Critical Care Explorations

Background. Capillary refill time is a resuscitation target in septic shock,1-4 but bedside measurement is examiner-dependent. An ICU monitor co-records a photoplethysmogram on the pulse oximeter and intermittent noninvasive blood pressure cuff cycles; if the probe and the cuff share a limb, each cycle is an unplanned vascular occlusion test on the distal microvascular bed. Standard practice places the two on opposite limbs. Objective. To measure how often, in MIMIC-IV-WDB v0.1.0, charted cuff cycles show the photoplethysmographic morphology expected of a same-limb cuff and probe, and to characterize the candidate capillary refill-like signal when that morphology is present. Methods. MIMIC-IV-WDB v0.1.05 was linked to the MIMIC-IV clinical database.6 A pre-registered rule-based detector identified candidate occlusion-reperfusion signatures on the 1-Hz perfusion-index envelope around each charted cuff timestamp. The primary endpoint was the proportion of cuff cycles suitable for analysis that were detector-positive at a 15-second reperfusion threshold, with 95% confidence intervals estimated by resampling patients at a fixed seed. A secondary analysis used a locally hosted multimodal language model (a Gemma-3 derivative on a non-device server) to adjudicate the same signature on perfusion-index plots; no MIMIC-IV-WDB content left the workstation. Results. Of 9,224 charted cuff cycles, 8,909 had a usable pulse-oximeter waveform, and 268 cycles in 15 patients (4.30% of the 6,236 cuff cycles suitable for analysis, 95% CI 2.60 to 6.03) met the primary 15-second threshold. The language model adjudicated the same cycles and called 1,367 of the 8,909 cycles with a usable waveform (15.34%) signature-present, roughly five times the detectors count. Because no laterality ground truth exists, agreement with a single blinded reader served as the comparator rather than accuracy. The two methods were about equally concordant with the reader: precision was 0.25 (95% CI 0.14 to 0.39) for the detector and 0.24 (95% CI 0.10 to 0.35) for the language model, although reweighting to the full population of cycles with a usable waveform lowered the language model to 0.030 (95% CI 0.009 to 0.053). These estimates are reference-limited: a blinded re-read of a 150-card subsample showed only moderate intra-rater reliability (Cohen {kappa} 0.46 to 0.59) with systematic undercalling on the first pass, and rescoring against the corrected re-read roughly doubled precision for both methods. Conclusions. Opportunistic extraction of capillary refill-like signals from archived ICU pulse oximetry is limited in two distinct ways. First, sensor geometry limits how often the signal is recordable: cuff cycles rarely show the morphology expected of a same-limb cuff and probe pair, consistent with opposite-limb placement, so the bottleneck is geometry rather than signal processing. Second, the modest reliability of morphology adjudication limits how well any single flagged cycle can be confirmed: against a blinded reader the detector is a usable screen but a noisy confirmer, the reference is itself only moderately reliable, and the language model is no more concordant despite flagging many more cycles. The minority of cycles in which the morphology appears contain a candidate signal that may merit prospective study under controlled placement with laterality recorded.

ABSTRACT Background: Corticosteroids reduce mortality in severe COVID-19 requiring oxygen or invasive mechanical ventilation, yet emerging data suggest that SARS-CoV-2-associated acute lung injury is biologically heterogeneous and that treatment response may vary across molecularly defined disease states. Lung-derived molecular endotypes of severe COVID-19-associated acute lung injury have been described, but direct molecular profiling is not routinely available at the bedside. We evaluated whether a clinical predictor of previously defined lung molecular endotype identifies heterogeneity in corticosteroid treatment effect among mechanically ventilated patients with COVID-19. Methods: We utilized a single-center cohort of 5,000 patients with COVID-19 treated at the University of North Carolina Hospital between January 1, 2020, and December 31, 2022, to emulate a target trial assessing the effect of corticosteroid receipt on mortality, length of stay, and incident organ support. Confounding was addressed through inverse probability of treatment weighting (IPTW). Outcomes for severely ill patients requiring mechanical ventilation were compared to the RECOVERY trial results, with subsequent moderation analysis and stratified analysis by clinically predicted lung molecular endotype and vaccination status. The primary outcome was 28-day mortality. Secondary Outcomes were time to discharge alive and progression to additional organ support. Results: This emulated target trial showed a directionally favorable but non-statistically significant association between corticosteroid treatment and reduced 28-day mortality in patients requiring mechanical ventilation for SARS-CoV-2 infection. A clinical predictor of lung molecular endotype moderated the effect of corticosteroids on 28-day mortality (p-value for interaction 0.038) and identified distinct predicted endotype-specific treatment effect. Corticosteroid treatment was associated with lower 28-day mortality in the predicted Hyper-Inflammatory endotype (OR 0.62, 95% CI 0.39, 0.99) but not in the predicted Metabolic Dysregulation endotype (OR 1.15, 95% CI 0.82, 1.61). We did not detect significant effect modification by vaccination status (p-value for interaction 0.65), although inference was limited by the small, vaccinated subgroup (28-mortality OR 0.78, 95% CI 0.37, 1.65 in vaccinated vs 0.94, 95% CI 0.70, 1.26 in unvaccinated). Conclusions: In this target trial emulation of mechanically ventilated patients with severe COVID-19, corticosteroid treatment showed a directionally favorable but non-statistically significant association with reduced 28-day mortality in the overall cohort. However, a clinical predictor of lung molecular endotype identified significant heterogeneity in treatment effect, with benefit concentrated in the predicted Hyper-Inflammatory endotype and no apparent benefit in the predicted Metabolic Dysregulation endotype. These findings support prospective validation of clinically deployable endotype-guided corticosteroid treatment strategies in acute lung injury and ARDS.

