Journal of Biomedical Informatics

Metabolic dysfunction-associated steatotic liver disease (MASLD) is highly prevalent among individuals with type 2 diabetes (T2D), yet accurate identification from electronic health records (EHRs) remains challenging because clinically relevant information is fragmented across structured data and unstructured clinical documentation. We developed a rule-based computable phenotype (CP) framework that integrates structured EHRs with clinical notes for scalable MASLD identification in patients with T2D. The framework comprises five complementary CP algorithms derived from validated fibrosis indices and liver injury biomarkers, together with a lightweight natural language processing module that identifies MASLD-related evidence using keyword matching, contextual filtering, and negation detection. The proposed framework was systematically evaluated across multiple threshold settings, with additional validation through age- and sex-stratified analyses and manual chart review. The findings demonstrate that integrating structured and unstructured EHR data provides an accurate and scalable framework for real-world MASLD phenotyping and offers a practical strategy for developing CPs for other chronic diseases.

De-identifying temporal protected health information (PHI) is challenging because temporal expressions appear in diverse formats and must be transformed while preserving clinically meaningful timelines. This study proposes and evaluates an LLM-based framework for temporal PHI de-identification in real-world clinical notes. Using 1,148 notes from 30 sarcoma patients (28,431 annotated temporal entities), we evaluated multiple modern LLMs for temporal entity extraction and surrogate generation. Proprietary models achieved strong extraction performance, with GPT-4o achieving the best overall results. Most models preserved temporal formatting (>99%), but surrogate generation remained challenging. GPT-5.4 achieved the highest shift correctness (90.2% on a shared-entity subset) and the lowest order violations. Error analysis revealed systematic shift deviations ({+/-}1, {+/-}30/31, {+/-}365 days), highlighting persistent limitations in LLM temporal reasoning. These findings suggest that reliable temporal de-identification will require hybrid or multi-agent approaches beyond standalone LLMs.

Biomedical artificial intelligence (AI) increasingly relies on heterogeneous data sources, yet differences in data structure, terminology, evidence quality, and provenance complicate their integration into reliable knowledge resources. Here, we develop an evidence-aware framework for processing and integrating heterogeneous pharmacovigilance data from FDA Adverse Event Reporting System (FAERS) reports, biomedical literature, clinical trials, and electronic health records in Mass General Brigham. The framework combines large language model (LLM)-based information extraction, ontology-driven semantic normalization, cross-source evidence assessment, statistical and machine learning-based prioritization, and ingredient-based adverse drug event (ADE) relationship inference to transform heterogeneous evidence into a structured and evidence-based knowledge resource. The constructed knowledge system contains 11,588 drug nodes, 44,843 ingredient nodes, 3,275 ADE nodes, 993 indication nodes, and more than 2.47 million relationships, with each drug-ADE relationship assigned a confidence score. We further developed a drug-level web interface for accessing the knowledge resource and validated its utility in downstream AI tasks. Specifically, incorporating the developed knowledge resource improved the comprehensiveness, relevance, and overall quality of LLM-based drug-specific ADE synthesis and improved the precision and F1 score of clinical ADE relation extraction. These results demonstrate that AI-assisted processing combined with evidence-aware curation can transform heterogeneous biomedical evidence into reusable, AI-ready knowledge with explicit provenance and evidence strength information. The framework provides a general strategy for integrating heterogeneous biomedical data to support reproducible knowledge construction and downstream biomedical AI applications.

