Medical Decision Making

Detailed agent-based simulations are increasingly used to support policy decisions, but their computational cost and complex uncertainty structure make systematic scenario analysis challenging. We present a data-driven, uncertainty-aware decision support (DDUADS) workflow for using stochastic simulation models as decision-support tools under limited computational budgets. The approach combines several established techniques-sensitivity screening, Bayesian calibration using simulation-based inference, and multi-surrogate model integration for translational efficiency-into a coherent pipeline that enables uncertainty-aware policy analysis. Rather than producing a single baseline, the calibration stage yields a posterior distribution over plausible model parameterizations, allowing flexible, uncertainty-aware forward projections. We demonstrate the DDUADS workflow on the INFORM-HIV agent-based model of HIV transmission in Chicago to evaluate potential disruptions in antiretroviral therapy (ART) and pre-exposure prophylaxis (PrEP) use. While the specific application is HIV modeling, the challenges and techniques described here arise in other simulation studies and can be applied to decision support in other domains.

Lightweight epidemic calculators are widely used for teaching and rapid scenario exploration, yet many omit the methodological detail needed for scientific reuse. We present a browser-native SIR calculator that exposes forward Euler and classical fourth-order Runge--Kutta (RK4) integration alongside epidemiologically interpretable outputs and a population-conservation diagnostic. The implementation is anchored to analytical properties of the deterministic SIR system, including the epidemic threshold, the peak condition, and the final-size relation. Benchmark experiments show that RK4 is essentially step-size invariant over practical discretizations, whereas Euler at a coarse one-day step overestimates peak prevalence by 3.97% and final size by 0.66% relative to a fine-step RK4 reference. These results demonstrate that browser-based tools can support publication-quality computational narratives when solver choice, diagnostics, and assumptions are treated as first-class outputs.

Large language models are increasingly used for clinical guidance while their parent companies introduce advertising. We tested whether pharmaceutical ads embedded in the prompts of 12 models from OpenAI, Anthropic, and Google shift drug recommendations across 258,660 API calls and four experiments probing distinct epistemic conditions. When two drugs were both guideline appropriate, advertising shifted selection of the advertised drug by +12.7 percentage points (P < 0.001), with some model scenario pairs shifting from 0% to 100%. Google models were the most susceptible (+29.8 pp), followed by OpenAI (+10.9 pp), while Anthropic models showed minimal change (+2.0 pp). When the advertised product lacked evidence or was clinically suboptimal, models resisted. This reveals a structured vulnerability: advertising does not override medical knowledge but fills the space where clinical evidence is underdetermined. An open response sub analysis (2,340 calls across three representative models) confirmed that advertising restructures free-text clinical reasoning: models echoed ad claims at 2.7 times the baseline rate while maintaining high stated confidence and rarely disclosing the ad. Susceptibility was provider dependent (Google: +29.8 pp; OpenAI: +10.9 pp; Anthropic: +2.0 pp). Because this bias operates within clinically correct answers, it is invisible to accuracy based evaluation, identifying a class of AI safety vulnerability that standard testing cannot detect.

Automating Hierarchical Condition Category (HCC) assignment directly from unstructured electronic health record (EHR) notes remains an important but understudied problem in clinical informatics. We present HCC-Coder, an end to end NLP system that maps narrative documentation to 115 Centers for Medicare & Medicaid Services(CMS) HCC codes in a multi-label setting. On the test dataset, HCC-Coder achieves a macro-F1 of 0.779 and a micro-F1 of 0.756, with a macro-sensitivity of 0.819 and macro-specificity of 0.998. By contrast, Generative Pre-trained Transformer (GPT)-4o achieves highest score of a macro-F1 of 0.735 and a micro-F1 of 0.708 under five-shot prompting. The fine-tuned model demonstrates consistent absolute improvements of 4%-5% in F1-scores over GPT-4o. To address severe label imbalance, we incorporate inverse-frequency weighting and per-label threshold calibration. These findings suggest that domain-adapted transformers provide more balanced and reliable performance than prompt-based large language models for hierarchical clinical coding and risk adjustment.

