Med

Clinical metagenomic sequencing detects hundreds of organisms whose relevance to infection requires clinicians' manual interpretation. We retrospectively reviewed 4,855 specimens from 19 sampling sites, linked 1.33 million detected organisms to clinical culprit-pathogen determinations, and fine-tuned a large language model on these labels. The model recovered 92.44% of reference culprit calls in the held-out test set and 84.49% in a later cohort. In a blinded re-adjudication of 215 cases, the model matched 90.5% of culprit calls on which both expert reviewers agreed. Adjudication accuracy varied with organism identity and specimen types. mNGS-derived metrics such as read counts or genome coverage alone were insufficient to identify culprit pathogens, and organisms classified as common colonizers could still be judged causative. Certain organisms were frequently co-reported as causative in the same specimen. These findings support further evaluation of human and AI-assisted report review, particularly regarding missed pathogens, excess calls, and differences across specimen types.

The gut-brain axis has emerged as a critical modulator in neurological diseases, including stroke. While acute post-stroke dysbiosis is well established, it remains unclear whether these alterations persist long-term and can be therapeutically modulated. This study addresses this gap through integrated metagenomic and metabolomic profiling in two independent clinical stroke cohorts, complemented by experimental models. Experimental stroke induced persistent changes in gut microbiota composition and altered short-chain fatty acid profiles. In patients, microbiota composition remained distinct from controls up to three months after ischemic stroke or transient ischemic attack (TIA), accompanied by persistent reductions in propionic and valeric acid and reduced abundance of Prevotella-associated microbial signatures. We performed a randomized, double-blind interventional trial (PRISE) testing a multispecies probiotic versus placebo, using household members as additional controls. Probiotic intervention led to microbiome compositional changes with increased abundance of beneficial taxa and significantly improved functional recovery. The tested probiotic strains showed anti-inflammatory and antioxidant effects in vitro. These findings suggest that targeted modulation of the microbiome represents a promising therapeutic avenue to improve chronic stroke recovery, warranting validation in larger Phase III clinical trials.

Background: Host response and biomarker research to advance sepsis therapeutics requires well-characterized cohorts with longitudinal biospecimens across a range of hosts and infections. Automated screening based on SIRS criteria is sensitive but nonspecific and identifies more candidates than research teams can approach, sample, and follow longitudinally. We evaluated a structured enrichment review by non-physician study staff, done after automated screening in a hospital-wide prospective sepsis cohort. Methods: The CISID cohort is a prospective observational study of hospitalized patients with suspected sepsis, enrolling adults from the ED, inpatient wards, and ICUs at a tertiary academic medical center. Patients were identified through: (1) automated electronic health record screening for two or more SIRS criteria with a 24-hour lookback, documentation of microbiologic culture or PCR, and receipt of a new antibiotic; (2) a sepsis enrichment step of structured chart review by trained non-physician study staff; and (3) discussion with the primary clinical team. Enrolled patients underwent physician adjudication into four diagnostic categories. To evaluate the enrichment step, we analyzed all patients screened during a 9-week period, assigned a composite "sepsis-likely" variable (CDC Adult Sepsis Event criteria or ICD-10 code R65.20/R65.21 within 48 hours before or after screening), and assessed mortality across each stage of screening. We also assessed concordance of the composite with physician adjudication among enrolled patients. Results: From November 2023 through February 2026, 4,318 patients screened in through EHR-based automated screening; 1,220 (28%) were flagged for enrollment, and 288 were enrolled. Physician adjudication classified 209 (73%) as sepsis, 46 (16%) as infection without sepsis, 14 (5%) were classified as undifferentiated critical illness, and 19 (6.6%) as non-infectious sepsis mimics. Patients with sepsis were severely ill (median SOFA score, 9; ICU admission, 80%; shock, 48%), and 51% were immunocompromised; mortality was 14% in hospital and 27% at 90 days. In a 9-week sample of 907 screened encounters, 21% of those deferred and 70% of those flagged for enrollment met the sepsis-likely composite, a 3.3-fold increase in prevalence after structured study team review. Conclusions: A brief structured enrichment review after automated screening tripled the prevalence of EHR-based sepsis indicators among patients selected for enrollment, while concentrating enrollment efforts on roughly one-quarter of screened-in patients in this hospital-wide, all-source suspected-sepsis cohort. This approach may help prospective cohorts direct limited biospecimen resources toward patients most informative for host-response and biomarker research.