Background. Capillary refill time, an examiner-dependent bedside test of distal microvascular perfusion, has become a resuscitation target in septic shock,1,2,3,4 motivating a continuous surrogate computed from the photoplethysmogram (PPG, the optical waveform the pulse oximeter on every ICU patient already records).5,6,7,8 Objective. We attempted three PPG-derived candidate measures on the MIMIC-IV Waveform Database (MIMIC-IV-WDB v0.1.0) and asked, by inspecting randomly drawn examples, whether each captured its intended physiology before any downstream modeling. Methods. MIMIC-IV-WDB v0.1.09 was linked to MIMIC-IV.10 The signals were a cuff-anchored perfusion-index recovery (reactive hyperemia when the cuff shares an arm with the probe), a slow Mayer-wave-band power ratio of the perfusion index (sympathetic vasomotor tone), and a per-beat diastolic exponential decay time constant (a refill-like recovery time). For each signal we drew 10 random examples at a fixed seed and checked them against a checklist fixed in advance. Each was read by the author and, separately, by MedGemma 1.5, a multimodal medical language model run locally. A synthetic test with a known time constant checked the third signal. Results. The cuff-anchored signal showed the expected occlusion-reperfusion shape on 268 of 6,236 evaluable cuff cycles (4.30%) in 15 of 19 patients, consistent with opposite-limb placement of the probe and cuff. The slow-band ratio returned a stable cohort value, but a clear, stationary peak appeared in only4 of 10 random windows. The per-beat fit met its goodness-of-fit threshold in 10 of 10 beats, yet a cardiac-frequency heuristic flagged a possible fit on the heart-rate oscillation in 7 of 10, and in 5 of 17 patients the time constant lay where an exponential is indistinguishable from a straight line. A 0.5Hz high-pass pre-filter implanted its own approximately 318 ms time constant regardless of truth. The language model tracked the human on clear positives but reported the pattern present on every call it returned, never absent. Conclusions. Two of the three candidate signals did not reflect their intended physiology in most examples, and the third was constrained by sensor placement. Inspecting a few random raw inputs against a checklist written in advance is an inexpensive upstream check before downstream inference on PPG-derived microvascular signals.

Background Electronic health record documentation patterns may reflect workflow complexity, monitoring intensity, and operational strain in intensive care settings. However, documentation-derived features can be sensitive to local documentation culture, data capture systems, and outcome definitions. Retrospective validation across multiple datasets is therefore needed before these signals are used in workflow intelligence or clinical AI governance tools. Objective To evaluate whether documentation-density and documentation-timing features show reproducible retrospective signal for ICU workflow complexity and long-stay proxy outcomes across de-identified critical care datasets, while distinguishing workflow and long-stay associations from unsupported claims about mortality prediction, burden reduction, or deployment readiness. Methods We synthesized retrospective validation results from de-identified ICU and workflow datasets generated through a prespecified documentation-density validation program. Feature families included Documentation Burden Score style features, Shift-End Documentation Rate style features, documentation reliability style metadata, and all-documentation feature sets where available. Outcomes included long ICU length of stay proxies, mortality where available, and workflow proxy endpoints. Models compared baseline feature sets with enhanced models containing documentation-density or workflow features. Performance was summarized using area under the receiver operating characteristic curve, Brier score where reported, delta AUROC, bootstrap confidence intervals where reported, and label-shuffle controls where available. Results The strongest external long-stay proxy evidence came from the NWICU chartevents analysis, which included 28,612 ICU stays, 20,267 stays with chart events, and 9,619,759 chart events. For ICU length of stay greater than the median, baseline AUROC was 0.5252. Enhanced AUROC was 0.9512 for Documentation Burden Score features, 0.9214 for Shift-End Documentation Rate features, 0.8470 for documentation reliability style features, and 0.9517 for all documentation features. Corresponding label-shuffle enhanced AUROCs were near random, ranging from 0.4897 to 0.5064. For ICU length of stay greater than the 75th percentile, baseline AUROC was 0.5155. Enhanced AUROC was 0.9433 for Documentation Burden Score features, 0.9194 for Shift-End Documentation Rate features, 0.8118 for documentation reliability style features, and 0.9427 for all documentation features, with label-shuffle enhanced AUROCs from 0.4836 to 0.4999. Additional retrospective support was observed in eICU workflow analyses, HiRID first-24-hour documentation-density analyses, MIMIC-IV HF ICU internal analyses, MIMIC-IV-Note metadata extensions, and nursing-chart or lab density proxy analyses. However, cross-institution discrimination transfer was weak without recalibration, and several analyses remained proxy validations rather than final clinical validations. Conclusions Documentation-density and documentation-timing features show promising retrospective signal for ICU workflow complexity and long-stay proxy outcomes, especially in NWICU chartevents and selected internal dataset-specific analyses. These findings support further preregistered, prospective, silent-mode validation of documentation-derived workflow intelligence. They do not establish prospective clinical performance, mortality reduction, clinician burden reduction, autonomous deterioration prediction, or deployment readiness.

