JAMA

In 2027, the National Institutes of Health (NIH) will no longer support costs associated with clinical trials (CT) for career development (K) awards. NIH's CT definition encompasses many types of research, from early-stage pilot work through definitive studies. The types and sizes of trials K awardees conduct have not been described, which could inform ongoing policy changes relevant to the next generation of clinical trialists. We conducted a cross-sectional analysis of NIH RePORTER data to characterize the types and sizes of CTs supported by NIH K01, K08, and K23 awards funded in fiscal years 2024-2025 under a CT Required or CT Optional funding opportunity announcement (FOA), with non-missing abstract text. A rule-based classification scheme applied to free-text abstracts categorized each award into 5 mutually exclusive trial types: pilot/feasibility/proof-of-concept, efficacy/effectiveness, trial confirmed with type unclear, other (mechanistic, ancillary, or phase I/II), or no trial-specific language identified. Trial size was assessed among abstracts reporting an explicit, manually verified sample size. Of 1336 unique awards funded under CT Required (n=1051) or CT Optional (n=285) FOAs, 815 (61.0%) were classified as pilot, feasibility, or proof-of-concept studies, and 55 (4.1%) as efficacy or effectiveness studies; 204 (15.3%) had a confirmed trial of unclear type, 41 (3.1%) were other, and 221 (16.5%) had no trial-specific language. Among CT Required awards, 730 (69.5%) were pilot, feasibility, or proof-of-concept work. Among 273 awards (20.4%) with a verified sample size, the median was 60 participants (IQR, 44-80), with 236 (86.4%) enrolling [≤]100 participants. Given these findings that CT within NIH K awards are generally small and early-stage, NIH's decision to no longer support trial costs may adversely affect the scientist development pipeline.

Background: The 2025/26 influenza season included A(H3N2) subclade K predominance with A(H1N1)pdm09 co-circulation and subsequent prolonged influenza B activity. The Canadian Sentinel Practitioner Surveillance Network reports vaccine effectiveness (VE) in the context of genetic and antigenic relatedness to circulating viruses. Methods: We estimated VE by test-negative design among outpatients with medically-attended acute respiratory illness between November and April. We characterized case viruses by whole genome sequencing and hemagglutination inhibition assay. Results: Among 2809/9047 (31%) influenza-positive specimens, 82% (2306/2809) were influenza A with 18% (504/2809) influenza B. Of subtyped influenza A, 85% (1848/2169) were A(H3N2). We sequenced 59% (1099/1848) of A(H3N2) viruses, with 87% (956/1099) subclade K, antigenically-mismatched to subclade J.2 vaccine. VE against A(H3N2) and subclade K was 38% (95% confidence interval: 27%-47%) and 33% (19%-45%). We sequenced 69% (222/323) of A(H1N1)pdm09 viruses, virtually all (n=218) clade 5a.2a.1, including subclades D.3.1 (21%; n=47) and D.3.1.1 (77%; n=171), antigenically-matched to subclade D vaccine. VE against A(H1N1)pdm09, subclades D.3.1 and D.3.1.1 was 28% (4%-47%), -21% (-144%-40%) and 44% (14%-64%). We sequenced 82% (413/504) of influenza B viruses, all B/Victoria, including antigenically-mismatched subclade C.3 (18%; n=74) with D197N gain-of-glycosylation, and antigenically-matched subclade C.5 (82%; n=339) relative to subclade C vaccine. VE against influenza B, subclades C.3 and C.5 was 55% (41%-66%), 10% (-60%-49%) and 68% (51%-79%). Conclusions: The 2025/26 influenza vaccine reduced the risk of A(H3N2), A(H1N1)pdm09 and B/Victoria by over one-third, one-quarter and one-half, respectively. Laboratory-based indicators of vaccine match remain uncertain proxies for actual vaccine protection, reinforcing importance of annual epidemiological monitoring of VE.

