BMC Medical Research Methodology

Background: Conventional discrimination metrics, such as the area under the receiver operating characteristic curve (AUROC), do not characterize the composition of outcomes among patients flagged as high risk by clinical prediction models. This limitation is particularly relevant in critical care, where competing deaths and incomplete follow-up can complicate the interpretation of prediction alerts. We developed the Balanced Risk Integrity Index (BRII), a stratum-level metric defined as the proportion of target events among target events, competing deaths, and early-censored observations in a flagged high-risk stratum. BRII is algebraically related to positive predictive value (PPV) and is reported alongside its component decomposition to characterize the outcome composition of flagged patients. Methods: We evaluated BRII using seven acute kidney injury (AKI) prediction models developed in MIMIC-IV and externally validated in MIMIC-III and eICU-CRD (116,010 patients in total). Additional simulations examined the effects of competing deaths and early censoring on BRII and conventional model performance metrics. Results: For 48-hour AKI prediction in the top 10% risk stratum, BRII ranged from 0.90 to 0.96 in MIMIC-III, compared with less than 0.26 across all models in eICU-CRD, where 61.5% of evaluated patients were censored before the prediction horizon. In the eICU-CRD top 10% strata for the same target, 37.0% to 50.2% of patients were censored early, 12.2% to 15.5% had AKI, and 3.4% to 8.1% died before AKI. BRII and PPV ordered at least one model pair differently in 34 of 48 cohort-outcome-top-k groups. Under simulated random censoring for 48-hour AKI prediction, BRII declined by up to 0.80 while binary-label AUROC changed by less than 0.02. Conclusions: BRII provided complementary information on high-risk stratum composition, yielding different model rankings from PPV and revealing substantial differences across external validation cohorts. Interpreted alongside PPV and its component decomposition, BRII offers a standardized approach to evaluating clinical prediction alerts across models and observation environments.

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.

Purpose: Accurately capturing a patient's health journey is essential for ensuring the validity of scientific and public health studies. Electronic health records (EHR) and administrative claims are commonly used in healthcare research, albeit separately. Previous research has demonstrated that EHR and claims capture different aspects of patient history and care, potentially introducing bias when used alone. In this study, we leveraged the largest public US EHR resource, the National Clinical Cohort Collaborative (N3C), alongside Centers for Medicare & Medicaid Services (CMS) claims data to compare their information coverage and agreement. Methods: We integrated N3C EHR and CMS data to create one cohesive dataset and quantified the proportion of key variables present in EHR alone, CMS alone, and the overlap present in both data sources. Overlap in information coverage was quantified pre- and post-landmark diagnosis of COVID-19. To evaluate the impact of data coverage differences, we conducted a case study evaluating the effect of COVID-19 vaccination on ischemic stroke using both data sources alone and in combination. Results: CMS data helped identify many individuals that were duplicated among EHR identifiers. Examination of the overlap showed that CMS data more completely captured conditions, while EHR data more completely captured medications after COVID-19 diagnosis. Some variables exhibited marked differences over time after landmark COVID-19 diagnosis. Using target trial emulation and self-controlled case series methods, we found that data capture variations altered effect estimates: EHR data alone yielded lower risk of ischemic stroke after vaccination in the target trial emulation (odds ratio 0.47, 95% CI 0.28-0.79), with a smaller, non-significant reduction in the self-controlled case series (incidence rate ratio 0.81, 95% CI 0.64-1.02), whereas analyses using CMS data alone or both sources combined yielded estimates close to the null in both designs. Conclusions: Results indicated that when using either a national database combining multiple EHR sources, or CMS claims data alone, patients' medical histories are less completely captured compared to when these data sources are combined. Case study results highlighted the value of combining data sources for not only yielding fuller patient health histories and valid scientific results, but for identifying and explaining biases that may emerge in research studies that rely on a single data source.