Background: Recent work has brought large language models (LLMs) into clinical variant interpretation. In most existing approaches, the model is confined to text-level operations, retrieving, extracting, compacting, and summarizing heterogeneous evidence, while the genuine genetic evaluation is delegated to a separate deterministic engine. Other tools employ LLMs to mine functional evidence for a single ACMG criterion, to rerank a classifier's output, or to answer isolated variant questions. Integrative genetic reasoning, by contrast, demands multi-source evidence weighing for each variant, reconciliation of inheritance with observed zygosity and parental genotypes, joint interpretation of variants in the same gene, and the derivation of a case-level conclusion. Moreover, to protect patient privacy, this must be done on hospital premises. Methods: o build an end-to-end variant interpreter that runs on local infrastructure, we benchmarked 13 small language models (SLMs) on 150 multiple-choice genetics questions and selected Qwen3.5-9B. In IVA (Intelligent Variant Analyser), a four-stage pipeline, the model assigns ACMG/AMP criteria, applies de novo, dominant, recessive, and X-linked inheritance logic to trios, and writes a structured report, while deterministic code handles retrieval, phasing, point summation, citation checking, and formatting. We evaluated IVA on synthetic scenarios, 16 solved trios, and 327 solved singleton cases of visual impairment, on a single GPU. Results: On the benchmark, Qwen3.5-9B reached a mean accuracy of 0.81, compared with 0.89 for a 1.6-trillion-parameter cloud model. In the 16 trios, IVA reported every previously identified pathogenic, likely pathogenic, or uncertain-significance variant of interest (sensitivity 100%; precision 39.5%). In the 327 singleton cases, it recovered 93% of the expert-reported causative variants and classified 82% of them as pathogenic or likely pathogenic (precision 43%). Failures fell into four recurrent classes: omission, unverified mechanism, mismatch between reasoning and label, and arithmetic errors. Conclusions: A small language model running on local hardware can serve as the reasoning layer for case-level and trio-based variant interpretation when it is surrounded by deterministic safeguards. Because its outputs are not deterministic, analyst review and prospective multicenter validation remain necessary.

BACKGROUND & AIMS: Inflammatory bowel disease (IBD) is often diagnosed months to years after symptoms begin, and diagnostic delay is associated with complications. We tested whether longitudinal machine learning using electronic health record (EHR) data can identify incident IBD before diagnosis, and quantified lead time and deployment burden. METHODS: We studied adults with an IBD-compatible symptom, sign, or finding at UCSF (2012-2024). Incident cases were confirmed by structured filtering and validated large language model note review. We trained a state space model (Mamba) and comparators on pre-diagnostic EHR data, excluding diagnostic-workup features, and assessed discrimination at 6- and 12-month landmarks. We externally tested Mamba without recalibration at ZSFG, a safety-net hospital serving a demographically and socioeconomically distinct population, and simulated diagnosis-code-only deployment at UCSF. RESULTS: Six months after presentation, among patients still undiagnosed (67% of UCSF test cases), Mamba achieved an AUROC of 0.969 (95% CI, 0.928-0.999) at UCSF and 0.965 (0.952-0.976) at ZSFG. Median lead time was 20 months at UCSF and 15 months at ZSFG. Among future cases with an initially normal colonoscopy (N=32) or a histological finding of acute inflammation without chronicity (N=63), the model flagged 84% and 89% of these patients, respectively, before those procedures, indicating the potential for pre-diagnostic detection. Among 358,886 symptomatic patients not used for training (60 incident cases), the diagnosis-code-only model achieved an AUROC of 0.926; reviewing 20 alerts/month identified 34 cases (PPV, 1.09%; 64-fold above baseline). CONCLUSIONS: Longitudinal machine learning can identify increased IBD risk months to years before diagnosis and may define a population for prospective studies of early IBD biology. Prospective evaluation should determine whether earlier identification reduces diagnostic delay.