Background World Symposium on Pulmonary Hypertension (WSPH) Group 2 pulmonary hypertension (PH) is a clinically integrated phenotype attributed to left heart disease, whereas pre- versus post-capillary classification is operationalized primarily by pulmonary capillary wedge pressure (PCWP). Although current recommendations emphasize contextual interpretation and provocative testing for intermediate PCWP values, the relationship between PCWP-based classification and underlying phenotype has not been systematically evaluated. We aim to quantify phenotype-hemodynamic discordance across the PCWP spectrum and evaluate a staged physiology-guided framework incorporating inhaled nitric oxide (iNO), ventricular geometry, and provocative testing. Methods We studied 1,032 participants from the NHLBI-sponsored PVDOMICS cohort with multidisciplinary adjudicated phenotypes integrating clinical, imaging, physiologic, and hemodynamic data. Stage-specific PCWP thresholds classified pre- versus post-capillary physiology at rest, during iNO, and during provocation (fluid challenge or invasive cardiopulmonary exercise testing [iCPET]). Echocardiographic right ventricular-to-left ventricular (RV/LV) ratio was evaluated as a marker of ventricular interdependence. Restricted cubic spline and staged concordance analyses defined certainty-based PCWP ranges and incremental diagnostic yield. Results Adjudicated Group 2 phenotype was present in 37.0% of participants. Resting PCWP demonstrated good discrimination (AUC 0.86), but substantial bidirectional phenotype-hemodynamic discordance persisted across intermediate PCWP ranges. At a resting PCWP of 12 mmHg, 25% of participants classified as pre-capillary had adjudicated Group 2 PH, whereas at 18 mmHg, 35% classified as post-capillary remained discordant non-Group 2. Concordance did not approach 90% until PCWP values were <9 mmHg or >24 mmHg. Dynamic testing incrementally improved concordance within these overlap zones. Nearly half of adjudicated Group 2 PH participants (46.5%) were not identified by resting PCWP alone; incorporation of iNO and provocative testing increased cumulative Group 2 identification by 63.4% and improved sensitivity from 79.9% to 83.7%. Model discrimination improved from an AUC of 0.863 to 0.908 (likelihood-ratio P<0.001). iNO increased PCWP in discordant Pre/G2 participants, unmasking latent left-sided limitation, while lowering PCWP in discordant Post/NonG2 participants, consistent with ventricular interdependence. RV/LV ratio [&ge;]0.94 reduced discordant Post/NonG2 classification by 70.5%, and incorporation of PCWP/cardiac output slope improved physiologic specificity during exercise. Conclusions Group 2 PH is a dynamic, load-dependent phenotype inadequately characterized by resting PCWP alone. Intermediate PCWP values represent continuous probabilities of bidirectional discordance rather than discrete diagnostic states. A staged physiology-guided approach integrating iNO, ventricular geometry, and provocative testing improves concordance between hemodynamic classification and clinically integrated phenotype assignment.

High-frequency physiological monitoring in ICUs can identify impending deterioration hours before clinical recognition yet extracting reliable early-warning signals from noisy vital-sign streams remains challenging. We present SIgnose, an interpretable prediction framework for early detection of abnormal shock index (SI), built from routinely monitored vital signs using physiologic variability and nonlinear time-series features. SIgnose was developed on the eICU Collaborative Research Database and externally validated on the MIMIC-III adult database and a pediatric SafeICU cohort (AIIMS New Delhi), with additional prospective validation in the pediatric ICU. We benchmarked three representation strategies: (i) engineered physiologic variability and nonlinear time-series features, (ii) deep learning, and (iii) Llama-3.1-8B embeddings with low-rank adaptation. Physiologic variability features consistently demonstrated superior cross-cohort generalization. The final model used 3,970 features from five vital signs to predict abnormal SI up to 8 hours ahead, achieving AUROC 0.861 (95% CI 0.859-0.863) and AUPRC 0.927 (95% CI 0.925-0.929) on eICU. External validation yielded AUROC 0.870 (95% CI 0.863-0.876) and AUPRC 0.935 (95% CI 0.930-0.940) on MIMIC-III, and AUROC 0.875 (95% CI 0.863-0.888) and AUPRC 0.915 (95% CI 0.898-0.930) on SafeICU; prospective pediatric validation (n = 88) achieved AUROC 0.885 (95% CI 0.868-0.902) and AUPRC 0.911 (95% CI 0.882-0.936). SHAP interpretability analysis identified heart rate variability, respiratory trend dynamics, and multi-scale blood pressure variability as key early-warning signatures. These findings establish SIgnose as a reproducible, low-compute, early-warning framework and demonstrate that physiologic variability features provide robust, generalizable representations for early deterioration detection across adult and pediatric critical care.

