Tag: pubmed

JMIR Med Inform. 2025 Sep 29;13:e74116. doi: 10.2196/74116.

ABSTRACT

BACKGROUND: Data scarcity and dispersion pose significant obstacles in biomedical research, particularly when addressing rare diseases. In such scenarios, synthetic data generation (SDG) has emerged as a promising path to mitigate the first issue. Concurrently, federated learning is a machine learning paradigm where multiple nodes collaborate to create a centralized model with knowledge that is distilled from the data in different nodes, but without the need for sharing it. This research explores the combination of SDG and federated learning technologies in the context of acute myeloid leukemia, a rare hematological disorder, evaluating their combined impact and the quality of the generated artificial datasets.

OBJECTIVE: This study aims to evaluate the privacy- and fidelity-related impact of horizontally federating SDG models in different data distribution scenarios and with different numbers of nodes, comparing them with centralized baseline SDG models.

METHODS: Two state-of-the-art generative models, conditional tabular generative adversarial network and FedTabDiff, were trained considering four different scenarios: (1) a nonfederated baseline with all the data available, (2) a federated scenario where the data were evenly distributed among different nodes, (3) a federated scenario where the data were unevenly and randomly distributed (imbalanced data), and (4) a federated scenario with nonindependent and identically distributed data distributions. For each of the federated scenarios, a fixed set of node quantities (3, 5, 7, 10) was considered to assess its impact, and the generated data were evaluated, attending to a fidelity-privacy trade-off.

RESULTS: The computed fidelity metrics exhibited statistically significant deteriorations (P<.001) up to 21% in the conditional tabular generative adversarial network and up to 62% in the FedTabDiff model due to the federation process. When comparing federated experiments trained with diverse numbers of nodes, no strong tendencies were observed, even if specific comparisons resulted in significative differences. Privacy metrics were mainly maintained while obtaining maximum improvements of 55% and maximum deteriorations of 26% between both models, although they were not statistically significant.

CONCLUSIONS: Within the scope of the use case scenario in this paper, the act of horizontally federating SDG algorithms results in a loss of data fidelity compared to the nonfederated baseline while maintaining privacy levels. However, this deterioration does not significantly increase as the number of nodes used to train the models grows, even though significative differences were found in specific comparisons. The different data partition distribution configurations had no significant effect on the metrics, as similar tendencies were found for all scenarios.

PMID:41021276 | DOI:10.2196/74116

JAMA Netw Open. 2025 Sep 2;8(9):e2530132. doi: 10.1001/jamanetworkopen.2025.30132.

ABSTRACT

IMPORTANCE: The tobacco industry spends more than $8 billion annually in the US on marketing at the point of sale. Exposure to tobacco retail has been associated with smoking outcomes, but substantially less is known about how objectively logged everyday tobacco retail exposure is associated with smoking outcomes.

OBJECTIVE: To assess preregistered hypotheses that individuals would report (1) greater craving and (2) more cigarettes smoked on days when their exposure to tobacco retail is higher than usual.

DESIGN, SETTING, AND PARTICIPANTS: This multimodal, within-person cohort study combined objectively logged geolocation tracking, public tobacco retail location records, and ecological momentary assessment data. Eligible participants recruited from the GeoSmoking Study were aged 21 to 65 years, smoked at least 5 cigarettes per day over the previous 6 months, owned a smartphone, and were a resident of Pennsylvania, New Jersey, or Delaware. Data were collected from May 25, 2022, to June 10, 2024.

EXPOSURE: Exposure to tobacco retail stores was assessed using mobility data matched with locations of tobacco retailers across Pennsylvania, New Jersey, and Delaware.

MAIN OUTCOMES AND MEASURES: Daily mean craving and daily number of cigarettes smoked were computed using ecological momentary assessment.

RESULTS: A total of 273 participants were included in the final analyses (mean [SD] age, 42.5 [10.7] years; 151 women [55.3%]). Multilevel models revealed support for both preregistered hypotheses. On days when individuals had more tobacco retail exposure than their own average, they reported significantly higher levels of craving (b = 0.04; 95% CI, 0.01-0.07; t3457 = 2.72; P = .01) and smoking significantly more cigarettes (b = 0.01; 95% CI, 0.0002-0.01; t3469 = 2.05; P = .04).

CONCLUSIONS AND RELEVANCE: In this cohort study of individuals who smoke cigarettes daily, exposure to tobacco retail in their everyday lives was associated with increases in craving and smoking. These findings highlight the importance of retail exposure and smoking outcomes, information that is critical for developing effective tobacco control interventions and lays the foundation for broader health research on environmental factors that shape health behaviors.