Introduction No widely adopted guidelines exist for collecting and reporting donor-level metadata in tissue-based research, limiting interpretability, reproducibility, and potentially introducing bias. This study aimed to inform ethical and appropriate metadata practices. Methods Semi-structured interviews were conducted with 16 investigators from the Human BioMolecular Atlas Program. Thematic analysis using inductively derived codes identified metadata elements and perspectives on their collection and reporting. Results Participants identified 80 metadata variables across six domains: demographic, sociodemographic, medical history, personally identifying information, cause of death, and tissue/organ data. Most supported routine collection and reporting of demographics and medical history, whereas views on cause of death and sociodemographic data were mixed. Conclusion We recommend routinely collecting and reporting demographics and medical history, while restricting cause of death and sociodemographic variables to situations with explicit consent or justification. These findings provide initial evidence to inform ethical donor metadata guidelines, with further stakeholder engagement and consensus-building needed.

Introduction Health care proxies serve as surrogate decision-makers for patients with limited decisional capacity, yet only one in three U.S. adults have designated one. This trial will evaluate low-cost, scalable behavioral interventions embedded in previsit mailed outreach to increase health care proxy documentation. Methods and analysis In this 2 x 2 + 1 factorial superiority trial, approximately 20,000 patients with an upcoming in-person primary care visit at a Northern California health system will be randomized over a 12-month period into one of four intervention mailers encouraging designation of a health care proxy, or to no mailer under usual care. Eligible patients will have no documentation of a health care proxy in the electronic health record (EHR) and will be aged 55 years or older, or 18 years or older with cancer, congestive heart failure, or chronic kidney disease. The four mailers differ by the presence or absence of a pre-commitment prompt, and whether the mailer is sent from the patient's primary care provider or the health system. The primary outcome is EHR-documentation of a health care proxy by 30 days after the index scheduled visit date. Secondary outcomes include proxy documentation through 90 days, legal advance care planning documentation through 30 and 90 days, and visit attendance. We will use logistic regression to estimate the intention-to-treat effect of receiving any intervention mailer versus no mailer, a pre-commitment mailer versus a mailer without pre-commitment, and a provider-signed versus health system-signed mailer. Ethics and dissemination The University of Southern California Institutional Review Board approved the protocol and granted a waiver of informed consent for this minimal-risk trial. Results will be posted on ClinicalTrials.gov within one year of the primary completion date. Trial Registration The trial was prospectively registered on December 12, 2025 at ClinicalTrials.gov (NCT07290478).

Background: Diabetes is a major global public health challenge, associated with substantial medical, economic, and social burden worldwide. As clinicians working at the intersection of medicine and nutrition, we observed that despite advances in pharmacologic therapy, many individual, particularly those living in underserved urban communitie, continue to experience suboptimal diabetes control due to persistent barriers related to food insecurity, cultural food practices, and limited access to practical nutrition education. In response, we implemented a community-based pilot study in an underserved urban area that focused on dietitian-led nutrition education grounded in Food Is Medicine principles. Aims: The intervention emphasized culturally responsive, cost-conscious, and client-centred approaches designed to make healthy eating simple and intuitive, with the goal of complementing standard diabetic medical therapy. Methods: We used conventional diabetic marker responses, fasting blood sugar (FBS), HgbA1c (percentage of glycosylated hemoglobin, subsequently documented as A1c), and urinary albumin:creatinine ratio (UACR) in a dual cohort design study measuring whether groups receiving dietitian-led nutrition education in addition to conventional therapy would show improvements in diabetic markers relative to groups receiving conventional therapy alone. Summary: Our pilot study summed over both cohorts support the hypothesis that additional dietitian-led nutrition education synergizes with conventional pharmaceutical therapy to improve diabetic control as gauged by improved FBS, A1c and UACR, implying a significant role for this combined approach to community based diabetic health care.