Background: Infants with critical congenital heart disease (CHD) are at high risk for abnormal brain development and later neurodevelopmental impairment. We hypothesized that the trajectory of perioperative whole-brain network development would predict neurodevelopmental outcomes in early childhood. Methods: This prospective longitudinal cohort of neonates with critical CHD (n = 97) underwent preoperative and/or postoperative brain MRI with diffusion imaging. Whole-brain network measures were derived from structural connectomes. Neurodevelopment was assessed between 1 and 4 years using the Bayley Scales of Infant and Toddler Development. Results: White matter injury was associated with slower perioperative growth in global efficiency (p = 0.013), a measure of network integration, whereas cardiac physiology was not associated with network development. Infants with greater perioperative increases in global efficiency had higher cognitive (p = 0.001), language (p < 0.001), and motor (p = 0.008) scores. For each 1-standard deviation increase in the trajectory of global efficiency, cognitive scores increased by 8.2 points (95% CI, 3.64-12.78), independent of brain injury and socioeconomic factors. Conclusion: In infants with critical CHD, longitudinal whole-brain network development was associated with neurodevelopment across multiple domains. Early network development may represent a candidate biomarker of neurodevelopmental risk and resilience in this population.

Background Emergency department (ED) triage is a high-stakes clinical decision process that determines patient prioritization and resource allocation under time pressure. Large language models (LLMs) have recently been proposed as decision-support tools for triage, yet most evaluations rely on simulated scenarios or curated datasets. Evidence from real-world clinical environments remains limited. The objective of this project was to systematically evaluate the performance, calibration, and reproducibility of multiple contemporary large language models for Emergency Severity Index (ESI) classification and sectoral allocation (ED vs. urgent care practice, UCP) using a comprehensive real-world triage dataset. Material and Methods Retrospective cross-sectional benchmarking study conducted at a tertiary academic emergency ED in Germany with an integrated central point of assessment (CPA). The study included all consecutive adult walk-in encounters (>18 years) presenting between October 2023 and February 2024 (N = 16,107). Data were collected from a structured clinical decision support system capturing presenting complaints, vital signs, and triage decisions recorded by specialized nursing staff. Structured clinical variables routinely collected at triage, including presenting complaint categories (CEDIS-PCL), vital signs according to the ABCDE framework, and additional structured or free-text clinical information. Results The primary outcome was the agreement between LLM-predicted and nurse-assigned ESI levels measured using quadratic-weighted Cohen's k. Secondary outcomes included sectoral assignment agreement, misclassification patterns (over- and under-triage), calibration metrics, and output reproducibility. Quadratic-weighted k values ranged from 0.18 to 0.75 across models. Only a structured stepwise prompting strategy achieved substantial agreement (k_qw = 0.747), approaching reported human inter-rater reliability. Most models demonstrated moderate or lower agreement and systematic overconfidence, with expected calibration errors (ECE) based on verbalized confidence ranging from 0.099 to 0.355. Sectoral assignment agreement (i.e. ED vs. urgent care practice, UCP) was uniformly low (k < 0.30). Reproducibility testing revealed substantial variability in 23% of cases, indicating non-deterministic output behavior for clinically relevant decisions. Conclusions Current large language models demonstrate heterogeneous and generally limited performance in real-world emergency triage tasks. Structured algorithm-guided prompting appears more influential than model architecture or size. Before clinical implementation, improvements in calibration, reliability, and workflow integration are required, alongside regulatory-compliant validation in prospective clinical settings.

Background: Most approaches to suicide risk assessment consider clinical conditions as independent risk factors, potentially overlooking prognostic information in the order in which conditions accumulate. We applied temporal sequence mining to linked claims and mortality data to identify ordered clinical diagnostic trajectories associated with suicide death. Results: The cohort included 3 647 059 insured Maryland residents aged 10 years or older with available claims records in the Maryland Suicide Data Warehouse from January 1, 2016, to December 31, 2020, among whom 768 suicide deaths were ascertained through medical examiner linkage. Sequential pattern mining of ICD-10-CM diagnoses grouped into Clinical Classifications Software Refined categories identified 89 221 candidate sequences, of which 1 816 remained significantly associated with suicide death in time-varying Cox models. Adjusted hazard ratios (AHRs) ranged from 2.4 to 134.1. Two-thirds of significant trajectories ended in physical conditions, and approximately half crossed from psychiatric to physical endpoints. Among suicide decedents, 62% were exposed to at least 1 significant sequence (median, 16 per case); median sequence duration was 18.7 months, and median time from completion to death was 13.1 months. In landmark analyses, among patients with depression who later developed suicidal ideation (n = 26 356), the path through anxiety, then anemia, was associated with higher risk (AHR, 4.6; 95% CI, 2.2-9.5), whereas the anxiety-only path was not (AHR, 1.3; 95% CI, 0.8-2.1). Among patients with anxiety who later developed hypertension (n = 149 215), the path through history of self-harm was associated with higher risk (AHR, 32.0; 95% CI, 16.6-61.6). Associations were generally consistent across sex and age. Conclusions: Temporal ordering of clinical conditions may carry prognostic information for suicide death. Clinical trajectories incorporating physical illness within psychiatric sequences identified higher-risk groups. These findings suggest that opportunities for risk detection may extend beyond psychiatric settings and that suicide risk signals may be fragmented across care settings and not apparent within isolated encounters.