Clinical AI deployment rarely reaches the bedside because both workstation integration and governed backend provisioning are challenging. We introduce a workflow that allows rapid and governed access to AI functionalities through a browser interface or API, enabling privacy-compliant data analysis. The complex setup for compliant healthcare applications such as a secure backend, private AI model endpoint, and managed identity configuration is encoded by itself, a concept known as Infrastructure as Code (IaC). IaC is a concept in which networks, access rights, and service endpoints are written down as machine-readable files that can be reviewed once and executed identically in any institution. Here, we apply this concept to a clinical use case. The process of deploying a governed, private clinical AI solution using this approach in a tertiary care center took, on average, 310.3 +- 196.6 seconds (n=25, mean +- SD). To further demonstrate the clinical value of this approach, we developed and deployed an LLM-based inconsistency checker, which identifies textual inconsistency in medical documents. When evaluating the tool on a set of 1,347 physician-approved clinical documents, the tool flagged a clinically severe inconsistency in 105 clinical documents (7.8%, approximately 1 in 13). In a severity-stratified physician audit, 149 of 150 sampled findings were confirmed (population-weighted PPV 98.2%; CI 94.8 - 100.0). In summary, we present the application of IaC frameworks to a healthcare use case and demonstrate that clinically useful AI applications can be deployed to physician workstations in hospital-governed cloud environments within a few minutes.

Biologically rich clinical trials generate heterogeneous datasets whose relationships and clinical context must be preserved. Here we present the POETIC-AI environment, a trial-centric digital environment for integrating, analysing, and exploring multimodal data from POETIC, a randomised phase III trial of peri-operative aromatase inhibition in hormone receptor-positive breast cancer (NCT02338310). We implemented a graph-based ontology linking patients, treatments, sampling timepoints, specimens, assays, images, spatial regions, and derived measurements; and designed code-based and graphical analytical workflows. Using proliferation as a test case, we reproduced POETIC's established reduction in Ki67% and examined variation underlying sample-level measurements. In response-enriched cohorts, mean Ki67% decreased while relative variation among pathologist-assessed fields increased (paired Wilcoxon, P < 0.0001). Pathologist-assessed Ki67% correlated strongly with an 18-gene transcriptomic proliferation signature (r = 0.834), and both decreased significantly during treatment (P < 0.0001). An annotation-free pixel-wise Ki67 H-score showed weaker correlations and a similar but non-significant reduction (r = 0.327-0.472; P = 0.15), highlighting differences between tumour-focused and whole-section image quantification. Digital pathology and spatial molecular measurements remained linked to their source images and annotated tissue regions, enabling exploration within their morphological and spatial context using bespoke interactive applications. POETIC-AI demonstrates how the structure, provenance, and tissue context of evolving multimodal clinical-trial data can be preserved within a connected research environment, supporting reproducible analyses and the laying the groundwork for future development of biologically interpretable computational biomarkers.

Sepsis causes approximately 11 million deaths globally each year. Whole-genome sequencing (WGS) of founder populations offers a unique opportunity for discovery of genes related to susceptibility of sepsis. We performed GWAS and EXWAS on WGS data for n=210 Arab ICU patients (n=68 sepsis cases, n=142 controls, other ICU-related disorders) and prioritized n=40 associated deleterious founder variants in n=31 genes (REGENIE Firth regression p < 0.05; CADD score > 20; enriched AAF vs AllofUs/gnomAD), including a novel start-loss variant in UBE2L3 (c.2T>G; p.Met1?; CADD = 25.6; SIFT = 0.001) in sepsis cases (3/68 vs 0/142 controls; Fisher's exact P = 0.026). In All of Us (AoU), UBE2L3 missense burden was associated with bacterial infection (n=26,939 cases; SKAT-O P = 2.38x10-3; OR=23.10; 95% CI = [0.63-845.27]). LLM AI models fine-tuned for drug development scored UBE2L3 a 7.8 of 10 as a drug target for sepsis, consistent with mouse UBE2L3 KO showing increased IL-1Beta mediated inflammation in sepsis.