PMID:41021230 | DOI:10.1001/jamanetworkopen.2025.30132

JAMA Netw Open. 2025 Sep 2;8(9):e2533912. doi: 10.1001/jamanetworkopen.2025.33912.

ABSTRACT

IMPORTANCE: Suboptimal child passenger safety behaviors are common despite the availability of effective child restraint systems (CRS), with national guidelines promoting-and state laws requiring-their use.

OBJECTIVE: To compare 6-month guideline adherence following the Tiny Cargo, Big Deal, Abróchame Bien, Cuídame Bien (TCBD/ABCB) intervention vs enhanced usual care (EUC); assess 12-month maintenance effect; and evaluate sequential, multiple-assignment, randomized trial (SMART) phase 2 outcomes by intervention intensity.

DESIGN, SETTING, AND PARTICIPANTS: This randomized clinical trial using a SMART study design was conducted at 1 pediatric emergency department (ED), 1 community ED, and 2 urgent care centers in metropolitan Chicago, Illinois, from February 2020 to August 2022. Eligible caregivers spoke English or Spanish, lived in Illinois, owned a smartphone, traveled weekly with their child (aged 6 months to younger than 11 years), and reported suboptimal child passenger safety behaviors. Of 5416 caregivers who were screened, 1517 were potentially eligible. Data analysis occurred from October 2024 to March 2025.

INTERVENTIONS: Phase 1 included basic TCBD/ABCB (remote motivational interviewing [MI] plus tailored mobile health [mHealth]) vs EUC (text message photograph requests, follow-up activity reminders, and corrective feedback only for predefined critical errors). In phase 2, participants exposed to TCBD/ABCB but who were nonadherent at 6 months received high-intensity (second remote MI session plus extra text messages) or low-intensity (mHealth alone) interventions.

MAIN OUTCOME AND MEASURES: The primary outcome was guideline adherence, a dichotomous measure of all 3 criteria: (1) appropriate CRS, (2) back seat location, and (3) never unrestrained. Outcomes were assessed at 6 and 12 months using logistic regression with prevalence differences, adjusted for randomization strata, caregiver gender, race, and ethnicity.

RESULTS: In phase 1, there were 474 participants (mean [SD] caregiver age, 36.1 [6.2] years; 443 mothers [93.5%]; 52 [11.0%] preferred Spanish]) of whom 342 were randomized to basic TCBD/ABCB and 132 to EUC. Among participants with 6-month follow-up data, 131 of 278 caregivers receiving TCBD/ABCB (47.1%; 95% CI, 41.3% to 53.0%) vs 38 of 118 receiving EUC (32.2%; 95% CI, 23.8% to 40.6%) were guideline adherent (absolute change in treatment response, 13.1%; 95% CI, 3.6% to 22.6%; P = .007) with 12-month maintenance effects (adjusted prevalence difference, 39.2%; 95% CI, 26.5% to 51.9%; P < .001). In phase 2, 185 participants who received TCBD/ABCB and were not guideline adherent were rerandomized to high-intensity (87 participants) or low-intensity (84 participants) intervention. High-intensity TCBD/ABCB had no effect compared with low-intensity TCBD/ABCB (adjusted prevalence difference, -3.9%; 95% CI -17.9% to 10.1%; P = .59).

CONCLUSIONS AND RELEVANCE: In this randomized clinical trial of caregivers, TCBD/ABCB was associated with improved adherence to child passenger safety guidelines with lasting effects among those caregivers who were adherent at 6 months. These findings suggest that this remote precision prevention intervention against a leading cause of death may be adaptable to other settings.

TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT04238247.

PMID:41021229 | DOI:10.1001/jamanetworkopen.2025.33912

JAMA Netw Open. 2025 Sep 2;8(9):e2533934. doi: 10.1001/jamanetworkopen.2025.33934.

ABSTRACT

IMPORTANCE: In-flight medical events are an inevitable challenge in commercial aviation. Managing these events is complicated by constrained medical resources and delayed access to definitive care.

OBJECTIVE: To characterize the epidemiology of in-flight medical events and identify factors associated with aircraft diversion, hospital transport, and in-flight mortality.

DESIGN, SETTING, AND PARTICIPANTS: This cohort study included 77 790 in-flight medical events reported to a global ground-based medical support center from January 1, 2022, through December 31, 2023. All passengers experiencing an in-flight medical event across 84 participating airlines during the study period were included. Data were collected from consultations initiated by flight crew via radio or satellite communication with a dedicated ground-based physician. No demographic or clinical exclusions were applied.

