Nevin Manimala

Biometrics. 2026 Jul 1;82(3):ujag140. doi: 10.1093/biomtc/ujag140.

ABSTRACT

For the analysis of interval-censored data, we propose a deep generalized accelerated hazards model. This model is designed to facilitate a detailed exploration of the relationship between various risk factors and the hazard associated with failure time. We develop a sieve maximum likelihood estimation procedure that combines deep neural networks and monotonic splines. By employing deep neural networks, we can effectively capture nonparametric effects, enabling a flexible and adaptive modeling approach for complex relationships. Under certain regularity conditions, we derive a nonasymptotic error bound for the resulting estimator and show that the finite-dimensional estimator is asymptotically normal and achieves the semiparametric efficiency. We conduct simulation studies to evaluate the finite-sample performance of the proposed approach. Furthermore, the proposed method is applied to the Atherosclerosis Risk in Communities study for practical illustration.

PMID:42574000 | DOI:10.1093/biomtc/ujag140

Biometrics. 2026 Jul 1;82(3):ujag138. doi: 10.1093/biomtc/ujag138.

ABSTRACT

Neyman-Pearson (NP) classifiers, which aim to maximize the clinical benefit while adhering to risk constraints, are crucial in many practical fields, including early cancer detection. However, applying these classifiers can be challenging due to discrepancies between the data distributions of the source and target populations. The potential impact can be disproportionately severe for under-represented groups. We propose a semi-parametric model-based approach for adapting NP classifier decision rules to different populations while equitably controlling classification errors specific to clinical applications. Our method involves a shift-adjustment strategy that leverages a small unlabeled sample from the target population, along with minimal auxiliary information and the labeled source data. This approach enhances the applicability of the learned decision rules and ensures they are consistently tailored for the target population. We demonstrate the performance through theoretical studies and simulations and illustrate the approach with an example of a prostate cancer study.

PMID:42573999 | DOI:10.1093/biomtc/ujag138

JAMA Intern Med. 2026 Aug 10. doi: 10.1001/jamainternmed.2026.3332. Online ahead of print.

ABSTRACT

IMPORTANCE: Medicare Advantage (MA) costs 22% more than original Medicare (OM) for a given individual ($83 billion in annual excess public costs). However, MA may improve type 2 diabetes (T2D) outcomes compared with OM by providing financial protections (eg, annual out-of-pocket spending caps) and supplemental benefits (eg, healthy food assistance) that OM cannot.

OBJECTIVE: To determine whether MA coverage is associated with better T2D outcomes than OM.

DESIGN, SETTING, AND PARTICIPANTS: A longitudinal cohort study using target trial emulation principles for design and analysis in adults aged 18 years or older receiving OM or MA with T2D, followed up before and after Medicare coverage in community-based health centers (January 2021 to June 2024) across 44 states. Analyses were conducted from September 2025 to May 2026.

EXPOSURES: MA or OM coverage.

MAIN OUTCOMES AND MEASURES: Hemoglobin A1c (HbA1c) (primary outcome), systolic blood pressure (SBP) and diastolic blood pressure (DBP), low-density lipoprotein (LDL) cholesterol, food insecurity, housing instability, and transportation barriers at 12 months after Medicare coverage (primary time point) and at 6, 18, and 24 months. Statistical analysis accounted for pre-Medicare coverage factors that may influence selection of MA vs OM using targeted minimum loss estimation. Covariates were age, sex, race and ethnicity, comorbidities, income, Social Vulnerability Index, pre-Medicare insurance, Medicaid coverage, and pre-Medicare coverage values for HbA1c, SBP, DBP, LDL cholesterol, body mass index, food insecurity, housing instability, and transportation barriers.