Electronic health records provide a source of real-world information for disease risk prediction, but many machine learning models learn implicit associations that are difficult to inspect or use for clinical reasoning. Causal Bayesian networks (CBNs) offer a form of causal-interpretable machine learning by representing conditional dependencies, putative directional relations, and probabilistic evidence propagation in a directed acyclic graph. However, conventional CBN structure learning becomes unstable and computationally expensive when applied to large-scale EHR data. To address this limitation, we propose UPEBNL, a scalable framework for causal-interpretable cancer prescreening based on parallel CBN learning. UPEBNL integrates adaptive data slicing, quality-aware structure aggregation, and global DAG construction to learn stable and interpretable dependency structures from large-scale observational data. We evaluated UPEBNL in high-dimensional and multi-million-sample simulations and applied it to EHR-based prescreening for esophageal and colorectal cancer. In simulations, UPEBNL improved structural recovery accuracy by nearly 40% and achieved up to a 221.28-fold speedup over conventional CBN learning strategies. For cancer risk prediction, the learned CBNs provided interpretable evidence paths and achieved validation AUCs of 0.8171 for esophageal cancer and 0.784 for colorectal cancer. Calibration and decision curve analyses further supported the reliability and clinical utility of the models. These findings suggest that scalable CBN learning can support interpretable cancer prescreening from large-scale EHR data.

Tabular synthetic data generation (SDG) can facilitate research and development in the health domain, as access to health data is restricted. To accommodate for its structural complexity, such as multimodal distributions and fragmentation, transformer models provides an alternative to conventional SDG methods since they are better equipped to infer global spatial coherence due to textual context of tabular records. However, methods for optimising their adaption to tabular SDG are still largely unexplored. We contribute to the understanding of transformer-based SDG by evaluating fidelity and privacy of three prompting regimes, zero shot, one shot and few shot training, for tabular health SDG using small transformer models. We evaluated llama 3.2 (1B and 3B parameters) and Phi (1.3B and 2.7B parameters), on a real clinical dataset. Our results showed that zero shot learning was less prone to hallucinations and performed overall better for fidelity, but at the loss of privacy in comparison to the other regimes. Our results showed no differences between model sizes and families. Future studies should extend the analysis to other models and datasets, and explore strategies to reduce hallucinations.

Objective: To develop and evaluate a framework for characterizing clinician editing of Artificial Intelligence (AI)-generated[KS1.1] documentation and assess its feasibility for health system-level monitoring of AI scribes. Materials and Methods: We analyzed outpatient encounters in which an AI scribe was used at a single academic health system, examining the History of Present Illness (HPI) and Assessment and Plan (A&P) sections of notes. We characterized edits on three dimensions[KS2.1][AG2.2]: lexical edit intensity based on Levenshtein distance, embedding edit intensity BERT[KS3.1]Score, and clinical edit intensity based on removed and added UMLS[KS4.1][AG4.2] concepts.[KS5.1][AG5.2] We analyzed the Positive Predictive Value (PPV) of clinical edit intensity as a measure of clinically meaningful editing using clinicians as the gold standard, examined correlations among dimensions, and designed exponentially weighted moving-average control charts to monitor longitudinal changes in clinician editing behavior.[KS6.1][AG6.2] Results: 268,379 encounters were included (267,654 HPI, 267,594 A&P). Clinical edit intensity [≥]1 had an 88.9% PPV for clinically meaningful editing. Lexical and embedding edit intensity were highly correlated (Spearman {rho} 0.95), while clinical edit intensity was less strongly correlated with both ({rho} 0.77-0.81). Longitudinal monitoring detected changes coinciding with system-wide rollout.[KS7.1][AG7.2] Discussion:[KS8.1][AG8.2] Clinicians edited A&Ps more heavily than HPIs, potentially reflecting greater attention to content involving clinical decision-making. Over one-third of sections involved clinical concept changes, and clinical edit intensity identified clinically meaningful edits while providing information complementary to lexical editing measures. Conclusion: Clinician editing can be characterized at scale using complementary editing dimensions, providing a scalable signal for post-deployment surveillance of the human-AI documentation process.