EXPOSURES: Medical conditions occurring during commercial flights that prompted contact with the ground-based support center. Data included clinical presentation, in-flight management, passenger demographics, involvement of volunteer medical professionals, and disposition.

MAIN OUTCOMES AND MEASURES: Primary outcome was aircraft diversion, and secondary outcomes were hospital transport and in-flight mortality. Descriptive statistics, univariate analyses, and multivariable analyses were used to identify clinical and operational variables associated with these outcomes.

RESULTS: Among 77 790 in-flight medical events, the overall incidence was 39 events per 1 million enplanements, with 1 event per 212 flights, or 17 events per billion revenue passenger kilometers. The median (IQR) age of affected passengers (42 316 females [54.4%]) was 43 (27-61) years. Aircraft diversion occurred in 1.7% of cases, most frequently due to neurologic (41%) and cardiovascular (27%) conditions. Suspected stroke (adjusted OR [AOR], 20.35; 95% CI, 12.98-31.91) and acute cardiac emergencies (AOR, 8.16; 95% CI, 6.38-10.42) were the factors associated with the highest odds of diversion. The involvement of a physician volunteer was also associated with increased odds of diversion (AOR, 7.86; 95% CI, 4.49-13.78).

CONCLUSIONS AND RELEVANCE: In this cohort study of 77 790 in-flight medical events, these events occur more frequently than previously reported. Serious neurologic conditions, cardiac events, and physician volunteer involvement are each associated with higher odds of diversion. These findings contribute to the understanding of in-flight medical event frequency and outcomes and may inform policy, flight crew training, and diversion protocols.

PMID:41021228 | DOI:10.1001/jamanetworkopen.2025.33934

JAMA Netw Open. 2025 Sep 2;8(9):e2533971. doi: 10.1001/jamanetworkopen.2025.33971.

ABSTRACT

IMPORTANCE: The integration of microlearning tools, such as short social media videos, into medical education may offer novel strategies to enhance student engagement and knowledge acquisition. However, evidence regarding their effectiveness remains limited.

OBJECTIVE: To evaluate short-term academic performance and student engagement associated with short social media videos as a supplementary educational resource in a neuroanatomy course.

DESIGN, SETTING, AND PARTICIPANTS: This nonrandomized clinical trial was conducted at the Faculty of Medicine, Damascus University, during the 2024 to 2025 academic year. Third-year medical students who responded to an open invitation were enrolled and divided into an intervention group and a control group based on voluntary participation and access to the social media platform.

INTERVENTION: The intervention group engaged with 50 educational Instagram reels (each 90 seconds) over 2 months through a study-specific Instagram account. The control group continued traditional study methods without exposure to the videos.

MAIN OUTCOMES AND MEASURES: Primary outcomes included changes in neuroanatomy knowledge scores, measured by standardized preintervention and postintervention assessments. Secondary outcomes included student perceptions of the educational value and usability of the videos.

RESULTS: A total of 167 third-year medical students (84 men [50.3%]) were enrolled and divided into an intervention group (n = 84) and a control group (n = 83). Mean (SD) baseline assessment scores were 5.5 (2.8) of 20 in the intervention group and 5.5 (2.3) of 20 in the control group. Postintervention scores were significantly higher in the intervention group compared with the control group (mean [SD] score, 9.9 [2.1] of 20 vs 7.5 [2.4] of 20; P < .001). No significant differences were observed between the intervention group and the control group in final neuroanatomy examination scores (mean [SD] score, 17.2 [1.3] of 20 vs 16.5 [1.5] of 20; P = .09). Among intervention participants, 84.5% (71 of 84) reported that the videos were a beneficial supplement to traditional learning, although 27.4% (23 of 84) noted that the content delivery could feel monotonous.

CONCLUSIONS AND RELEVANCE: In this nonrandomized clinical trial, the use of short social media videos was associated with improved short-term academic performance and positive student engagement but not with long-term benefit. The sustained impact of short social media videos demands ongoing evaluation, evidence-based modifications, and a focus on balancing creativity with educational outcomes.

PMID:41021227 | DOI:10.1001/jamanetworkopen.2025.33971

J Womens Health (Larchmt). 2025 Sep 29. doi: 10.1177/15409996251383009. Online ahead of print.