RESULTS: In this study in 34 648 adults (19 054 in OM, 15 594 in MA) with T2D, followed up before and after Medicare coverage, the mean (SD) age was 65.24 (9.73) years and 53.42% were women. Twelve months after Medicare coverage, MA was not associated with better HbA1c (mean difference, 0.01; 95% CI, -0.04 to 0.05, P = .74), SBP (-0.15; 95% CI, -0.54 to 0.24; P = .44), DBP (0.06; 95% CI, -0.15 to 0.27; P = .58), or LDL cholesterol (-0.41; 95% CI, -1.24 to 0.42; P = .33), with similar results at other time points. MA was also not associated with a lower risk of food insecurity (relative risk [RR], 1.00; 95% CI, 0.94-1.05), housing instability (RR, 1.00; 95% CI, 0.91-1.09), or transportation barriers (RR, 1.00; 95% CI, 0.93-1.07) at 12 months or any other time point.

CONCLUSIONS AND RELEVANCE: In this study, when accounting for factors that may drive MA selection, MA was not associated with better T2D outcomes or fewer health-related social needs than OM. Given substantially higher spending for MA, it is important to ensure this spending is being used effectively to improve health.

PMID:42573997 | DOI:10.1001/jamainternmed.2026.3332

Obes Surg. 2026 Aug 10. doi: 10.1007/s11695-026-08895-9. Online ahead of print.

ABSTRACT

OBJECTIVE: To evaluate whether metabolic bariatric surgery (MBS) influences the risk of alcohol-related liver disease (ARLD), particularly alcoholic cirrhosis, and assess its long-term hepatic implications.

METHODS: This study followed PRISMA guidelines and was prospectively registered in PROSPERO. PubMed, Embase, Web of Science, Scopus, Cochrane Library, and ClinicalTrials.gov were systematically searched from inception to 26 December 2025 for observational cohort studies comparing the incidence of alcoholic cirrhosis between MBS patients and matched non-surgical controls. Data extraction and quality assessment were independently performed, and pooled effect sizes were calculated using random-effects models. Subgroup, sensitivity, and publication bias analyses were conducted. Statistical analyses were performed using Stata 17.0.

RESULTS: Five studies were included. The overall association between MBS and alcohol-related cirrhosis was not statistically significant (OR 1.18, 95% CI 0.91-1.53; p = 0.214), with substantial heterogeneity (I² = 97.5%). However, significantly increased odds were observed in studies with sample sizes ≥ 1,000,000 and in those with follow-up durations ≥ 10 years. Leave-one-out analysis showed that no single study substantially altered the pooled estimate, and Begg’s and Egger’s tests detected no significant publication bias.

CONCLUSIONS: Current evidence does not show a significant overall association between MBS and alcohol-related cirrhosis. However, substantial heterogeneity and increased odds in large-sample and long-term studies preclude firm conclusions regarding long-term hepatic safety.

PMID:42573970 | DOI:10.1007/s11695-026-08895-9

Dermatol Ther (Heidelb). 2026 Aug 10. doi: 10.1007/s13555-026-01884-x. Online ahead of print.

ABSTRACT

INTRODUCTION: Scalp psoriasis is often associated with poor adherence to topical therapy. A novel formulation of calcipotriol and betamethasone dipropionate based on polyaphron dispersion (CAL/BDP PAD-cream) showed improved outcomes and satisfaction in the PRO-SCALP study, particularly in patients with high adherence. We evaluated how adherence to CAL/BDP PAD-cream influences patients- and clinicians-reported outcomes and treatment preferences in mild-to-moderate scalp psoriasis under real-life conditions in Europe.

METHODS: PRO-SCALP patients reported their adherence level using a visual analogue scale (VAS). Outcomes were compared between low- and high-adherence subgroups.