Objective Large language models (LLMs) are increasingly used to generate, summarize, and evaluate clinical documentation, yet standardized methods for evaluating nuanced, context-dependent clinical language remain limited. Existing approaches rely primarily on annotated clinical corpora or keyword-based methods, restricting reproducibility, public dissemination, and systematic comparison across models. We present ClinStigma, a clinician-guided framework for constructing, annotating, and evaluating synthetic benchmarks of contextual stigmatizing language, demonstrated through ClinStigma-SUD, its inaugural implementation for substance use disorder documentation. Materials and Methods Candidate clinical sentences were generated using a frontier language model and iteratively reviewed according to a clinician-developed annotation rubric refined through expert review in stigmatizing language. The resulting benchmark intentionally incorporates explicit stigmatizing, neutral, and contextually ambiguous ("gray-zone") documentation requiring clinical reasoning rather than keyword recognition. Three locally deployed open-weight LLMs were evaluated using zero-shot inference, followed by external evaluation using independently annotated clinical documentation from MIMIC-IV. Results ClinStigma-SUD consists of 500 clinician-adjudicated synthetic clinical documentation sentences. Inter-rater agreement was 98.8% across the full benchmark (Cohen's {kappa} = 0.976) and 94.0% within the intentionally ambiguous subset ({kappa} = 0.880). Under zero-shot prompting, binary accuracy ranged from 84.8% to 91.8% across evaluated models, while F1 for the ambiguous (gray) class ranged from 0.038 to 0.423. Zero-shot external evaluation on 3,716 annotated examples from the independently annotated Harrigian MIMIC-IV corpus yielded 84.4% accuracy (84.4% balanced accuracy; macro F1 0.824), against a 69.9% majority-class baseline. Stigmatizing-class F1 was 0.879 in the substance-related subgroup and 0.883 in the full external corpus. Discussion and Conclusion ClinStigma establishes a reproducible clinician-guided framework for developing synthetic evaluation benchmarks for contextual clinical language. By combining structured clinical expertise, explicit annotation guidance, synthetic benchmark generation, and external evaluation, the framework enables open, privacy-preserving evaluation of language models while providing a scalable foundation for benchmark development across additional clinical specialties, including obstetrics, emergency medicine, and mental health.

The rapid growth of technologies such as wearable devices and high-throughput sequencing has generated unprecedented volumes of longitudinal data, often with multiple correlated features. This trend highlights the need for robust clustering methods tailored to multi-feature longitudinal datasets. Although previous reviews have benchmarked clustering methods in this context, most evaluations relied heavily on simulated and overly simplified datasets, which substantially limit their interpretability and applicability to real-world data. Missing data is also ubiquitous in longitudinal data, yet understudied in the context of clustering. To address these gaps, we systematically evaluate clustering methods across three distinct types of real-world longitudinal datasets: (1) a high-dimensional, low-timepoint metagenomic study, (2) a low-dimensional, high-timepoint multi-site repeated-measure study, and (3) a high-dimensional, high-timepoint repeated-measure study. We further investigate how missing data and different imputation strategies influence clustering performance. In addition, we introduce a novel resampling framework based on the empirical cumulative distribution function (eCDF) and a copula. This framework enables us to simulate multi-feature longitudinal data that preserves the correlation structure of real datasets, while also allowing for controlled variations such as sample size. Our simulation framework naturally accommodates missing data, thereby providing a flexible tool for benchmarking and comparing clustering methods in more realistic longitudinal settings.

This paper shows that it is possible to predict which newly published biomedical research papers will be connected to subsequent clinical trials. We use CAR-T cell therapy as a proof-of-concept setting because its development can be followed from published research through clinical testing, regulatory approval, and commercialization. We reconstructed connections between research papers and clinical trials and trained machine-learning models to predict which papers were associated with research that progressed to clinical testing. Using 65 paper-inherent indicators, including scientific topic, language, biomedical orientation, and team structure, the models achieved a cross-validated AUC of 0.908. The resulting rankings were also practically informative. In a retrospective analysis using no information generated after 2016, 80% of the top 20 papers came from laboratories associated with one of 13 predefined major CAR-T clinical or commercial outcomes. As an additional finding, first-year citations alone achieved an AUC of 0.797, while provenance indicators based on author, journal, and institutional reputation performed less strongly and added little to the paper-inherent model. Our results show that whether biomedical research papers will be connected to subsequent clinical trials can be predicted from newly published papers with high discrimination in CAR-T. Our findings could help research organizations and pharmaceutical companies identify papers, laboratories, and therapeutic programs that warrant closer evaluation before clinical outcomes are known.

