{"schema_version":1,"research_id":"eoa_inverse_innovation_exp03_external48_20260801","source_assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","cell_id":"layer_decay_and_expiration_management__statistics_experimental_design","selection_stratum":"DEPLOYABLE_PRIORITY","search_queries":["site:fda.gov clinical trial computerized systems version control audit trail source data guidance records retention","site:ema.europa.eu ICH E6 R3 essential records version history trial master file superseded document","clinical trials wrong version protocol statistical analysis plan document version errors study","Veeva Vault eTMF document lifecycle superseded archived retention legal hold restore","site:ich.org E6(R3) final guideline essential records version control superseded 2025 PDF","site:gov.uk MHRA trial master file version control superseded documents guidance","empirical study clinical trial protocol version control wrong version amendments implementation errors","site:platform.veevavault.help document lifecycle superseded current version search Veeva Vault","site:platform.veevavault.help retention policy document retention rules legal hold deletion restore Vault","empirical study obsolete code version wrong dataset clinical research reproducibility artifacts version control"],"sources":[{"source_id":"S1","title":"ICH E6(R3) Guideline for Good Clinical Practice","publisher":"International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use","url":"https://database.ich.org/sites/default/files/ICH_E6%28R3%29_Step4_FinalGuideline_2025_0106.pdf","source_class":"STANDARD","publication_date":"2025-01-06","accessed_at":"2026-08-02","claims_supported":["Essential records should be identifiable and version controlled when appropriate.","Trial-record storage and archive systems should support identification, version history, search, and retrieval.","Sponsors and investigators/institutions share responsibility for access, management, location tracking, retention, and availability of essential records."]},{"source_id":"S2","title":"Archiving and retention of clinical trial records","publisher":"UK Medicines and Healthcare products Regulatory Agency","url":"https://www.gov.uk/government/publications/clinical-trials-for-medicines-archiving-and-retention-of-clinical-trial-records/archiving-and-retention-of-clinical-trial-records","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026-04-28","accessed_at":"2026-08-02","claims_supported":["UK trial essential documents must remain complete, legible, readily available, and retained for at least 25 years under the stated rules.","Sponsors must appoint a named person responsible for archiving and restrict archive access appropriately.","Electronic archives should preserve audit trails and metadata, support long-term readability, and document validated migrations.","Documents must not be destroyed before applicable retention periods expire."]},{"source_id":"S3","title":"Frequency of multiple changes to prespecified primary outcomes of clinical trials completed between 2009 and 2017 in German university medical centers: A meta-research study","publisher":"PLOS Medicine","url":"https://doi.org/10.1371/journal.pmed.1004306","source_class":"PRIMARY_RESEARCH","publication_date":"2023-10-31","accessed_at":"2026-08-02","claims_supported":["The study examined complete registry-version histories for 1,746 randomized trials and a random publication-comparison subsample of 292 trials.","Fifty-five percent of the subsample had a primary-outcome change at some time, and 23% had a major change.","For 14% of sampled trials, changes occurred only within registry history and would not be detected by comparing the publication with the latest registry entry.","The study supports preserving and examining historical specification versions, but did not measure analysts selecting obsolete local artifacts."]},{"source_id":"S4","title":"Guidance for Industry: Computerized Systems Used in Clinical Trials","publisher":"US Food and Drug Administration","url":"https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/fda-bioresearch-monitoring-information/guidance-industry-computerized-systems-used-clinical-trials","source_class":"OFFICIAL_GUIDANCE","publication_date":"1999-04","accessed_at":"2026-08-02","claims_supported":["Electronic-record audit trails should permit reconstruction of creation, modification, and deletion events.","The guidance addresses software version control, change control, retrieval, study reconstruction, and backup and recovery.","Computerized trial systems should be designed to preclude errors in data creation, modification, maintenance, archiving, retrieval, or transmission."]},{"source_id":"S5","title":"Document State Types & Special States","publisher":"Veeva Systems","url":"https://platform.veevavault.help/en/lr/14560/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2026-03-30","accessed_at":"2026-08-02","claims_supported":["Veeva Vault implements starting, steady, superseded, obsolete, and deleted document-state types.","A configured entry action can automatically move the previous steady-state version to superseded.","Steady-state documents are ordinarily protected from deletion, although a power-delete permission can override the rule.","The deleted state did not yet have special behavior in the documented release."]