ABSTRACT

Introduction: Whether the history of adverse pregnancy outcomes (APOs) contributes to cognitive decline in women is unclear. Methods: Among parous female Adult Changes in Thought (ACT) participants (aged ≥ 65 years without dementia at enrollment), we abstracted APO history for those born 1940 or later and enrolled between 2005 and 2020 (n = 444). Generalized estimating equations estimated the association between APO history and cognition score, measured using the Cognitive Abilities Screening Instrument assessment using item response theory (CASI-IRT), and cognitive decline. Results: Among all participants, 13% had a history of APO. In adjusted models, women with an APO history had 0.23-point lower CASI-IRT score at any age (95% confidence interval: -0.54, 0.07); this small difference was not statistically significant. We found no evidence of an association between APO history and 4-year cognitive change. Conclusion: Among women in the ACT study born in 1940 or later, we found no evidence of a significant association between history of APO and lower cognition or cognitive decline in older adulthood.

PMID:41021219 | DOI:10.1177/15409996251383009

J Endocrinol Invest. 2025 Sep 29. doi: 10.1007/s40618-025-02712-3. Online ahead of print.

ABSTRACT

BACKGROUND: The expanding clinical use of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) for type 2 diabetes and obesity has raised concerns about underrecognized ocular adverse events (AEs). Despite their metabolic benefits, fragmented evidence and methodological limitations in pharmacovigilance systems hinder comprehensive risk characterization. This study aimed to systematically evaluate ocular toxicity signals associated with GLP-1 RAs using advanced signal mining across the FDA Adverse Event Reporting System (FAERS) and VigiBase.

METHODS: Data from FAERS (2005-2025 Q1) and VigiBase (1987-2025 Q1) were analyzed for ocular AEs linked to five GLP-1 RAs (exenatide, liraglutide, dulaglutide, semaglutide, lixisenatide). Disproportionality analyses using Bayesian (Information Component, IC₀₂₅ > 0) and frequentist (Reporting Odds Ratio, ROR 95% CI > 1) methods identified significant signals. Duplicate reports and incomplete data were excluded, with statistical processing performed in R packages and Excel software.

RESULTS: Across both databases, ocular AEs predominantly occurred in individuals aged 45-64 and females. Semaglutide exhibited the strongest signals, notably for NAION (ROR = 40.18, IC₀₂₅ = 2.06 in VigiBase and ROR = 31.46, IC₀₂₅ = 4.84 in FAERS), with 2,878 cases in VigiBase and 2,047 in FAERS. Dulaglutide showed high rates of visual impairment and diabetic retinopathy. Early-onset AEs (≤ 30 days) were common with dulaglutide and semaglutide, while exenatide-related events peaked after > 360 days. Discrepancies between databases emerged, such as absent liraglutide/lixisenatide signals in VigiBase but moderate activity in FAERS. Furthermore, review of the prescribing information for these GLP-1RAs indicates that ocular adverse events are incompletely characterized. Diabetic retinopathy is reported solely for semaglutide and dulaglutide, while blurred vision is the only ocular event listed for lixisenatide. Other potential ocular risks are not mentioned in any of the respective drug labels.

CONCLUSION: This large-scale pharmacovigilance analysis suggests underrecognized ocular safety signals associated with GLP-1 RAs, most prominently NAION with semaglutide and visual impairment/diabetic retinopathy with dulaglutide. The distinct AE profiles, temporal patterns (early vs. late onset), and database discrepancies highlight the complexity of GLP-1 RA ocular toxicity. These findings underscore the need for enhanced clinical vigilance, targeted post-marketing surveillance, and consideration of label updates to reflect these emerging ocular safety concerns.

PMID:41021211 | DOI:10.1007/s40618-025-02712-3

Sports Med. 2025 Sep 29. doi: 10.1007/s40279-025-02326-y. Online ahead of print.

ABSTRACT

BACKGROUND: The epidemiology of injury in soccer has traditionally focused on soccer players, rather than match officials. Although injury data on referees exist, no comprehensive review has summarized injury profiles in this population.

OBJECTIVE: To conduct a systematic review and meta-analysis of injury epidemiology in elite and amateur soccer referees, focusing on injury rates, types, locations, severity, and causes.

METHODS: PubMed (Medline), Web of Science, Scopus, CINAHL, and SPORTDiscus, covering their entire history up to 19 April 2025 were searched. This review included prospective and retrospective studies reporting injury incidence or prevalence among football match officials, with a study period of at least one season. Studies needed to specify injury definitions and include data on injury location, type, mechanism, or severity. Both male and female officials were eligible. Systematic reviews, commentaries, and letters were excluded. Study quality and risk of bias were evaluated using the STROBE-SIIS, in addition to the Newcastle-Ottawa Scale and funnel plots. Injury incidence rates were estimated using a random effects Poisson regression, accounting for heterogeneity and moderators. Heterogeneity was assessed with the I² statistic.

