JAMA Network Open

Objective: To estimate the adjusted relative risk (RR) of administrative grant disruption faced by first-time and mechanism-first principal investigators (PIs) during the 2025 National Institutes of Health (NIH) grant disruptions. Design: Retrospective cohort study linking NIH RePORTER data to a publicly curated registry of grants disrupted in 2025. Setting: All NIH active research grants in fiscal years 2024 to 2025. Participants: 80,976 active projects: 4,961 disrupted during the wave that peaked in May 2025, 76,015 non-disrupted controls. Main outcome measures: Adjusted RR of disruption by two pre-specified first-time PI constructs: absolute first-time PI (no prior NIH grant) and mechanism-first PI (no prior NIH grant with the same activity code). Modified Poisson regression with institution-clustered standard errors adjusted for project, institutional, and geographic covariates. A pre-specified fiscal year 2024 common-anchor analysis addressed year-of-disruption confounding. Results: Of 4,961 disrupted grants, 237 (4.8%) had an absolute first-time PI and 396 (8.0%) had a mechanism-first PI. After adjustment, absolute first-time PIs faced 77% elevated risk of disruption (RR 1.77, 95% CI 1.34 to 2.32) and mechanism-first PIs faced 57% elevated risk (RR 1.57, 1.16 to 2.11). Under the common-anchor analysis, the absolute first-time effect attenuated to RR 1.22 (0.95 to 1.58); the mechanism-first effect persisted (RR 1.48, 1.07 to 2.06). The elevated risk was concentrated in research-mechanism grants (RR 1.78, 1.26 to 2.52) and was robust across 8 of 9 pre-specified sensitivity analyses. The Track A start-time construct, which asks whether the disrupted project was the PI's debut grant, yielded null estimates (RR 0.98, 0.93 to 1.04), with any effect concentrated entirely in newly started projects. Conclusions: First-time and mechanism-first PIs faced disproportionately elevated risk of disruption during the 2025 NIH wave, concentrated in research-mechanism grants and robust to year-confounding-free identification. The relevant exposure was being early-career at the moment of administrative action, not at project initiation. The findings have direct implications for workforce equity in US biomedical research.

ObjectivesTo examine the extent to which heat-related causes of death are recorded in fatal drug overdoses, how these patterns vary across states and over time, and how overdose characteristics differ between deaths with, versus without, heat involvement recorded. MethodsDeath certificate data for all drug overdose deaths in US residents from 2001 to 2024 (from the National Center for Health Statistics) were analyzed to identify whether a heat-related cause of death was also listed on the death certificate. Joinpoint regression, descriptive statistics, and nonparametric tests were used to examine temporal trends and compare overdose deaths with versus without recorded heat involvement. ResultsIn 2001, fewer than 10 drug overdose deaths with recorded heat involvement were identified, but this number increased to 558 in 2024. From 2013 to 2024, mortality rates increased significantly, with an estimated annual percent change of 30.1 (95% Confidence Interval, 26.5-47.1). The highest mortality rates and numbers of deaths were observed in residents of Arizona and Nevada. American Indian/Alaska Native, Mexican-heritage, and foreign-born populations accounted for larger shares of overdose deaths with, compared to without, heat involvement recorded. A street or highway was more frequently identified as the place of injury in overdose deaths with (18.9%), versus without (2.2%) heat involvement reported. Psychostimulants such as methamphetamine were involved in 85.9% of overdose deaths with, compared to 28.9% without, recorded heat involvement. ConclusionsAlthough representing only a fraction of all overdose deaths, fatal overdoses involving heat exposure have increased markedly over time and disproportionately impact certain states and demographic groups.

IntroductionAlthough prior studies suggest e-cigarette use is associated with worse mental health, it remains unclear whether these associations persist independent of diagnosed depression and how tobacco use and depression jointly affect health-related quality of life. Although the long-term health risks of vaping are still unknown, self-reported health is a reliable measure of quality of life. This study provides the first health utility estimates of the independent and combined effects of cigarette use, e-cigarette use, and depression on health-related quality of life. MethodsWe analyzed 2022-2023 Behavioral Risk Factor Surveillance System data on health-related quality of life as measured by self-reported physically or mentally unhealthy days in the past 30 days. The average number of unhealthy days was estimated by age, gender, smoking status (current versus non-smoking), depression status (received a prior diagnosis), and current e-cigarette use status (every day or some day use). We converted the number of overall healthy days into EQ-5D utility scores by age-specific percentile matching of BRFSS and MEPS distributions, a method developed by Jia and Lubetkin. ResultsCigarette use, e-cigarette use, and depression were each associated with worse health-related quality of life. Mentally unhealthy days increased with the accumulation of these conditions. Associations with physically unhealthy days followed a similar pattern, particularly among younger adults, although the magnitude of association was smaller. E-cigarette use alone was associated with 2.0-4.2 (95% CI: 2.0-4.6) additional mentally unhealthy days per month across all age groups. Notably, e-cigarette use was independently associated with poorer mental health among adults aged 18-64 with or without diagnosed depression. After accounting for smoking and depression status, e-cigarette use was associated with disutility scores of 0.011 in men and 0.015 in women among young adults, with the largest losses observed when multiple conditions co-occurred. ConclusionE-cigarette use is associated with poorer health-related quality of life, particularly among younger adults, and these effects are amplified when combined with cigarette use and depression. Quantifying these joint impacts as health utility losses highlights the importance of addressing e-cigarette use within integrated tobacco control and mental health policies, especially for young populations.