},{"source_id":"S6","title":"Archiving Documents","publisher":"Veeva Systems","url":"https://platform.veevavault.help/en/gr/34126/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2025-07-09","accessed_at":"2026-08-02","claims_supported":["Vault archive status excludes documents from standard search and ordinary library views while retaining an authorized archive search path.","Archiving is reversible through unarchive actions and blocks many document-modification operations.","Archive access and management are permission-controlled."]},{"source_id":"S7","title":"About Legal Holds","publisher":"Veeva Systems","url":"https://platform.veevavault.help/en/lr/36467/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2025-04-23","accessed_at":"2026-08-02","claims_supported":["Vault legal holds can be scoped and applied automatically or manually.","A hold prevents deletion while permitting controlled lifecycle activity and preserving audit history.","Legal-hold records themselves use an approval lifecycle and permissions."]},{"source_id":"S8","title":"Deleting Documents","publisher":"Veeva Systems","url":"https://platform.veevavault.help/en/lr/1529/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2025-12-17","accessed_at":"2026-08-02","claims_supported":["Vault retains an administrative deletion log and can generally restore deleted documents through support within approximately 30 days.","Deletion is blocked by conditions including steady state, approved legal hold, incoming relationships, binder membership, and object references.","Restoration is incomplete for some relationships, attachments, binders, and legal holds, showing that recovery is not equivalent to verified reconstruction."]}],"problem_evidence":{"support":"MODERATE","rationale":"Official standards and guidance establish that trial records require version control, traceable change, reliable retrieval, reconstruction, and long retention, while primary meta-research shows that historically material outcome-definition changes can be hidden when only the latest version is considered. This supports the preservation-versus-current-authority tension. However, no opened source quantified how often analysts retrieve or apply obsolete local protocols, SAPs, code, or derived datasets, so the candidate's specific operational prevalence remains unverified.","source_ids":["S1","S2","S3","S4"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"ICH and MHRA assign concrete record-management and retention responsibilities to sponsors, investigators/institutions, service providers, and a named archive custodian, establishing credible authorizers and compliance-driven interest. The bounded search did not find direct survey, procurement, complaint, or adoption evidence specifically requesting a cross-repository lifecycle-gating overlay.","source_ids":["S1","S2"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Veeva Vault regulated-document lifecycle and archive","similarity":"It already implements effective/steady, superseded, obsolete, and deleted states; automatic supersession; standard-search exclusion for archived documents; reversible unarchiving; permissions; legal holds; audit logs; relationship-based deletion restrictions; and limited post-deletion restoration. These cover most of the proposed mechanism.","remaining_difference":"The documentation does not establish a single cross-repository identity and dependency layer spanning protocols, SAPs, randomization specifications, source code, environments, and derived datasets, nor routine end-to-end reconstruction tests or controlled evidence that gating reduces wrong-version selection. Its documented restore process also fails to recover some relationships and holds.","source_ids":["S5","S6","S7","S8"]},{"name":"ICH E6(R3) essential-record management and regulated TMF practice","similarity":"The standard requires identification, appropriate version control, version history, search, retrieval, location records, traceable alteration, retention, and regulator access, making governed lifecycle management an established requirement rather than a new objective.","remaining_difference":"ICH specifies outcomes and responsibilities, not the candidate's proposed lifecycle vocabulary, cross-system dependency checks, simulated gated-discovery interface, or comparative retrieval experiment.","source_ids":["S1","S2"]},{"name":"Clinical-trial registry version histories","similarity":"Registry histories preserve earlier outcome specifications and permit reviewers to detect changes that are invisible when only the latest entry is considered.","remaining_difference":"Registry history preserves evidentiary versions but does not govern operational authority across local documents, code, datasets, and execution dependencies.","source_ids":["S3"]}],"distinctive_claim_remaining":"For heterogeneous analysis artifacts distributed across document, code, and data repositories, a metadata-only overlay that resolves cross-system identity and dependencies, gates ordinary discovery to the approved contextual version, and verifies bundle reconstruction will reduce wrong-version selection beyond an existing regulated-document lifecycle without increasing required-artifact misses or reconstruction failures.