Background and Objectives: Dozens of projects around the world are sequencing the genomes of healthy babies near birth. Genomic information becomes actionable across infancy, childhood and adulthood, making timing of return a design choice for programs. We report the first study of health care professionals' views on timing. Methods: We conducted a qualitative interview study of US-based clinical geneticists, genetic counselors, laboratory personnel, pediatric primary care clinicians and genomic screening implementers. Participants responded to three strategies: staged throughout childhood when information becomes actionable, all at birth with adult-actionable results deferred, and all at birth. Transcripts were analyzed using framework analysis. Results: We interviewed 52 individuals; 39 were asked directly about timing. Giving parents a choice, raised by participants rather than presented, was the most endorsed position (18), ahead of staging across childhood (12). Many viewed staging as preferable in theory, but feasibility concerns weighed against it, including that 'actionability' was not a robust enough concept. A further concern was that parents would not grasp the distinction between data generated and data examined, on which staging depends. Where information is staged, participants saw a role for adolescent assent; where it is not, disclosure to the developing child becomes important, and participants identified a lack of support for parents. Conclusions: Tying the return of information to the age of actionability is intuitive but hard to operationalize. Parental choice was the most endorsed position but will only be viable with decision support and guidance for disclosure to children.

The US spends more on health care than any other nation, yet tens of millions of Americans are uninsured or underinsured, and coverage retractions enacted in 2025 are widening these gaps. The misalignment between the for-profit insurance architecture and optimal patient care, together with the inefficiencies of a fragmented system, contributes to both unnecessary costs and preventable mortality. We update our previous analyses with the most recent data to project the economic benefits and the number of lives saved that would be achieved by single-payer universal coverage, as proposed in the Medicare for All Act. We estimate that such a system would reduce national health expenditure by $1,041 billion annually. Sources of savings include reductions in administrative overhead, pharmaceutical prices, fraudulent billing, and avoidable emergency care. Combined with the reversal of recent retractions, universal coverage would save over 114,000 lives annually.

Fairness audits of chest radiograph artificial intelligence usually score subgroup performance against labels extracted from radiology reports by natural language processing (NLP). We tested whether these labels can make a model look fairer than it is, a failure we call phantom fairness. On 4,376 NIH ChestX-ray14 images with both NLP and radiologist-adjudicated labels, we audited classifiers from three backbones by sex, age, projection and older women. NLP labels missed 54% to 59% of radiologist-confirmed airspace opacity, pneumothorax and nodule or mass, and every fracture. These misses were not random: models scored them lower than NLP-caught positives for every finding and backbone (median rank-biserial correlation 0.20 for the primary backbone and 0.16 for the others; all adjusted p [≤] 0.038), so the labelling errors hid the models' own mistakes. Subgroup gaps were larger on adjudicated labels in 16 of 24 comparisons, none individually significant. For pneumothorax in women older than 60 years, the NLP audit showed a sensitivity gap of 0.064 against a true gap of 0.184; removing the hardest positives reproduced this concealment (0.124 against 0.120 observed), whereas random label errors produced less (0.077). Fairness audits based on NLP-extracted labels should be treated as provisional until checked against radiologist-adjudicated labels.

