{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp05_complete_proposal_portfolio20_20260803","cell_id":"layer_decay_and_expiration_management__psychology","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_id":"psychology_clinical_formulation_leasebook_v0","proposal_index":1,"version":0,"title":"Clinical Formulation Leasebook for Retiring Stale Patient Hypotheses","problem":"During longitudinal psychological care, provisional formulations—such as suspected triggers, maintaining factors, risk interpretations, and explanations for behavior—are repeatedly deposited in notes and handoffs. Later clinicians may copy, retrieve, or encounter these layers without seeing their age, evidentiary basis, supersession status, or continuing relevance. Unsupported formulations can therefore remain cognitively active after the circumstances that produced them have changed, while indiscriminate removal would erase the history needed to understand prior decisions.","actors":["Patient receiving longitudinal psychological care","Treating psychologist or other authorized clinician","Clinical supervisor or multidisciplinary reviewer","Health-record or clinical-workspace administrator"],"observable_state":"For each patient, multiple formulation statements addressing the same behavior or treatment question remain visible as current guidance even though some have not been corroborated recently, conflict with newer observations, cite departed informants or ended circumstances, or have been superseded without an explicit marker. Auditable indicators include formulation age, last corroboration date, contradictory later evidence, reuse in recent plans, unresolved source dependencies, and whether current and superseded statements are visually distinguished.","consequence":"Stale hypotheses can continue framing attention, interpretation, questioning, and treatment planning, while the growing stack makes current reasoning harder to locate. Conversely, deleting old formulations without lineage can prevent reconstruction of why earlier assessments or interventions were chosen.","affected_objective":"Keep current clinical reasoning easy to identify and inspect without sacrificing historical accountability, reconstruction, or patient safety.","intervention":"Add a clinician-governed Clinical Formulation Leasebook beside the immutable signed record. Each patient-specific inferential claim receives an identity, creation date, author, supporting observations, referenced decisions, sensitivity class, last corroboration date, and lifecycle state: active, review-due, superseded, quarantined, or archived. A class-specific review lease makes an uncorroborated claim review-due rather than automatically false or destroyed. At review, an authorized clinician may renew it with evidence, supersede it with a linked formulation, archive it from routine views, or place a proposed removal in reversible quarantine after checking dependencies and retention obligations. Routine views privilege active claims; archived claims remain recoverable, and a compact marker preserves the identity, disposition reason, and successor link of any retired claim.","structural_mapping":[{"archetype_element":"Sequential deposits","domain_realization":"Patient-specific provisional formulations accumulate across sessions, clinicians, assessments, and handoffs."},{"archetype_element":"Changed usefulness or validity","domain_realization":"A formulation may lose standing when circumstances change, contradictory observations appear, its source becomes unavailable, or it remains uncorroborated."},{"archetype_element":"Explicit age and lifecycle state","domain_realization":"Every formulation records creation and corroboration dates plus an active, review-due, superseded, quarantined, or archived state."},{"archetype_element":"Decay or expiration trigger","domain_realization":"A formulation-class review lease expires after a preset interval without corroboration, triggering review and demotion from default visibility rather than automatic clinical rejection."},{"archetype_element":"Dependency and exception check","domain_realization":"Before retirement, the reviewer checks whether the formulation supports an active safety plan, assessment conclusion, treatment rationale, supervision issue, complaint review, or required record-retention obligation."},{"archetype_element":"Differentiated disposition","domain_realization":"Review can renew, supersede, archive, quarantine, or retain by exception; expiration never implies automatic destruction."},{"archetype_element":"Reconstruction and audit preservation","domain_realization":"Archived claims retain their evidence and decision links, while disposition markers record what changed, who authorized it, why, and which claim replaced it."},{"archetype_element":"Revalidation loop","domain_realization":"Scheduled sampling tests whether archived formulations can be restored and whether exceptions and lease lengths remain justified."}],"mechanism_mapping":[{"mechanism_slug":"stale_layer_detection_dashboard","role":"Maintains the formulation inventory and surfaces review candidates using age, absent corroboration, contradiction, changed context, and supersession signals; it does not alter clinical status.","counterfactual_removal":"Without it, stale formulations remain distributed across notes and cannot be reliably identified or ranked for review."},{"mechanism_slug":"time_to_live_ttl_policy","role":"Assigns a review lease when a formulation is created; lease expiry changes the claim to review-due and removes default-current presentation until a clinician renews or disposes it.","counterfactual_removal":"Without it, every formulation remains active by default and retirement depends on irregular manual cleanup."},{"mechanism_slug":"dependency_safe_delete_check","role":"Blocks retirement when a formulation is still referenced by an active safety process, assessment conclusion, treatment rationale, legal review, or required reconstruction path.","counterfactual_removal":"Without it, cleanup could hide or remove a claim that remains load-bearing for care or accountability."},{"mechanism_slug":"retention_schedule","role":"Defines lifecycle rules by formulation class and records safety, evidentiary, complaint-related, and other authorized preservation exceptions.","counterfactual_removal":"Without it, reviewers would apply inconsistent lifetimes and could mistake clinical staleness for permission to destroy a required record."},{"mechanism_slug":"soft_delete_quarantine_window","role":"Makes any approved removal from the working layer reversible for a bounded period, with the window sized by dependency and reconstruction impact.","counterfactual_removal":"Without it, an erroneous retirement could become irreversible before a missed dependency or clinical need is discovered."},{"mechanism_slug":"tombstone_or_deletion_marker","role":"Leaves a compact disposition record containing the retired claim's identity, date, reason, authorizer, and successor or archive location.","counterfactual_removal":"Without it, later reviewers could not distinguish a retired formulation from one that never existed or determine what replaced it."},{"mechanism_slug":"archive_restore_test","role":"Periodically samples archived formulations across ages and formats and exercises retrieval, reading, and reconnection to their evidence and decision lineage.","counterfactual_removal":"Without it, archival could silently become loss while the system continued to promise reconstructability."