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"First-party Veeva documentation demonstrates production implementations of the principal primitives: lifecycle states, automatic supersession, archive search gating, unarchiving, scoped legal holds, permissioned deletion, relationship checks, audit logs, and limited restoration. This strongly supports implementability of the components, but not the incremental cross-repository layer or its effectiveness.","source_ids":["S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Specification versions can encode major outcome changes, and regulators treat version history, retrieval, and reconstruction as important to result reliability. The exact incidence of wrong-version use in active workflows is still unknown.","source_ids":["S1","S3","S4"]},"stakeholder_pull":{"score":4,"rationale":"Sponsor, investigator, institutional, service-provider, and archive-custodian duties create strong compliance pull for reliable version and retention controls, although direct demand for this particular overlay was not found.","source_ids":["S1","S2"]},"incremental_advantage":{"score":2,"rationale":"Veeva already supplies most proposed lifecycle, gating, hold, dependency-restriction, audit, and recovery primitives. Incremental value is confined mainly to heterogeneous cross-repository identity, dependency-aware bundle reconstruction, and measured workflow performance.","source_ids":["S5","S6","S7","S8"]},"distinctiveness_plausibility":{"score":2,"rationale":"The cross-repository experimental-artifact scope and routine reconstruction test remain differentiable hypotheses, but the core lifecycle-state and gated-discovery mechanism substantially collides with existing regulated-document functionality.","source_ids":["S1","S5","S6","S7","S8"]},"technical_implementability":{"score":4,"rationale":"Existing product documentation demonstrates that the main states and controls are technically feasible. Reliable identity resolution across code, datasets, external registrations, and document systems remains the difficult unproven component.","source_ids":["S5","S6","S7","S8"]},"adoption_authority_feasibility":{"score":4,"rationale":"Official sources identify sponsors, investigators/institutions, delegated providers, and a named archive custodian as responsible actors. Cross-institutional or multi-sponsor studies may still divide authority.","source_ids":["S1","S2"]},"evidence_readiness":{"score":4,"rationale":"A read-only comparison can use the externally documented current/steady-state and archive controls as a concrete rival and directly measure selection errors, task time, misses, and reconstruction. Representative prevalence evidence is not yet available.","source_ids":["S1","S5","S6"]},"safety_net_benefit":{"score":5,"rationale":"The candidate's reversible gating and no-deletion first step fit the regulatory need to preserve complete, available histories. Holds, audit logs, archive permissions, and restoration checks directly address the risk of destroying evidentiary records.","source_ids":["S1","S2","S6","S7","S8"]},"scalability":{"score":3,"rationale":"Commercial systems show that lifecycle controls can scale within regulated repositories, but heterogeneous repository integrations, local retention rules, identity resolution, and human adjudication limit straightforward institution-wide reuse.","source_ids":["S1","S2","S5","S6"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Four-week, one-study metadata-only inventory; adjudication of 30-50 artifact identities and states; construction of baseline and simulated cross-repository views; counterbalanced retrieval tasks; dependency and reconstruction simulation; and analysis by statistical, records, and technical staff.","confidence":"MODERATE","assumptions":["No participant-level content is copied into the evaluation.","Existing repositories and sandbox tools are available without new enterprise licenses.","Approximately six to ten users can complete scripted tasks within ordinary work schedules.","The estimate includes labor, coordination, secure workspace, scripting, analysis, and review but no production validation."],"source_ids":["S1","S2","S5","S6"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Configure a metadata overlay for a small number of repositories, lifecycle rules, authority workflows, hold checks, search gating, audit logging, identity links, dependency warnings, test fixtures, security review, and acceptance evaluation for several studies.","confidence":"LOW","assumptions":["An existing repository or platform is extended rather than replaced.","Licensing is mostly pre-existing and the scope is one institution or coordinating center.","Records, privacy, statistical, security, and system-validation labor is included.","Legacy identity remediation is limited to selected active studies."],"source_ids":["S1","S2","S5","S6","S7","S8"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Institutional rollout across several document, code, and data repositories, including connectors, metadata migration, role and hold configuration, validated workflows, training, governance approvals, monitoring, evaluation, and launch support.","confidence":"LOW","assumptions":["The launch covers one institution or coordinating center, not a multinational sponsor network.","Automatic deletion remains out of scope.","Commercial platform replacement and mass remediation of all historical studies are excluded.","Compliance validation, security testing, data stewardship, change management, and evaluation labor are included."],"source_ids":["S1","S2","S5","S6","S7","S8"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Ongoing