Purpose: High clinical workload is associated with worse patient and hospital outcomes and is a well-established driver of clinician burnout. Trainees may be particularly exposed, shouldering both clinical and educational responsibilities. Evidence-based work design offers a data-driven approach to healthcare work but relies on robust workload measurements. Trainee workload remains poorly characterized, as commonly used metrics (e.g., duty hours, patient census) overlook cognitive and contextual dimensions. This pilot evaluated the feasibility of combining survey-based and electronic health record (EHR) data to characterize internal medicine (IM) trainee workload. Methods: A pilot study was conducted including IM and Medicine-Pediatrics residents (postgraduate years 1-4) between March 31 and June 22, 2025. Participants completed daily surveys during a seven-day inpatient schedule assessing workload and work experience domains, including environment, professional fulfillment, psychological safety, autonomy, and rounding experience, using validated instruments where available. Concurrently, EHR data captured chart review, documentation, orders, and secure messaging activity. Associations between survey and EHR data were assessed. Results: Among 37 eligible residents, 28 (76%) participated in the pilot capturing 166 shifts. Trainees spent 4.4 +/- 1.6 (mean +/- SD) minutes completing daily surveys and 8.6 +/- 2.3 minutes completing the final survey. Trainees reported working 11.6 +/- 1.0 hours/day and a median census of 9.0 (IQR 6.0-11.0). NASA-TLX score was 50.8 +/- 12.6. Positive shift ratings were associated with lower NASA-TLX scores and perceived rounding length. First-to-last EHR login duration was 15 +/- 2 hours/day, and EHR data showed 204 +/- 46 active minutes/day. Login duration correlated with self-reported hours (r=0.43, p<0.0001), and notes signed correlated with self-reported team (r=0.19, p=0.013) and personal census (r=0.34, p<0.0001). Conclusions: Integrating survey-based and EHR-derived workload measures provides multidimensional insight into trainee work. This novel approach supports scalable measurement and evidence-based work design interventions to improve trainee well-being, education, and clinical efficiency.

Background Persons experiencing homelessness (PEH) have a 2-3-fold greater risk for cardiovascular disease (CVD) mortality compared with domiciled counterparts. Evidence has repeatedly shown elevated chronic disease burden, reduced access to many types of care, and lower utilization of medication to control CVD risk factors in clinical settings dedicated to providing health care to PEH. There are federally funded health clinics targeting barriers to access for patient populations experiencing homelessness in place. These clinics are frequently overwhelmed and limited by their scope to primary care despite well documented burdens of co- and tri-morbid conditions. There is scarce evidence on differences between access, quality, and experiences of care delivered relative to other safety-net models. Method The 2022 Health Center Patient Survey (HCPS) was collected on behalf of the Health Resources and Services Administration (HRSA). The HCPS is a nationally representative, three-staged, sample-based survey collected via 1:1 interview with clinic patients. The survey assessed sociodemographics, health conditions and behaviors, access to and utilization of care, and patients? experiences with comprehensive services they received at HRSA-funded Federally Qualified Health Centers (FQHCs), including community health centers (CHC), healthcare for the homeless (HCH) clinics, and public housing primary care (PHPC) clinics. One hundred and three unique awardees and 318 health center sites were recruited, and 4,414 patient interviews were completed. Investigators analyzed patient characteristics and multiple survey items related to AHA?s Essential 8 metrics for differences between HCH and CHC patient responses. Results HCH clinics had fewer elderly patients (~7%) than CHCs (~17%). Reported 7-day physical activity measures, average sleep below 7 hours per day, and Lifetime smoking (>100 cigarettes; OR=4.2, p<0.001) were all greatest among HCH patients. Fewer HCH patients reported ever having or recent lipid tests (both p<0.001). HCH patients were more likely to report hypertension (p=0.003) but less likely to report receiving nutrition advice (all p<0.05). HCH patients were less likely to be taking medication even if it was prescribed (p<0.001). Adjustments for differences in age or CVD history were able to explain some observed differences but increased the magnitude of other disparities. Conclusions CVD burden differs across the various HRSA funding mechanisms for clinics, as do demographics and multiple metrics of health behaviors and biomarkers of cardiovascular health. Greater disease burden in HCH patients is likely compounded by increased risk factors and underperformance in providing health education interventions.

