Clinical Trials

Background: Adolescent substance-use rehabilitation is a care-continuum problem spanning detection, engagement, active treatment, relapse prevention, aftercare, family support, and equity-oriented implementation. Existing reviews are often modality-specific and do not show how evidence aligns with substances, populations, outcomes, stages of care, or policy needs. Objectives: To map and synthesise the 2015-2025 adolescent and transitional-age youth SUD rehabilitation literature across intervention domains, stages, substances, outcomes, equity/disadvantage, geography, and economics, and to perform meta-analysis only where pooling was clinically defensible. Methods: PubMed, Scopus, and Web of Science records were harmonised to 2015-2025 and deduplicated. Two reviewer roles applied a predefined charting codebook for substance focus, technique family, rehabilitation stage, equity/disadvantage flags, outcome family, and study-design signal. Evidence was synthesised across AI/digital, psychiatric/psychotherapeutic, pharmacological, family/social, behavioural, residential/continuing-care, school/community, harm-reduction, and policy domains. Random-effects meta-analysis was restricted to comparative youth OUD medication-supported trials with extractable binary outcomes. Results: The search identified 1,676 records; 554 duplicates were removed, leaving 1,122 unique records. Metadata screening retained 579 records for evidence-map charting: 112 high-confidence records and 467 conservative metadata-supported records requiring full-text verification before final selective-journal submission. The charted evidence was concentrated in active treatment (n=433) and relapse prevention (n=114); aftercare/follow-up was weak (n=8). Intervention-family signals were led by pharmacological/MOUD (n=72), psychotherapy/psychiatric care (n=65), school/community/brief interventions (n=46), residential/continuing care (n=41), family/social therapy (n=30), AI/digital/telehealth (n=25), harm-reduction/policy (n=24), and CM (n=22). The primary youth OUD retention/completion meta-analysis favoured medication-supported treatment (OR 7.67, 95% CI 3.98-14.78; I^2=0%; k=2; n=188). An exploratory favourable-outcome analysis produced a similar estimate (OR 7.94, 95% CI 4.24-14.89; I^2=0%; k=3; n=229). Conclusions: The strongest pooled quantitative claim supports medication-supported treatment for youth OUD. For non-opioid substances, digital care, family therapy, CM, residential care, aftercare, and equity-oriented implementation, the literature is clinically important but not yet consistently synthesis-ready. Future trials should evaluate complete care pathways, adopt core outcomes, report age-banded and equity subgroup effects, and include economic and implementation endpoints.

Background: Manual review of 30-day hospital readmissions can identify actionable quality and safety problems, but it is labor-intensive. We developed and evaluated an agentic AI workflow for evidence-grounded readmission review. Materials and methods: We studied adult patients with unplanned 30-day readmission after discharge from a medicine hospitalist service at a single academic health system. An AI agent using a large language model queried a database containing notes, encounters, procedures, laboratory results, and other clinical data, and completed the same structured readmission-review rubric used by physicians. In the primary comparative evaluation, 20 randomly selected readmissions from 2025 were each reviewed by two physicians and the AI system. Blinded physician evaluators rated review quality. After rubric refinement, the AI workflow was applied to 100 recent readmissions in an exploratory expanded-cohort analysis of recurring improvement opportunities. Results: In the primary comparative evaluation, the AI classified 9/20 readmissions (45%) as preventable, compared with 19/40 physician reviews (47.5%). Blinded overall quality ratings were similar for AI and physician reviews (4.35 vs. 4.20 on a 1-5 scale; mean difference 0.15, 95% CI -0.20 to 0.48; p=0.49), as were factuality/support and usefulness/actionability ratings. No AI hallucinations were identified during factuality review. Agreement on preventability and primary readmission category was low for both AI-human and human-human comparisons. The AI system cost $0.23 per chart; physician reviewers took a median of 15 minutes, corresponding to an estimated $42.43 per chart. In the exploratory expanded-cohort analysis, AI-assisted review identified recurring vulnerabilities in post-discharge follow-up plans, incomplete inpatient workups, medication-safety transitions, and indwelling-device transitions. Conclusions: Agentic AI produced readmission reviews with similar blinded quality ratings to physician reviews in this small single-center primary comparative evaluation and supported identification of recurring quality-improvement themes in the exploratory expanded-cohort analysis. Preventability judgments remained variable among both AI and physicians, underscoring the need for human oversight and prospective evaluation before operational use.

