly defined outputs. This audit examined the TGA's February 2021 Safety Plan — 19 outputs (17 Plan strategies plus 2 governance outputs) Step 2: Establish Benchmark. Use the plan's stated outputs as assessment criteria in a rating matrix. Criteria must derive from the plan's own language Step 3: Systematic Evidence Collection. Submit targeted FOI requests and log all responses. Search agency websites, parliamentary testimony, and independent databases using 14 strategic terms derived from plan language. Step 4: Apply Assessment Protocol Evidence Hierarchy (strongest to weakest): Tier 1: OAIC decisions, FOI decisions and direct agency communications Tier 2: Senate testimony, Questions on Notice, and agency reports Tier 3: Partner-organisation and publicly accessible surveillance data Tier 4: International guidance and contextual material Rating Criteria: Fully Implemented/Partially Implemented/Not Documented (assigned on the basis of available documentary evidence consistent with ISO 19011:2018 clause 6.4.8). Cross-reference findings across independent sources. Where agency searches produced zero results, corroborate against FOI releases, Senate testimony, and OAIC decisions. Step 5: Declare Limitations. Document all constraints explicitly, including statutory powers not held. Step 6: Falsifiability. Publish all findings, primary sources, and version history in a version-controlled repository. State explicitly what evidence would close each finding. Step 7: Permanent Archiving. Anchor documents to public blockchain. Archive with national library. Assign persistent DOIs. Apply LOCKSS preservation principle. 3. Reproducible Analytical Operations For reproducible analytical operations on the matrix dataset — including status distribution, status-by-objective cross-tabulation, and severity rating aggregation — see the data paper at SSRN: doi.org/10.2139/ssrn.6610438 and Zenodo: doi.org/10.5281/zenodo.20175557 and dataset codebook. The dataset is compatible with spreadsheet software, Python, R, and Stata. Audit governance: https://github.com/paulrekaris/TGA-COVID19-Vaccine-Safety-Monitoring-Audit/tree/main/audit-governance