RESULTS: Among 252 patients, 59.9% reported high adherence (VAS 80-100). Older patients and those with moderate disease reported high adherence (both p < 0.05). High-adherent patients reported higher scores in the Treatment Satisfaction Questionnaire for Medication Version 9, for Convenience of use (p = 0.0019) and Global Satisfaction (p = 0.0166) domains, and in the Psychosocial Effects of Scalp Psoriasis Questionnaire (p = 0.0004) at week 8. Both adherence subgroups showed significant reductions in the scalp Worst Itch Numeric Rating Scale (WI-NRS), scalp-modified Psoriasis Area and Severity Index (S-mPASI), and Scalpdex scores (all p < 0.0001 vs. baseline), although high-adherent patients achieved greater improvements in WI-NRS (p < 0.0001), S-mPASI (p = 0.0153), and the Scalpdex Symptoms domain (p = 0.001) than low-adherent. Each 10% increase in adherence corresponded to a 0.10- and 0.35-point reduction in S-mPASI and WI-NRS (both p < 0.05) at week 8. Scalp-Physician Global Assessment success rates were comparable in low- vs. high-adherence subgroups (65.0% vs. 70.5%; p = 0.3633). Sleep quality improved significantly in both subgroups (p < 0.0001). High adherence was associated with higher Patient Preference Questionnaire scores (p = 0.0012), better Cream Usability Scalp Psoriasis Questionnaire ratings (p = 0.0455) and greater product consumption (p < 0.0001), despite similar once-a-day usage.

CONCLUSION: High adherence to CAL/BDP PAD-cream was associated with greater effectiveness, satisfaction, preference, and QoL. While patients with low adherence still benefited, maximizing adherence is key for optimal real-world outcomes in scalp psoriasis.

TRIAL REGISTRATION NUMBER: ClinicalTrials.gov identifier NCT05811234.

PMID:42573965 | DOI:10.1007/s13555-026-01884-x

Esophagus. 2026 Aug 10. doi: 10.1007/s10388-026-01238-8. Online ahead of print.

ABSTRACT

BACKGROUND: Upper gastrointestinal endoscopy traverses the full upper aerodigestive tract, unlike radiography, potentially enabling incidental detection of non-gastric malignancies. However, its population-level incidental detection rate for non-gastric upper aerodigestive tract cancers has not been systematically quantified. The aim of this study was to compare endoscopic screening with radiography and quantify detection of non-gastric upper aerodigestive tract cancers in a population-based setting.

METHODS: This population-based cohort study was conducted by linking the Okayama City municipal gastric cancer screening registry with the Kokuho Database (KDB) for fiscal years 2016-2021. Among 36,326 participants contributing 64,822 screening examinations (40,832 radiography; 23,990 endoscopy), diagnoses of oral cavity, pharyngeal, laryngeal, and esophageal cancer occurring within 2 months of screening were ascertained from the KDB. Generalized estimating equations with modified Poisson regression were used to estimate adjusted risk ratios (aRRs) comparing endoscopy with radiography.

RESULTS: Endoscopic screening was significantly associated with higher composite incidental detection rates for non-gastric upper aerodigestive tract cancers (95.9 vs. 34.3 per 100,000 examinations; aRR 2.94, 95% confidence interval [CI] 1.50-5.76; p = 0.002). Specifically, esophageal cancer detection was markedly higher with endoscopy (83.4 vs. 17.1 per 100,000; aRR 5.16, 95% CI 2.16-12.32; p < 0.001). No statistically significant differences were observed for oral cavity, pharyngeal, or laryngeal cancers.

CONCLUSIONS: Endoscopic gastric cancer screening is associated with substantially higher incidental detection of non-gastric upper aerodigestive tract cancers, particularly esophageal cancer. These findings suggest an additional detection value for endoscopy that extends beyond its primary gastric cancer target in organized screening programs.

PMID:42573963 | DOI:10.1007/s10388-026-01238-8

Environ Sci Pollut Res Int. 2026 Aug 10. doi: 10.1007/s11356-026-38124-w. Online ahead of print.