Objective A diversity of approaches and cohorts have been published in research literature for suicide risk estimation, leading to some uncertainty around which factors and conditions are most strongly contributing to risk. Methods exist for weighting estimates by propensity score, removing simplifying the interpretation of model effects. Few studies have explored the use of these tools to assess when the effect of certain factors is uncertain or inconsistent. Method A mix of 1,001,737 suicide decedents and living patients were evaluated for suicide risk, relying on a subset of commonly identified factors: demographics, psychiatric conditions, substance use disorders, and social needs. Care utilization patterns were then assessed controlling for these factors in a hospital care setting. Four methods were considered for controlling multivariate confounding of odds predicting death by suicide: multivariate logistic regression, 1-to-1 sample matching, inverse propensity weighting (IPTW), and entropy balancing (EBAL). Results Test statistics showed suicide decedents were much more likely to have various psychiatric and substance use diagnoses, social needs and more likely to heightened care utilization across sampling intervals. Greater utilization was found close to date of death. Effect sizes after IPTW and EBAL weighting were also generally greater than would have been documented by regression alone. Conclusions Propensity weighting is a useful alternative to exploring nuanced effects, especially when the direction and strength of an association differ from traditional statistics. By weighting, the overall imbalance of cases remains consistent and effect terms might reflect more nuanced associations than regression or sample matching alone.

Objective: Case-finding models for underdiagnosed conditions are usually trained on diagnosis codes, which are missing from the records of many patients with a clinical diagnosis. We asked whether models trained on self-reported diagnoses rank uncoded self-reporters higher than models trained on later codes, and how a score fit only from released summary counts compares with a participant-level model. Methods: In the All of Us Research Program, we followed survey respondents with no qualifying diagnosis code by the survey date for later codes. For chronic fatigue, fibromyalgia, and endometriosis (participants recorded as female at birth), we fit a naive Bayes score with empirical-Bayes shrinkage, from released feature-by-label counts for the self-reported diagnosis and from exact counts inside the data enclave for a later code, and an elastic net on participant-level data for each label. In a random 20% validation split, we measured discrimination of self-report, of uncoded self-reporters (no qualifying code observed), and of later codes. Results: For chronic fatigue (149,860 respondents, including 7,770 self-reporters, 95% with no code observed), the summary-count score trained on self-report ranked uncoded self-reporters with an area under the receiver operating characteristic curve (AUROC) of 0.74 (95% CI, 0.72-0.75), against 0.67 (95% CI, 0.65-0.69) when trained on later codes. For later codes, the two scores were similar (0.69 vs 0.70). Fibromyalgia showed the same pattern (0.73 vs 0.64 and 0.74 vs 0.72). Elastic nets trained on self-report ranked uncoded self-reporters higher than the summary-count scores (0.80 vs 0.74 and 0.78 vs 0.73) but discriminated later codes less well than code-trained elastic nets. For endometriosis, the self-report score also ranked uncoded self-reporters above the code-trained score (0.61 vs 0.53), but discrimination was low and it did not discriminate later codes (0.50 vs 0.70). Conclusion: The choice of training label changes which patients a case-finding model ranks highly. For all 3 conditions, self-reported diagnoses were the more useful label for finding uncoded self-reporters. For chronic fatigue and fibromyalgia, the self-report score also discriminated later codes as well as the code-trained score. Independent clinical validation is needed before these models guide case finding.

Estimating the correlation between subject-specific rates of change is a common objective in longitudinal studies. A natural approach is to fit ordinary least squares regressions per subject and correlate the resulting slope estimates directly. Under complete data, this slopes-first estimator is algebraically equivalent to a contrast-based plug-in estimator derived from the joint variance-covariance matrix of the two signals across time. Under independent dropout, however, the slopes-first approach implicitly applies heterogeneous weights across missing-data patterns, introducing bias whenever the cross-covariance between the two signals is not constant over time. We formalise this result and propose an unbiased alternative that estimates the joint covariance matrix before applying the contrast. We further extend our results to include covariates as well. Simulations confirm that the slopes-first estimator is biased under non-constant cross-covariance while the proposed alternative remains unbiased. Further, we show in a real-world example that the two estimators can lead to different scientific conclusions and recommend that researchers working with longitudinal data inspect the missing data pattern to make an informed decision about the estimator of choice.