ICD coding is the backbone of medical-record structuring, insurance payment, and clinical research. Today's best automatic coders are deep models, but they assume tens of thousands of labeled records, GPU servers, and machine-learning engineers. Many hospitals, especially small ones, have none of these. We study Chinese electronic medical records (EMRs) under a hard budget: 2,200 labeled records, CPU only, and fully interpretable decisions. Our method is lightweight and follows the way human coders work. First, codebook knowledge injection marks standard disease and operation names in the text as useful clues. Second, multi-view field weighting gives more weight to the clinical sections that matter most. Third, character n-gram TF-IDF features feed linear classifiers that score each candidate code. Fourth, a small set of clinical rules acts as a final safety check. Our main methodological finding is about this last step: naively applied rules do more harm than good (overall F1 drops from 0.8070 to 0.7995, and main-operation accuracy from 0.9727 to 0.9432). We therefore audit each rule on held-out data by counting how often it fixes a wrong prediction or breaks a correct one, and we remove rules that do more harm than good. This turns rule post-processing into a safety check that never hurts. On the CCL2026-Eval Task 6 dataset, our method reaches macro-F1 of 0.899 (main diagnosis) and 0.966 (main operation) in 5-fold cross-validation. A frequency-stratified analysis shows a clear long-tail limit: codes seen fewer than three times score zero F1. In a head-to-head test, a fine-tuned Chinese pretrained encoder scores lower than our method on all four tasks at this training size, while needing a GPU. Ablation tests show that the gains come mainly from field-weighted text. The framework trains in about 40 minutes on an ordinary CPU, and every prediction can be broken down into interpretable feature contributions.

Personalized medicine aims to tailor treatments to individual patients, but its implementation is limited by scarce clinical resources and heterogeneous treatment effects. Electronic health records now offer longitudinal data that can support data-driven therapy selection. We focus on SGLT2 inhibitors, medications with broad metabolic and cardiovascular benefits, whose clinical impact varies significantly based on individual patient profiles. We propose a counterfactual Structural Causal Model (SCM) to estimate individual treatment effects (ITE) of SGLT2 inhibitors from observational data. The model assumes counterfactual stability and uses a Gumbel-Max sampling mechanism to capture stochastic treatment assignment while enabling differentiable inference. Optimal treatments are defined by maximizing the counterfactual value of future diagnostic outcomes. We evaluate the Counterfactual Gumbel-Max SCM using a doubly robust value estimator and show that, across diagnostic categories, it outperforms real-world prescriptions in predicted treatment value, highlighting subpopulations where SGLT2 inhibitors yield superior outcomes. By integrating causal inference with electronic health data, this framework provides individualized counterfactual reasoning about SGLT2 treatment effects and can support clinicians in identifying optimal therapies for type 2 diabetes and comorbid conditions.