**Abstract** **Background:** Transcatheter edge-to-edge repair (TEER) is an established treatment for mitral regurgitation but remains highly dependent on operator experience and complex transesophageal echocardiography (TEE)-guided intraprocedural imaging. Artificial intelligence (AI)-based semantic segmentation may improve procedural reproducibility and intraprocedural guidance; however, no TEER-specific segmentation framework has been reported. **Objectives:** To develop and evaluate AutoClip, a clinician-driven AI-guided TEE semantic segmentation model designed for simultaneous delineation of mitral valve anatomy and in-vivo TEER device components. **Methods:** A retrospective proof-of-concept study was conducted using 987 intraprocedural TEE frames derived from 10 video clips in 3 patients undergoing MitraClip G4 implantation. Seven semantic labels, including mitral leaflets and device components, were manually annotated using ITK-SNAP. Following standardized preprocessing and region-of-interest extraction, an Attention U-Net architecture was trained frame-wise on bicommissural and corresponding X-plane TEE views. Model performance was assessed using mean intersection-over-union (IoU) and Dice coefficient on an independent test set. **Results:** The Attention U-Net demonstrated improved sensitivity to small device structures compared with conventional U-Net architectures. Preliminary training performance achieved a mean IoU of approximately 0.93, while independent test performance reached a mean IoU of 0.46 across foreground classes. Qualitative assessment demonstrated feasible simultaneous segmentation of mitral leaflets, clip arms, grippers, and delivery shaft during TEER procedures. **Conclusions:** AutoClip represents a proof-of-concept TEER-specific TEE semantic segmentation framework initiated through a clinician-oriented workflow without formal computer science expertise. Although preliminary accuracy remains modest due to limited sample size, this study establishes a reproducible pathway for future AI-assisted intraprocedural guidance systems and larger multicenter development efforts in structural heart interventions.

Background and Purpose: Hemorrhagic transformation (HT) is a serious complication of endovascular thrombectomy (EVT), yet dedicated prediction models for young adults are lacking. We aimed to develop and externally validate a simplified risk score for HT in young adults with acute ischemic stroke undergoing EVT. Methods: This multicenter retrospective study included patients aged 18 to 49 years with acute anterior circulation large vessel occlusion who underwent EVT. The primary outcome was any HT within 24 hours after EVT. Multivariable logistic regression was used to identify independent predictors of HT, from which the NO?PAIN Score was derived. External validation was performed in an independent cohort of 138 patients. Results: Among 598 patients in the derivation cohort, HT occurred in 176 (29.4%). Five independent predictors were identified: admission NIHSS, number of thrombectomy passes, atrial fibrillation, alcohol consumption, and mTICI grade. The mTICI grade demonstrated a non-linear, inverted U-shaped relationship with HT risk, peaking at partial recanalization. The NO-PAIN Score showed acceptable discrimination in both the derivation (C-index, 0.737; optimism-corrected C-index, 0.748) and external validation cohorts (C-index, 0.726), with satisfactory calibration. Conclusions: The NO-PAIN Score is a simple risk prediction tool for HT after EVT in young adults with acute anterior circulation large vessel occlusion. It may assist in individualized risk stratification in this population.

INTRODUCTION Severe acute brain injury (stroke, traumatic brain injury or hypoxic-ischemic encephalopathy; SABI) is increasingly recognized as a chronic condition with care and communication needs beyond the initial hospitalization. This study aimed to characterize post-acute care patterns among SABI survivors, focusing on healthcare utilization and outpatient communication. METHODS Data were collected from a prospective cohort of hospitalized SABI patients using surveys, chart reviews, and the ED Information Exchange database. Socioeconomic disadvantage was assessed using the Area Deprivation Index (ADI), and qualitative analysis of outpatient notes examined conversations around palliative care needs and goals-of-care. RESULTS Two-thirds of patients (140/222) survived until discharge, primarily to nursing facilities (39%) or inpatient rehabilitation (38%). Among 109 with one-year follow-up, there were 89 hospitalizations, 104 ED visits, and 28 deaths. Patients from the most disadvantaged neighborhoods had significantly higher odds of rehospitalization or ED use within 30 days (OR 3.37, p=0.036). ADI was not linked to one-year utilization. seen outpatient by primary care (40%), neurology/neurosurgery (57%), and palliative care (1%), but conversations rarely revisited prognosis or goals-of-care. CONCLUSIONS Our findings highlight the need for improved long-term care planning and communication, particularly for socioeconomically disadvantaged survivors of SABI.