Background Randomized controlled trials (RCTs) often have incomplete methods reporting despite widespread adoption of the CONSORT guideline. The editorial process is supposed to detect these shortcomings and request clarifications from authors, which is time-consuming. We developed an LLM-based CONSORT Rohe Nordberg Report that highlights which CONSORT items appear fully or partially reported and checks page references claimed by authors, and then creates follow up questions for authors to more easily correct missing information. Methods This parallel-arm, superiority RCT will randomize eligible RCT submissions (after desk screening) 1:1 into intervention (editorial team and authors receive the Rohe Nordberg Report) or control (standard editorial review only). The primary outcome is whether manuscripts improve their reporting of CONSORT items in the Methods and Results sections between the original submission and first revision. This will be assessed by blinded human reviewers who evaluate the textual changes for improvements between the original and revised manuscripts for each relevant CONSORT item. Secondary outcomes include time to editorial decisions, rejection and non-resubmission rates, if authors can correctly identify where CONSORT items are reported, and extent of revisions. Human evaluators will be blinded to whether the manuscript was in the intervention or control group. Discussion By providing authors and the editorial team with specific follow up questions for each underreported CONSORT item, we hypothesize that basic underreporting will be more efficiently detected and corrected. Using blinded human reviewers as the primary outcome assessors ensures a rigorous, unbiased evaluation. If successful, this approach may help align manuscripts more closely with CONSORT standards, ultimately benefiting evidence synthesis.

Introduction The link between Human Papillomavirus (HPV) and cancer is well-established. In Singapore, bivalent HPV vaccines are subsidised for females, but not males. Economic analysis of HPV vaccination has generally assessed the costs to the health system, but this may not be as relevant to individual decision-making as potential lost income. We estimated the impact of bivalent HPV 16/18 vaccination on sick leave, unemployment, and premature mortality as a function of age and sex to understand the broader impact of HPV-related cancers. Methods We developed a population-level economic model to estimate lifetime income losses by diagnosis age, sex and cancer type. We applied sex- and cancer-specific Cox regressions to the Singapore Cancer Registry for annual predicted survival from 1992 to 2022. These were combined with census and employment data to estimate HPV-associated income losses in Singapore. Attributable fractions and vaccine effectiveness data for HPV 16/18 from the literature were used to estimate the effectiveness of bivalent HPV vaccination. Structural sensitivity analysis examined the role of 80% population coverage conferring herd immunity. Results The registry contained 17,294 individuals with an HPV-associated cancer diagnosis. Lost income was greatest for cervical cancer due to its high prevalence, however the losses per diagnosis were highest for oropharyngeal cancer. Bivalent HPV vaccination led to income benefits of $SGD1,397 [$895 to $1,838] in girls and -$62 [-$76 to -$48] in boys. A gender-neutral HPV vaccination of 80% of 15-year-old Singaporeans, conferring herd immunity, would have lifetime income protective benefits of $24.4m [$14.2m, $33.7m] per cohort, a five-fold return on investment. Conclusions In addition to avoiding healthcare costs and lost quality of life, parents should consider vaccination as a means of avoiding potential income losses. A national policy of gender-neutral HPV vaccination could deliver substantial income protection due to both individual vaccine protection and herd immunity.