Multi-agent prompting of large language models has produced contradictory diagnostic results, and it remains unclear whether any benefit comes from specialist personas or from perspective independence. We compared a single direct call, five personas in one context, and the same five roles as isolated agents integrated by a moderator, with five repeat runs per case, on 87 CPC cases, 406 MedCaseReasoning cases, and 364 emergency department encounters, under an LLM judge validated against clinicians. On the external benchmark the team beat the single call on both pre-specified recall endpoints (top-3 +3.0 points, p = 0.0079; top-5 +3.9, p = 3.8 x 10^-5;); a factorial attributes the gain to independent generation plus moderated synthesis, not the specialist roles. On real emergency presentations the benefit reversed (top-1 40.1% versus 34.3%, p < 0.0001), carried by the specialist role lists and surviving added objective results. Deployment should key on the question and the input at hand.

Looping is a major cause of difficult and painful insertion during unsedated colonoscopy, yet strategies are largely guided by trial and error; prolonged insertion may jeopardise completion, discourage future surveillance and is associated with lower adenoma detection. In this randomised trial of 141 patients, protocolised review of pre-existing abdominal-pelvic computed tomography (CT) anticipated colonic anatomy and shortened caecal intubation time by 136 s, while reducing pain, discomfort and operator workload and increasing willingness to repeat colonoscopy. Effects were largest in N-loop and alpha-loop configurations, which accounted for 83% of patients. Routinely acquired CT could support anatomy-informed colonoscope insertion especially when anesthesia services are limited or sedation poses an elevated risk.

Ischemic stroke is a leading cause of long term disability worldwide. Although timely reperfusion can improve outcomes, realizing this benefit requires clinicians to integrate fragmented clinical and imaging information with evolving evidence under the time pressure of emergency care, often with limited access to specialist expertise. Here we present StrokeAgent, a multimodal agentic decision-support framework designed around temporal alignment and continuous information synthesis across the acute stroke pathway. It processes clinical records, laboratory results, raw electrocardiographic (ECG) waveforms and source multimodal CT images to support emergency triage, the initial thrombolysis decision and final reperfusion planning. As information becomes available, StrokeAgent formulates and updates recommendations at each decision point, grounding each recommendation in an auditable chain of supporting findings and evidence. We evaluated the system in a curated set of 100 patients with multimodal clinical and imaging data, deliberately enriched for clinically complex reperfusion scenarios alongside standard presentations. StrokeAgent achieved 86% concordance with expert-adjudicated final decisions, exceeding the mean performance of three tool-free large language models (LLMs) and six resident and attending neurologists by 13 and 17 percentage points, respectively, under matched information content and timing. Complete pathway concordance (agreement at every applicable decision point) was 76%, with even larger margins over both comparator groups. On average, StrokeAgent completed the full decision pathway 73.9 s faster than the case level median neurologist time. In 1,200 blinded clinician assessments, StrokeAgent received higher overall clinical utility ratings than each LLM (common odds ratios, 10.8 to 16.3) and higher reasoning quality ratings, with fewer outputs judged to pose severe potential harm. StrokeAgent demonstrates the feasibility of a decision support system that integrates native multimodal data processing with auditable, evidence supported recommendations aligned with the sequential nature of acute stroke care. Prospective studies are needed to determine whether physician-supervised use improves clinical care.

In medicine, artificial intelligence (AI) safety has focused on discrete flaws such as hallucination, bias and miscalibration. Conversational systems pose a distinct hazard: in dialogue, a model can covertly advance a misaligned objective by redirecting the clinical judgment it should support. Whether clinical expertise equips physicians to detect and resist such subversion is unknown. In a participant-blinded, randomized experiment, 225 physicians from 42 countries and eight clinical disciplines made sequential diagnostic and treatment decisions with AI assistance on anonymized cases from neurocritical care, their shared field of expertise. Unknown to them, each physician used the same model under two conditions: instructed to steer them towards harmful targets (adversarial) or correct clinical options (aligned), each established by expert consensus. With adversarial AI, 190 physicians made at least one harmful decision, compared to 21 with aligned AI. Across 2,485 targeted decision points, adversarial AI raised the harmful-decision rate from 2% with aligned AI to 49% (adjusted difference 47 percentage points, 95% CI, 43-51) and reduced the correct-decision rate from 83% to 25%, with no clear protection associated with clinical experience. Of the 190 physicians who made harmful decisions with adversarial AI, 141 did not report anything unusual about that AI. Two physicians reported suspecting systematic steering or manipulation. Physicians cannot safeguard the clinical use of conversational AI through expertise and decision-making authority alone.