Background: Papillary thyroid microcarcinoma (PTMC, equal to or smaller than 1 cm) is increasingly diagnosed, and a subset behaves aggressively. Preoperative ultrasound (US) and fine needle aspiration (FNA) accuracy in this size range, particularly with multinodular disease, is poorly defined outside East Asian screening cohorts. Materials and Methods: We retrospectively analyzed 180 consecutive PTMC cases (2017 to 2026) with preoperative US and FNA and compared them with 191 benign thyroidectomy controls. ROC analysis identified an optimal size cutoff. We compared sensitivity, specificity, PPV, NPV, and accuracy across tumor-size subgroups, before/after 2023, and with/without dominant nodules (DNs) equal to or larger than 2 cm. Results: ROC analysis identified 5 mm as the optimal cutoff (AUC 0.80). FNA sensitivity was 44.4% overall, falling to 14.1% for tumors <5 mm versus 61.2% for [&ge;]5 mm, with specificity uniformly high (94.8%). US showed a similar pattern (sensitivity 23.4% vs. 64.7%) but lower specificity (69.6%), driven largely by calcification and Hashimoto's-related hypoechogenicity. Combined FNA+US improved sensitivity to 59.4%. DNs equal to or larger than 2 cm markedly reduced FNA true-positive rates (67.9% vs. 10.8%; OR 17.4). Excluding DN cases raised FNA sensitivity to 67%, NPV to 84.2%, and accuracy to 85.2%. Post-2023 experience significantly improved all metrics (sensitivity 34.4% to 54.4%). Conclusion: FNA and US sensitivity for PTMC decline sharply below 5 mm and in multinodular glands with dominant nodules, while specificity/PPV remain high. A 5 mm threshold, operator experience, and selective avoidance of FNA in dominant-nodule bearing multinodular goiters (favoring direct surgical referral) can meaningfully improve diagnostic performance and guide management.

Background: The 2026 American Heart Association/American College of Cardiology (AHA/ACC) guidelines replaced the 2019 European Society of Cardiology (ESC) four-tier pulmonary embolism (PE) risk scheme with five clinical categories (A-E) and subcategories. These categories were set by expert consensus and have not been validated against outcomes. How patients are reclassified relative to ESC, or how the two systems compare prognostically, is unknown. Methods: We utilized three cohorts of patients with confirmed PE using structured electronic health record data, laboratory biomarkers, and large-language-model abstraction of radiology reports: Duke University Health System (n=12,992, drawn from 95,760 consecutive inpatient CT pulmonary angiography studies, 2014-2025, with no referral or registry enrollment step between imaging and cohort entry), INSPECT (Stanford; n=3,870), and MIMIC-IV (Beth Israel Deaconess; n=361). Patients were assigned AHA/ACC categories B through E, subcategorized where data allowed, and mapped to 2019 ESC risk strata. The primary outcome was 30-day mortality; discrimination was assessed with Harrell C-index. Results: Among 17,223 patients with confirmed PE, pooled 30-day mortality rose monotonically across categories: 1.5% (B), 8.9% (C), 15.5% (D), and 31.9% (E), with the ordering preserved in all three cohorts despite differing baseline mortality. Subcategory-level discrimination was reliable only at the high-acuity extreme (D2-E2); across subcategories C1 through D1, mortality did not order monotonically (9.2%, 10.8%, 8.1%, 10.9%), and adding subcategories to category C did not improve discrimination at Duke (C-index 0.699 vs 0.699). Category C patients lacking both echocardiography and biomarker testing (12.7% of category C) had mortality (10.4%) equal to or exceeding classified peers. Relative to ESC, the frameworks were concordant at the extremes, but 5.7%of ESC intermediate-risk patients were reclassified to category D, with modestly higher but non-significant 30-day mortality than those remaining in category C (10.8% versus 8.9%). Conclusions: Across a three-health-system cohort, the 2026 AHA/ACC framework produced a reproducible mortality gradient at the category level, with added subcategory granularity refining risk chiefly at the highest-acuity tiers. Discrimination across the broad intermediate band was limited, and reclassification from ESC fell almost entirely within this range.