BackgroundEarly initiation of substance use is linked to later adverse outcomes, and risk factors come from multiple domains and are shared across substances. In our previous work, traditional time-to-event Cox models identified individual risk factors, but these models are not designed to jointly model multiple outcomes or capture complex non-linear relationships. Multi-task learning (MTL) can leverage shared structure across related outcomes to improve prediction and distinguish common versus substance-specific predictors. However, most MTL studies rely on baseline features and focus on single outcomes, which limits their ability to capture shared risk and temporal changes. Substance use initiation is a time-dependent process that unfolds during development and reflects changing exposures over time. Baseline-only models cannot capture these changes or represent risk dynamics. Discrete-time modeling provides a practical approach by estimating interval-level initiation risk and combining it into cumulative risk at the subject level. By integrating multi-task learning with dynamic modeling, it is possible to share information across outcomes while capturing how risk evolves over time, which may improve prediction performance. MethodsUsing the Adolescent Brain Cognitive Development (ABCD) Study(R) (release 5.1), we developed two complementary multi-task learning (MTL) frameworks to predict initiation of alcohol, nicotine, cannabis, and any substance use. A baseline MTL model predicted fixedhorizon (48-month) initiation using one record per participant, while a dynamic discrete-time MTL model incorporated longitudinal interval data to model time-varying risk. Both models used multi-domain environmental exposures, core covariates, and polygenic risk scores (PRS). Performance was evaluated on a held-out test set using AUROC, PR-AUC, and calibration metrics, and compared with single-task logistic regression (LR). Feature importance was assessed using permutation importance and compared with Cox proportional hazards models. ResultsMTL showed comparable or improved performance relative to LR, with larger gains for low-prevalence outcomes (cannabis and nicotine). Incorporating longitudinal information led to consistent improvements across all outcomes. Dynamic models increased AUROC by +0.044 to +0.062 for MTL and +0.050 to +0.084 for LR, indicating that temporal information was the primary driver of performance gains. Feature importance analyses showed modest overlap across methods, with higher agreement between dynamic MTL and Cox models than static MTL. A small set of features, including externalizing behavior, parental monitoring, and developmental factors, were consistently identified across all approaches. ConclusionsDynamic multi-task learning improves the prediction of substance use initiation by leveraging longitudinal structure and shared information across outcomes. While MTL provides additional gains, incorporating time-varying information is the dominant factor for improving performance. Combining baseline and dynamic frameworks offers a comprehensive strategy for identifying robust risk factors and modeling adolescent substance use initiation.

Introduction: Even though state minimum wage (MW) is a policy lever that affects income and poverty and can prevent of violence, no prior study has comprehensively evaluated its impact in the United States (US). In this study, we estimated the impact of at least a $1 USD increase in state MW above the federal MW on overall, firearm, and non-firearm homicide mortality and examined its impact on racialized inequities. Methods: We conducted a quasi-experimental study using controlled interrupted time series (CITS) and synthetic controlled interrupted time series (SCITS) approaches to examine immediate and sustained impact of state MW increases. We used state-month level homicide victimization mortality data from 2010-2019. Homicide victimization death was identified using International Classification of Disease codes, 10th revision. State MW data was obtained from the Bureau of Labor Statistics. Results: Demographic and social variables from intervention, never-exposed, and always-exposed states were similar to each other and representative of the total US population from all 50 states. The CITS results show that after MW increases in the intervention states, these states experienced a sustained decline of -0.22 (-0.37, -0.07) homicide victimizations/ 100,000 person-years/ year relative to the never-exposed states and -0.39 (-0.59, -0.18) relative to always-exposed states. This resulted in 5,657 fewer homicide victimization deaths in the intervention states over four years of post-MW increase period compared to the never-exposed states. SCITS results were similar to the CITS results, and the majority of sustained declines were observed in firearm-related deaths and among Black population. Conclusion: MW increase was associated with a reduction in homicide victimization rates, which were robust in multiple sensitivity analyses, more pronounced for firearm-related homicide deaths, and reduced homicide victimization inequities for Black Americans.