RESULTS: A total of 17 studies were included, encompassing 3621 referees. The most frequent injuries were strains and sprains in the knee and ankle. The overall injury incidence was 2.19 injuries per 1000 h of exposure (95% CI 1.30-3.69). On-field referees experienced an incidence rate of 1.46 injuries per 1000 h of exposure (95% CI 0.76-2.81), while assistant referees had a lower rate of 0.84 per 1 h of exposure (95% CI 0.36-1.97). During matches, the injury incidence was 2.24 per 1000 h of exposure (95% CI 1.38-3.64), compared with 0.67 injuries per 1000 h of exposure during training sessions (95% CI 0.36-1.24). However, despite sensitivity analysis, there were still high levels of heterogeneity across included studies.

CONCLUSIONS: Findings noted higher injury incidence during matches compared with training, and on-field referees compared with assistants. The variation in injury profiles highlights the importance of implementing targeted preventive strategies tailored to the unique demands of refereeing. However, there is still a lack of research in this population, especially in female referees.

PROSPERO REGISTRATION NUMBER: CRD42024497970.

PMID:41021207 | DOI:10.1007/s40279-025-02326-y

Discov Oncol. 2025 Sep 29;16(1):1752. doi: 10.1007/s12672-025-03536-6.

ABSTRACT

BACKGROUND: Platelet indices are non-invasive biomarkers used to assess disease status. Observational studies have shown that platelet indices are related to liver cancer. However, the specific relationship between the two is still inconclusive. We conducted a Mendelian randomization (MR) study to investigate the causal relationship between platelet indices and liver cancer.

METHODS: We identified 2 cancer outcomes (liver cell carcinoma, malignant neoplasm of liver and intrahepatic bile ducts) from genome-wide association studies (GWAS). We used inverse variance weighted (IVW) method as the main method. In order to evaluate the potential horizontal pleiotropy or heterogeneity, sensitivity analyses were examined via MR-Egger regression, heterogeneity test, pleiotropy test.

RESULTS: Two-Sample MR analysis showed there were causal relationships between platelet indices and liver cancer. The mean platelet volume was positively associated with hepatocellular carcinoma (HCC) risk, while platelet count was inversely correlated with HCC risk, platelet endothelial cell adhesion molecule levels were significantly related to an increased malignant neoplasm of liver and intrahepatic bile ducts risk. In addition, we conducted reverse MR analyses and found that malignant neoplasm of liver and intrahepatic bile ducts was positively associated with platelet count and platelet crit, but negatively correlated with platelet distribution width. Following multivariable mendelian randomization adjustments for potential confounding variables, the independent effects of platelet indices on HCC were not statistically significant.

CONCLUSIONS: The bidirectional causal relationship between platelet indices and liver cancer may be mediated by confounding factors such as liver cirrhosis. Further validation through mediation MR analyses or larger-scale replication studies is required to elucidate these causal pathways.

PMID:41021163 | DOI:10.1007/s12672-025-03536-6

Discov Oncol. 2025 Sep 29;16(1):1774. doi: 10.1007/s12672-025-03579-9.

ABSTRACT

BACKGROUND: Liver cancer remains one of the most prevalent and lethal malignancies worldwide, while autoimmune hepatitis is an immune-mediated inflammatory liver disease characterized by circulating autoantibodies and distinctive histopathological features. Mendelian Randomization (MR), as an innovative epidemiological research method, providing a reliable approach for exploring causal relationships between gene expression and liver disease risk.

METHODS: This study employed a comprehensive two-sample Mendelian randomization analytical approach to systematically evaluate the causal relationships between druggable gene expression levels and liver disease risk, including both liver cancer and autoimmune hepatitis. The study incorporated expression quantitative trait loci (eQTL) data from 4,479 druggable genes as exposure variables, with liver cancer genome-wide association study (GWAS) data and autoimmune hepatitis GWAS data serving as outcome variables.

RESULTS: Mendelian randomization analysis revealed significantly different impacts of multiple gene expression levels on liver disease outcomes. For liver cancer, high expression of CCL7 and CCL11 genes demonstrated protective effects, with CCL11 gene showing statistical significance (HR = 0.600, 95%CI: 0.409-0.953, p = 0.027), suggesting these genes may exert protective functions through immune response regulation mechanisms. Conversely, high expression of CASP8, CD5, and FGF21 genes significantly increased patient mortality risk.

CONCLUSIONS: This study provides sufficient and reliable scientific evidence for liver disease risk-related genes through multidimensional Mendelian randomization validation analysis.

PMID:41021153 | DOI:10.1007/s12672-025-03579-9