Background: Five nonprescription statin attempts have failed; a sixth attempt, under the FDA's 2025-issued Additional Condition for Nonprescription Use (ACNU) pathway, is under way. Self-selection evidence is evaluated by whether ineligible consumers are excluded; how often eligible consumers are denied has never been reported. Objective: To quantify denial of access to eligible consumers across the published self-selection record. Methods: Secondary analysis of the three published United States self-selection studies (SELECT, CREST, TACTiC), the FDA guidance, and four advisory committees' questions and votes. We recomputed diagnostic accuracy from published cross-tabulations, derived TACTiC's prespecified unmitigated endpoints, and compared access strategies by decision curve analysis. Results: The guidance's endpoints contain no measure of denial, and every committee question addressing a selection error concerns admission of the ineligible. CREST's sensitivity was 59.0% (95% CI 42.1%-74.4%), 29.5 percentage points below its reported positive predictive value; 41% of clinician-eligible participants were not approved. TACTiC's reference standard reached only application-approved consumers; none of the 10,332 rejected was assessed; sensitivity is inestimable and its prespecified unmitigated endpoints, 84.2% and 80.9% against the published mitigated 90.7% and 98.1%, remain unpublished. Decision curve analysis favors the application over both alternatives across reasonable thresholds at the sample prevalence. Conclusions: For nonprescription statin applications, the self-selection evidence framework counts errors of admission and leaves errors of denial unmeasured; its public health objective, fewer eligible adults untreated, cannot be evaluated on evidence counting only admissions. Four design and reporting changes, none requiring new methodology, would let the next application measure it.

BACKGROUND Extensive research has documented medication abortion safety. Large-scale insurance claims data offer an important complement to clinical studies by assessing real-world outcomes in routine practice. In claims data, researchers must infer mifepristone use, clinical outcomes, and event severity from diagnosis, treatment, and utilization codes. Understanding how these analytic choices influence safety and effectiveness estimates is important for interpreting evidence used to inform care delivery and medication policy. METHODS We conducted a retrospective cohort study of mifepristone use with national multipayer claims data, both medical and pharmacy, from 2017 through 2023. We identified both a broad cohort and a more specific medication-abortion cohort after excluding evidence of other indications. We classified potential safety events, effectiveness outcomes, and health care utilization within 45 days of mifepristone use. Serious adverse events required claims-observable consequences consistent with FDA seriousness criteria. RESULTS The broad cohort included 869,409 mifepristone uses among 698,780 unique patients for medication abortion or other indications. After restriction to uses with additional evidence of medication abortion, 775,931 uses among 626,443 unique patients remained. In the medication-abortion cohort, 3,306 uses were followed by at least one serious adverse event, resulting in an absolute 45-day risk of 0.43%. Serious bleeding or hemorrhage occurred after 0.24% of uses, serious infection or sepsis after 0.05%, serious hypersensitivity or anaphylaxis after 0.01%, and serious ectopic-pregnancy outcomes after 0.05%. CONCLUSIONS Serious adverse events after mifepristone use were rare in this national study. Applying established, clinically grounded criteria markedly reduced the number of events classified as serious.

Background. Giant cell arteritis (GCA) is the commonest systemic vasculitis and a medical emergency. High-dose glucocorticoids are started on suspicion, before confirmation, and cause substantial dose-dependent harm in the many patients who prove not to have GCA. Diagnostic confirmation by temporal artery biopsy or ultrasound remains imperfect. The performance of both approaches is influenced by the clinical context and, for ultrasound, operator expertise, creating the potential for both unnecessary treatment and delayed or missed diagnosis. A laboratory-based, multi-analyte proteomic blood test may help, but its intended use and performance requirements are undefined. Objective. To develop, by structured expert consensus, a Target Product Profile (TPP) for a laboratory-based, multi-analyte proteomic blood test to aid the diagnosis of cranial GCA in secondary care, used alongside clinical assessment, imaging and/or biopsy, with the primary objective of reducing unnecessary glucocorticoid exposure without increasing missed cases, from a UK health-system perspective. Methods and analysis. A multidisciplinary committee of approximately 15 to 30 members, including patient and public representatives, will follow a three-phase process: a scoping phase that confirms the intended use and the prioritised list of characteristics; a drafting phase that specifies minimum (acceptable) and preferred (desirable) values across the TPP domains, using outcome-based methods where feasible; and a consensus-building phase using an up-to-three-round modified Delphi (75% consensus threshold) with interim consensus meetings and a 28-day public consultation. Reporting follows ACCORD and CREDES.