AO_SCPLOWBSTRACTC_SCPLOWO_ST_ABSBackgroundC_ST_ABSLung donor-recipient matching requires balancing recipient medical urgency with expected post-transplant benefit. Conventional prediction models estimate outcomes for historical donor-recipient pairs but do not directly compare alternative candidates for the same donor. We developed an offline contextual-bandit framework to assess the feasibility of learning donor-centered lung-matching policies from historical registry data. MethodsWe used linked United Network for Organ Sharing (UNOS)/Organ Procurement and Transplantation Network (OPTN) Standard Transplant Analysis and Research files to identify 8,529 lung transplant allocation events from 2023 through 2025. Each event included the historical recipient and nine date-matched, ABO-compatible pseudo-candidates. The reward combined 72-hour respiratory-support-free survival (weight, 0.75) with a normalized initial Composite Allocation Score waitlist medical-urgency score (weight, 0.25). We trained a pessimistic neural lower-confidence-bound contextual-bandit policy (NeuraLCB) and evaluated it on held-out donor events using matched-action evaluation and self-normalized inverse propensity scoring (SNIPS). ResultsThe reconstructed dataset contained 85,290 donor-candidate rows. In the held-out set of 1,706 allocation events, the observed clinician reward was 0.534 (95% CI, 0.518-0.549). NeuraLCB agreed with the historical recipient in 174 events (10.2%), with a matched-action reward of 0.616. The SNIPS-estimated value of the NeuraLCB policy was 0.649 (95% CI, 0.568-0.717), corresponding to a paired difference of +0.116 (95% CI, +0.035 to +0.181) relative to observed clinician reward. However, the effective sample size was 51.5 events, indicating limited overlap between the learned and estimated historical policies. ConclusionsThis study demonstrates the feasibility of constructing and evaluating an offline contextual-bandit framework for lung donor-recipient matching using registry data. The estimated policy value was higher than the observed historical reward, but this finding is exploratory because candidate sets were reconstructed and off-policy evaluation had limited overlap. Verified offer sets, time-stamped candidate data, explicit compatibility constraints, and prospective evaluation are required before clinical use.

Background: Unstructured free-text narratives in electronic health records (EHRs) contain critical clinical information that is not captured by structured standard codes. In Catalonia, primary care notes follow a semi-structured format known as MEAP (in Catalan). However, extracting structured variables using cloud-hosted commercial large language models (LLMs) raises substantial data privacy concerns and incurs high operational costs. Objective: To develop and validate a privacy-preserving local hybrid pipeline combining a lightweight small language model (SLM) with deterministic regular expressions for extracting six uncoded urinary tract infection (UTI) clinical features (fever, nitrites, leukocytes, lumbar pain, abdominal pain, and haematuria) from primary care EHR narratives. Methods: We deployed the open-weight model (Phi4-mini, 3.8B parameters) within the institutional firewall using Ollama. Prompts were iteratively refined in collaboration with clinicians, and regular expressions were applied post-inference. Validation was conducted across two independent arms: (1) a double-blind clinician gold standard review of 60 real patient records (consensus-resolved cases), and (2) an adversarial synthetic dataset of 720 notes enriched with linguistic noise, generated using GPT-4.1 and Grok-4.1. Point estimates and 95% confidence intervals (CIs) were computed for key performance metrics. Results: The pipeline processed 15,498 MEAP narratives from a matched cohort of 2,962 primary care patients and identified 4,663 clinical feature occurrences. Patients progressing to acute pyelonephritis (cases) presented a higher features burden than non-progressing controls (71.3% vs. 55.8%; standardized mean difference [SMD] = 0.326), particularly for fever (33.0% vs. 9.0%; SMD = 0.616) and lumbar pain (29.0% vs. 9.8%; SMD = 0.502). In real-world clinician validation, overall performance yielded 93.8% accuracy (95% CI 90.6%-96.1%), 83.6% sensitivity (95% CI 73.0%-91.2%), 96.6% specificity (95% CI 93.6%-98.4%), 87.1% positive predictive value (PPV; 95% CI 77.0%-93.9%), and 95.5% negative predictive value (NPV; 95% CI 92.3%-97.7%). In synthetic stress-testing, the pipeline demonstrated 87.2% accuracy (95% CI 84.6%-89.6%), 73.9% sensitivity (95% CI 68.9%-78.5%), near-perfect specificity (99.5%, 95% CI 98.1%-99.9%), and 99.2% PPV (95% CI 97.2%-99.9%). Conclusions: A privacy-preserving local hybrid framework combining a compact open-weight SLM with regular expression rules achieves high specificity and precision for extracting six uncoded clinical features from multilingual primary care narratives. Running entirely within institutional servers, this approach keeps patient data secure, meets privacy standards, and avoids Application Programming Interface (API) costs, making it a practical tool for EHR research, though with lower sensitivity for narratively complex descriptions.