Randomized controlled trials (RCTs) play a central role in assessing the benefits and harms of interventions. Incomplete reporting in RCT publications can compromise the verifiability and usefulness of RCTs. SPIRIT and CONSORT reporting guidelines aim to improve the completeness of RCT protocols and results publications, respectively. However, many RCTs are not reported completely. Checking manuscripts automatically could help authors improve the completeness of reports prior to publication. We previously annotated SPIRIT-CONSORT-TM, a corpus of 200 articles (comprising 100 protocol-results publication pairs) using 83 checklist items drawn from SPIRIT 2013 and CONSORT 2010. We also trained machine learning models to automatically assess reporting at the item level. Each checklist item can include multiple constituent elements (i.e., specific details required for that item), and an item might be considered fully reported when all of its elements are present. However, prior work does not explicitly capture or evaluate reporting at the element level. To address this gap, we extended SPIRIT-CONSORT-TM by incorporating element-level annotations and using them to assess reporting completeness (SPIRIT-CONSORT-ELM). We formulated element-level assessment as a machine reading comprehension task, operationalized through 119 questions, where each question targets a specific reporting element within a checklist item. Using the 200 articles included in SPIRIT-CONSORT-TM, two annotators independently answered 119 questions for 50 articles (25 protocol-results pairs) and resolved any discrepancies through discussion; the remaining 150 articles (75 protocol-results pairs) were assessed by a single annotator. We then developed an automated pipeline for element-level assessment using SPIRIT-CONSORT-ELM. The pipeline first applies a PubMedBERT-based model to identify sentences containing item-level reporting information, then it uses a generative large language model (LLM; GPT-5) with chain-of-thought reasoning to answer element-level questions based on the retrieved evidence. Agreement between the two annotators was high (Gwet's AC1: 0.782) and our pipeline achieved high accuracy in identifying element-level reporting evidence (F1: 0.822, Gwet's AC1: 0.796). Ablation studies indicate that chain-of-thought reasoning and the inclusion of illustrative in-context examples modestly improve LLM performance on the machine reading comprehension task. SPIRIT-CONSORT-ELM provides a benchmark for evaluating reporting guideline completeness at the element level, enabling assessment of RCT transparency beyond the simple presence or absence of checklist items and is publicly available at https://osf.io/kznx4/. The automated pipeline establishes a robust baseline for assessing RCT reporting and demonstrates potential as a practical aid for authors, reviewers, and editors to identify and address gaps in completeness and transparency of RCT reports.

Clinical trials often enrol populations that differ from those who ultimately receive the interventions, raising concerns about external validity and health equity. Trial registries could provide an early opportunity to assess representativeness, but it is unclear whether registry data contain sufficient information to enable such assessments. This study evaluated the feasibility of using registry data to assess representativeness in Phase II and III pharmacological randomised controlled trials. A search of ClinicalTrials.gov from December 2024 to January 2025 identified trials with results posted after 1 January 2023 across cardiovascular disease (CVD) excluding stroke, diabetes mellitus, and selected mental health disorders. Of 1,328 records screened, 98 trials met inclusion criteria (51 Phase III, 47 Phase II). Reporting completeness was variable, particularly in Phase II studies. CVD and diabetes trials predominantly included middle-aged to older adults, while mental health trials recruited mainly individuals aged 36 to 50 years. Across CVD and mental health trials, participants were largely male. Reporting of BMI, contraception, and comorbidity criteria was inconsistent, though available data suggested these factors influenced sample composition. Fewer than 10% of trials reported equity-relevant characteristics beyond age and sex, and none addressed intersectionality. Assessing equity using registry data is feasible but constrained by incomplete and inconsistent reporting.