Introduction. Empirical evidence linking specific national structural policies to the provision of key HIV services in low- and middle-income settings remains scarce. This study addresses the research gap by quantifying the within-country relationships between six national structural policy indicators and the presence of the HIV prevention service component targeted at sex workers in Southeast Asia. Methods. We constructed a balanced panel dataset covering eight Southeast Asian countries from 2018 to 2025 from the UNAIDS Global AIDS Monitoring (GAM) framework. We used Fixed-Effects (FE) and Random-Effects (RE) models to analyse the relationships, with the FE model selected as the more statistically appropriate estimator. We enhanced robustness by using clustered standard errors and one-period lagged explanatory variables. Results. The primary finding from the FE model indicated a statistically significant and positive contemporaneous association between the existence of legal or administrative barriers to social protection (barriers_spi,t) and the presence of HIV prevention services for sex workers ({beta} = 0.8531; p < 0.001). However, the robustness check revealed a statistically significant negative association between the two when using the lagged barrier variable (barriers_spi,t-1), suggesting a decline in HIV prevention service availability over time ({beta} = -0.3540; p < 0.05). We did not find any other policy variable's coefficient to be statistically significant in the FE models. Conclusions. While the immediate recognition (contemporaneous effect) of structural barriers to access social protection may occur alongside prioritised HIV prevention service provision, the sustained presence of these impediments acts as a long-term constraint that undermines the effectiveness and sustainability of targeted HIV programmes. National HIV programmes must urgently prioritise the removal of structural barriers to ensure long-term service stability for key populations.

BackgroundIn the absence of structured donor registries, social media platforms have become a dominant mechanism for blood donor recruitment in many low-resource settings. However, the implications of this shift for transfusion timeliness and system reliability remain unclear. ObjectiveTo evaluate the impact of social media-sourced donors on transfusion delay, donor reliability, and hemovigilance-related outcomes compared with conventional donor pathways. MethodsThis prospective analytical study included 400 transfusion episodes across tertiary hospitals in Bangladesh. Donor sources were categorized as social media (SM) or conventional (CON). The primary outcome was delay-to-transfusion. Secondary outcomes included donor-related irregularities, documentation completeness, near-miss events, and acute transfusion reactions. Multivariable logistic regression identified predictors of delay [&ge;]4 hours. ResultsSocial media-sourced donors were associated with significantly longer transfusion delays (5.98 vs 2.97 hours; p<0.001). Delay [&ge;]4 hours occurred in 83.6% of SM cases versus 17.6% of CON cases (OR 23.78). Donor-related irregularities were observed in 85% of SM episodes and absent in CON donors. Safety outcomes did not differ significantly between groups. Social media donor sourcing remained the strongest independent predictor of delay (adjusted OR 18.09). ConclusionUnregulated social media-based donor recruitment introduces substantial delays and undermines system reliability without improving access. Integration of digital tools into regulated donor systems is essential to strengthen transfusion timeliness and hemovigilance in resource-limited settings.

Transgender women are routinely recruited for HIV prevention research and describe feeling over-researched, undervalued, and disconnected from the benefits of research. Research fatigue refers to the adverse impacts of research participation from the volume, frequency, or intensity of research engagement. Research beneficence, an underdeveloped construct, refers to perceptions that research participation is empowering, appreciated, and beneficial to individuals and communities. This study sought to develop and psychometrically evaluate a research fatigue and beneficence scale and examine associations with cohort retention and study procedures among transgender women in the US and Puerto Rico. We developed a novel 7-item measure of research fatigue and beneficence informed by prior literature and qualitative work with transgender women. We assessed internal consistency reliability, factor structure, convergent and divergent validity, and predictive validity with 6-month study retention outcomes and procedures among 2189 transgender women enrolled in a US nationwide cohort (April 2023-December 2024) for the full 7-item research fatigue and beneficence scale, a 4-item research beneficence subscale, and a single-item research fatigue measure. Research beneficence items demonstrated good internal consistency (0.78) and excellent model fit. Research fatigue and beneficence varied by race/ethnicity with participants of color reporting both greater empowerment and greater concerns about community-level benefits. The item "I feel that I am asked to participate in research too frequently" was associated with lower 6-month retention, greater survey missingness, and preference for less invasive HIV testing modalities. Findings highlight multiple dimensions of research experience and the need for reduced participant burden, culturally tailored study designs, and intentional dissemination efforts to improve participant-centered research practices.