Objective: To determine how cohort versus nested case-control (NCC) design affects EHR-based prediction of a rare outcome. Materials and Methods: We compared two cohort and five NCC designs for predicting early-onset gastrointestinal cancer using L1-regularized logistic regression (LASSO), XGBoost, and multilayer perceptron (MLP). Models underwent nested five-fold cross-validation with model-specific hyperparameter tuning and temporal evaluation in a held-out prospective cohort representing intended deployment. Model discrimination was assessed using area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). Model calibration was assessed using Brier scores and calibration plots before and after recalibration with isotonic regression. Feature-ranking similarity was quantified using rank-biased overlap. Results: The cohort development datasets included 256,854-415,529 patients, with 2-yr cumulative incidence of 0.12%-0.17%, compared with approximately 9% in the 1:10 matched NCC datasets. NCC designs generally produced higher internal discrimination, but temporal performance varied by algorithm and index-date definition. Cohort designs achieved the highest temporal AUPRC for LASSO (0.068), whereas diagnosis-anchored NCC achieved the highest temporal AUPRC for XGBoost (0.086) and MLP (0.070). Temporal AUPRC generally decreased as the required lead time before diagnosis increased. Raw predicted probabilities substantially overestimated observed risks in the temporal cohort. Isotonic recalibration improved probability scaling while largely preserving discrimination, although residual miscalibration remained. Cohort-based feature rankings were more consistent across algorithms, whereas NCC-based rankings depended more strongly on algorithm and index-date anchoring. Conclusion: Study design should be selected and evaluated according to the intended clinical prediction task and deployment setting.

Accurate prediction of individual medical outcomes is essential for optimizing treatment allocation amid rising costs, coverage denials, and limited clinical resources. Traditional predictive models, including linear regression and neural networks, rely on average effects and cannot tailor predictions to the specific circumstances of individual cases. We present relevance-based prediction (RBP), a model-free and task-specific method that predicts outcomes as weighted averages of observed cases, with weights determined by a rigorously defined and theoretically justified measure of relevance. Unlike model-based methods that rely on fixed calibrated parameters, relevance-based prediction revisits the original data for each prediction and customizes both the cases and variables used. We illustrate RBP by applying it to predict opioid treatment outcomes and demonstrate that it provides case-specific insights unavailable from conventional models, including how each prior case informs a prediction, how each variable affects each predictions reliability and value, and how reliable each prediction is before it is made. These individualized insights may prevent misleading average-based decisions and reduce harmful or suboptimal treatment.

