{
 "standard": "data/dossiers.json — living topic dossiers. Each aggregates the KPSGILL knowledge graph around one subject: current state, law, litigation, enforcement, evidence, organisations, timeline, reform. The page list is derived from data/entities-registry.json appearsIn, so a new page tagged to the topic appears without editing the dossier.",
 "asOf": "2026-09-01T21:00:00-07:00",
 "rule": "A dossier is published only where the graph already holds enough pages to synthesise. Four qualify. A dossier never contains a fact that is not on one of the pages it aggregates or in a cited primary source.",
 "dossiers": [
  {
   "id": "prior-authorization",
   "entity": "topic.prior-authorization",
   "title": "Prior authorization",
   "url": "/dossiers/prior-authorization/",
   "currentState": "A prior authorisation is valid for a period the payer sets, and an adverse determination can be produced by an automated system without a named clinician who owns it. California AB 539 would impose a twelve-month duration floor; it completed legislative passage on 31 August 2026 and is not law. Federal interoperability rules set decision timelines for certain plans but reach neither duration nor automation.",
   "whatTheLawProvides": "California law sets response timeframes and requires that denials on medical necessity be made by a licensed reviewer with appropriate expertise. Federal rules impose decision timelines on certain Medicare Advantage and marketplace plans. Neither fixes authorisation duration, neither guarantees continuity across a plan transition, and neither bars an automated final denial.",
   "theGap": "Three things are unaddressed: how long an approval lasts, what happens to a stable therapy when coverage changes, and who is accountable when software generates the denial.",
   "kpsgillPosition": "/reform/proposals/prior-authorization-duration-and-algorithmic-denial/",
   "positionSummary": "Authorisation valid for the course of treatment or twelve months; ninety-day continuity on coverage transition; a named accountable clinician on every adverse determination; disclosure of automated involvement. Preferred sequencing publishes plan-level overturn rates first, because the published data is what makes the rest enforceable.",
   "litigation": "/reform/litigation-questions/algorithmic-denial/",
   "litigationNote": "A feasibility analysis rather than a docket: which theories could test an automated denial, and where each one runs into ERISA preemption or the absence of a duty running to the patient.",
   "openQuestion": "If automation may recommend a denial but a named clinician must own it, what stops the naming from becoming a formality — a single reviewer attesting to thousands of machine determinations?",
   "bills": [
    "AB 539"
   ],
   "metrics": [
    "Adverse determinations per thousand covered lives",
    "Overturn rate on internal appeal, by plan",
    "Therapy interruptions at coverage transition",
    "Median days from request to determination"
   ]
  },
  {
   "id": "corporate-practice-of-medicine",
   "entity": "topic.corporate-practice-of-medicine",
   "title": "Corporate practice of medicine",
   "url": "/dossiers/corporate-practice-of-medicine/",
   "currentState": "California prohibits lay control of clinical judgement. No regulation enumerates the levers through which control is actually exercised, so the prohibition is unenforceable in the structural cases and simultaneously over-readable against ordinary administrative services.",
   "whatTheLawProvides": "The corporate-practice doctrine and the professional-corporation provisions prohibit lay interference with professional judgement. The Attorney General reviews certain health-facility transactions. Management-services agreements are now the standard structure and are not themselves unlawful.",
   "theGap": "Eight levers determine independence in practice — productivity targets, panel size, referral steering, template design, coding pressure, staffing ratios, schedule control, termination without cause — and none is named in any instrument.",
   "kpsgillPosition": "/reform/proposals/corporate-control-of-clinical-judgement/",
   "positionSummary": "Implement the statute by regulation: enumerate the eight levers, expressly permit ordinary administrative services, require disclosure of the management agreement to each licensee practising under it, and make the target-setter traceable. Preferred sequencing omits a private right in the first instrument — enumerate, disclose, trace, then measure enforcement before adding a cause of action.",
   "litigation": null,
   "litigationNote": "No KPSGILL litigation object yet. Enforcement under the doctrine is rare, which is the point the proposal turns on.",
   "openQuestion": "If capital exits underserved markets when the acquisition model becomes more expensive, does enumeration protect physician independence at the cost of access? The proposal names this as a monitoring obligation rather than answering it.",
   "bills": [],
   "metrics": [
    "Complaints alleging enumerated levers, and disposition",