ImportanceGuideline-concordant care for young children with attention-deficit/hyperactivity disorder (ADHD) includes recommending parent training in behavior management (PTBM) as first-line treatment. However, assessing guideline adherence through manual chart review is time-consuming and costly, limiting scalable and timely quality-of-care measurement. ObjectiveTo evaluate the accuracy and explainability of large language models (LLMs) in identifying PTBM recommendations in pediatric electronic health record (EHR) notes as a scalable alternative to manual chart review. Design, Setting, and ParticipantsThis retrospective cohort study was conducted in a community-based pediatric healthcare network in California consisting of 27 primary care clinics. The study cohort included children aged 4-6 years with [&ge;] 2 primary care visits between 2020-2024 and ICD-10 diagnoses of ADHD or ADHD symptoms (n=542 patients). Clinical notes from the first ADHD-related visit were included. A stratified subset of 122 notes, including all cases with model disagreement, was manually annotated to assess model performance in identifying PTBM recommendations and rank model explanations. ExposuresAssessment and plan sections of clinical notes were analyzed using three generative large language models (Claude-3.5, GPT-4o, and LLaMA-3.3-70B) to identify the presence of PTBM recommendations and generate explanatory rationales and documentation evidence. Main Outcomes and MeasuresModel performance in identifying PTBM recommendations (measured by sensitivity, positive predictive value (PPV), and F1-score) and qualitative explainability ratings of model-generated rationales (based on the QUEST framework). ResultsAll three models demonstrated high performance compared to expert chart review. Claude-3.5 showed balanced performance (sensitivity=0.89, PPV=0.95, and F1-score=0.92) and ranked highest in explainability. LLaMA3.3-70B achieved sensitivity=0.91, PPV=0.89, and F1-score=0.90, ranking second for explainability. GPT-4o had the highest PPV [0.97] but lowest sensitivity [0.82], with an F1-score of 0.89 and the lowest explainability ranking. Based on classifications from the best-performing model, Claude-3.5, 26.4% (143/542) of patients had documented PTBM recommendations at their first ADHD-related visit. Conclusions and RelevanceLLMs can accurately extract guideline-concordant clinician recommendations for non-pharmacological ADHD treatment from unstructured clinical notes while providing clear explanations and supporting evidence. Evaluating model explainability as part of LLM implementation for medical chart review tasks can promote transparent and scalable solutions for quality-of-care measurement.

BackgroundDigital health plays an indispensable role in facilitating data analysis and use for enhancing healthcare delivery across health settings. However, there is scant information on the extent to which digital health influences the improvement of primary health services delivery through data use. This study examined the determinants that influence the use of digital health to improve health service delivery in council hospitals in Tanzania. MethodsA cross-sectional design was employed in six regions, involving 12 council hospitals. We used a self-administered questionnaire to collect data from 203 members of hospital quality improvement teams. Descriptive analysis was used to determine the frequency, proportion, and mean of responses, while bootstrapping analysis was conducted to test the statistically significant influence of digital health factors on data use for improving health service delivery. ResultsResults show moderate agreement on data compatibility for planning and decision-making, with 40.4% of respondents agreeing it supports ordering commodities, 43.8% for staff allocation, and 38.4% for planning. However, dissatisfaction was higher for user-friendliness (47.8%), reliability (up to 65.5%), and usefulness (up to 63.5%). Overall, 50.2% (M=2.74{+/-}0.87) disagreed that digital systems effectively support data use. Structural model analysis confirmed significant positive influence of usefulness ({beta}=0.199, p<0.001) and access to quality data ({beta}=0.729, p<0.001) on data use, which strongly impacted service delivery ({beta}=0.593, p<0.001), despite some factors showing no direct influence. ConclusionThe study finds that current digital health initiatives only modestly improve the user-friendliness, reliability, and usefulness of data systems, partly due to fragmented, non-interoperable platforms that burden data management. However, compatibility, usability, reliability, and usefulness of digital tools significantly enhance access to quality data and data-driven decisions. The study recommends strengthening and integrating existing systems and providing continuous digital health training to institutionalize data-informed decision-making.