Genetic testing for cardiomyopathy and arrhythmia (CM/ARRH) provides diagnostic information and informs screening for at-risk relatives. Clinical guidelines recommend genetic testing for these conditions; however, data on how genetic test results influence clinical management recommendations are limited. Here, we determined the frequency of cardiologist-recommended management changes for patients following CM/ARRH genetic testing. This was a retrospective cross-sectional study of patients referred for multigene panel testing between April 2016 and April 2024. Genetically-experienced cardiologists at multicenter academic clinical practices were recruited for participation to complete surveys indicating clinical decision making on patients who had genetic testing. Cases for review were randomly selected to have both positive and non-positive results. Among 249 patients (138 positive, 111 non-positive), 136 (54.6%) received clinical management recommendations for their own or their at-risk relatives? care. Of these, 75 (55.1%) received recommendations for the patient?s own care, most frequently additional diagnostic tests/procedures (n=33). Compared to non-positive results, patients with positive results were more likely to receive recommendations for their own management (66/138, 47.8% vs 9/111, 8.1%; P<0.00001). Patients with positive results in arrhythmogenic cardiomyopathy genes had 263% higher odds of recommended management changes compared to those with TTN (OR=3.63, CI:1.40-9.84, P=0.009). The results from CM/ARRH genetic testing on affected patients influenced cardiologists? medical decision making and management recommendations. Additional research is needed to evaluate how genetic testing for CM/ARRH impact health outcomes.

Background The phase 3 ADVOCATE trial evaluated the efficacy and safety of avacopan in patients with granulomatosis with polyangiitis (GPA) or microscopic polyangiitis (MPA). Concerns raised regarding the 2019 primary endpoint adjudication process prompted a blinded, independent readjudication of all participants' primary outcomes, the results of which are described here. Methods Patients with GPA or MPA were randomized 1:1 to receive oral avacopan 30 mg twice daily or oral prednisone on a scheduled taper, each in combination with rituximab- or cyclophosphamide-based standard of care. In 2026, the Duke Clinical Research Institute Clinical Events Classification group conducted an independent, blinded committee re-adjudicated the Birmingham Vasculitis Activity Score (BVAS), relapse, and remission from weeks 26 through 52 using procedures aligned with the original adjudication charter. The primary endpoints were remission at week 26 and sustained remission at week 52. As per the original analysis plan, noninferiority and superiority were declared if the lower bounds of the 95% confidence interval (CI) for the difference in the primary outcome rates between avacopan and a prednisone taper were greater than -20.0 and 0.0 percentage points, respectively. Results Among 330 participants in the intent-to-treat population, remission at week 26 was achieved by 68.1% (113/166) and 67.1% (110/164) of participants in the avacopan and prednisone taper groups, respectively, in the 2026 readjudication (adjusted difference: 2.2%; 95% CI, -7.5, 11.9), compared with 72.3% (120/166) and 70.1% (115/164) in the 2019 primary outcome adjudication (adjusted difference: 3.4%; 95% CI, -6.0, 12.8). Sustained remission at week 52 was achieved by 61.4% (102/166) and 52.4% (86/164) of participants, respectively, in the 2026 readjudication (adjusted difference: 9.8%; 95% CI, -0.3, 19.9), compared with 65.7% (109/166) and 54.9% (90/164) in the 2019 readjudication (adjusted difference: 12.5%; 95% CI, 2.6, 22.3). Concordance between the 2019 and 2026 adjudications was 95.2% for remission and 93.6% for sustained remission. Conclusion The re-analysis of ADVOCATE based on the 2026 readjudication further supports the efficacy of avacopan for GPA/MPA. Non-inferiority of avacopan versus a prednisone taper was confirmed at weeks 26 and 52 despite a median 81% reduction in glucocorticoid exposure observed in the avacopan versus prednisone taper groups. While a consistent numerical difference favoring avacopan at week 52 was observed in the 2019 and 2026 analyses, this difference did not reach statistical superiority.