Background: Cost-effectiveness analyses (CEA) built on simulation models are usually delivered through project-specific scripts. Each new model, and often each new question, rebuilds the same scaffolding: parameter management, calibration, base-case analysis, sensitivity analysis and export. This is costly, and is where traceability between an estimate and the model version producing it is lost. We present THALASSA, a web application providing this infrastructure once, for any model satisfying a small programming contract. Methods: A simulation model is connected by one file implementing a fixed contract of R functions that determine its behaviour. THALASSA exposes the complete CEA workflow in a browser: a generated overview of the connected model, parameter editing, calibration, base-case comparison with the efficiency curve, and deterministic and probabilistic sensitivity analysis at a willingness-to-pay threshold, each with timestamped file export carrying the parameter values it was run with. Every session is bound to a repository version, so results carry their generating model version. Results: We demonstrate THALASSA on two public models of contrasting difficulty: a simple tutorial model and a generic cancer screening model with two competing natural-history pathways and a 21-dimensional calibration deliberately built to be more complex. Both ran on the same unmodified setup, implementing the same contract, with no code changes. Conclusions: Separating analysis infrastructure from model logic makes the CEA workflow reusable across projects and reproducible by construction. A modelling team can move from a versioned repository to audit-ready, exportable CEA outputs without writing analysis code. The integration effort does not grow with model size and one installation serves unrelated projects.

Background: TB vaccine development is hampered by the natural history of M. tuberculosis: most infected people rarely progress to disease, so trials must be very large and long to accrue sufficient disease endpoints. Current trials enrol IGRA-positive participants (IGRA: Interferon-Gamma Release Assays), but IGRAs remain positive for many years and cannot distinguish recent from remote infection. Tests for recent infection are now in development, and we argue they could make trials substantially more efficient. Here, we investigate how screening participants prior to enrolment with a test for recent M.tb infection could impact the total number of enrollees required for an adequately-powered TB vaccine trial. Methods: Using a statistical model of vaccine trial recruitment, we estimated the number needed to contact, test, and enrol to achieve an 'effective' trial (defined as 50 disease endpoints), under IGRA-gated enrolment (life-long duration of positivity) vs. enrolment with a test for recent infection. We considered five recruitment strategies: non-test-based enrolment for ages 15-49, ages 12-34, and household contacts of recent cases; and test-based enrolment using IGRA, and a test for recent infection with two-year duration of positivity. The model explored trial recruitment in South Africa, Kenya, and India. Results: Compared with an extended-positivity test, a test for recent infection reduced the median number needed to enrol by 4.2- to 6-fold under favourable conditions. It also outperformed non-test-based strategies, including enrolling household contacts. This comes at the cost of substantially more testing, but crude estimates suggest recency-based enrolment would still be cheaper overall provided the new test does not cost more than roughly 5-10 times an IGRA. Conclusions: Diagnostics for recent M.tb infection have the potential to improve vaccine trial efficiency, but more research is now needed to understand test characteristics and to evaluate their use in this setting. Our Viewpoint sets out the test attributes that would need to be established, such as time to positivity, duration of positivity, sensitivity, and specificity, before such tests could be used in vaccine trials.

Background and Objective: Adding belzutifan to adjuvant pembrolizumab improved disease-free survival (DFS) in LITESPARK-022, but its durability beyond the 1-year treatment is unknown. Methods: Following a registered plan, we reconstructed individual patient data from published Kaplan-Meier curves of ten phase 3 adjuvant renal cell carcinoma trials. Within each trial, we estimated restricted mean survival time differences ({Delta}RMST), hazard ratios (HRs) before and after planned treatment ended, and DFS differences. Key Findings and Limitations: In LITESPARK-022 (central review), the 24-month {Delta}RMST was 1.34 months (95% confidence interval 0.73-1.95), and the belzutifan arm remained ahead after month 12 ({Delta}RMST increment 0.74 months/yr, 0.32-1.19). The HR was 0.55 (0.43-0.71) before month 12 and 1.16 (0.87-1.54) afterwards (post/pre ratio 2.10); investigator assessment showed similar attenuation. Similar ratios occurred in KEYNOTE-564 at its first interim analysis (2.02), whose benefit persisted, and in SORCE and PROTECT (2.09, 2.59), whose benefit faded. From months 12 to 30 (post hoc), the DFS difference changed by -3.7 percentage points (-8.1 to 0.8) with central review and 0.6 (-4.0 to 5.0) with investigator assessment. Data were reconstructed, and follow-up was short. Conclusions and Clinical Implications: Relative hazard reductions with belzutifan were greater before month 12; durability of the absolute DFS advantage is uncertain and assessment-dependent, so single-HR projections should be avoided.