Background To anticipate the impact of new pneumococcal vaccines and guide future updates, accurate forecasts of changes in non-vaccine serotypes (NVTs) are needed. We developed and evaluated three models that incorporated different assumptions about the way in which NVTs will increase and generated ensemble predictions for the frequency of NVTs in different post- pneumococcal conjugate vaccines (PCV) periods. Methods We analyzed age- and serotype-specific invasive pneumococcal disease (IPD) cases from the United States CDCs Active Bacterial Core surveillance during the pre-PCV (1998-1999), early post-PCV7 (2000-2004), late post-PCV7/pre-PCV13 (2005-2009), early post-PCV13 (2010-2014), and late post-PCV13 (2015-2019) periods. These data were augmented with IPD cases from several countries and combined with serotype-specific invasiveness to infer serotype-specific carriage prevalence. Three models (Ranking, Proportionate, NFDS-lite) generated independent predictions of post-PCV IPD frequencies, which were integrated using an accuracy-weighted ensemble. Model performance was evaluated using the normalized root mean square error (NRMSE). Results A total of 23,959 non-PCV7 and 15,580 non-PCV13 cases were analyzed. NVT cases increased from the pre-PCV7 to the late post-PCV7 and post-PCV13 periods. The accuracy of predictions across age groups and models was consistent and high during the post-PCV13 periods but varied during the post-PCV7 periods. The Proportionate model (NRMSE=0.70-3.95) outperformed the NFDS-lite (NRMSE=0.93-8.91) and Ranking (NRMSE=1.51-5.37) models during the early-post-PCV7 period, whereas the NFDS-lite model (NRMSE=1.55-9.82) was superior to the Proportionate (NRMSE=1.45-10.22) and Ranking (NRMSE=1.86-11.35) models during the late post-PCV7 period. The Ensemble model improved on these individual models. Conclusions The Ensemble model offers a tool for forecasting serotype patterns to inform pneumococcal vaccines impact and future pneumococcal vaccine formulation.

Background While a mobile lung cancer screening (mLCS) program can mitigate barriers to access, this study conducted a survey study to assess barriers and facilitators to mLCS which could inform the implementation of new mLCS programs or inform modifications to existing programs. Methods Patient eligibility included current age of 50 to 80 and had undergone any cancer screening at Moffitt Cancer Center (MCC) between January 1, 2023 and December 1, 2024. A web-based survey was administered from May 2025 to June 2025 which collected data on health behaviors, barriers, facilitators, screening preferences, and demographics. Descriptive statistics were used to quantify survey responses. Results Among participants who completed the survey, 73.4% reported no concerns about getting screened in a mobile screening unit, 67.9% reported concerned about the cost or if insurance covered mobile lung cancer screening, and 82.4% reported they would be screened if a voucher or insurance would pay for it. For preferences, 54.1% reported no preference for the time of year for a mobile screening event, 59.6% reported they will be willing to wait up to 30 minutes to get screened, and 44% would travel more than 20 minutes to get screened. There were no statistically significant differences in barriers and facilitators when the analyses were stratified by LCS eligibility. Conclusions We found acceptability of mobile lung cancer screening and preferences that are actionable including daytime weekday events, indoor waiting, short waits, proximity to home, clear cost coverage, and streamlined clinician recommendation.