ImportanceDrug-related pregnancy-associated mortality is a leading contributor to the US maternal mortality crisis, yet whether it follows the persistent rural disadvantage documented for all-cause maternal mortality--or is restructured by the geographic dynamics of drug markets--has not been established. ObjectiveTo characterize geographic variation in pregnancy-associated overdose (OD) and substance use disorder (SUD) mortality across the rural-urban continuum and by US Census region from 2016 through 2022. Design, Setting, and ParticipantsNational population-based surveillance study using individual-level National Vital Statistics System (NVSS) mortality and natality records. Pregnancy-associated deaths (occurring during pregnancy or within one year of the end of pregnancy) were ascertained among 25,007,723 live births during 2016-2022 using the NVSS 2018 algorithm. ExposuresRural-urban classification cross-classified by four US Census regions. Main Outcomes and MeasuresRates of pregnancy-associated OD mortality and SUD mortality per 100,000 live births. Post-COVID excess OD mortality was estimated using a Bayesian hierarchical Poisson model. ResultsThere were 516 OD deaths (2.06 per 100,000 live births) and 1,080 SUD deaths (4.32 per 100,000) nationally; SUD exceeded OD mortality more than two-to-one in all strata, and both outcomes were concentrated in the late postpartum period (43 days to 1 year). OD mortality converged across the rural-urban gradient during the COVID era (2020-2022)--the inverse of the persistent rural disadvantage in all-cause maternal mortality--with metropolitan areas falling below pre-pandemic trajectory expectations while non-metropolitan areas exceeded theirs. Credible excess OD mortality was identified in non-metropolitan Southern and Northeastern counties. SUD rates were non-monotonic across urbanicity, with metro-adjacent counties carrying elevated rates in all regions. Conclusions and RelevanceDrug-related pregnancy-associated mortality follows a distinct geographic logic from all-cause maternal mortality, shaped by drug supply dynamics and harm reduction geography rather than obstetric care infrastructure alone. The convergence of OD mortality across the rural-urban gradient, the dominance of SUD over acute overdose, and the concentration of deaths in the late postpartum year point to care and surveillance gaps requiring integrated obstetric and addiction treatment, extended postpartum insurance coverage, and rural harm reduction capacity.

Importance: Parent/caregiver-infant early relational health (ERH) is known to play a critical role in the promotion of socio-emotional functioning and wellbeing across the life course. The negative impact of the COVID-19 pandemic on maternal mental health and secondarily on ERH and child socio-emotional functioning is clear. However, the direct impact of maternal viral exposure during pregnancy on ERH has not been investigated. Objective: The goal of this study was to determine the impact of prenatal SARS-CoV-2 exposure on ERH and infant socio-emotional functioning in the first 6 months of life. Design: Mothers with and without SARS-CoV-2 exposure during pregnancy who gave birth from 02/2020 to 09/2021 were enrolled from 05/2020 to 09/2021 in one of two parallel prospective studies (the COVID-19 Mother Baby Outcomes [COMBO] Initiative or the Respiratory Syndrome Coronavirus 2 in Pregnancy and Infancy [ESPI] COMBO sub-study). Mothers reported on their health and the socio-emotional functioning of their infant via online surveys (REDCap) at enrollment, 1, 2, 4, and 6 months. At 4 to 6 months, dyads were invited to participate in a video-based, remote assessment of ERH. Participants: 884 mother-infant dyads from three U.S. States (Alabama, New York, and Utah). Exposure: Prenatal SARS-CoV-2. Main Outcomes and Measures: Maternal-reported ERH (parental stress, parenting confidence and bonding) and observer-based ERH (video-coded quality of maternal caregiving behaviors and mother-infant emotional connection). Infant socio-emotional development assessed using the 6-month Ages and Stages Questionnaire: Socio-Emotional 2nd Edition (ASQ:SE-2). Results: 316 (36%) mothers had a positive prenatal SARS-CoV-2 exposure. Prenatal SARS-CoV-2 exposure was associated with an adjusted estimate of ~5% reduction (incidence rate ratio=0.95, 95% confidence interval [0.90, 1.00], p=0.03) in observed maternal caregiving quality, after accounting for postnatal maternal mental health and sociodemographic factors. We found no evidence of effect on other ERH constructs or infant socio-emotional functioning. Conclusions and Relevance: In this large prospective cohort study, prenatal SARS-CoV-2 was associated with a small decrement in caregiving quality, but not other ERH constructs or infant socio-emotional functioning. These findings should be interpreted as hypothesis generating and will require replication in independent studies.