Purpose As genomics permeates all healthcare specialties, genetic counseling in conjunction with genetic testing is broadly recommended. Improving access to genetics professionals is crucial for Americans insured by Medicaid. The purpose of this study was to conduct a comprehensive review of Medicaid policies for genetic counseling performed by Certified Genetic Counselors (CGC(C)). Methods Fee-for-service Medicaid policies across 50 states and Washington DC were reviewed and coded. Four states with exemplary policies were identified, and CGC managers in two of these were surveyed regarding the real-world effects of these policies. Results As of 2024, 20 states (39%) had a published policy for genetic counseling with most (N=16, 80%) expressly covering genetic counseling in connection with any covered genetic test. Of the states without a policy, 12 (24%) mention genetic counseling in the context of scenario-specific policies, and 19 (37%) have no published policy. Twenty states explicitly cover CGC services, while 2 exclude CGCs as service providers. CGC managers in Indiana and Michigan confirmed the policies identified as exemplary successfully led to reimbursement of CGC services. Conclusion There is significant variability in Medicaid coverage for genetic counseling. Comprehensive policies are needed to support patient access to genetics professionals, including CGCs.

Stigmatizing language in medical documentation may reflect and perpetuate bias, but its prevalence in obstetrics has not been systematically quantified. We applied a keyword-guided BERT classifier to 640,345 obstetric notes from 26,178 pregnancies at an academic medical center. Stigmatizing language was detected in 47% of 26,178 pregnancies. Black pregnancies had significantly higher odds of stigmatizing language compared with Asian (aOR=1.5, p=3x10-8) or White (aOR=1.4, p=6x10-6). Indicated and spontaneous preterm births were also significantly associated with stigmatizing language compared to term (aORs=1.5, 1.2; p=7x10-12, 0.01). Pregnant individuals with only 12th-grade maternal education were more likely to experience stigma than those with college (aOR=1.5; p=4x10-14). These findings provide evidence of differences in clinical documentation across race, education levels, and clinical conditions. They also demonstrate how automated natural language processing can enable systematic monitoring of bias in healthcare language at scale.

Background: Clinical language models must use evidence to produce the number and action required by a particular case. Whether numerical knowledge reliably becomes a correct clinical response is unclear.`n`nMethods: NUMBERS evaluated 16 model configurations on 1,300 questions linked to public clinical evidence. Linked experiments tested prevalence updating, patient-specific estimates and clinical actions. The direct-action experiment compared cutoff recall plus action selection with a supplied complete rule across 50 rules, five models and 7,500 calls.`n`nResults: Among diagnostic estimates outside the source-result tolerance, 77.1% remained within the evidence's central 80% predictive range. Models updated prevalence-dependent quantities correctly in 78.5% of comparisons with inputs and a calculation request, versus 33.7% with clinical wording. Supplying inputs and requesting calculation raised patient-level near-target answers from 29.7% to 78.8% across 2,340 pairs. Models selected an incorrect action despite stating a cutoff that implied the correct action in 413 of 3,750 recall-arm calls (11.01%; 95% confidence interval, 8.93 to 13.17). Supplying the complete rule raised action accuracy from 85.63% to 99.41%, an improvement of 13.79 percentage points (95% confidence interval, 11.65 to 15.95).`n`nConclusions: Models often produced evidence-consistent numbers but failed to adapt them to a case or act consistently with their own stated cutoff. Explicit inputs, calculations and complete rules substantially improved performance in controlled prompts.`n`nFunding: National Academy of Medicine, Agreement No. 2026A008797.