Importance: Assessment of cardiovascular health (CVH) during may unmask latent metabolic vulnerability and indicate long-term disease risk. However, the prognostic value of the AHA's Life's Essential 8 (LE8) framework during pregnancy remains uncertain. Objective: To evaluate CVH during using a modified Life's Essential 8 (mLE8) score in association with time to incident cardiometabolic disease. Design: Prospective cohort study with electronic medical record (EMR) surveillance for 7 years postpartum (August 2018-March 2026). Adjusted accelerated time-to-failure models estimated mLE8 associations with incident conditions. Setting: A population-based prenatal cohort recruited from a large academic medical system in South Carolina. Participants: Singleton pregnancies in individuals aged 18 to 44 years without pre-existing diabetes or cardiovascular disease (CVD) Exposures: A 7-component mLE8 score assessed during pregnancy, incorporating hypertensive disorders of pregnancy (HDP), 50-g glucose tolerance test results, pre-pregnancy body mass index, smoking status, sleep adequacy, diet quality, and physical activity. Scores ranged from 0 to 100, with higher scores indicating more favorable CVH. Main Outcomes and Measures: Post-delivery incident cardiometabolic conditions captured through EMRs and classified as chronic hypertensive conditions, chronic metabolic conditions (e.g., dyslipidemia, impaired glucose regulation), and CVD (e.g. cardiac arrest, cardiomyopathy). Time to incident diagnosis was measured in days from delivery. Results: Among 1,225 pregnancies (mean age, 25.0 [5.3] years), 499 incident cardiometabolic events occurred over a median follow-up of 6.2 (2.8) years. Each 10-point higher mLE8 score was associated with a longer time to incident diagnosis of chronic hypertensive conditions (time ratio [TR], 1.26; 95% CI, 1.11, 1.42) and chronic metabolic conditions (TR, 1.20; 95% CI, 1.11, 1.29). Healthier HDP, glucose, BMI, and sleep scores were most strongly associated with longer time to diagnosis of chronic metabolic disease. Results were robust to sensitivity analyses excluding individuals who developed gestational diabetes or HDP. Conclusions and Relevance: In this racially diverse, low-income cohort study of 1,225 pregnancies, better CVH during pregnancy was associated with a longer time to incident post-delivery diagnosis of cardiometabolic conditions. Pregnancy-based CVH assessment may help identify individuals with elevated and emerging cardiometabolic risk who could benefit from early, targeted intervention and enhanced longitudinal surveillance.

Introduction: Monitoring trends in transitions in the use of electronic nicotine delivery systems (ENDS) and cigarettes among youth is important for understanding the potential public health impacts of these products. Methods: Using a weighted Markov multistate transition model accounting for complex survey design, we estimated transition rates and one-year transition probabilities between never, non-current, ENDS-only, and cigarette use (with or without dual use of ENDS) among 26,744 youth aged 12-17 years who participated in at least two consecutive waves from Waves 2-7.5 (approximately 2015-2023) of the nationally representative Population Assessment of Tobacco and Health (PATH) Study. We also estimated transitions stratified by ages 12-14 and 15-17 years. Results. The one-year probability of ENDS-only initiation from never use among youth peaked in 2017-19 (Waves 4-5) at 4.0% (95%CI: 3.6-4.3%) and was higher for 15-17-year-olds at 5.8% (95%CI: 5.2-6.4%) than 12-14-year-olds at 2.2% (95%CI: 1.8-2.6%). In the following years, ENDS-only initiation rates declined and plateaued, with 2.6% (95%CI: 2.3-3.0%) initiation in 2022-23. Cigarette initiation from never use decreased over 2015-23 from 0.8% (95%CI: 0.6-1.0%) in 2015-16 to 0.1% (95%CI: 0.0-0.2%) in 2022-23. There was an increase in the fraction of youth who transitioned from non-current product use to ENDS-only use from 13.7% (95%CI: 7.5-20.0%) in 2015-16 to 35.1% (95%CI: 25.4-44.8%) in 2022-23, paired with a decrease in non-current use to cigarette use from 20.9% (95%CI: 11.8-30.0%) to 6.3% (95%CI: 1.7-10.8%). Transitions from ENDS-only or cigarette use to non-current use remained relatively constant over time at around 25% and 15% per year, respectively. Conclusion. ENDS-only use initiation has changed over time, peaking around 2019 and subsequently decreasing and plateauing, but cessation rates for both ENDS and cigarettes have remained relatively stable. Thus, interruption of tobacco product initiation may be the most effective approach to reducing tobacco product use among youth.