Background. Fairness-aware machine learning increasingly targets demographic performance disparities in clinical prediction, yet whether standard bias mitigation strategies genuinely improve equity in physiological signal analysis remains unclear. Age-based disparities in photoplethysmography (PPG)-based heart rate prediction present a particular challenge, as age-related performance differences may reflect context-dependent physiological structure rather than correctable artifacts. Methods. We evaluated three fairness interventions, inverse-frequency weighting (IF), Group Distributionally Robust Optimization (GroupDRO), and adversarial debiasing (ADV), applied via fine-tuning of a PPG foundation model across three clinical datasets spanning intensive care unit, laboratory, and consumer wearable contexts. Outcomes were assessed using a 2x2 framework classifying each intervention-dataset combination by the joint direction of change in mean absolute error (MAE) and fairness gap (FG) across age groups, yielding four outcome types: genuine improvement (G), leveling down (L), selective benefit (S), and both worse (W). Results. Across nine intra-domain conditions, no intervention simultaneously improved both MAE and FG (0/9 genuine improvement). The dominant pattern was leveling down (5/9): FG decreased but was accompanied by MAE degradation, indicating that apparent fairness gains were achieved at the cost of overall predictive performance. Age-group difficulty ordering varied across clinical contexts at baseline and was not preserved under intervention. In 18 cross-domain transfer conditions, genuine improvement was rare (4/18) and observed exclusively in non-MIMIC source configurations; models fine-tuned on MIMIC-sourced data yielded no genuine improvements (0/6). Embedding-level representation changes following fine-tuning did not reliably predict fairness outcomes. Conclusions. Age-based fairness interventions in PPG heart rate prediction indicate a leveling-down pattern rather than genuine equity improvement, suggesting that age-related performance gaps reflect context-dependent physiological structure not fully addressable through standard bias mitigation. Cross-domain transfer further amplifies this instability. These findings suggest that fairness evaluation frameworks for age-stratified physiological prediction should account for context-dependent performance structure rather than treating observed gaps as correctable bias.

Objective: Clinical phenotyping methods that rely on clinical and informatics expertise can be time-intensive and costly. We tested both manual and highly automated approaches using electronic health record (EHR) data to identify an FDA Sentinel Initiative health outcome of interest, acute pancreatitis. Materials and Methods: We trained and evaluated machine learning algorithms using EHR data with two approaches: a custom approach that included manually curated features and trained on outcomes data validated with medical record review, and a highly automated approach that greatly simplifies and automates feature engineering and relies on low-cost silver-standard outcomes for model training. Results: Custom algorithms using manually curated structured claims data discriminated cases from non-cases with a high degree of accuracy (cv-AUC 0.89 [95%CI 0.84-0.94]); the inclusion of natural language processing (NLP)-derived covariates from clinical notes increased performance slightly (cv-AUC 0.91[95%CI 0.86-0.97]). The automated algorithm trained on the outcome count of diagnosis codes performed less well (AUC 0.80 [95% CI 0.75-0.85]) but improved using maximum lipase value as an outcome (AUC 0.88 [95% CI 0.84-0.92]). At a positive predictive value of 90%, the custom algorithm had a sensitivity of 92%, the automated algorithm trained on diagnosis code count had a sensitivity of 45%, and the automated algorithm trained on maximum lipase value had a sensitivity of 84%. However, a prediction rule derived by clinicians during chart review was nearly as accurate (maximum lipase value [&ge;] 3 times upper limit of normal; AUC 0.86, PPV 85%, sensitivity 92%). Discussion: Machine learning algorithms with manually curated structured data and NLP features trained on validated outcomes data successfully identified validated events. Use of an outcome in the automated model based on specific phenotype knowledge (maximum lipase value) allowed for performance similar to the custom model and with considerably less resources.