Objective: To compare the accuracy, cost, and speed of a low-cost, non-generative structured-judgment model and four general-purpose large language models (LLMs) for extracting smoking status, pack-years, and quit date from clinical notes to support lung cancer screening (LCS) clinical decision support (CDS). Materials and Methods: We built a synthetic, shareable benchmark of 3,000 outpatient notes in three conditions: 1,000 template-generated notes, 1,000 realistic, ``messy'' notes written by an LLM from structured facts, and 1,000 ``messy'' notes that required complex arithmetic to determine pack-years and quit dates. Ground truth reference labels were programmatically generated before each note was created. We compared TypeSafe Jev 1.13 with Claude Haiku 4.5, Claude Sonnet 5, GPT-6 Luna and GPT-6 Sol, each using an identical structured output schema. We determined United States Preventive Services Task Force (USPSTF) 2021 and American Cancer Society (ACS) 2023 eligibility from each system's output in code. Results: Jev made the correct eligibility decision for 99.1%, 99.9%, and 94.4% of notes in the three conditions, compared with 97.8%, 98.4%, and 98.1% for Haiku, 100.0%, 99.6%, and 99.8% for Sonnet, 99.7%, 98.8%, and 98.5% for Luna, and 100.0%, 99.8%, and 99.6% for Sol. In the complex condition, Jev produced 27 false positive and 14 false negative screening flags per 1,000 notes, the most of any system. Luna cost $0.12 to $0.21 per 1,000 notes and Jev $0.61 to $0.64, compared with $2.44 to $4.22 for Sol, $3.37 to $4.43 for Haiku, and $3.76 to $6.61 for Sonnet. Jev was the fastest system (median 0.45 to 1.21 seconds per note, compared with 1.71 to 3.50 seconds for the others). Discussion: All models tested had high accuracy. The structured-judgment model had the lowest latency, which could allow for synchronous use in clinical decision support systems rather than batched, cached invocation. The cost of the structured-judgment model was low, but the newest cost-efficient LLM, Luna, was less expensive and more accurate on the complex notes, though its latency was higher. Accuracy, latency, and cost are all moving quickly as new models are released. Conclusion: Fast, low-cost language models could make real-time, note-based CDS practical, and current models are highly accurate. Prices and capabilities are changing rapidly, so continuous benchmarking and system updates are essential.

Objective: To measure adherence to a national inpatient hyperglycemia guideline at the level of its individual recommendations across a hospital network, to determine whether it changed after the guideline was issued, and to identify which recommendations track subsequent outcomes. Materials and Methods: Retrospective multicenter study of 48,354 admissions among 29,577 adults with diabetes or inpatient hyperglycemia at six Israeli general hospitals, August 2022 to December 2025. Each of 14 recommendations was formalized as a quality-assessment temporal pattern over knowledge-based temporal abstractions of the record, scored wherever it applied. Pre-post adherence was compared by Welch t-test with false discovery rate correction. Associations with 10 complications were estimated by bias-reduced logistic regression over 24-, 48-, and 72-hour windows, and by inverse probability of treatment weighting contrasting above- with at-or-below-median adherence. Results: Adherence ranged from near-universal admission and medication-management actions to substantial gaps in insulin management (means, 0.15-0.28). Overall adherence rose modestly after issue (0.682 to 0.703, P < .001), with marked inter-hospital heterogeneity. After weighting, above-median adherence carried 44% lower odds of 30-day mortality (odds ratio [OR], 0.557; 95% CI, 0.514-0.603), 82% lower odds of hyperglycemia (OR, 0.175) and 52% lower odds of severe hyperglycemia (OR, 0.485), against 20% higher odds of hypoglycemia (OR, 1.204). Discussion: The least-implemented recommendations were largely those whose adherence tracked outcomes: repeated insulin decisions spread across a stay, rather than discrete actions on admission. Conclusion: Guideline adherence can be assessed automatically and per-recommendation at national scale, identifying where fuller implementation would most plausibly matter. Keywords: inpatient diabetes; clinical guideline adherence; quality assessment; temporal abstraction; electronic health records; propensity score; causal inference