INTRODUCTION: The digital clock drawing test (dCDT) is a cognitive screening tool capturing high-resolution pen-movement data. Existing approaches for classifying cognitive impairment from dCDTs rely on selected features or image-based methods. We propose representing dCDTs as structural networks to classify cognitive impairment and characterize spatial organization in clock drawings. METHODS: Weighted networks were constructed from 3,362 Framingham Heart Study participants, with and without cognitive impairment. Network representations were generated separately for command and copy conditions, with nodes representing pre-defined clock components (digits, hands, clock face) and edges capturing pairwise spatial relationships between components. Random forest models and graph convolutional networks were trained to predict cognitive impairment from the networks. Models were benchmarked against demographics-only and established summary feature baselines. RESULTS: Network-based models achieved AUC values comparable to the established summary feature baseline across command-only and copy-only conditions. Differences in spatial organization between the cognitively intact and impaired groups were greatest in the upper-right clock region, involving digits 1 and 2, the clock hands, and the surrounding clock face. Differences between the command and copy networks were larger among cognitively impaired participants. DISCUSSION: The proposed network representations provide an objective, scalable, and flexible framework for representing clock drawings for use in classification models. dCDT networks matched the predictive performance of established summary feature-based approaches and identified clock regions where spatial organization differed by cognitive status. After external validation, this approach has the potential to be incorporated within a clinical workflow to support automated and interpretable cognitive screening.

BACKGROUND Timely cause-of-death (COD) information guides health policy, yet many deaths in low- and middle-income countries are not medically certified. In Africa, fewer than 10% of deaths are. Verbal autopsy (VA) yields a structured symptom questionnaire and a free-text narrative. Mortality surveillance increasingly records the narrative alongside the questionnaire, but the algorithms used to produce VA-based statistics were designed for structured responses only, so narratives inform physician COD assignment but do not contribute to algorithmic COD assignment. We evaluated whether supervised models could use both modalities to improve COD assignment. METHODS We analyzed 4,999 VA records for deaths at ages 12 years or older from a national sample of registered deaths in South Africa (RSA), using physician-assigned CODs as the reference standard. We trained unimodal and multimodal pipelines on the labelled records: pretrained language models (PLMs) fine-tuned on narratives alone, machine learning (ML) models on structured responses alone, and the two combined at the data, feature, and decision levels. We also prompted a locally deployed generative large language model (LLM) on the same unimodal and multimodal inputs, without parameter updates. All were compared with InSilicoVA, the operational comparator representing current automated practice. We evaluated individual-level assignment, population-level agreement, calibration, and generalization across provinces in RSA, reporting these metrics rather than ranking pipelines on any one of them. RESULTS Every supervised pipeline improved on InSilicoVA by 8.5-15.9 percentage points in top-1 accuracy and 3.5-7.1 points in cause-specific mortality fraction (CSMF) accuracy. The narrative and structured pipelines performed similarly in aggregate but differed by cause: 63.9% of deaths received identical predictions, 11.6% were correct only from the narrative pipeline, and 13.6% only from the questionnaire pipeline, with the narrative pipeline more accurate for noncommunicable causes and the questionnaire pipeline for communicable causes. Against the unimodal approaches, the decision-fusion pipelines added 4.2-5.4 percentage points in top-1 accuracy, concentrated in the most common causes, whereas feature fusion with cross-modal attention improved the less common causes, adding 6.6 points (95% CI, 1.7-11.5) in balanced accuracy and 5.5 points (95% CI, 0.6-10.3) in macro F1. Feature fusion reached the highest CSMF accuracy (95.7%). Paired differences from the better unimodal representative ranged from -0.2 to +3.4 points, with all intervals including zero. The LLM exceeded InSilicoVA on individual-level metrics and reached the highest balanced accuracy of any pipeline (60.6%), while the supervised pipelines retained higher CSMF accuracy. Supervised pipelines had lower expected calibration error than InSilicoVA. Supervised gains held across all nine held-out provinces, including every top-1 comparison and nearly every macro F1 and CSMF comparison. CONCLUSIONS PLMs using narratives alone substantially improved on current best practice and matched ML pipelines using structured responses, while multiple multimodal strategies produced further gains in individual COD assignment. The zero-shot LLM showed particular strength on rare causes, whereas supervised pipelines better recovered the population cause distribution. These findings support incorporating narratives as a core input to automated VA systems and evaluating representative fusion strategies for the intended output. External validation across settings and age groups will support broader implementation.