Background: Red cell distribution width (RDW), a readily available hematological parameter reflecting erythrocyte size heterogeneity, has been increasingly recognized as a prognostic marker in congestive heart failure (CHF), with elevated levels independently associated with adverse outcomes. However, RDW-derived composite indices-particularly the RDW-to-platelet ratio (RPR) and RDW-to-hemoglobin ratio (RHR), which integrate inflammatory, hemostatic, and oxygen-delivery pathways-remain largely unexplored in CHF populations. Whether these indices provide incremental prognostic value beyond RDW alone in critically ill patients with CHF has not been established. Methods: This retrospective cohort study included 30,409 participants from the MIMIC-IV and eICU-CRD databases. Multivariable logistic regression, restricted cubic spline (RCS) analysis, and subgroup analyses were employed to evaluate the associations between RDW, RDW-derived indices (RPR and RHR), and in-hospital mortality in patients with congestive heart failure. Results: Based on a pooled cohort of 30,409 patients with CHF from the MIMIC-IV and multi-center eICU-CRD databases (15,983 and 14,426, respectively), 16,295 (53.6%) were male and 14,114 were female, with a median age of 71.7 years. The mean RDW was 16.0 {+/-} 2.5, and the overall in-hospital mortality rate was 12.6%. Higher RDW quintiles were associated with progressively increased in-hospital mortality. In the fully adjusted model, RDW, RPR, and RHR were all significantly associated with increased in-hospital mortality, with adjusted odds ratios (ORs) of 2.46 (95% CI: 2.17-2.79) for RDW, 1.55 (95% CI: 1.38-1.73) for RPR, and 2.43 (95% CI: 2.09-2.82) for RHR. Sensitivity analyses using restricted cubic splines demonstrated that the association between RDW and RHR with in-hospital mortality was linear (P for nonlinearity > 0.05), whereas that for RPR exhibited a non-linear pattern (P = 0.02 for non-linearity). Conclusions. Elevated RDW, RPR, and RHR were independently associated with increased in-hospital mortality in patients with congestive heart failure. Notably, RPR exhibited a non-linear threshold association with in-hospital mortality.

Background: Machine learning models trained on population health surveys offer scalable tools for cardiovascular screening, but recurring methodological weaknesses undermine their credibility and equity: data leakage from synthetic oversampling, qualitative rather than quantitative explainability evaluation, and the absence of demographic fairness auditing at the clinical operating threshold. Methods: We present EXHEART, a leakage-free stacked ensemble pipeline trained on BRFSS 2015 (n = 253,680) and validated on BRFSS 2020 (n = 319,795; temporal transport and retrain) and a clinical cardiovascular examination dataset (n = 68,730). The pipeline combines XGBoost, LightGBM, Random Forest, and a multi-layer perceptron as base learners with 5-fold out-of-fold logistic regression stacking and Platt scaling calibration. A quantitative SHAP-LIME consistency framework, based on Kendall-tau rank correlation and Jaccard overlap, accompanies a decision-curve analysis, a subgroup-stratified SHAP interaction analysis, and an intersectional fairness audit (Sex x Age x Income) with threshold-shifting mitigation and a frontier of the fairness-utility trade-off. The framework also adds cross-instrument fairness-disparity attribution, an empirical diagnostic that provides evidence on whether an observed subgroup disparity is more consistent with a measurement-induced or a substantive explanation by re-validating it on a dataset that measures the same clinical construct objectively. On heart disease, this diagnostic associates 89% of the sex TPR gap (95% CI [0.65, 0.99]) with the self-reported survey outcome rather than with a substantive risk difference. Results: On BRFSS 2015, EXHEART achieves AUC-ROC = 0.850, AUPRC = 0.371, Brier score = 0.071, and reduces ECE by 96% (0.256 to 0.011) via Platt scaling. Global SHAP-LIME rank agreement is moderate-to-strong (Kendall-tau = 0.580, Spearman-rho = 0.818) with a substantial top-3 divergence (Jaccard@3 = 0.200), where Stroke flips from SHAP rank 8 to LIME rank 1. The Sex TPR gap is 0.124 at the screening threshold; intersectional Sex x Age disparities reach 0.649 among adequately-powered cells, 5.2x the single-attribute gap. Temporal transport to BRFSS 2020 collapses sensitivity from 0.776 to 0.267, while retraining restores AUC = 0.840 and ECE = 0.012. On clinical examination data, the Sex TPR gap collapses to 0.014; the attribution test indicates this gap is instrument-dependent, consistent with a measurement or outcome-definition explanation rather than a substantive risk difference. Cross-domain SHAP analysis identifies four instrument-independent CVD risk factors and two major portability failures. Conclusions: EXHEART combines three practices that population-scale cardiovascular classifiers usually apply in isolation: leakage-free training with calibrated probabilities, a test of whether the model's explanations are stable, and a fairness audit that examines intersecting subgroups rather than single attributes. Bringing them together proved worthwhile. The intersectional audit revealed disparities that single-attribute auditing missed, and the cross-instrument comparison indicated that much of the sex gap reflects how the outcome is measured in survey data rather than a substantive difference in risk. The temporal transport findings indicate that deployed BRFSS models warrant periodic monitoring and retraining to maintain clinical utility. EXHEART is a retrospective methodological evaluation on public de-identified data; it is not validated for direct clinical decision-making, diagnosis, or treatment recommendation without prospective clinical validation.