Importance: Combinations of VEGFR tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs), such as antibodies to programmed cell death-1 (PD-1), or to its ligand PD-L1, are now first-line standard of care for renal cell carcinoma (RCC), but the pivotal clinical trials excluded patients with common comorbidities, leaving their real-world effectiveness uncertain. Objective: To determine whether adding PD-1/PD-L1 inhibitors to VEGFR-TKIs therapy is associated with improved overall survival in a real-world RCC cohort. Design, Setting, and Participants: This retrospective cohort study used a target trial emulation framework and real-world electronic health records data from the University of Florida Health Integrated Data Repository (IDR). Data was analyzed from September 2009 through June 2023. Adult patients ([≥]18 years) with confirmed RCC and at least one VEGFR-TKIs prescription were eligible. The date of the first VEGFR-TKIs prescription was defined as the index date, and patients were followed for up to 24 months. Variable-ratio propensity score matching (up to 2:1) across 13 baseline covariates was used to emulate randomized treatment assignments. Of 107,783 patients screened, 387 met eligibility criteria, and 319 remained in the matched cohort. Exposures: VEGFR-TKIs monotherapy (control group) versus VEGFR-TKIs combined with PD-1/PD-L1 inhibitors (experimental group). Main Outcomes and Measures: Overall survival (OS), analyzed by weighted Kaplan-Meier estimation, cluster-robust Cox regression, and restricted mean survival time (RMST) at {tau} = 24 months, prespecified given anticipated non-proportional hazards. Results: Among 319 matched patients (mean [SD] age, 62 [12] years; 76% male), 107 deaths occurred (33.5%). Twelve-month OS was higher in the combination arm (81.8%; 95% CI, 74.7--89.6%) than VEGFR-TKIs monotherapy (68.1%; 95% CI, 61.1--76.0%), converging by 24 months (61.1% vs 56.7%). The Cox hazard ratio was 0.718 (95% CI, 0.484-- 1.064; P = 0.0986). RMST was 2.79 months greater with combination therapy (95% CI, 0.93-- 4.65; P = 0.0033). Conclusions: Adding PD-1/PD-L1 inhibitors to VEGFR-TKIs therapy was associated with a statistically significant and clinically meaningful gain in restricted mean survival, supporting the real-world generalizability of combination therapy and the importance of appropriate treatment effect measures under non-proportional hazards.

Meta-analysis usually reduces each study to an effect estimate with a standard error and pools these by inverse-variance weighting: fixed effect (FE), random effects (RE), or unrestricted weighted least squares (UWLS). We propose information-geometric meta-integration (IGMI), representing each study by its sampling distribution, the Gaussian N(theta_i, Sigma_i), and pooling studies as a weighted Frechet mean (barycenter) under Bures-Wasserstein (BW), Fisher-Rao, or Wasserstein-Fisher-Rao (WFR) geometry. In the scalar fixed-variance case the BW barycenter mean is exactly the FE estimate; the minimized Frechet functional reproduces the Higgins-Thompson I^2 and DerSimonian-Laird tau^2 heterogeneity statistics; and a Frechet-scatter pivot reproduces the Hartung-Knapp-Sidik-Jonkman interval at m = 1 and yields an exact Hotelling F(m, K-m) region for m outcomes under proportional total covariances. WFR adds a robust outlier-resistant pool: as its length scale delta grows without bound it converges monotonically to BW, whereas finite delta gives a redescending M-estimator with rejection point exactly pi*delta. Simulations show calibrated multivariate coverage at small K, where Wald intervals undercover, and strong resistance of the equal-weight WFR pool to contamination. In 2,445 Cochrane meta-analyses, WFR most often wins leave-one-out predictive scoring. In 835 bivariate meta-analyses, the closed-form BW barycenter matches REML multivariate meta-analysis predictively and is exactly invariant to the unreported within-study correlation, unlike the likelihood estimate.