ABSTRACT

The transfer of pesticides to Aiguillon Bay, a major coastal ecosystem on the Atlantic coast of France, was investigated in relation to watershed characteristics, agricultural pressure, and hydrological dynamics. The bay receives inputs from three main rivers (Sèvre Niortaise, Lay, and Curé) as well as from the Vieux channel, a downstream branch of the Lay watershed characterized by distinct land-use and drainage features. This study combined spatial land-use analysis with contamination indicators to clarify pesticide transfer pathways and associated ecological risks in intensively cultivated sub-basins. Monthly surface water samples were analyzed using LC-MS/MS and GC-MS/MS. A Proximity Indicator was developed to identify high-pressure agricultural zones adjacent to watercourses. Individual Risk Quotients (RQs) were also calculated by comparing measured environmental concentrations with Predicted No-Effect Concentrations (PNECs) to evaluate the ecological risk associated with selected pesticide compounds, and the Cumulative Toxic Pressure Index (CTPI) was applied to assess mixture toxicity. CTPI analysis revealed recurrent exceedances of the toxicity threshold (CTPI > 1) across all monitored systems, with strong seasonal variability linked to hydrological conditions. The Sèvre Niortaise and Vieux channel exhibited sustained mixture pressure, whereas the Lay showed pronounced event-driven peaks associated with rainfall episodes. Despite its smaller size, the Curé watershed displayed disproportionately high toxic pressure, reflecting strong hydrological connectivity and cereal-dominated land use. Herbicides and their metabolites were the primary contributors to mixture toxicity, including the persistent metabolite chlorothalonil R471811, frequently detected despite its regulatory ban in 2020, suggesting legacy contamination and progressive remobilization. Although individual Risk Quotients indicated negligible ecological risk for the selected compounds, CTPI revealed repeated mixture toxicity exceedances, demonstrating that cumulative effects represent the primary ecological pressure within the watershed. By integrating land-use characterization, hydrological analysis, statistical comparison, mixture toxicity assessment, and ecological risk evaluation, this study provides a comprehensive framework for understanding pesticide transfer to protected coastal ecosystems and supports the development of more effective watershed management and monitoring strategies.

PMID:42573961 | DOI:10.1007/s11356-026-38124-w

Sleep Breath. 2026 Aug 10;30(4):236. doi: 10.1007/s11325-026-03781-1.

ABSTRACT

INTRODUCTION: Obstructive sleep apnea (OSA) commonly coexists with Restless Legs Syndrome (RLS). Positive Airway Pressure (PAP) treatment for coexisting OSA may alleviate RLS symptoms. However, RLS often contributes to insomnia, leading to difficulties with sleep initiation and maintenance that may compromise PAP adherence. Objective long-term adherence data in patients with comorbid RLS and OSA remains limited. This study evaluated PAP therapy adherence and its determinants in patients with comorbid RLS and OSA.

METHODS: This retrospective cohort study included patients with RLS who reported snoring and underwent overnight polysomnography (PSG) from Jan 2022 to July 2024, at the Division of Sleep Medicine in the Peking University People’s Hospital. Clinical features and objective PAP adherence were collected. And factors influencing adherence were assessed through semi-structured interviews. Patients were followed up 1 year after PAP initiation. Good adherence was defined as device use for ≥ 4 h/night on ≥ 70% of nights.

RESULTS: Among 114 patients with RLS, 84 (73.7%) had comorbid OSA. Of these, 70 initiated PAP therapy, 37 acquired devices and only 25 provided objective adherence data. The proportion of adherent users declined progressively: 76% at 1 week, 36% at 1 month, 20% at 3 months, 16% at 6 months and 12% at 1 year. Main influencing factors included a perceived lack of need or unwillingness to continue PAP treatment, device intolerance and socioeconomic burden.

CONCLUSIONS: Long-term PAP adherence appears suboptimal in this cohort of patients with comorbid RLS and OSA. Interventions targeting sleep symptom relief, device tolerability and socioeconomic factors may improve treatment adherence.

PMID:42573935 | DOI:10.1007/s11325-026-03781-1

Drug Saf. 2026 Aug 10. doi: 10.1007/s40264-026-01696-0. Online ahead of print.