Background: The first thousand days are a sensitive window for child health and social health inequalities. When routine care falls short, parents of young children turn to online sources and, recently, to generative artificial intelligence (AI). Evidence on their concerns, child-health literacy, and AI use is scarce. Objective: To describe health informational needs and concerns, child-health literacy, and use of AI for child-health questions, including differences by parental role and parity, among parents of young children. Methods: In this cross-sectional study, two anonymous self-administered online questionnaires were disseminated through a digital health application between May 12 and June 11, 2026, to French-speaking parents of at least one child aged 0 to 47 months in France. Questionnaire 1 covered unmet information needs, concerns across 16 topics, perceived intra-couple concordance, and confidence in routine care. Questionnaire 2 covered a four-item child-health literacy quiz and AI use (frequency, topics, 12-month trend, barriers, trust conditions, perceived effect on consultations, and disclosure). Results: In total, 12,634 parents responded (Questionnaire 1, n=5,410; Questionnaire 2, n=7,224), mostly mothers (94.6% and 92.9%). Most (61.8%) had searched for information without finding a clear, reliable answer, selecting a median of 4 (Interquartile range [IQR] 3-7) over 16 topics. Leading concerns were sleep (51.3%), nutrition (48.1%), and education (i.e., upbringing and behavior) (47.8%). Mothers were nearly twice as likely as fathers to select accident prevention (26.5% vs 16.3%) and parental mental health (38.7% vs 28.6%); first-time parents selected most topics more often than experienced parents, most markedly dental hygiene (43.8% vs 27.6%). The median child-health literacy score was 3 (IQR 3-4) on a 0-to-4 quiz, with 19.9% of parents scoring 2 or below; tooth-brushing age was the weakest item (68.4% correct). Mothers outscored fathers on every item. More than half of parents (54.2%) had used AI for a child-health question, 11.2% weekly; among those reporting a trend, 68.3% had increased AI use for child-health questions over the year and 6.2% decreased it. The main barriers to AI use were a preference for professionals (55.9%) and reliability concerns (48.7%); trust depended most on medical oversight of the tool (76.6%). AI use appeared additive, with 66.8% reporting no change in healthcare visit frequency. One third (35.5%) had told an AI something shared with neither a professional nor their network, rising to 64.1% among regular users. Conclusions: Parental informational needs were widespread and frequently unmet. AI has become a common and fast-rising source for child-health questions, at times a confidential one, used alongside rather than instead of professionals. Child-health literacy was generally high but uneven, with oral health the clearest gap. Strengths include the sample size and early-generative-AI timing; limitations include the cross-sectional design, self-report, digital-recruitment selection bias, and the small proportion of fathers.

Introduction: Raised blood pressure is an important cardiovascular risk factor in Cambodia; however, national averages may obscure substantial local geographic heterogeneity. We assessed the spatial distribution of raised blood pressure and its demographic, metabolic, behavioural, built-environment, and environmental correlates using the 2023 Cambodia STEPwise approach to noncommunicable disease (NCD) risk factor surveillance (STEPS) survey. Methods: We analysed 289 georeferenced primary sampling units (PSUs) from the 2023 Cambodia STEPS Survey. PSU-level case counts and numbers of adults with valid blood pressure data were modelled using a frequentist binomial generalised additive model with a two-dimensional Gaussian-process smooth. An explanatory model included selected PSU-level and contextual covariates; a separate mapping model used spatially continuous predictors to generate predictions on a 5-km grid. Design-based national and residence-specific prevalence estimates were calculated separately using the survey design and Step-2 weights. Results: Among 4,225 adults with valid blood pressure data, unweighted prevalence of raised blood pressure was 28.6%, compared with 18.8% after Step-2 weighting; weighted prevalence was 17.5% in rural areas and 21.7% in urban/suburban areas. Mean observed unweighted PSU-level prevalence was 29.0%. In the adjusted spatial explanatory model, the proportion aged 60-69 years ({beta}=1.969, SE=0.327; p<0.001) and overweight/obesity prevalence ({beta}=0.996, SE=0.265; p<0.001) were positive ecological correlates. The spatial model had lower Akaike information criterion and residual deviance than its non-spatial counterpart, although Bayesian information criterion was higher. Prediction uncertainty was greater in some sparsely sampled areas. Conclusions: Raised blood pressure showed significant geographic heterogeneity across Cambodia. Older age composition and overweight/obesity prevalence were positive PSU-level correlates, not individual-level causal effects. Predicted prevalence should be interpreted alongside uncertainty and should complement, rather than replace, design-based STEPS prevalence estimates for geographically informed NCD planning.