In this work, we investigate the homogeneity of performance in binary classification between the Rotterdam phenotypes and phenotype prediction. For this purpose, a multi-task neural network architecture is designed to improve the prediction accuracy of phenotypes while maintaining the accuracy of binary classification. Using a clinical PCOS dataset which contains unlabeled samples of patients that cannot be attributed to any of the known phenotypes, three models are evaluated: optimized binary Random Forest (RF), six-class weighted class RF, and multi-task neural network (with a common hidden layer and two heads for binary and six-class classifications) optimized by a genetic algorithm. Techniques used for evaluation include ten-fold out-of-fold predictions, paired McNemar tests, bootstrap tests, ablation study, and clustering (Adjusted Rand Index, ARI). Binary RF model reached 90.2% accuracy; however, recall of different phenotypes was significantly unbalanced - from 75.0% of phenotype B to 97.8% of phenotype A. This difference of 22.8 % points was observed (23.3 points) even after excluding phenotype-specific features. While direct phenotype prediction improved recall for common phenotypes, it decreased recall for phenotype B to 43.8% and unclassified cases to 31.5%. Our multi-task model preserved 90.2% binary accuracy and significantly recovered phenotype B recall to 87.5% and unclassified cases to 79.6% (p<0.05). The cluster analysis confirmed the weak correlation between the results and labels (ARI=0.045). The ablation analysis demonstrates that feature selection and class weighting drive this improvement, isolating the specific components responsible for performance gains. We establish that aggregate binary accuracy does not reflect the heterogeneous detection of PCOS phenotypes. Direct phenotype prediction alone redistributes the classification errors, but the multi-task architecture successfully improves the detection of rare phenotypes without compromising binary accuracy. Phenotype-disaggregated evaluations, supported by significance testing and ablation, are essential for clinical classifier research.

Objective: To estimate Long COVID undercoding in electronic health records (EHRs) against federal survey prevalence, and to evaluate a predictor of later coding that anyone can rebuild from counts a research enclave releases. Materials and Methods: Among 23,790 All of Us adults with COVID-19, we compared coded shares with Household Pulse Survey prevalence. Among 23,300 adults uncoded at day 90, we predicted a first U09.9 code by day 455. We fitted naive Bayes and linear discriminant models outside the enclave from released counts and compared them with an elastic net fitted inside to the same patients and features. Tests were a random 20% of patients (4,680, 50 coded) and 18 held-out EHR sites (5,760, 60 coded). Results: Under survey-comparability assumptions, about 1 in 9 adults who ever had Long COVID carried its code, and the women-to-men coding ratio was 0.79 (95% sensitivity interval, 0.65-0.95). The released models' areas under the receiver operating characteristic curve (AUROCs) were 0.68 (95% CI, 0.60-0.75) on the patient test and 0.64 (0.55-0.72) at held-out sites, 0.049 (0.001-0.104) above the elastic net. On the patient test, the elastic net had a higher area under the precision-recall curve (difference 0.062, 0.003-0.122). Released probabilities overestimated coding risk at held-out sites. Discussion: Fitting from released counts did not consistently lower discrimination. Coded outcomes inform disease discrimination only under stated coding assumptions. Conclusion: Long COVID is substantially and unevenly undercoded, and a model rebuilt from public counts predicts its later coding with moderate discrimination.

In hypertension screening, previously diagnosed cases provide reliable positive labels, whereas individuals without a prior diagnosis comprise both undiagnosed hypertensive and non-hypertensive individuals, creating a canonical positive-unlabeled (PU) learning setting. We propose PU-Boost, a two-stage reliable-sample reconstruction framework that uses only readily obtainable demographic, behavioral, medical-history, and anthropometric variables. During training, only previously diagnosed hypertension cases are treated as labeled positives (P), while all remaining individuals are treated as unlabeled (U); no additional reference labels are supplied to the two-stage reconstruction procedure. Elastic Net first provides a global ranking of P versus U, after which a random forest refines the uncertain region and highly ambiguous observations are rejected. XGBoost is then trained on the reconstructed class-balanced pseudo-labeled sample. In the independent test set, PU-Boost achieved a sensitivity of 0.817 and a balanced accuracy of 0.682, the highest observed values among the compared models. It produced 729 false negatives, 41.1% fewer than standard XGBoost, although specificity decreased to 0.547. SHAP analysis indicated that established hypertension-related features, including age, body mass index, waist circumference, and family history, contributed strongly to model output. PU-Boost therefore provides a statistically motivated framework for reducing missed cases when screening for undiagnosed hypertension under incomplete diagnostic labeling.