}],"causal_chain":["Longitudinal care deposits successive provisional formulations about the same patient and clinical questions.","The Leasebook gives each formulation an identity, age, evidentiary links, dependencies, and explicit lifecycle state.","Expiry of an unrenewed review lease moves a formulation to review-due and removes its default-current presentation without declaring it false or deleting it.","A clinician evaluates current support, contradictory evidence, patient context, dependencies, and preservation exceptions.","The clinician renews, supersedes, archives, quarantines, or exceptionally retains the formulation, and the system records the disposition and successor relationship.","Routine clinical views contain fewer unmarked stale hypotheses competing with current formulations.","Archives and disposition markers preserve the ability to reconstruct earlier reasoning, with restore drills testing that preservation.","Current reasoning becomes easier to identify while lifecycle decisions remain reversible, attributable, and reviewable."],"baseline":"Signed notes and handoffs remain chronologically stored, while clinicians manually copy important formulations forward or search prior notes. Old and current hypotheses may appear with similar visual authority; reconciliation occurs opportunistically during case review, and deletion or archival is generally organized around whole documents rather than the lifecycle of individual inferential claims.","nearest_rivals":["Periodic clinician-led chart or case review that reconciles formulations without persistent per-claim lifecycle metadata","Problem-list reconciliation that marks broad clinical problems active or resolved but does not govern the supporting formulation layers","Automated note summarization that compresses prior material into a current synopsis but may obscure supersession, dependencies, and reconstructable lineage","Simple age-based hiding of old notes that reduces visibility without distinguishing still-valid, superseded, safety-critical, or legally retained content"],"remaining_contrastive_claim":"The proposal's testable contrast is the joint management of individual clinical hypotheses as expiring-but-recoverable layers: review leases control current visibility, dependency and exception checks constrain disposition, and markers plus tested archives preserve reconstruction. A rival that only reviews, summarizes, or hides whole notes does not instantiate that full causal path.","authority_safety":{"decision_authority":"The treating clinician may renew or supersede routine formulations; archival or quarantine follows organizational policy, and contested, safety-linked, complaint-linked, or retention-sensitive cases require a designated supervisor or records authority. The system itself has no authority to declare a formulation false, change a diagnosis, alter treatment, or destroy a record.","authorized_first_step":"Run a read-only retrospective inventory and a non-clinical shadow-interface evaluation on previously documented or synthetic formulation layers; no lifecycle label is shown in live patient care and no record is moved or deleted.","excluded_actions":["Automatic diagnosis, risk classification, treatment recommendation, or treatment change","Automatic hard deletion of any clinical record or formulation","Demotion of an active safety-plan dependency without human review","Using age alone as evidence that a formulation is false","Allowing patients, clinicians, or administrators to lose access to records contrary to governing policy or authorized requests","Rewriting signed historical notes","Applying the system covertly in live care"],"halt_rollback":"Halt if the inventory exposes information beyond authorized reviewers, if lifecycle labels are mistaken for truth judgments, if dependency detection misses a safety-linked reference, or if archived material cannot be reconstructed. In a later reversible pilot, restore all quarantined layers to their prior visibility, disable lease-triggered demotion, preserve the audit log, and return to the chronological baseline."},"negative_tests":{"strongest_counterevidence":"Clinicians may already interpret dated formulations in context, and removing older hypotheses from routine visibility could eliminate useful longitudinal patterns or reduce productive reconsideration more than it reduces stale framing.","problem_falsifier":"In a bounded audit, older uncorroborated or contradicted formulations are already consistently marked as historical, are rarely reused as current premises, do not impede retrieval of current reasoning, and are not associated with measurable disagreement about which formulation governs present care.","intervention_falsifier":"In blinded vignette or shadow-interface testing, the Leasebook does not improve identification of the currently supported formulation and its provenance, or it increases missed safety dependencies, unjustified confidence, reconstruction failures, review time, or clinically material disagreement relative to the chronological baseline.","risks":["A lease-expired label could be misread as evidence that a formulation is false.","Clinicians could renew claims mechanically, turning exceptions into permanent holds.","Composite staleness signals could encode documentation habits rather than clinical validity.","Demoting old material could conceal recurrence or long-horizon behavioral patterns.","Dependency checks may miss narrative, external, or tacit uses that are not machine-readable.","Extra review work could divert clinician attention from patient contact.","Lifecycle metadata could expose sensitive interpretations to unauthorized viewers.","Quarantine may conflict with a binding destruction requirement if policy is configured incorrectly."]},"next_evidence_step":"Using a bounded set of synthetic cases or appropriately authorized, de-identified historical case timelines, have independent clinicians identify the formulation that should govern a specified decision and list its supporting and contradicting evidence under two interfaces: ordinary chronology and a read-only Leasebook mock-up. Measure agreement on current formulation, stale-premise selections, missed safety-linked dependencies, time to locate evidence, and ability to reconstruct a prior decision. Separately perform one scripted archive-and-restore drill on copied test data. Do not alter source records or patient care; proceed beyond this step only if dependency misses and reconstruction failures remain within predeclared halt criteria.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"No comparison with other proposals was performed. This candidate is defined specifically around the lifecycle of patient-level clinical formulation claims rather than whole-document storage or generic information cleanup.","revision_record":{"parent_version":null,"progress_targets_addressed":["Produce one complete reverse-innovation candidate","Preserve the layer-decay and expiration causal structure in psychology","Specify authority, safeguards, falsifiers, rivals, and bounded first evidence"],"conceptual_changes":[],"operational_changes":[],"evidence_changes":[],"claim_changes":[]}}