Retention in antiretroviral therapy care remains a major challenge in high-burden settings such as Malawi, where substantial loss to follow up undermines treatment outcomes and long-term epidemic control. Although machine learning models can accurately identify patients at high risk of disengagement, there is limited evidence on how these predictions can be translated into improved retention outcomes in practice. This study addresses this gap by linking machine learning-based risk stratification to the targeted allocation of retention interventions, providing a framework for evaluating their expected impact on ART retention outcomes. We developed a patient-level Monte Carlo simulation model that integrates individual predicted probabilities of loss to follow up from a validated Extreme Gradient Boosting model with intervention effect sizes derived from a meta-analysis of ART retention interventions conducted in sub-Saharan Africa. The study population included 1,705 ART patients receiving care at Lighthouse Trust clinics in Lilongwe, Malawi. Patients were stratified by predicted risk, and the highest-risk group (n = 512) was targeted for intervention. Six interventions were evaluated, including Expert Client support, psychosocial support, two-way text messaging, adherence clubs, community ART groups, and teen clubs, followed by subgroup-specific and combined approaches allocated based on predicted risk. The primary outcome was twelve-month ART retention, estimated over 5,000 simulation iterations. Subgroup and post-simulation analyses were conducted to assess heterogeneity in intervention response. Among patients classified as high risk (n = 512), baseline retention was 44.1%. Individual interventions improved retention to 52.7% with two-way texting (RR = 1.19; p < 0.001) and 55.0% with Expert Client support (RR = 1.25; p < 0.001). A combined intervention package produced larger gains, increasing retention to 64.0% (RR = 1.45; p < 0.001), corresponding to an absolute improvement of 19.9 percentage points. Intervention effects varied across subgroups, with significant improvements observed among newly initiated patients (43.0% to 58.9%; RR = 1.37; p < 0.001) and clinically unstable patients (28.3% to 39.1%; RR = 1.38; p = 0.01), while effects among adolescents were more modest (34.3% to 45.6%; RR = 1.33; p = 0.03). Despite these improvements, 46% of high-risk patients remained hard to retain after receiving multiple interventions. In this subgroup, expected retention increased only marginally from approximately 0.15 at baseline to 0.20 after intervention, with poor outcomes observed among patients who were virally unsuppressed, had depressive symptoms, or were younger. Machine learning-guided targeting of ART retention interventions can substantially improve retention outcomes, particularly when interventions are combined. However, a substantial subgroup of patients remains hard to reach and vulnerable to disengagement, indicating that existing strategies may be insufficient for individuals with complex clinical and psychosocial needs. This study contributes to knowledge by introducing an integrated framework that combines machine learning risk prediction, meta-analytic intervention effects, and patient-level Monte Carlo microsimulation to quantify twelve-month ART retention outcomes under risk-based targeting with subgroup-specific intervention allocation before real-world implementation. These findings highlight the potential of using individual risk to guide the delivery of retention interventions within routine ART programs to enable more efficient, proactive, and patient-centered allocation of retention resources.

Objectives. To present DrugSet, a validated R Shiny application supporting the construction medicinal products codelists based on the Anatomical Therapeutic Chemical (ATC) system and their mapping to Clinical Practice Research Datalink (CPRD) Aurum prodcodes within a single interactive workflow. Materials and Methods. DrugSet comprises four modules: data preparation, ATC-based hierarchical code selection, string-based CPRD Aurum prodcodes mapping, and codelist export. Validation was conducted against World Health Organization (WHO) ATC reference codelists and manually curated prodcodes mappings across three drug classes: metformin, beta-blocking agents, and topical salicylic acid. Sensitivity, specificity, and Positive Predictive Values (PPV) were calculated for ATC codelist generation. Agreement proportions (overlapping against total identified codes) were calculated for prodcodes mapping. Time needed for codelist construction using DrugSet was recorded and compared to manual approaches. Results. DrugSet ATC codelist generation against WHO manual reference achieved 100% sensitivity, specificity, and PPV across all medicinal products. Prodcodes mapping agreement ranged from 89.2% to 98.3% with discrepancies due to missing data in the prodcodes input vocabulary. DrugSet completed codelist construction in 9 minutes compared to 3 hours and 10 minutes manually, across all medicinal products classes. Discussion. DrugSet provides a unified workflow that runs directly on ATC and source CPRD Aurum vocabulary files. The reduction in codelist construction time and export of the generated codelists supports reproducibility in pharmacoepidemiologic studies where codelist creation can represent a significant proportion of study setup time. Conclusion. DrugSet is an open-source, validated tool that improves accuracy, and efficiency of codelist construction for medicinal products based on ATC codes towards CPRD Aurum prodcodes.

