AI
SMARtCARE: Privacy-Preserving Agentic AI Systems for Bounded-Autonomy Clinical Decision Support
Publication: cs.AI updates on arXiv.org
Author: Srini Ramaswamy, Deveeshree Nayak
Source published: 29/09/2026, 04:00:00 UTC
Briefing created: 29/09/2026, 05:21:45 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 29/09/2026, 05:18:23 UTC.
Briefings
SMARtCARE is a privacy-preserving AI system designed to detect early vital-sign drift in ICU settings by using a lossy six-channel fingerprint of patient history rather than automatic record retrieval. The system operates in four states: Stable, Meta-cognitive, Assisted, and Regulated (Revoked). It identifies one prior-pattern recurrence among 14 two-admission patients on MIMIC-III, but found no matches among 9 two-admission patients on MIMIC-IV, highlighting a limitation of a fixed canonical pattern library. All logged decisions remain fully traceable and correctly attributed across data loading, logging, and audit layers.
Key points
- SMARtCARE uses a lossy six-channel fingerprint of patient trajectory to detect drift, triggering a Meta-cognitive escalation when current data matches a prior pattern but prior records are absent from context.
- The system operates in four states (Stable, Meta-cognitive, Assisted, Regulated) and successfully logs all decisions with full traceability and correct attribution across data layers.
Context and limitations
Evaluation relies on synthetic Monte Carlo studies and real-data demos on MIMIC-III and MIMIC-IV; it does not claim clinical efficacy. Results are limited to two-admission patient cohorts and do not account for all clinical scenarios.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 29/09/2026, 05:21:45 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeib6r7ajpatfavbc5ga5or5tmcfgbgfaod3qa2t5gedzswfekuhffe