Importance: Coccidioidomycosis, also known as Valley fever, is a serious fungal disease caused by inhaling Coccidioides spores aerosolized from the soil. Despite rising public health concern, the distribution and incidence of coccidioidomycosis in the US remain poorly characterized as not all states are required to report cases, and case under-ascertainment is presumed to be high, particularly outside of the endemic region. Objective: To estimate monthly, county-level coccidioidomycosis incidence across the US while accounting for widespread under-ascertainment. Design, Setting, and Participants: We integrated multisource data, including human cases diagnosed in electronic health records, serological testing of dogs, and ecological and sociodemographic predictors, in a machine learning model estimating monthly, county-level coccidioidomycosis incidence. The model was trained on surveillance data from Arizona and California from 2013 to 2023, then applied to predict incidence in states across the contiguous US over the same time period. Main Outcomes and Measures: Incidence of coccidioidomycosis at the monthly county level. Results: The model predicted coccidioidomycosis incidence from 2013 to 2023 for 3,108 counties from all states in the contiguous US. The model identified 60 counties in non-reporting states where predicted incidence was [&ge;]2 cases per 100,000 population (considered moderate to high incidence). In states with mandatory reporting outside of Arizona and California, the model predicted county-level incidence rates that were 28% higher (95% bootstrap interval, 0.1%-69.8%) than those captured in national surveillance, equating to a total of 608 potentially missed cases per year (95% bootstrap interval, 340-964). Conclusions and Relevance: In this cross-sectional study, the model quantified case under-ascertainment for coccidioidomycosis relative to Arizona and California, and highlighted the potential for alternative data sources to fill in gaps in mandatory reporting. Estimating incidence while accounting for under-ascertainment and non-reporting may help efforts to improve public and physician awareness and to detect emergence in new regions.

This systematic review aimed to synthesize evidence on the association between DNA methylation (DNAm) age and all-cause and disease-specific mortality risks in adults. A systematic search of six databases from Jan 1 2011 to April 15 2026 identified articles assessing the association of DNAm age measures and all-cause or disease-specific mortality in adults from general or disease-specific populations. Effect sizes were pooled as hazard ratios (HRs) with 95% confidence intervals (CIs) using inverse-variance random-effects models. Of 42 023 records screened, 125 articles (380 705 participants) met the inclusion criteria, with 53 included in meta-analyses. Overall, 48 distinct DNAm clocks were evaluated. Higher DNAm ages were associated with higher risks of all-cause, cardiovascular, and cancer mortality risks. In general populations, GrimAge epigenetic age acceleration (EAA) showed consistent associations across mortality outcomes, with each 1 year higher GrimAge EAA associated with ~6-10% higher all-cause (31 cohorts, n=53 642, HR per SD: 1.46 [95% CI 1.38, 1.55], HR per 1-year: 1.09 [95% CI 1.08, 1.11], I^2=84.1%, p-value=1.21 x 10^(-14)), cardiovascular (6 cohorts, n=30 751, HR per SD: 1.55 [95% CI 1.38, 1.73], HR per 1-year: 1.10 [95% CI 1.07, 1.13], I^2=28.0%, p-value=1.71 x 10^(-4)), and cancer (5 cohorts, n=27 187, HR per SD: 1.31 [95% CI 1.15, 1.50], HR per 1-year: 1.06 [95% CI 1.03, 1.10], I^2=33.6%, p-value=4.76 x 10^(-3)) mortality. Clocks incorporating additional surrogates and technical refinements showed numerically larger pooled associations than their original counterparts. GrimAge2 (HR per SD: 1.65 [95% CI 1.53, 1.77] vs 1.46 [1.38, 1.55]) and PCGrimAge (1.56 [1.36, 1.78] vs 1.51 [1.32, 1.74]) showed higher pooled HR association with all-cause mortality than GrimAge. Further work is warranted to evaluate the responsiveness of DNAm clocks to interventions and potential clinical utility.

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.