Background: Store-and-forward teledermatology commonly relies on several patient-submitted photographs of the same concern, but most dermatology artificial intelligence models classify single images independently. Objective: To develop and internally validate a case-level diagnostic-support model that aggregates multiple patient-submitted photographs for common dermatologic conditions. Methods: We conducted a retrospective diagnostic-modeling study using the Skin Condition Image Network, a public dataset of deidentified self-taken dermatology images from US adults. We curated 2,336 cases comprising 5,041 images across 10 common inflammatory, allergic, and infectious conditions. Cases were split at the submission level into training, validation, and held-out test sets. Frozen general-purpose and dermatology-specific encoders were compared with image-level classifiers and a gated-attention multiple instance learning model that generated one case-level output from 1-3 images. Results: The strongest image-level baseline, dermatology-specific embeddings with random forest classification, achieved macro/micro ROC-AUCs of 0.797/0.854. Case-level aggregation improved discrimination, with dermatology-specific embeddings plus multiple instance learning achieving mean macro/micro ROC-AUCs of 0.819/0.863 across repeated stratified experiments. The locked final model achieved macro/micro ROC-AUCs of 0.800/0.849 on the held-out test set. Balanced-threshold sensitivity/specificity examples were 0.702/0.688 for eczema and 0.818/0.826 for urticaria. Limitations: Internal validation used a 10-condition subset from a US volunteer dataset; external validation, calibration, subgroup performance analysis, and prospective workflow studies are required. Conclusion: Modeling the teledermatology submission as a multi-image case better reflects asynchronous dermatology workflow than single-image classification. The model is preliminary clinician-facing support for structured review and triage, not autonomous diagnosis.

Introduction Cesarean delivery accounts for nearly one-third of U.S. births and is associated with substantial maternal morbidity and health care costs. Persistent racial disparities have been documented, yet the structural factors contributing to these differences remain incompletely understood. The extent to which insurance coverage shapes racial disparities in cesarean delivery remains unclear. Objective To evaluate the independent and interactive associations of race/ethnicity and insurance coverage with cesarean delivery in the United States. Methods Population-based retrospective cohort study using singleton live births recorded in the United States Vital Statistics Natality files from 2014 to 2024. Multivariable logistic regression was used to estimate the independent effects of race/ethnicity and insurance status on cesarean delivery, including interaction terms to test effect modification, using national birth certificate data. Models were adjusted for maternal demographics, clinical factors, and temporal covariates. Adjusted odds ratios, predicted probabilities, and absolute risk differences were derived from post-estimation marginal effects. The main outcome measure was cesarean delivery (yes vs no). Results Among 41,543,568 deliveries from 2014 to 2024, 13,312,221 (32.0%) were cesarean deliveries. After adjustment, both race and ethnicity and insurance status were independently associated with cesarean delivery. Compared with non-Hispanic White women, non-Hispanic Black women had higher odds of cesarean delivery (odds ratio [OR], 1.22; 95% CI, 1.22-1.23). Relative to uninsured women, those with private insurance had 59% higher odds of cesarean delivery (OR, 1.59; 95% CI, 1.58-1.60). Significant interaction effects were observed, indicating that insurance coverage modified racial and ethnic differences in cesarean delivery. Non-Hispanic Black women had the highest predicted probabilities across all insurance categories, with the largest absolute disparities observed among uninsured women. Conclusion Racial and ethnic differences in cesarean delivery persist in the United States and are modified by insurance coverage, suggesting that coverage-related differences may contribute to inequities in obstetric care.

