cs.AIAug 14, 2026

Generated Context versus Governed State: Functional Conditions for Accountable Longitudinal Clinical Reasoning

Authors: Augusto Bernardo Pissarra, Victor Lorena de Farias Souza

Organizations: MyndwareMed

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Large language models (LLMs) have become the dominant interface of clinical artificial intelligence, yet the interface they expose (text in, text out, one context window at a time) maintains no explicit, persistent, governed representation of what is currently true about a patient. This paper argues that longitudinal clinical reasoning is a state-estimation problem under partial observability, and that the axis on which clinical AI succeeds or fails is not the fluency of the model reading the record but the governance of the patient state it reasons over. We distinguish generated context from governed state; separate five objects that clinical AI habitually conflates (true state, observations, evidence, belief, and simulated state); define a tiered governance standard against which any clinical AI system can be audited; and show that an operational definition of accountability decomposes into four information requirements: an immutable evidence ledger with awareness-time versioning, a belief state distinct from accumulated evidence, an observation-process model, and claim-level causal typing. We are explicit that this decomposition is analytic rather than a necessity theorem, and that its value is conceptual hygiene: it converts "accountable clinical AI" from a slogan into an audit instrument. A six-level maturity framework separates what a system makes governable from what it can compute, locating current LLM-centric practice at high capability but low maturity. The paper is fully self-contained: the four research questions the framework poses are stated in the introduction, and the conclusion records what the paper establishes toward each; future work develops the buildable core of the architecture and the research program toward full Clinical World Models. No empirical result is claimed here.

Figures & tables

Figure 1: Generated context versus governed state. Left: knowledge of the patient is reconstituted into a transient buffer on every call and vanishes with it. Right: one canonical, versioned belief state — fed by an immutable evidence ledger — is the object every component reads from and writes through, and the object an auditor queries.
TierRequirements
Coreidentity ⋅\cdot persistence ⋅\cdot explicit semantics ⋅\cdot provenance ⋅\cdot update governance ⋅\cdot historical query (served by deterministic replay)
Safety-criticalcore ++ explicit uncertainty typing ⋅\cdot deterministic constraint validation ⋅\cdot correction propagation
Predictivesafety-critical ++ outcome calibration ⋅\cdot model-version binding ⋅\cdot monitored validity under distribution shift
Table 1: The tiers of Definition 1 (Governed Clinical State), normalized. Each tier inherits the tier below it; a system is audited at the tier matching what its state is used for.
Architectural classPersistentGovernedReplayBelief ≠\neq evid.Quar. sim.
Electronic health record (EHR) / document store [ 19 ]yespartialnonono
RAG / long-context LLM [ 4 ]nonononono
LLM agent ++ memory [ 20 , 21 ]partialnononono
Knowledge graph (typical) [ 5 ]yespartialnonono
Learned world model [ 11 , 22 , 23 ]partialnononopartial
Governed clinical state (this paper)yesyesyesyesyes
Table 2: Architectural classes, not products, against the governance properties of Definition 1 (Governed Clinical State). The five columns compress the audit for readability rather than replacing it: Persistent summarizes identity and persistence; Governed summarizes explicit semantics, update governance, provenance, deterministic constraints, and calibration; Replay summarizes historical query and rollback/correction; the last two columns audit the object separations of Section 4 (belief distinct from evidence; quarantined simulation), which the tier requirements presuppose. Entries indicate properties guaranteed by the architectural class as such, not properties a particular implementation could add; “partial” marks properties implementations may add without a class-level guarantee; the final row states what this paper requires, not what any deployed system has demonstrated — the audit questions, not the grades, are the contribution.
Figure 2: The five objects and the epistemic boundary. The true state xtx_{t} sits above a boundary no system crosses; observations oto_{t} , evidence units eie_{i} , and belief bτ(xt)b_{\tau}(x_{t}) are successive constructions below it, and simulated state x~t branch\tilde{x}^{\,\mathrm{branch}}_{t} lives in a quarantined branch whose values never re-enter evidence or belief. The pathologies of LLM-centric practice are conflations of adjacent stages — most commonly, treating evidence as if it were belief.
ComponentGenerated context (RAG / long-context)Governed state
Renal functionTreats a 14-month-old creatinine as the patient’s current renal status — fluent, confident, staleBelief over renal function widened by 14 months without measurement: “unknown; last known value xx ; a point-of-care creatinine would resolve it”
MetforminWhichever fragment ranks higher wins; the conflict itself is invisibleContradicted-open: both evidence chains preserved, belief explicitly uncertain, conflict flagged for human closure
Prior reactionRenders absence as “no known allergies”Typed unknown/unavailable, not negative: prior contrast records exist at a source unavailable to this system
The gaps themselvesNo mechanism: retrieval returns documents, not the absence of documentsFirst-class output: the typed gaps and the measurements that would close them
Table 3: One clinical question, component by component. The failure on the left is not that the language model reads badly — it reads perfectly. The failure is that four of the five objects of Section 4 do not exist anywhere in that architecture, so no amount of reading can consult them.
Figure 3: One patient, six years, one query. Top: the timeline, with a documentation lag (2022), an availability exit (2023), and an informative measurement gap (2023–24). Bottom: the same record read at the same moment by the two architectures of Figure 1 . The governed answer is not produced by a better reader but by objects the generated-context architecture does not possess. The case is an illustrative construction, not an executed comparison: the claim is about mechanisms each architecture can and cannot express, not measured performance.
Figure 4: The six-level framework: increasing degrees of explicit, persistent, governable clinical structure. An architectural maturity framework, not a claimed law of intelligence.
More in this topic

AI Accountability

AI accountability concerns how responsibility for AI decisions and harms is assigned, evidenced, challenged, and remedied.