metadata stewardship, disputed-state adjudication, hold and retention reviews, connector and software maintenance, periodic reconstruction exercises, access reviews, training, monitoring, audit support, and outcome evaluation.","confidence":"LOW","assumptions":["Human review remains mandatory for disputed status and disposition.","Repository licensing is largely pre-existing.","Study volume and regulatory complexity are moderate.","Major migrations, platform replacement, and multinational operations are excluded."],"source_ids":["S1","S2","S6","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Regulatory and standards sources require version control, version history, retrieval, reconstruction, and long retention, while primary research demonstrates that materially different historical outcome specifications occur and can be obscured by latest-version views. Direct wrong-version selection prevalence remains a blocking measurement rather than negating the supported risk.","source_ids":["S1","S2","S3","S4"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"Sponsors and investigators/institutions have explicit essential-record responsibilities, and UK guidance requires a named archive custodian, providing credible authorizers for a read-only evaluation.","source_ids":["S1","S2"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"Despite substantial prior-art collision, the remaining claim is specific: a cross-repository identity, dependency, and reconstruction layer must outperform existing steady-state and archive controls on wrong-version selection and time without increasing misses or reconstruction failures.","source_ids":["S5","S6","S7","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A four-week, one-study, metadata-only, counterbalanced comparison can test the remaining claim without changing source artifacts, approved records, permissions, or production search behavior.","source_ids":["S1","S2","S5","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The evidence step is read-only, excludes deletion and participant-level copying, retains baseline visibility, and can halt on any status dispute, required-artifact miss, unresolved dependency, or reconstruction failure. These controls align with official retention obligations.","source_ids":["S1","S2","S7","S8"]},"credible_cost_scope_and_range":{"status":"YES","reason":"Broad bands are credible as resource-equivalent planning ranges because the scopes separately identify labor, coordination, software configuration, security, compliance, integration, training, and evaluation. Confidence is low beyond the first test because no vendor quote or institution-specific repository inventory was available.","source_ids":["S1","S2","S5","S6","S7","S8"]}},"next_evidence_step":"Over four weeks, use one completed study's metadata and non-participant artifacts, or synthetic stand-ins where access is restricted, to create two read-only views of the same 30-50 artifacts: A, the institution's existing repository/eTMF steady-state and archive behavior; and B, a simulated overlay that additionally links protocol, SAP, randomization, code, environment, and derived-data identities and dependencies. Have six to ten analysts or reviewers complete 24 counterbalanced retrieval and reconstruction tasks. Compare wrong-version selections, median task time, required-artifact misses, unresolved-status disagreements, dependency-warning accuracy, and successful bundle reconstruction. Falsify advancement if B does not reduce wrong-version selection or task time relative to A, produces any material increase in required-artifact misses, or fails reconstruction at least as often as A. Change no source artifact, approved record, permission, retention state, or production search setting.","blocking_evidence":["Representative measurement of how often superseded protocols, SAPs, code, or derived datasets are encountered and wrongly selected in active workflows.","Comparative evidence that the cross-repository overlay improves selection or review outcomes beyond existing eTMF/Vault-style steady-state, supersession, archive, and hold controls.","Evidence that artifact identity, approval state, external dependencies, retention duties, and holds can be resolved consistently across document, code, data, and registration systems.","Evidence that archived bundles can be reconstructed with required relationships, environments, and holds intact; existing product restoration is explicitly incomplete for several relationship types.","Institution-specific platform pricing, validation burden, integration inventory, and recurring stewardship effort.","Direct stakeholder acceptance thresholds for false labels, retrieval delay, required-artifact misses, and reconstruction failure."],"research_disposition":"PRIOR_ART_DIFFERENTIATION_STUDY","world_novelty_boundary":"This bounded search found substantial prior art in regulated-document lifecycle management, especially Veeva Vault's state, supersession, archive-search, hold, deletion-control, audit, and restoration features. It did not establish whether any deployed system combines those controls across heterogeneous experimental documents, code, environments, and derived datasets with dependency-complete reconstruction testing and a controlled wrong-version retrieval evaluation. The remaining difference is therefore a testable differentiation claim, not a claim of world novelty."}