Pregnant and postpartum people who use drugs in the United States are trying to survive at the intersection of two of the gravest public health crises of the 21st century US: epidemics of (1) maternal mortality and (2) the overdose epidemic. Although extensive evidence documents racial/ethnic disparities in each of these epidemics separately, comparatively little research has characterized racial/ethnic patterns in their collision, that is, in maternal overdose mortality. We analyzed individual-level mortality records from the National Vital Statistics System (NVSS) for 2016-2022 to describe racial/ethnic disparities in pregnancy-associated overdose deaths (PA-OD) and pregnancy-associated substance use disorder-related deaths (PA-SUD). Racial/ethnic-specific mortality rates were calculated per 100,000 live births with exact Poisson confidence intervals. Temporal trends were summarized using annual percent change (APC), and disparities were quantified using rate ratios and differences relative to non-Hispanic White individuals. Overdose-related maternal mortality increased substantially during the study period across multiple racial and ethnic groups. Rates increased nearly threefold among non-Hispanic White individuals and rose more steeply among non-Hispanic Black individuals, producing a Black-White disparity that emerged over time. Rates among Hispanic individuals remained lower but increased rapidly, while estimates among American Indian and Alaska Native individuals were often high but unstable because of small counts. Substance use disorder-related maternal mortality exhibited a pronounced surge during 2019-2021 across several racial and ethnic groups. These findings highlight rapidly evolving racial/ethnic patterns in maternal overdose mortality and underscore the need for targeted prevention and harm-reduction strategies to reduce overdose-related deaths during pregnancy and the postpartum period. FundingWe are grateful to the following NIH grants for supporting this research: U54HD113292 and R01DA059182.

Introduction: COVID-19 transmission has not been evenly distributed across racial groups, with exposure being shaped by social and structural factors. The emergence of highly transmissible variants (i.e., Omicron) dramatically increased infection rates. However, it remains unclear whether racial disparities in transmission disappeared or persisted over the course of the pandemic. Objective: To understand how SARS-CoV-2 transmission differed by race in Canada and whether those disparities changed with the Omicron variant. Methods: We analyzed cross-sectional SARS-CoV-2 seroprevalence data from the Canadian Blood Services serosurveillance program (June 2020 to April 2023) using a previously described dynamic susceptible-infection model, while accounting for seroreversion. Race-specific force of infection was estimated for the pre-Omicron and Omicron periods (with the emergence of Omicron defined as beginning December 26, 2021). Results: Prior to Omicron, racialized individuals had a 74% higher force of infection (IRR = 2.205; 95% CI: 2.115-2.299). During the Omicron period, infection rates rose significantly within each racial group relative to the pre-Omicron period, with a 55.52-fold increase among White individuals and a 31.27-fold increase among racialized individuals. Despite this, racialized individuals remained disproportionately affected following the emergence of Omicron, with 24% higher infection rates than those of their White counterparts (IRR = 1.242; 95% CI: 1.231-1.253). Conclusion: Widespread transmission during Omicron did not result in epidemiologic equity, as racialized populations continued to experience higher infection risk despite crude seroprevalence depicting convergence.

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 [≥] 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.

BackgroundImplementing proven vehicle safety standards recommended by the UN World Forum for Harmonization of Vehicle Regulations is among the most cost-effective strategies to reduce road traffic deaths. In 2022, Mexico approved updated vehicle safety standards, including side pole testing, electronic stability control, seatbelts, airbags, side structures, and anchorage child restraint systems. However, pedestrian protection and advanced driver-assistance technologies, such as autonomous emergency braking systems (AEBS), were excluded. These exclusions are critical, given that more than half of road traffic deaths involve vulnerable road users. Local evidence on the expected benefits of implementing comprehensive vehicle safety standards is needed to guide policy decision-making. ObjectiveTo estimate the potential public health impact of increasing the availability of recommended vehicle safety technologies in Mexico. MethodsWe conducted a comparative risk assessment analysis to estimate the impact of improving vehicle safety standards on road traffic deaths, injuries, and disability-adjusted life years. Counterfactual analyses were defined using traffic statistics for 2019 as baseline, relative risk estimates associated with each safety technology, and technology penetration within Mexicos vehicle fleet. Three scenarios were modeled: (1) full implementation of Mexicos 2022 standards; (2) addition of crashworthiness, AEBS, and motorcycle ABS/ESC; and (3) inclusion of expanded AEBS crash configurations, lane departure warning (LDW), and lane keeping assistance (LKA) systems. ResultsScenario 1 reduced deaths by 18%, injuries by 16%, and DALYs by 18%, with the greatest benefits for car occupants. Scenario 2 reduced deaths by 29%, injuries by 27%, and DALYs by 30%, benefiting motorcyclists and pedestrians the most. Scenario 3 reduced deaths, injuries, and DALYs by 41%, 38%, and 41%, respectively, benefiting car occupants and motorcyclists. ConclusionsCurrent vehicle safety standards in Mexico are expected to reduce deaths, injuries, and disabilities, yet existing guidelines focus largely on protecting car occupants. Mexico should strive to update and strengthen its current legislation by adding technologies that protect vulnerable road users, such as pedestrians and cyclists, and to focus on technologies for motorcycle users to further reduce the burden of road traffic injuries.