Introduction: Primary care electronic health records (EHRs) are an important resource for health research, but death recording in primary care EHRs may be incomplete or differ from Office for National Statistics (ONS) death registrations. We used OpenSAFELY-TPP to assess the completeness and concordance of death recording and provide recommendations for using date-of-death information. Methods: Using linked primary care EHRs and ONS death registrations, we assessed whether ONS-registered deaths from 2020-2025 had a primary care date of death and, where present, calculated the difference between dates. Where a primary care date was absent, we assessed death-related SNOMED CT codes. Analyses were conducted overall and by year, population subgroup, and general practice. We also examined GP registration end dates and primary-care-recorded deaths without a corresponding ONS record, including monthly patterns near the data extraction date. Results: Among 1,409,275 ONS-registered deaths, 88% had the same date recorded in primary care and 8% had no corresponding primary care date. Among 1,324,995 deaths recorded in primary care, <2% had no corresponding ONS record. Deaths recorded in only one source were more frequent among people aged <45 years and those from Mixed, Asian or Asian British, Black or Black British, and Chinese or Other ethnic groups. ONS-only dates were also more frequent in London and the West Midlands. Primary-care-only deaths increased close to date of data extraction, reaching 7% in February 2026. Same-day GP deregistration increased from 9% in 2020 to 85% in 2025. Conclusions: Death recording in primary care EHRs showed high completeness and concordance with ONS registrations, but discrepancies remained for recent deaths and some population groups. Data recency and population-specific completeness should therefore be considered when using linked mortality data for research.

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.

Awareness of comorbidities in people with dementia is vital context for dementia research and clinical management. Natural Language Processing (NLP) applied to free text in the electronic health record (EHR) may help ascertain comorbidity presence. Our aim was to assess the validity of the MedCAT Comorbidity NLP app in a cohort of people with dementia. We used text from the EHR of South London and Maudsley NHS Mental Health Trust via Clinical Records Interactive Search (CRIS), linked to structured data from local primary care (Lambeth DataNet) and hospital admission data in a cohort of 3,543 people with dementia. We evaluated MedCAT Comorbidity, examining internal validity via manual review of text and external validity by cross-validation against linked primary care and hospital data; and comparison of the predictive value of comorbidity measures for mortality. Internal precision was highest for atrial fibrillation (100%) and lowest for diabetes (78%). Cross-validation with linked data varied substantially across conditions, with precision from 96% to 59% and recall 77% to 4%. The median number of comorbidities identified was two using any single source, or three when combined. MedCAT Comorbidity demonstrated predictive ability for mortality (HR 1.11 per condition), although the model fit was better using comorbidity from hospital admission data. In people with dementia, the MedCAT Comorbidity app showed high precision for most conditions during internal validation. The variability of precision and recall on external validity assessment partly reflects incomplete documentation. Results from the MedCAT comorbidity app can assist researchers in understanding comorbidity of cohorts, particularly when primary care or hospital data is not available. Our findings highlight the need to develop quality assurance protocols that include internal and external validity checks before deploying general data extraction apps for specific clinical subpopulations.

Background: Ambient artificial-intelligence (AI) scribes are being deployed beyond medicine into allied-health practice, but comparative evidence mainly concerns physician documentation. Objective: To compare quality and transcript-grounded clinical-risk error burden of AI-generated and allied health professional notes from identical simulated consultations. Methods: Paired simulation across dietetics, physiotherapy, occupational therapy and speech and language therapy (five consultations each; 20 pairs; 40 notes). Heidi generated an unedited first-draft note and a professional documented each encounter. Professionals rated PDQI-9 quality. Professional and automated review identified unsupported statements and omissions; two language models graded clinical risk. Co-primary outcomes were PDQI-9 score and Critical+High error density. Results: Mean PDQI-9 scores were 38.45 and 35.05 for AI and professional notes (paired difference +3.40, 95% CI +0.01 to +6.79; Wilcoxon p=0.042; dz=0.47). The Bayesian estimate was +3.47 (94% equal-tailed credible interval +0.66 to +6.30; P[AI>professional]=0.989). Automated-reviewer all-error density was 3.35 vs 8.55 per note (professional-minus-AI difference 5.20, 95% CI 3.85 to 6.55); omission density was 1.45 vs 5.20 (difference 3.75, 2.80 to 4.65). Critical+High density was 0.35 vs 0.75 (difference 0.40, -0.10 to 1.00). The pooled quality result was driven principally by dietetics. Conclusions: AI-generated notes achieved modestly higher pooled quality and fewer transcript-grounded errors, chiefly omissions; serious-error evidence was inconclusive. Small discipline samples and incomplete ground truth preclude safety or discipline-wide superiority claims.