Background: The UK Government aims to reduce smoking rates by implementing new, and investing in existing, tobacco control strategies including increased funding for Stop Smoking Services (SSS) in England. This study examined whether the additional funding starting in April 2024 was associated with a detectable increase in quit attempts supported by SSS and whether it was cost-effective. Methods: We used data from the Smoking Toolkit Study, a repeat cross-sectional survey conducted in 2021 to 2025. Adults aged [&ge;]18 years who smoked cigarettes and had made a quit attempt in the past year were included (weighted n=5,076). The outcome was monthly prevalence of past-year quit attempts supported by SSS. We fitted general additive models with a step change in April 2024 to represent the start of the increased funding. We adjusted for tobacco tax increases, the Swap-to-Stop scheme, age, gender, and a measure of socioeconomic position. In an unplanned analysis, we extended the time series back to 2006. For the cost-effectiveness, we estimated incremental cost-effectiveness ratios for the total population and age groups, accounting for future lifetime cessation. Results: In the primary model, the April 2024 step change was not statistically significant (adjusted odds ratio: 1.13; 95% CI: 0.52, 2.49). The cost-effectiveness analysis ranged from cost-effective to extremely ineffective (incremental cost-effectiveness ratio (ICER): GBP 104,126, 95% CI: 939,398 to 8,293). When using the extended time series, the adjusted odds ratio for the step change was 2.70 (95% CI: 2.03, 3.60) and the intervention was cost-effective (ICER: GBP 13,857; 21,393 to 9,620). Conclusions: Compared with the long-term trend, increased funding to SSS in England in 2024 appeared to lead to an increase in quit attempts supported by SSS at the population level. This result is somewhat uncertain because our primary pre-planned analyses assessing the impact relative to a more recent trend were insensitive.

Clinical oncology datasets often store systemic therapy as a regimen label with a start date and an end date. Those records are clinically recognizable but can be analytically incomplete when the research question concerns whether a patient was exposed to a concurrent CNS-active drug (cCNS-aD) or an adjuvant CNS-active drug (aCNS-aD) around radiotherapy. Contemporary CNS-oncology studies usually define CNS activity by empiric drug lists and define concurrency by fixed calendar windows, although the literature shows substantial heterogeneity across both concepts. This paper proposes a generalizable framework for converting raw systemic therapy records into reproducible cCNS-aD and aCNS-aD variables, useful in subgrouping for clinical studies. The framework uses a transparent CNS scoring model based on three clinical evidence components: intracranial objective response rate, consensus CNS endorsement, and intrathecal route of administration. It then defines a pharmacokinetic exposure proxy as the recorded end date plus five half-lives. Concurrent exposure is classified by overlap with the radiotherapy interval, while post-radiotherapy exposure is classified by overlap with a prespecified post-RT attribution window. The framework separately identifies post-RT pharmacokinetic persistence and post-RT treatment initiation, allowing investigators to distinguish continued exposure from true adjuvant initiation. This is a methodological framework and reference implementation. Implementation audits and endpoint-specific sensitivity analyses remain necessary before use as a definitive exposure classifier

Background: Retrieval of adverse event reports based on coded drug-event co-occurrence enables large-scale pharmacovigilance analyses, but yields candidate reports rather than validated cases, risking misinterpretation if used alone. Aim: To develop and apply a framework for identification and characterization of clinically meaningful case series in pharmacovigilance. Methods: We conducted two case studies. The first developed and refined the framework in an information-rich setting, focusing on drug-induced impulsivity across selected drugs; the second tested its applicability in a more routine, information-poor setting, focusing on drug-induced suicidality. Results: In Case 1, non-relevant reports were frequent for drugs with uncertain evidence and negative controls ({approx}20-40%) compared to drugs with established causal roles (4%). The emerging framework assessed relevance based on exposure, event, drug-event relationship, and population. For suspected adverse drug reactions, relevant reports were further characterized by reporter suspicion and evidentiary qualifiers supporting or refuting causality; higher suspicion was associated with more supportive qualifiers. Applied to Case 2, the framework ruled out 69% of reports as non-relevant but highlighted substantial non-assessability (17%). Conclusions: In pharmacovigilance, retrieval is not equivalent to case identification. Relevance is question-specific and shaped by how reports are captured, processed, and retrieved. This can be especially critical for emerging or bias-prone safety questions. Transparent and reproducible case definition and adjudication are essential for interpretable analyses.