BackgroundArtificial intelligence and machine learning (AI/ML) are among the fastest-growing domains in NIH research funding, but whether children have shared equitably in this expansion is unknown. We characterized pediatric representation in NIH AI/ML funding from fiscal years (FY) 2020 to 2024. MethodsNIH grant data were obtained from Research Portfolio Online Reporting Tools Expenditures and Results bulk files for FY2020 to FY2024. AI/ML grants were identified using the NIH Research, Condition, and Disease Categorization "Machine Learning and Artificial Intelligence" category, and pediatric grants using the "Pediatric" category. Subprojects were excluded. Grants were deduplicated within each fiscal year by core project number for trend analyses and across all years retaining the most recent fiscal year for cross-sectional totals. Disease areas were identified by keyword searches of titles and abstracts. ResultsAcross FY2020 to FY2024, 5,624 unique NIH AI/ML grants totaling $3,371 million were identified. Of these, 836 grants (14.9%) were classified as pediatric, representing $401 million (11.9%) of total NIH AI/ML funding. Although this share was consistent with the historically reported overall NIH pediatric funding baseline of approximately 10% to 12%, it remained substantially below the US pediatric population share of approximately 22%. The pediatric share of NIH AI/ML funding declined from 12.3% in FY2020 to 10.8% in FY2024, despite growth in absolute pediatric funding. Indexed to FY2020, pediatric AI/ML funding grew approximately 2.6-fold compared with 3.0-fold growth in the total portfolio. Across disease areas, unadjusted adult/general-to-pediatric funding ratios ranged from 2.0-fold in mental health to 9.8-fold in cancer. ConclusionsPediatric representation in NIH AI/ML funding remained low and declined over time as the overall portfolio expanded. These findings suggest that growth in NIH AI/ML investment has not been matched by proportional gains for pediatric research.

Importance: Problematic use of the internet (PUI) behaviors, including problematic gaming, social media use, smartphone use, and general internet use, have been increasingly studied worldwide. So far, it is unclear what the global prevalence of PUI is. Objective: To critically appraise existing systematic reviews and meta-analyses on the prevalence of PUI behaviors and generate aggregated global prevalence estimates across different manifestations and definitions. Data Sources: MEDLINE (Ovid), Embase (Ovid), Scopus, Web of Science, CINAHL, and the Cochrane Review Library were searched for relevant articles from database inception to the most recent available search prior to manuscript preparation. Searches targeted systematic reviews and meta-analyses reporting prevalence for PUI-related behaviors. Study Selection: Systematic reviews and meta-analyses of observational studies reporting prevalence estimates for problematic gaming, problematic internet use, problematic smartphone use, problematic social media use, or sexting were included. Scoping reviews were retained for descriptive synthesis only. Data Extraction and Synthesis: An umbrella review methodology was used. Data extraction and methodological appraisal were conducted using AMSTAR-2 to assess the quality of included systematic reviews up to February 2026. Primary studies included in each review were extracted and pooled using random-effects meta-analysis. Analyses were conducted to estimate pooled prevalence with 95% confidence intervals (CIs) and heterogeneity across non-overlapping primary studies. Small-study effects were examined. Main Outcomes and Measures: Global pooled prevalence estimates for PUI behaviors, including problematic gaming, problematic internet use, problematic smartphone use, problematic social media use, and sexting. Results: Eleven reviews, including 10 systematic reviews and 1 scoping review, met inclusion criteria, representing data from 3,145,428 individuals, of whom 3,030,023 were included in pooled prevalence analyses. Across regions, pooled prevalence estimates were 6% (95% CI, 5%-7%) for problematic gaming, 16% (95% CI, 15%-17%) for problematic internet use, 32% (95% CI, 28%-35%) for problematic smartphone use, and 23% (95% CI, 19%-28%) for problematic social media use. Substantial heterogeneity (I2 > 99%) was observed across primary studies, reflecting variation in study methodologies, sampled populations, and definitions of PUI behaviors. Conclusions and Relevance: PUI behaviors appear to affect a substantial proportion of the global population. However, methodological concerns were common, with 9 of 10 systematic reviews rated as having low or critically low confidence according to AMSTAR-2. Evidence remains concentrated in East Asia and Europe, and many reviews combine heterogeneous populations and sampling strategies. Additional high-quality epidemiological research, including studies in underrepresented regions, is needed to refine prevalence estimates, clarify risk factors, and support the development of standardized criteria for PUI behaviors.