    "Management agreements disclosed to practising licensees",
    "Documented clinical overrides and any adverse action following one"
   ]
  },
  {
   "id": "healthcare-ai",
   "entity": "topic.healthcare-ai",
   "title": "Healthcare AI governance",
   "url": "/dossiers/healthcare-ai/",
   "currentState": "California passed four AI-related health measures in the 2025–2026 session and none is law: AB 2575 (tool inventories, clinical decision support, worker override) and AB 1979 completed passage; SB 503 (developer and deployer duties for biased outputs) and SB 903 (AI barred from providing or advertising therapy) are enrolled or enrolling. WHO/Europe published evidence that governance readiness, not deployment speed, is the binding constraint — as a convened expert report, not WHO policy. The European Commission designated a general AI service under platform law with obligations due end-November.",
   "whatTheLawProvides": "Medical-record content and retention are governed by state record requirements and electronic-record certification criteria; neither requires that machine-generated clinical text be identifiable as such. FDA’s clinical-decision-support policy addresses device status, not authorship. Attestation law assumes a human author.",
   "theGap": "No instrument requires provenance capture at generation, model-version recoverability, or a face-of-the-record distinction between physician and machine text. Every surveyed instrument ends at the clinician, because the clinician is the only actor already licensed, insured and disciplinable.",
   "kpsgillPosition": "/reform/proposals/ai-medical-record-integrity/",
   "positionSummary": "Require machine-readable provenance for AI-generated clinical text at the point of generation, and move physician responsibility from authorship to clinically material adoption. The twelve-dimension readiness framework treats validation, provenance, bias monitoring, accountability allocation and model change control as separate questions.",
   "litigation": null,
   "litigationNote": "No docket yet. The KPSGILL forecast is that contested board discipline over AI-assisted decisions arrives before any board publishes risk-tiered guidance.",
   "openQuestion": "Responsibility should follow control, and a clinician using a system she cannot inspect does not control it. Every instrument surveyed nonetheless lands responsibility on her. What allocation would survive a case where the model was right, the clinician overrode it, and the patient was harmed?",
   "bills": [
    "AB 2575",
    "SB 503",
    "SB 903",
    "AB 1979"
   ],
   "metrics": [
    "Share of certified record systems capturing generation-time provenance",
    "Share of AI-drafted notes with recoverable model and version identity",
    "Board cases in which provenance was available to the adjudicator"
   ]
  },
  {
   "id": "credentialing",
   "entity": "topic.credentialing",
   "title": "Credentialing and privileging",
   "url": "/dossiers/credentialing/",
   "currentState": "California AB 1199 (health care provider credentialing) completed legislative passage on 31 August 2026 with Assembly concurrence 62–0. It is not law and the chaptered text has not been read. Credentialing itself remains six separate determinations by six different bodies, each verifying substantially the same facts.",
   "whatTheLawProvides": "Licensure is state law. Certification is private and not legally required to practise. Credentialing, medical-staff membership and privileging are institutional determinations governed by bylaws and accreditation requirements. Payer enrolment is contractual. Adverse actions reach the National Practitioner Data Bank under federal law.",
   "theGap": "Nothing requires that a primary-source verification performed to a standard be relied upon by the next institution. The same licence, diploma and training record are verified from scratch at every institution and every payer, and the delay — not the scrutiny — is the harm.",
   "kpsgillPosition": "/reform/credentialing-modernization/",
   "positionSummary": "No position registered. The Lab entry separates the six gates, identifies duplication rather than scrutiny as the binding constraint, and states the measurement gap: this site has no published data on credentialing delay or its consequences, so a recommendation to restructure verification would be an opinion wearing a proposal’s clothes.",
   "litigation": null,
   "litigationNote": "No KPSGILL litigation object yet.",
   "openQuestion": "If the same facts are verified six times, what is the marginal safety value of verifications two through six — and has anyone measured it? A shared verification record must carry adverse findings with the same fidelity as favourable ones, or it becomes a laundering mechanism.",
   "bills": [
    "AB 1199"
   ],
   "metrics": [
    "Days from application to privileging decision, by institution",
    "Re-verification events per physician per year",
    "Posts unfilled during credentialing",
    "Adverse findings surfaced by second and subsequent verifications"
   ]
  }
 ]
}