Importance. Large language models (LLMs) increasingly inform mental health decisions by patients and clinicians. Inference-time activation steering can shift model behavior on a target dimension without altering weights or prompts and without disclosure to users, allowing treatment recommendations to be silently changed for commercial or ideological reasons. Objective. To determine whether directional activation steering can shift an open-weights LLM's depression treatment recommendations. Design, Setting, and Participants. This non-human subjects study applied directional activation steering to an open-weights LLM (DeepSeek V4 Flash) responding to 12 depression-advice scenarios (4 favoring medication, 4 favoring avoidance, 4 neutral), generated at 30 amplitudes from -1.5 to +1.5 in 0.1 increments plus an unsteered baseline. Exposures. A single steering direction contrasting antidepressant medication with self-directed approaches (diet, exercise, meditation, dietary supplements), constructed from 16 paired training prompts and applied at the attention output of every transformer block; weights and system prompt were held constant. Main Outcomes and Measures. The extent to which medication and four self-care categories were addressed, scored 0 to 3 by a human-validated LLM rater (Claude Opus 4.7), the medication-versus-self-care balance, and clinician referral, estimated per unit of amplitude using mixed-effects models with a scenario random intercept. Results. Across 372 generations, steering produced a graded, dose-dependent shift in the medication-versus-self-care balance, which declined by 0.32 per unit of amplitude (beta=-0.32; 95% CI, -0.39 to -0.25; P < .001); medication extent fell and self-care extent rose. The shift was largest for scenarios with no stated treatment preference (beta = -0.44; 95% CI, -0.54 to -0.34; P < .001). A clinician referral appeared in 322 of 372 responses (87%) and did not vary with steering amplitude (P = .63). Conclusions and Relevance. In this open-weights LLM providing depression treatment information, inference-time activation steering shifted treatment recommendations without altering weights, prompt structure, or safety outputs, with the largest effect among users expressing no treatment preference. These findings suggest a need for LLM disclosure standards and independent auditing as such models inform clinical decisions.

Introduction: Primary care clinicians (PCC) commonly make decisions in the context of time delay and uncertainty. Delay discounting (DD) and probability discounting (PD) are cognitive biases related to delay and uncertainty that are minimally explored in PCC. We assessed DD and PD in PCC and evaluated their association with low-value care (LVC) decision-making. Methods: We administered a survey to PCC in a Southeastern U.S health system and within the American Academy of Family Physicians networks. The survey comprised standardized psychometric assessments of DD and PD and four LVC clinical vignettes. Outcomes included DD and PD discounting rates for two monetary rewards ($100 and $10,000) and ratings of LVC likelihood (0-100). We used regression analysis with model selection to evaluate the relationship between variables. Results: 225 PCC (89% physicians, 11% advanced practice providers) participated. Heterogeneity in DD and PD rates was observed. For the $10,000 reward, ln k(DD)= -6.80, IQR:-7.60--6.10) and ln h(PD)= 1.75, IQR:1.75-2.36). The reward amount impacted DD and PD in opposing directions (i.e., lower DD/higher PD rates for $10,000 vs. $100). LVC likelihood was highest for low-value antibiotics and lowest for low-value cervical cancer screening (median 20, IQR:10-40 and 0, IQR:0-10, respectively). Model selection revealed demographic associations with LVC likelihood, but no association with DD or PD. Conclusions: Consistent with effects previously reported in non-clinicians, PCC exhibited a range of DD and PD, which ranged by reward magnitude. Neither DD nor PD predicted vignette-based LVC likelihood. Further research should investigate actual clinical practice patterns and other LVC scenarios.

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have been linked to heterogeneous, potentially pleiotropic effects across organ systems, motivating outcome-wide comparative risk profiling in real-world data. A central challenge in such analyses is \emph{residual bias} that remains after adjustment for observed confounders, which can distort effect estimates and mis-calibrate uncertainty. We present distributional diagnosis and calibration (DC), which uses panels of negative control outcomes (NCOs) to diagnose residual bias and calibrate uncertainty. DC evaluates null behavior via $p$-value uniformity and empirical coverage across NCOs, and uses the empirical distribution of NCO effect estimates to calibrate confidence intervals for prespecified primary outcomes. DC is modular: it can wrap around commonly used causal inference methods and operates directly on summary statistics, supporting collaborative research under data-sharing constraints. Using electronic health records from a large U.S. clinical research network (152.7 million patients), we compared GLP-1RAs with sodium--glucose cotransporter~2 inhibitors across 15 prespecified outcomes spanning cardiovascular, mental health, and genitourinary domains using four causal estimators. Across outcomes and methods, DC diagnostics revealed substantial and method-dependent residual systematic error. DC calibration attenuated systematic error signals observed in negative controls and yielded more stable, better-calibrated estimates for clinical outcomes, supporting DC as a practical strategy to strengthen the credibility of real-world comparative effectiveness research.