Background Six years after its emergence, SARS-CoV-2 continues to have a substantial burden. The impact of vaccination and the optimal timing of its rollout remain uncertain given existing population immunity and variability in outbreak timing between summer and winter. Methods The US Scenario Modeling Hub convened its 19th round of ensemble projections for COVID-19 hospitalizations and deaths in the United States, where eight teams projected trajectories in each US state and nationally from April 2025 to April 2026 under five scenarios regarding vaccine recommendations and timing. Recommendations had two eligibility scenarios (high-risk individuals only and all-eligible) and two timing scenarios (classic start: mid-August, earlier start: late June). These were crossed to create four scenarios and were compared against a counterfactual scenario with no vaccination. Findings Compared to no vaccination, our ensemble projections estimated 90,000 (95% PI 53,000-126,000) hospitalizations averted in the high-risk and classic timing scenario across the US. Expanding to all-eligible age-groups averted an additional 26,000 (95% PI 14,000-39,000) hospitalizations, which when coupled with the early vaccination timing, was projected to further reduce national hospitalizations by 15,000 (95% PI -3,000-33,000). The majority of teams projected both summer and winter waves. Implications We project COVID-19 will cause significant hospitalizations and deaths in the US in the 2025-26 season and estimate significant benefits from a broad all-eligible vaccination recommendation. The results also suggest an additional benefit is likely to be gained from an earlier vaccination campaign. Funding Centers for Disease Control and Prevention; National Institute of Health (US), National Science Foundation (US)

BackgroundHospital visiting-hour policies vary widely across the United States, yet the structural factors shaping this variation remain poorly characterized. ObjectiveThis study investigates how hospital-level financial characteristics, payer mix, and rurality relate to the restrictiveness of inpatient visiting-hour policies, and assesses whether these relationships differ across states with diverse Medicaid expansion statuses. DesignCross-sectional observational analysis of hospital visitor policies in four states (Massachusetts, Wisconsin, Tennessee, and South Carolina) selected based on Medicaid expansion status, population size, and hospital density. ParticipantsA total of 318 acute-care hospitals were included using publicly available data from the Centers for Medicare & Medicaid Services and the National Academy for State Health Policy. Main MeasuresThe primary outcome was total daily visiting hours in general inpatient wards. Predictors included volume/capacity, patient mix, financial performance/efficiency, geography and organizational structure. Key ResultsHospital-level characteristics including higher Medicaid payer mix, stronger financial margins, greater inpatient occupancy, and larger size were associated with shorter visiting hours in unadjusted analyses. Commercial payer mix and rurality predicted longer hours. Mean visiting duration was 14.1 hours/day (SD = 5.07; range 0-24), with Massachusetts having the shortest on average across states (10.5 hours/day) and Wisconsin the longest (16.3 hours/day). Medicaid payer mix was the only predictor associated with visiting-hour restrictiveness after multiple-testing correction. Each 10-percentage-point increase in Medicaid payer mix was associated with an approximately 11.3% decrease (p = 0.002) in visiting hours. Within-state variation exceeded the differences between-states. ConclusionsVisitation hours vary considerably, with correlations around rurality of the community served, size of the hospital, and the number of patients on Medicaid. Medicaid payer mix emerged as the most consistent predictor of restrictiveness after adjustment. Hospitals can use these findings to evaluate visitation practices to balance patient-centered care with operational demands.

ObjectivesUptake of evidence-based smoking cessation support remains limited. Digital interventions offer the prospect of scalable and highly accessible support. Smoke Free, a digital mobile application using established behaviour change techniques, has shown promise, but no large-scale randomised controlled efficacy trial has yet been conducted. We assessed its effectiveness for smoking cessation. DesignIn this prospective, randomised, controlled, two-arm, parallel clinical trial with 6-month follow-up, study personnel and patients were blinded. SettingThe trial was conducted nationwide in Germany, utilising a decentralised, fully remote trial design. Enrolment took place digitally after receiving brief advice from a healthcare professional, following guidelines for primary care. ParticipantsOut of a volunteer sample of 1850 patients assessed for eligibility, 1466 adult cigarette smokers who had at least moderate cigarette dependence (F17.2, FTCD[&ge;]3) were recruited between August 2023 and February 2024; 84.1% (1233 participants) completed the primary outcome measure. InterventionsThe intervention group (IG) received the Smoke Free app including behaviour-change missions and gamification elements, while the control group (CG) received a text-only cessation information app. Both groups received brief advice from a healthcare professional. Main outcome measuresThe prespecified primary outcome was self-reported 7-day point-prevalence abstinence from combustible tobacco at 6 months post-randomisation; secondary outcomes included biochemical validation of abstinence in participants providing a saliva sample (59% of eligible participants). ResultsSelf-reported abstinence (primary outcome) was significantly higher in the IG compared with the CG (283 [39.3%] vs. 182 [24.4%], OR=2.01, 95% CI 1.60 to 2.50, p<0.0001). The NNT was 6.7 (5.1 to 9.8). The effect was consistent with biochemical validation (OR=1.76, 95% CI 1.27 to 2.44, p<0.0001) and across secondary outcomes and sensitivity analyses. The 6-month follow-up rates for the primary outcome did not differ between groups (IG: 601 [83.5%]; CG: 632 [84.7%]; p=0.52). Eighty-four serious adverse events were reported by 75 participants (IG: 31, 4.3%; CG: 44, 5.9%; p=0.53); none were treatment-related. ConclusionsThe Smoke Free app is effective for aiding smoking cessation in at least moderately dependent cigarette smokers compared with an informational app when provided as an adjunct to brief advice from a healthcare professional. Trial registrationThe trial was registered with the German Clinical Trials Register (DRKS00031140). FundingSmoke Free 23 GmbH (for-profit company).