International travel can increase infection risk and spread antimicrobial resistance (AMR), yet clinical systems rarely capture travel history in a structured format. Extracting this information from free-text clinical documentation at scale could improve alerts for high-consequence infections and AMR surveillance. We developed an annotation framework covering key aspects of travel history, including destination, exposures, travel duration, and multiple temporal variables. The framework was used to annotate a corpus of 100 clinical case reports by five clinicians, yielding over 5,000 clinician annotations. Several LLMs were evaluated on the same extraction task, with their outputs compared against clinician consensus. Agreement among clinicians was high for straightforward variables such as presence of travel and destination, but dropped substantially for temporal variables such as time since return and symptom onset relative to return. LLMs showed a similar pattern. They performed well on high-level variables but struggled with the same temporal fields that challenged clinicians. Model errors were also concentrated in cases where clinicians disagreed. Across models, outputs were more sensitive to sampling temperature than to prompt wording. A smaller open-weight model performed similarly to substantially larger models on several variables but remained weaker on temporal fields. Overall, LLMs could extract some travel history variables, but reliable extraction of temporal information remained difficult.

The tumor microenvironment in colorectal cancer (CRC) is a heterogeneous ecosystem in which host cells and microbial communities interact dynamically, influencing disease progression. However, the clinical utility is limited by the lack of a scalable spatial host-microbiome technique and by insufficient integration of artificial intelligence for the interpretation of high-dimensional multi-omics data. To overcome the barriers, we present AlphaFISH, a platform technology integrating both technical and computational innovations for multi-omics spatial analysis of clinical biopsies at subcellular resolution. The system uses a sequencing-free, high-throughput, spatial profiling technique to construct, to date, the largest clinical spatial transcriptomics and spatial microbiome datasets acquired from 149 colorectal biopsies from 68 human subjects, supported by a comprehensive scRNA-seq atlas covering 4.27 million cells across 650 patients for robust cell annotation. Deep learning of the cohort-scale dual-omics data, consisting of more than 10 million subcellular sampling vectors, enables the development of a transformer model with joint embeddings of gene expression, spatial architecture, and the microbial microenvironment in colon tissues, achieving nearly 90% accuracy in predicting CRC-associated pathological features using unseen spatial omics inputs. The AI interrogation further predicts tumour recurrence at 81% accuracy in 28 patients followed within 1 year post tumor resection period. AlphaFISH reveals that spatial interactions between Fusobacterium and cellular niche consisting of tumor and T cells serve as key markers of CRC malignancy, progression, and recurrence, indicating the critical role of spatial bacterial-immune crosstalks.

BACKGROUND: White matter hyperintensities (WMHs), the principal imaging manifestation of cerebral small vessel disease, lack an animal model that produces pathology of sufficient magnitude to validate potential therapeutics. We characterized white matter (WM) degeneration in a renovascular hypertension (RVH) rat model, interrogated its mechanism, and tested its relevance in patients. METHODS: RVH was induced by two-kidney, two-clip surgery and assessed by T2-weighted and diffusion MRI, myelin and axonal histology, and blood-brain barrier (BBB) leakage markers. The influence of angiotensin II (Ang II) on barrier function in brain endothelial cells was examined using transendothelial electrical resistance and reactive oxygen species (ROS) assays. Losartan was infused intracerebroventricularly to block brain AT1R (Ang II type 1 receptor) without systemic blood pressure reduction. WMH burden was compared between patients with and without renal artery stenosis, and WMH progression across antihypertensive classes was assessed in the CHALLENGE (Korea), MSS2 (UK), LBC1936 (UK), and SPRINT-MIND-MRI (US) cohorts. RESULTS: Approximately two-thirds of RVH rats developed WMHs that expanded between 7 and 9 weeks, accompanied by reduced fractional anisotropy, demyelination, axonal and mature oligodendrocyte loss, and glial activation. IgG, albumin, and fibrinogen leakage was confined to affected WM. Ang II impaired endothelial barrier function and generated ROS through AT1R. Intracerebroventricular losartan attenuated WM pathology and improved behavior while blood pressure remained comparably elevated. Patients with renal artery stenosis had higher WMH volumes (10.52 versus 6.52 mL; P = 0.035). Angiotensin receptor blocker use was associated with the slowest tertile of annual WMH progression in the Korean cohort (OR, 2.57 [95% CI, 1.20-5.61]) and among White participants in SPRINT-MIND-MRI (OR, 3.52 [95% CI, 1.05-12.29]). CONCLUSIONS: RVH produces prominent ischemic WM degeneration driven by AT1R signaling, identifying a blood pressure-independent therapeutic target with corroborating clinical evidence.