See all 157 papers

Explore similar work

CardsList
Sep 30, 2026

Can LLMs Reason Over Long Horizons? An Empirical Evaluation of Context Strategies for Longitudinal Clinical Reasoning

Longitudinal clinical reasoning requires large language models (LLMs) to identify and integrate relevant evidence distributed across extended patient histories. Although long-context models can process increasingly large amounts of information, providing more history does not necessarily make relevant evidence more accessible or improve reasoning. We compare five context strategies (Full, Recent, Episodic, Semantic, and Hybrid) on MedLoCoMo across four open-weight LLMs, examining answer correctness, robustness to query-evidence distance, and abstention on questions with unsupported premises. Episodic and Hybrid generally achieve the strongest overall accuracy, while Recent Context degrades most as supporting evidence becomes more distant; Episodic and Hybrid maintain the highest accuracy at long distances. Analysis of adversarial questions further shows that strong performance on answerable questions does not necessarily translate to successful abstention when the available history does not support the requested conclusion. These findings show that reliable longitudinal reasoning depends not only on how much history an LLM can access, but critically on how relevant evidence is selected and presented for reasoning.

Taye Akinrele, Noorbakhsh Amiri Golilarz, Subash Neupane +2

Department of Computer Science, The University of Alabama

Long-Context Language Model EvaluationClinical ReasoningLLM Context ManagementMedical LLMs
AskPDF
Oct 1, 2026

A rubric landscape for evaluating clinical reasoning in large language models: what exists, what is missing, and what needs to be combined

Exam-style accuracy does not establish whether large language models (LLMs) reason well over clinical records. We define clinical reasoning as integrating and updating evidence across time and sources to form, revise and justify a patient's problem representation and a defensible plan. This structured narrative review maps three literatures: medical education assessment instruments, clinical LLM benchmarks published from 2023 onwards, and general-domain methods for evaluating long-form generation. We examine six dimensions: problem representation, temporal synthesis, differential and management reasoning, counterfactual reasoning, calibrated uncertainty, and reasoning faithfulness. Preprints are included and flagged. No single instrument covers all six dimensions. Problem representation and differential or management reasoning are reasonably covered, although reliability varies by instrument and setting. TIMER-Eval targets temporal synthesis, and ER-Reason assesses sequential diagnostic belief updating. Dedicated uncertainty and counterfactual evaluations are emerging, but their applicability to longitudinal free-text reasoning remains limited. Factual completeness is well theorised in general-domain evaluation, with early clinical evidence of important omissions. Faithfulness remains the weakest dimension, with one identified clinical causal-ablation study on multiple-choice questions. Existing tools should be combined through binary rubric items, separate completeness and correctness scores, case-specific importance weighting with non-compensable safety caps, temporal order-consistency checks, and chance-corrected reliability reporting. Further design work is needed for calibrated uncertainty, counterfactual reasoning and faithfulness over longitudinal free-text records. This review provides a design rationale, not a validated instrument.

Zhangshu Joshua Jiang, Zina Ibrahim, James T. Teo

DRIVE-Health CDT Department of Biostatistics and Health Informatics Institute of Psychiatry, Psychology and Neuroscience King’s College London London, UK · Cleveland Clinic London London, UK

Faithfulness of Language Model ExplanationsTemporal Reasoning in Language ModelsCounterfactual Reasoning in Language ModelsClinical ReasoningClinical Language Model Evaluation
AskPDF
Aug 27, 2026

LAAF: A Layered Accountability Architecture Framework for LLM Applications

Large Language Models (LLMs) operate in hospitals, courtrooms, banks, and public service desks, where fluent, confident outputs are treated as authoritative even when ungrounded or incorrect. When such an output contributes to harm, who is answerable, and through what mechanisms can responsibility be traced, explained, and acted upon? Following PRISMA guidance, five databases were searched from January 2022 to March 2026 against four review questions; of 4,512 records identified, 122 primary studies were included, together with 12 regulatory and standards documents analysed as primary sources. The review consolidates a sociotechnical account of accountability as an actor-forum relation resolved into five dimensions, and synthesises mechanisms across four families: technical controls, human oversight, organisational governance, and documentation and traceability, each with a maturity assessment. The corpus is read through a four-layer classification device spanning provenance, application logic, human oversight, and governance and redress, cross-cut by traceability, role clarity, and continuous monitoring. Both are mapped onto the EU AI Act, whose high-risk obligations have applied since 2 August 2026, the NIST AI RMF with its Generative AI Profile, ISO/IEC 42001, and sectoral guidance in healthcare, consumer finance, education, and the public sector. Four persistent gaps emerge: under-specification of human oversight, absence of shared accountability metrics, disciplinary disconnection, and limited empirical evaluation, alongside five structural tensions that no surveyed instrument resolves. The review closes by consolidating the classification device into an integrated accountability architecture, LAAF, with cybersecurity aligned to the OWASP LLM Top 10 (2025); it is a synthesis of the surveyed evidence rather than a validated artefact.

Prachi Chaturvedi, Shahnawaz Ahmad, Ehsan Nowroozi +5

School of Computer Science Engineering and Technology, Bennett University, Greater Noida, India · Centre for Sustainable Cyber Security (CS2), University of Greenwich, United Kingdom · School of Electrical Engineering, Computing & Mathematical Sciences, Curtin University, Australia +1

AI AccountabilityAlgorithmic AccountabilityResponsible AIAI Governance
AskPDF