Long COVID affects millions of people worldwide, yet no disease-modifying treatment has been approved, and existing interventions have shown only modest and inconsistent benefits. A key reason for this limited progress is that current computational drug repurposing pipelines do not match well with the clinical reality of Long COVID. These patients often have persistent, multisystemic symptoms and may already be taking multiple medications, making treatment safety a primary concern. However, most repurposing workflows still treat safety as a downstream filter and rely on disease-associated targets rather than causal drivers. They also assume that the findings of one analysis would generalize across the diverse presentations of Long COVID. We introduce SPLIT, a safety-first repurposing framework that addresses these limitations. SPLIT prioritizes safety at the start of the candidate evaluation, integrates complementary causal inference strategies to identify likely driver genes, and uses a counterfactual substitution design to compare drugs within specific cohort contexts. When applied to cognitive and respiratory Long COVID cohorts, SPLIT revealed three main findings. First, drugs with similar predicted efficacy could have very different predicted safety profiles. Second, the drugs flagged as unfavorable were often different between the two cohorts, showing that drug prioritization is phenotype-specific. Third, SPLIT flagged 18 drugs currently under active investigation in Long COVID trials as having unfavorable predicted profiles. SPLIT provides a practical framework to identify safer, more context-appropriate candidates earlier in the process, supporting more targeted and better-tolerated treatment strategies for Long COVID.

IntroductionIndividuals with COPD can be classified according to their levels of physical activity (PA) and physical capacity (PC). The relationship between nutrition and body composition within these classifications remains unclear. ObjectivesTo compare the body composition and food intake of people with COPD and verify the associations. MethodsCross-sectional exploratory analysis study in which body composition and food intake were assessed in individuals with COPD. Classification was based on six-minute walk test (PC) and accelerometry(PA): Quadrant "can do, dont do" (I-preserved PC, low PA); quadrant "can do, do do" (II-preserved PC, preserved PA). Results72 individuals with COPD, 39 in quadrant I and 33 in quadrant II, with mean ages of (69 {+/-} 6) (67 {+/-} 7), respectively. Group I had a higher proportion of males, whereas group II had a higher proportion of females. A positive trend in skeletal muscle mass (p=0.011) (B= 2.883) and a negative trend in basal metabolic rate (p=0.010) (B=-0.092) for group I. ConclusionBrazilians with COPD classified in quadrants I and II showed similar results in terms of body composition and food intake. A positive trend in skeletal muscle mass was observed for the group I. These findings align with the pathophysiological model of COPD, in which the preservation of muscle mass and adequate protein intake support functional capacity and the maintenance of higher physical activity levels.

Objective: To investigate the effects of breaking up prolonged sedentary behavior (SB) on daily movement behavior and energy balance in adults with overweight/obesity. Methods: Thirty participants (16F/14M; 34.2+-7.3y; 29.5+-3.2kg/m2) were randomized to either BREAK (nine hourly 5-min brisk walking bouts) or a duration-matched intervention, ONE (45-min brisk walking), both performed 5 days/week for 6 weeks. Pre- and post-intervention, daily SB and physical activity (PA; accelerometry), body composition (doubly labeled water [DLW]), total daily energy expenditure (TDEE; DLW), appetite, and fasting leptin were measured. Linear-mixed effects models tested time effects and time-by-group interactions. Results: Only BREAK reduced prolonged SB (-8%; interaction: p=0.043). Both groups shifted SB-PA composition toward greater moderate-to-vigorous PA with proportional reductions in SB and light PA (time: all p<0.012), which were associated with increases in TDEE (+0.67 MJ/d; time: p=0.040). Body and fat mass increased in ONE only (interaction: p=0.061 and p=0.055). No differences were noted in energy intake, appetite, or leptin levels. Conclusions: Spreading short PA bouts throughout the day increases MVPA and TDEE to the same extent as a traditional continuous PA bout. Future studies should investigate whether minor differences in body composition are driven by distinct behavioral/physiological compensations influenced by the daily pattern of PA/SB.