Elective surgery late cancellations and ``did not attend'' (LCDNA) events waste theatre capacity, lengthen waiting lists, and impose avoidable costs on NHS Trusts. We present a decision-support approach that ranks upcoming elective procedures by expected cancellation cost and supports capacity-constrained outreach by selecting the highest-risk Top-K cases for intervention. Using cost-sensitive learning and a clinically grounded cost model, the policy reduces expected cost from approximately 103 GBP per case under business-as-usual to 77.08 GBP per case in a hospital-holdout (cross-site) evaluation designed to mimic deployment to a new hospital. In a complementary time-forward evaluation, representing prospective use within the same service environment, expected cost falls further to 70.97 GBP per case. The 6.11 GBP per-case difference between the two regimes highlights the added uncertainty introduced by cross-site operational shift and supports a conservative roll-out with local calibration and monitoring. Explainability analyses suggest that booking-to-procedure lead time, specialty or service line, calendar effects, and prior cancellation history are the strongest drivers of prediction, helping to inform tiered intervention workflows that prioritise near-term bookings and use model--pathway mismatches as an audit signal. Overall, the framework turns predictive performance into practical, capacity-aware policy guidance for reducing avoidable cancellations while supporting safe and equitable implementation.

Clinical informatics pipelines increasingly compute validated clinical endpoints from upstream NLP outputs. Even when the endpoint is defined by an established rubric, translating that rubric across representations - natural language instructions, program logic, and reference implementations - can introduce specification drift, where ostensibly equivalent calculators yield meaningfully different scores. We study this phenomenon for the Expanded Disability Status Scale (EDSS), a standard measure of disability in multiple sclerosis. Holding constant a shared set of functional system (FS) subscores extracted by a large language model (LLM), we compare EDSS values computed across three representations of the same scoring rubric: prompt-executed natural language, LLM-generated code, and a canonical reference implementation. We characterize disagreement structure, distributional shifts, and clinically salient boundary flips, and we propose an audit workflow that treats endpoint computation as a first-class verification target in clinical NLP systems.

Medical practice is bottlenecked by the slow production of high-quality clinical evidence. Despite progress in automating selected stages, autonomous conduct of the entire research life cycle remains beyond reach. Here we present OpenEBM, the first autonomous system to generate decision-grade clinical evidence by conducting evidence-synthesis research end to end. To enable and evaluate this, we develop OpenEBM-Corpus, a foundation resource of expert-annotated research trajectories that enables training of a specialist model, and OpenEBM-Bench, a multidisciplinary benchmark that evaluates the entire research life cycle. Our compact specialist model generates valid clinical evidence in 90.7% of end-to-end evaluations and matches expert performance across the research trajectory, whereas GPT-5 falls to 3.8% as failures propagate through dependent stages. In blinded evaluations across clinical domains, independent evaluators prefer OpenEBM at multiple stages and cannot distinguish its reasoning traces from expert-conducted work above chance. Applied to a question left unresolved by current guidelines, OpenEBM produces de novo evidence addressing the efficacy and safety of neoadjuvant chemotherapy for locally advanced rectal cancer. OpenEBM brings within reach the founding aspiration of evidence-based medicine and establishes a paradigm for scalable evidence generation.