Importance BRCA genetic testing is critical for cancer risk assessment, treatment and personalization, yet substantial underutilization persists. Socioeconomic and clinical factors may strongly influence testing uptake; therefore, identifying the potential drivers to BRCA testing and treatment is essential for addressing gaps in access, increasing retention into care, and improving cancer outcomes. Objective To quantify the putatively causal effects of SDoH on BRCA genetic testing among individuals with breast, ovarian, pancreatic, and prostate cancers and to develop a predictive model to identify patients at risk for underuse of testing. Design, Setting, and Participants This observational case-control study used data from a large multistate clinical research data network covering southern US (2012-2023). The network contained records of more than 26 million individuals and was linked with ZIP code-level SDoH variables derived from national socioeconomic datasets. Adults diagnosed with breast, ovarian, pancreatic, or prostate cancer were eligible for cases (received BRCA testing) or controls (did not receive BRCA testing, matched by cancer diagnosis). Exposure SDoH categories, including economic conditions, education, healthcare access, neighborhood conditions, and social connectedness. Main Outcomes and Measures The primary outcome was receipt of BRCA genetic testing after cancer diagnosis. Results Among 3,279 people diagnosed with cancer, 748 received BRCA testing and 2,531 served as controls. Study population mean [SD] age was 66.8 [15.7] years; 1,758 were women [53.6%], 2,238 [69.6%] were White and 616 [18.8%] were Black or African American. Breast (1,420 [42.8%]) and prostate (1,342 [40.9%]) cancers were the most common diagnoses, followed by pancreatic (242 [7.4%]), ovarian (238 [7.2%]), and multiple cancers (55 [1.7%]). Upon adjusting for potential confounding, higher educational attainment (odds ratio [OR], 1.19), public-sector employment (OR, 1.42), neighborhood safety (OR, 1.28), and social participation (OR, 1.72) showed an increased likelihood of undergoing BRCA testing, whereas economic instability, including housing cost burden and reliance on public insurance, had an effect of reduced testing. A random forest classifier demonstrated good discriminative performance (AUROC, 0.776) to predict cancer patients who were likely to take BRCA testing, where nativity, language, and residential stability ranked among the most influential social determinants according to SHapley Additive exPlanations (SHAP) analysis. Conclusions and Relevance In this observational case-control study, SDoHs were strongly associated with receipt of BRCA genetic testing among people with cancer. These findings suggest that disparities in genetic testing may reflect structural and social barriers rather than differences in clinical eligibility alone. Efforts to improve equitable access to genetic testing may benefit from integrating social-context information into clinical workflows and targeting outreach or navigation strategies toward socially disadvantaged populations.

Prenatal depression is a substantial contributor to maternal morbidity, and screening is an entry point to psychiatric assessment and treatment during pregnancy. Following updated guidelines and quality metrics for prenatal depression screening, we evaluated whether screening uptake differed by preferred language within a large U.S. healthcare system. We used electronic health record data to identify a retrospective cohort of deliveries at or beyond 20 weeks gestation in 2019-2024. We used logistic regression with a language-year interaction to estimate the adjusted marginal probabilities of screening by language preference. Among 99,526 pregnancies (82,632 individuals), screening increased substantially over time but increases differed across language groups (p<0.001). In 2019, screening probabilities were similar (English 0.50; Spanish 0.48; Another Language 0.50). By 2024, probabilities diverged (English 0.81; Spanish 0.66; Another Language 0.71). Unequal screening uptake can systematically under-identify prenatal depression among patients with non-English language preference, with implications for equitable access to psychiatric care.