Background: Aortic stenosis (AS) is a progressive valvular disease associated with poor prognosis once symptoms develop, yet routine echocardiographic screening is impractical. While artificial intelligence (AI)-based electrocardiogram (ECG) models have shown promise for AS detection, it remains unclear whether they primarily reflect conventional left ventricular hypertrophy (LVH) voltage criteria or capture additional ECG features. Methods and Results: We developed a deep learning model using 244,816 ECGs from 51,713 patients across six academic institutions in Japan (CLIDAS database). AS labels were derived from inpatient Diagnosis Procedure Combination (DPC) codes. The model achieved an area under the receiver operating characteristic curve (AUC) of 0.849 (95% confidence interval 0.832-0.865) in the independent test cohort, with consistent performance across institutions, sex, and age. At a threshold of 0.1, sensitivity was 79.1%, specificity was 73.9%, and negative predictive value (NPV) was 98.0%. Conventional LVH voltage criteria (Sokolow-Lyon AUC 0.706; Cornell AUC 0.692) showed lower performance, and adding them to the AI model conferred no incremental benefit (AUC 0.849 vs. 0.847). Gradient-weighted class activation mapping (Grad-CAM) revealed predominant attention around QRS complexes in limb leads, beyond regions typically assessed in LVH evaluation. Conclusions: This multicenter AI-ECG model demonstrated strong discrimination for AS and captured ECG features beyond conventional LVH voltage criteria. The high NPV supports its use as a rule-out pre-screening tool.

In February 2026, the FDA announced that a single pivotal phase 3 (P3) trial would become the new default standard for drug approval - a regulatory direction that had been legally enabled since the FDA Modernization Act of 1997. This announcement has strategic, scientific, and economic implications for drug developers, contract research organizations (CROs), and biotech investors. We argue that the expansion of this framework, originally reserved for various niche submissions, represents a paradigm change, dramatically increasing the value of rigorous early phase (P1 and P2) trial design, requiring sponsors to establish both statistical efficacy signals and mechanistic biological understanding before entering phase 3. Using a CNS indication cost model, we show that single P3 approval can reduce total development expenditure from approximately $447 million over 14 years to $297 million over 12 years - a savings of $150 million and providing two years of additional commercial runway for a modeled CNS drug. Case examples including lecanemab, omaveloxolone, and tofersen illustrate how biomarker-informed early phase strategies can establish the confirmatory evidence necessary for single-trial approval. We provide practical guidance for maximizing the value of P1 and P2 under this evolving framework.

Objective: To evaluate whether a locally hosted open-weight large language model (LLM) can extract documented psychosocial factors from pediatric psychiatric intake notes and apply validated extraction to a large emergency psychiatry cohort. Materials and Methods: We identified emergency department presentations at Cincinnati Children's Hospital Medical Center from January 1, 2016, through December 31, 2024, among patients younger than 18 years with psychiatric billing diagnoses. Using full-text intake notes, gpt-oss:120b classified peer conflict, sleep disruption, and school-related academic, attendance, and disciplinary issues as detected, negated, or indeterminate. Four human raters independently reviewed 50 notes. We compared Fleiss' kappa among humans alone versus humans plus the LLM, assessed repeated-query stability across 50 independent calls per note, and applied the workflow to all eligible notes. Results: Among 37,315 eligible admissions, 22,284 had eligible intake notes; 22,270 produced parseable JSON. In detected-versus-not-detected coding, human-plus-LLM reliability did not differ significantly from human-only reliability across measures (human {kappa} 0.71-0.94; human-plus-LLM {kappa} 0.70-0.93). Stability was associated with human agreement: mean LLM-human agreement increased from 42.6% for classifications with less than 80% stability to 82.7% for classifications with 100% stability (Pearson r = 0.36). Full-cohort extraction showed frequent and overlapping documented factors: sleep disruption was most frequently detected (57.7%), followed by peer conflict (47.2%), academic issues (43.4%), disciplinary issues (43.3%), and attendance issues (16.9%). Discussion: Agreement varied by construct and was strongest when repeated model outputs were stable. Conclusion: Locally hosted open-weight LLMs can support scalable structured extraction of documented psychosocial factors from pediatric psychiatric intake notes after local validation.

End-of-rotation handoffs are critical for patient safety but add to documentation burden for hospitalists. Generative artificial intelligence (AI) may help automate handoff creation using electronic health record data, but its impact on quality and safety is unclear. Methods: We developed an AI handoff tool with a large language model using clinical notes as input and conducted a retrospective evaluation comparing AI-generated and clinician-authored handoffs. Handoffs were assessed across domains of quality and safety through a structured review. Results: Quality ratings were similar between AI and human handoffs (3.7 vs. 3.5, p=0.57). AI-generated handoffs were rated higher for organization (4.4 vs. 4.1, p=0.05) and completeness (4.1 vs. 3.6, p=0.01), but lower for conciseness (3.7 vs. 4.1, p=0.03) and accuracy (4.1 vs. 4.4, p=0.03). Error rates were comparable (0.3/handoff in both groups); however, AI-generated handoffs included inaccuracies (9% of AI errors) and hallucinations (1% of AI errors), while clinician-authored handoffs contained only omissions. Conclusion: Human and AI handoffs have differing error profiles and tradeoffs between completeness and conciseness. Prospective evaluation in clinical workflows is underway.