ABSTRACT

BACKGROUND AND AIM: Pharmacovigilance is essential to ensuring patient safety by enabling timely identification of adverse reactions in increasingly complex and voluminous data. Routine quantitative signal detection methods generate statistical alerts for product-event pairs based on predefined criteria; however, most alerts do not warrant further investigation, creating inefficiencies and significant time demands for pharmacovigilance teams. Manual triage of these alerts is often resource-intensive, prone to variability, and challenging to audit, highlighting the need for more reliable, transparent and efficient triage strategies. This study aimed to design, develop and prospectively evaluate an explainable Machine Learning for Intelligent Triage (MLIT) tool to assist pharmacovigilance teams in reviewing statistical alerts for vaccine and drug portfolios. The objective was to enhance signal detection performance without increasing the risk of missing signals, improving operational efficiency and maintaining decision traceability and regulatory compliance.

METHODS: Alert and individual case safety report data were retrieved from the company’s safety and signal management databases. Feature selection was guided by prior experience with a published case completeness tool, called Clinical Utility Score for Prioritisation (CUSP), and expert input. Of several ML methods explored, eXtreme Gradient Boosting (XGBoost) emerged as the optimal algorithm, with models trained and tested using a 75/25 split dataset. Iterative model refinement was conducted using Shapley Additive Explanations analyses to ensure explainability and alignment with safety reviewers’ decision-making processes. Refined models underwent prospective validation in two four-month prospective validation studies, covering over 20 products across vaccine and drug portfolios. The prospective validations assessed concordance between model predictions and reviewers’ decision under real-world conditions, as well as estimated time savings.

RESULTS: The vaccine model demonstrated robust predictive performance, achieving a weighted-average F1 score of 0.81 and an accuracy of 0.79. In the prospective validation phase, 92% of vaccine alerts were closed in alignment with the model’s top-ranked prediction, while 98% were closed within the top 3 predictions. The MLIT tool also identified inconsistencies and human errors in manual triage, highlighting its potential role as a quality-control mechanism. Safety reviewers reported a 24% reduction in time spent on triage activities, and explainability analyses confirmed that the model’s decision-making was conceptually aligned with safety reviewers’ logic. Comparable results were observed for the drug portfolio.

CONCLUSION: This study highlights the potential of ML-based tools to improve pharmacovigilance by enhancing signal detection performance, reducing the likelihood of missed signals, while increasing operational efficiency, and strengthening reproducibility and transparency. While MLIT demonstrated high concordance with expert decisions and provided meaningful time savings, human oversight remains essential, especially for low-confidence predictions. Ongoing refinement and user engagement will be critical for broader implementation and further automation, marking a significant step forward in ensuring safer and more efficient drug safety surveillance.

PMID:42573921 | DOI:10.1007/s40264-026-01696-0

Odontology. 2026 Aug 10. doi: 10.1007/s10266-026-01531-z. Online ahead of print.

ABSTRACT

This study aims to compare the diagnostic accuracy, appropriateness of treatment planning, and source citation performance of five large language models ChatGPT-4o (Free), ChatGPT-5.1 Plus, Microsoft Copilot, Google Gemini, and DeepSeek-R1 in root resorption scenarios. In December 2025, twelve clinical scenarios were created based on the classification of the European Society of Endodontology and each scenario was presented to all chatbots over four consecutive days. All responses were evaluated using a blinded assessment protocol and a binary scoring system. A total of 720 observations (12 cases × 4 repetitions × 3 criteria per model) were analyzed. The collected data were analyzed using chi-square, Fisher’s exact, and Cochran Q tests. In terms of diagnostic accuracy, Microsoft Copilot (79.2%), ChatGPT-5.1 (77.1%), and ChatGPT-4o (Free) (75%) showed the highest performance. Google Gemini (68.8%) demonstrated a moderate level of accuracy, while DeepSeek (39.6%) showed markedly low performance. All models exhibited high accuracy in treatment plan recommendations, and no statistically significant differences were detected. Regarding citation accuracy, Copilot ranked first with 100% accuracy. Although large language models present potential as supportive decision-making tools in the evaluation of root resorption, diagnostic inconsistencies, limitations in source accuracy, and variability in responses restrict their independent use in clinical applications. Therefore, the outputs generated by these models should be interpreted cautiously within clinical decision-making processes.

PMID:42573919 | DOI:10.1007/s10266-026-01531-z