In 2024, Chicago experienced a measles virus outbreak. Air samples collected as part of routine respiratory virus surveillance from three hospitals that received measles patients were retrospectively analyzed. Measles virus was detected in air samples from one emergency department across five weeks when measles patients were present and another when no known cases were present. Two of three viral sequences derived from air samples were identical to measles patients. The source of one diverging sequence could not be identified and may represent a case not identified by traditional surveillance methods. Air sampling may supplement measles investigation, surveillance and response.

Mesaconate (MES, an immunosuppressive metabolite) but not itaconate is upregulated in stool of Crohns Disease (CD) patients, albeit below therapeutic levels.. However, colonic MES level is difficult to maintain because of its high absorption rate, limiting its therapeutic potential for CD. Herein, we developed two strategies that elevate MES level and suppress inflammation: EcN-CAD, an engineered probiotic Escherichia coli Nissle 1917 (EcN) that overexpresses cis-aconitate decarboxylase (CAD, produces itaconic acid that is converted to MES); and MES+Zn, a Zinc-MES ionically crosslinked formulation. Mechanistically, MES modulates T cell function by downregulating glycolytic and mitochondrial activity and decrease ROR{gamma}t expression. In preclinical SAMP1/YitFC dexamethasone-withdrawal model and TNFDARE model, MES+Zn decreased gut permeability significantly compared to controls and reduced myeloperoxidase (MPO, an inflammatory marker) 1000-fold compared to controls. EcN-CAD probiotics decreased ROR{gamma}t expression in CD4+ and CD8+ T cells, gut permeability and inflammation in preventive SAMP1/YitFC model, and SAMP1/YitFC dexamethasone-withdrawal model. Overall, MES rewires T cell responses and technologies that enhance MES levels can be effective in gut-related inflammatory diseases.

Host genetic variation influences heterogeneity in infectious disease severity. Genome-wide association studies (GWAS) have identified 69 lead single nucleotide polymorphisms (SNPs) associated with critical COVID-19, but most lack known regulatory mechanisms. To date, the genetic cis-regulatory effects of critical COVID-19 GWAS lead SNPs have not been systematically characterized across acute illness and convalescent phases in a COVID-19 specific cohort. Here, performing severity-stratified, longitudinal cis-expression quantitative trait loci (cis-eQTL) mapping and colocalization with critical COVID-19 GWAS signals, we find that host cis-regulatory effects vary across clinical severity and disease phase. We analyzed peripheral blood mononuclear cells from 1,031 patients in the Immunophenotyping Assessment in a COVID-19 Cohort (IMPACC), a prospective multicenter study of adults hospitalized with COVID-19, profiled across 10 timepoints spanning acute illness and convalescence. We identified 23 GWAS lead SNPs that colocalized with cis-eQTLs at 26 candidate causal genes, with 16 detectable only when severity stratification and longitudinal sampling were incorporated. A colocalized, protective allele (chr19:4,717,660) at the TMIGD2 locus upregulated TMIGD2 expression exclusively during the acute phase in the prolonged-critical/fatal cohort, with no detectable effect in the rapid-recovery cohort. Our results show that host genetic effects on critical COVID-19 can be context-dependent, with cis-regulatory effects varying across clinical severity and disease phase. Severity-stratified, longitudinal cis-eQTL designs have greater sensitivity to identify candidate causal genes for critical COVID-19 GWAS associations than single-timepoint cross-sectional studies.

Decision models like Jev read text and return probabilities, categories, or ordinal levels, with confidence scores for the latter two, at a fraction of large language models' (LLMs) cost. We explored how Jev could be integrated into clinical workflows for all these output types: mortality risk estimation, categorical Choosing-Wisely decisions for patient questions, and ordinal colon cancer staging. With standard LLM-style inputs, Jev trailed five flagship LLMs: mortality AUROC was 0.822 versus 0.837-0.846; Choosing-Wisely accuracy, 87.7% versus 98.4-100%; and staging accuracy, 45.4% versus 94.2-100%. Developer-recommended input alterations improved staging accuracy to 75.8% with medical context, 82.1% with task decomposition, and 85.8% with both. None of Jev's staging errors had full confidence. Using Jev for fully confident staging decisions and LLMs for the rest was non-inferior to each LLM alone and cut flagship LLM costs by 33-44%. Prompted as designed, decision models may efficiently share structured clinical tasks with LLMs.