Background: GPs are often the first clinicians to assess acute chest pain in gatekeeper healthcare systems, but usually with limited diagnostic support. They must distinguish conditions that can be managed safely in primary care from those requiring immediate hospital evaluation, particularly acute coronary syndrome (ACS), while avoiding unnecessary over-referral to already pressured emergency departments. We evaluated HEART-GP, a novel GP-led strategy combining structured clinical assessment, fingerstick point-of-care high-sensitivity troponin I testing, and ECG when available. Methods: We performed a prospective, multicentre, paired diagnostic accuracy study in four Dutch out-of-hours primary care centres implementing the HEART-GP strategy. Patients with acute non-traumatic chest pain and possible ACS according to the treating GP were enrolled from March 2023 to August 2025. Before troponin results were available, GPs recorded intended management, allowing paired comparison with unaided clinical judgement. The HEART-GP strategy recommended urgent referral for high clinical suspicion, troponin of >=4 ng/L, or ischaemic ECG. The primary outcome was adjudicated 6-week major adverse cardiovascular events (MACE). The study was registered as ISRCTN11954040. Findings: Of 1104 patients who underwent HEART-GP, 917 were included. MACE occurred in 62 (6.8%) and myocardial infarction in 52 (5.7%). HEART-GP classified 478 (52.1%) as low risk. For MACE, sensitivity was 95.2% (95% CI 86.5-99.0) and negative predictive value was 99.4% (98.1-99.8); three MACE occurred in the low-risk group, with no cardiac deaths. For myocardial infarction, sensitivity was 98.1% (89.7-100) and negative predictive value was 99.8% (98.6-100). Compared with unaided GP judgement, HEART-GP improved sensitivity for MACE (95.2% vs 67.7%; p<0.001) and myocardial infarction (98.1% vs 67.3%; p<0.001), and reduced urgent referrals (439 [47.9%] vs 483 [52.7%]; p=0.042). Interpretation: HEART-GP showed high rule-out safety and may support GP diagnostic certainty and efficient referrals in patients with possible ACS. Implementation requires training, governance, safety-netting, and local pathways.

Introduction: Maternal morbidity and mortality in the US disproportionately affect some women, including the Medicaid-insured, rural, and Black, Hispanic, and Native American populations. The drivers of disparities are complex and there is an urgent need for multilevel interventions focused on pregnancy-related and associated morbidity and mortality (PRAMM). This study will be the first large-scale test of Community Health Worker (CHW) inclusive home visiting and provider/practice/system level improvement strategies that support empowered patients, pro-active providers, and integrates community and clinical care designed to reduce PRAMM and disparities among women who are disproportionately affected. Methods and analysis: The primary study outcome will be rates of PRAMM - a composite outcome including pregnancy-associated and pregnancy-related morbidity, severe maternal morbidity, and mortality - and disparities among women disproportionately affected. A quasi-experimental, stepped wedge design will be used. Participants will be Medicaid-insured women in three Michigan counties observed during pregnancy, at birth, and up to one year postpartum, who give birth between 2021-2028 (>101,000 births). Individual health outcomes will be assessed at three steps of the stepped-wedge design. Analyses will use a statewide linked data system including all Medicaid birth and death records, Medicaid claims, and other program data. In the context of Michigan extending pregnancy-related Medicaid, the study will evaluate the impact of the proposed multilevel intervention vs. usual care from early pregnancy through 12 months postpartum. The study was developed with full community engagement and participation including shared leadership with academic and community-based Principal Investigators representing intervention communities. Ethics and dissemination: An Institutional Review Board determined the study exempt from the human subject's regulations, for the following reasons. Two of the three intervention components will not interact with human subjects to collect data - we will only rely on secondary research using data or biospecimens not collected specifically for this study, and the data will be provided without identifiable information by someone without any role in this research study except providing the data. The third intervention component, medical provider surveys, is a benign behavioral intervention where information will be recorded using methods that prevent subjects' identities from being readily ascertained. Results will be published in peer-reviewed journals following EQUATOR guidelines.