The rapid expansion of the medical artificial intelligence (AI) literature has outpaced our ability to judge how far published models have progressed towards clinical use. We investigated whether the translational maturity of a study can be estimated automatically from its abstract. Using PubMed, we assembled a corpus of 11,024 candidate articles, reduced it to 1,816 AI-related articles by heuristic filtering, and manually double-annotated a balanced sample of 524 articles across five maturity classes (internal validation, external validation, prospective evaluation, implementation or governance, and not applicable). Abstracts were represented as TF-IDF features and classified using multinomial logistic regression with a Lasso penalty, chosen for interpretability and suitability for a small, imbalanced dataset. On a stratified held-out test set (n = 104), the model achieved 69.2% accuracy, Cohen's kappa of 0.495, macro-F1 of 0.458 and a weighted AUC of 0.820. Performance was strong for the frequent classes but poor for the rare implementation or governance class, which the model failed to recover. A balanced manual verification of 200 large-corpus predictions confirmed this pattern, with per-class precision ranging from 82.5% (internal validation) to 5.0% (implementation or governance). An interpretable, low-resource classifier can support literature mapping but requires human oversight for advanced maturity levels.

Objective: Online health information seeking is rising, and individuals increasingly act on peer advice without clinical oversight, adjusting doses, delaying care, and modifying treatment. Current misinformation detection assumes factually inaccurate content is what makes these decisions unsafe. We introduce VERITAS (Verification Engine for Risk-aware Information Trust Assessment in health Stories) and formalize the Risk Irrelevance Principle: divergence from accepted clinical practice and potential for harm are distinct, weakly associated dimensions that must be assessed separately. Materials and Methods: VERITAS transforms unstructured health narratives into Agent-Action-Outcome graphs and computes two continuous metrics: Narrative Truth Distance (NTD), quantifying epistemic divergence, and Narrative Risk Score (NRS), assessing harm potential. We evaluated VERITAS on 704 threads from four Reddit health communities. Two domain experts annotated 2,000 segments (Krippendorffs =0.78-0.81). NTD-NRS independence was validated using seven tests. Results: NTD and NRS shared under 5% of variance (r = 0.222; mutual information 0.096 bits): a posts divergence from consensus conveys little about whether acting on it will cause harm. On 435 labeled posts, VERITAS identified 62.2% of expert-labeled misinformation versus 57.5% for the strongest text classifier, the gain concentrated in factually plausible content describing unsafe self-management (27.6% of misinformation) that accuracy-focused classifiers approve. VERITAS assessed 37.8% of this misinformation as low-risk, pending clinical validation. Discussion: Fact-checking-based screening systematically approves the content most likely to prompt unsafe self-management while flagging content least likely to cause harm. Conclusion: Separating divergence from harm potential shifts verification from whether information is correct to whether it is safe to act upon.

Background: Longer follow-up periods in clinical trials for S. aureus bacteremia (SAB) may capture unrelated deaths, adding random noise that risks biasing trial results towards the null. Objective: To evaluate the timing and infection-relatedness of deaths within a large SAB clinical trial platform. Design: Blinded duplicate adjudication of trial deaths using a modified 7-point Likert-Scale. A third reviewer settled disagreements. Setting: 37 Canadian hospitals participating in the S. aureus Network Adaptive Platform (SNAP) Trial. Participants: 1515 adult patients recruited to SNAP between February 2022 and May 2026. Measurements: Timing and relatedness of 90-day deaths categorized as at least possibly SAB-related not likely to be SAB-related. Optimal follow-up cut-off was determined using Youden's index and graphically. Results: 247 deaths occurred; 97 (39.3%) were adjudicated as at least possibly SAB-related and 150 (60.7%) as not likely related. For probably/definitely related deaths, interrater agreement was 85.0% (Gwet's AC 0.73, substantial); for at least possibly related, it was 77.3% (Gwet's AC 0.55, moderate). Median survival was significantly shorter for SAB-related deaths (12 vs. 30.5 days; difference: 19 days earlier, 95% CI: 12-26, p<0.0001). Nearly 80% of SAB-related deaths occurred by day 30, whereas 50% of unrelated deaths occurred between days 30 and 90. Youden's index optimized follow-up at 20.5 days. Limitations: Potential for cause of death misclassification and data limited to Canadian sites. Conclusion: Deaths considered attributable to SAB cluster rapidly within the first month, while later deaths are predominantly unrelated. A 30-day all-cause mortality window may be more appropriate than 90 days for primary mortality outcomes in trials evaluating acute SAB therapies with longer follow up reserved for metastatic infection and recurrence.