BackgroundRoutinely collected health data can reveal recorded care around diagnosis, but disease-specific curation does not scale across thousands of conditions. We scaled a previously developed automated framework that reduces heterogeneous clinical events into standardised, diagnosis-centred summaries of real-world care patterns. MethodsIn this retrospective observational study, we used EST-Health-30, a pseudonymised 30% random sample of Estonian residents with health-care records from 2012 to 2024. We constructed cohorts for ICD-10 three-character diagnostic categories using the first eligible observed diagnosis after a 3-year diagnosis-free observable lookback. For each category, we summarised recorded clinical events from a 90-day pre-index window, a 30-day post-index window, and a 365-day post-index window. An enrichment-based workflow compared earlier self-comparator periods and matched population controls, then filtered and aggregated concepts across clinical domains. Diagnosis-concept relationships were assessed by rubric-guided language-model classification, and atlas plausibility and utility by ten experts. FindingsOf 1645 observed diagnostic categories, 1080 met inclusion criteria, yielding 3240 diagnosis-window summaries across 509,856 individuals. Each summary characterised diagnosis-associated events by prevalence, enrichment, timing, co-occurrence, and variation between patient groups after reducing candidate concepts by 97-98%. Among 296,303 retained diagnosis-concept pairs, the language model classified 24% as directly related, 55% as indirectly related, and 21% as noisy. Experts rated the atlas highly for generating research questions (median 6{middle dot}5 of 7 [IQR 6-7]), providing information difficult to obtain conventionally (6 [5-7]), and supporting observational study design (5{middle dot}5 [5-6]). InterpretationThe atlas makes multidimensional diagnosis-centred patterns in longitudinal health records inspectable at population scale in order to support clinical orientation, hypothesis generation, and observational study planning. Outputs represent recorded care rather than individual patient trajectories or causal effects. FundingEstonian Research Council; European Union; Estonian Ministry of Education and Research; Innovative Medicines Initiative 2 Joint Undertaking; EFPIA. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe searched PubMed and medRxiv from database inception to Aug, 2026, without language restrictions, using terms for electronic health records, real-world data, disease trajectories, treatment and care pathways, treatment patterns, clinical-event analysis, OMOP, and population-scale atlases. We also examined relevant reviews and reference lists. Previous population-scale studies have mapped disease trajectories, treatment pathways, or care intensity, generally focusing on diagnoses alone, prespecified events, selected disease cohorts, or a single summary measure of care. OMOP-based treatment-pattern studies have likewise concentrated on specific disease groups. We found no previous population-scale resource that automatically reduced multidomain clinical events surrounding an index diagnosis across more than 1000 diagnostic categories while retaining prevalence, enrichment, timing, co-occurrence, and patient-level heterogeneity. Added value of this studyWe applied a previously clinically benchmarked concept-reduction workflow to longitudinal OMOP-mapped health data across 1080 ICD-10 diagnostic categories and three observation windows, producing 3240 diagnosis-window summaries from 509,856 individuals. The workflow reduced candidate clinical concepts by 97-98% and generated compact summaries spanning diagnoses, procedures, medications, measurements, observations, visits, and deaths. The resulting interactive atlas combines concept prevalence and enrichment with timing, co-occurrence, and exploratory patient clustering. An internal formative expert evaluation supported its use for generating research questions and planning observational studies. Implications of all the available evidenceAutomated reduction of high-dimensional health-record data can provide a scalable starting point for clinical orientation, hypothesis generation, and observational study design without requiring a separately curated event set for every disease. Applying the same workflow to other OMOP databases could enable direct comparison of diagnosis-centred treatment and care patterns between regions and health systems, while helping distinguish differences in clinical practice from differences in data capture. The atlas is exploratory and should not be interpreted as evidence of causal effects, guideline adherence, or individual patient trajectories.