Tag: statistics

BMC Med Educ. 2026 Mar 2. doi: 10.1186/s12909-026-08892-7. Online ahead of print.

ABSTRACT

BACKGROUND: Reliable assessment and meaningful feedback are essential to effective learning in medical education. However, conventional unstructured evaluation of essay-type responses is highly vulnerable to rater bias, inter-rater variability, and nonspecific feedback. To address these limitations, we developed an Assessment-cum-Feedback Checklist to provide a structured, criterion-based approach to scoring and feedback. In this study, we evaluated the checklist’s effectiveness in enhancing the reliability, consistency, and clarity of assessment while exploring student and faculty perceptions of its educational value.

METHODS: We used a mixed-methods design. Sixty-two first-year MBBS students and four faculty members (two junior < 5 years’ experience; two senior > 10 years’ experience) participated. Two essay-type questions were assessed independently by all four teachers using both the conventional unstructured method and the checklist-based method. Quantitative analyses included descriptive statistics, Wilcoxon signed-rank tests, Levene’s test for equality of variance, intraclass correlation coefficients (ICC), and Bland-Altman analysis to compare variability and agreement across methods. Data were analysed using JASP (version 0.18.3.0) at a 5% significance level. Student perceptions were gathered using a structured questionnaire, and faculty perceptions were explored through in-depth interviews. Qualitative data were analysed using QCAmap (2020). Institutional Ethics Committee approval was obtained.

RESULTS: Checklist-based scoring demonstrated lower standard errors, standard deviations, and coefficients of variation, indicating improved precision and reduced subjective variability compared with the conventional method. Mean scores were lower with the checklist, and Bland-Altman analysis showed a negative bias, reflecting greater scoring stringency due to explicit criteria. ICC values increased notably with the checklist-particularly among senior teachers-demonstrating improved inter-rater reliability and tighter limits of agreement. Teachers reported that the checklist enhanced objectivity, reduced bias, clarified performance expectations, and standardized assessment practices. Students expressed strong support, citing improved clarity, transparency, and usefulness of feedback.

CONCLUSIONS: The Assessment-cum-Feedback Checklist was associated with measurable improvements in the reliability and consistency of essay-type assessment. Both faculty and students perceived the checklist-based approach to enhance clarity, transparency, and the usefulness of feedback by making assessment criteria explicit. With appropriate faculty orientation and iterative refinement, the checklist represents a promising and potentially adaptable tool for strengthening assessment and feedback practices in constructed-response formats in medical education.

PMID:41772543 | DOI:10.1186/s12909-026-08892-7

J Environ Manage. 2026 Mar 1;402:129027. doi: 10.1016/j.jenvman.2026.129027. Online ahead of print.

ABSTRACT

During sewage sludge composting, the dynamic and nonlinear interactions among composting parameters and ammonia (NH₃) emissions present challenges to conventional statistical methods, hindering stage-specific mitigation of NH₃ emissions. This study employed multi-stage machine learning to investigate the associations among composting parameters and cumulative NH₃-N emissions (cNH₃) across different composting stages. The stage-specific models demonstrated high predictive accuracy (R² = 0.85-0.91) on independent test sets. Shapley Additive Explanations analysis identified composting time, aeration rate, and pH as the features most strongly associated with cNH₃ during the mesophilic and thermophilic stage, while aeration rate, pH, and organic matter were the predominant factors during the cooling and mature stage. Bivariate partial dependence plots revealed optimal parameter ranges and interactions, including a synergistic relationship between organic matter and nitrate levels, which was linked to lower cNH₃ in the model outputs. These findings illuminate the evolving relationship between key composting parameters and NH₃ emissions throughout the composting process, providing a scientific basis for developing stage-specific strategies to minimize NH₃ emissions based on model-derived patterns.

PMID:41771229 | DOI:10.1016/j.jenvman.2026.129027

Reprod Biomed Online. 2025 Aug 14;52(4):105221. doi: 10.1016/j.rbmo.2025.105221. Online ahead of print.

ABSTRACT

RESEARCH QUESTION: Is the quality, relevance and empathy of the answers provided by large language models (LLMs) in response to the most frequently asked patient questions in reproductive medicine comparable to those provided by human specialists?

DESIGN: This monocentric, double blind, prospective study involved two clinicians and two embryologists who answered 13 frequently asked questions in their respective field. The same questions were asked to a free online LLM, with the same constraint of text length as practitioners. All answers were blindly evaluated by four assessors (two gynaecologists and two embryologists depending on the topic) for quality and accuracy. A psychologist also evaluated empathy.

RESULTS: The mean number of words per answer was significantly higher (P < 0.001) for LLM than for humans. The average quality of answers was not statistically different between LLM and professionals. No answer provided by LLM was evaluated as completely aberrant, and only a minority contained false or inappropriate information or was scored as being very poor by assessors. Answers provided by embryologists, but not clinicians, ranked significantly higher (P = 0.02) than LLM. The psychologist chose LLM answers as most empathetic, clear, or both, in 14 out of 26 questions.

CONCLUSIONS: LLMs could be used as an educational tool within assisted reproductive technology centres to answer frequently asked patient questions. Although the potential applications of LLMs’ capabilities in answering medical questions are numerous, this should be carefully evaluated and regulated to prevent the dissemination of inaccurate information to patients.

PMID:41771212 | DOI:10.1016/j.rbmo.2025.105221

Int Dent J. 2026 Mar 1;76(3):109404. doi: 10.1016/j.identj.2025.109404. Online ahead of print.

ABSTRACT

INTRODUCTION AND AIMS: Automated dental report generation faces significant challenges in multimodal fusion, often resulting in suboptimal semantic quality and risks of hallucination, where AI generates clinically unsupported content. Current approaches that rely on simple feature concatenation or bidirectional attention mechanisms fail to effectively capture visual-textual relationships in medical imaging. This study aims to develop MedFusionT5, a unidirectional cross-modal alignment framework that (1) achieves superior clinical report quality through focused attention between visual patches and clinical text representations, and (2) ensures exceptional factual consistency by minimising hallucination rates.

METHODS: We implemented a novel architecture that integrates vision transformer (ViT) for patch-based visual feature extraction with Bio_ClinicalBERT for clinical text encoding. The core innovation is a unidirectional multihead attention alignment module that selectively maps textual embeddings to relevant visual patches before multimodal fusion. A T5-base decoder then generates diagnostic reports from the aligned representations. We evaluated performance on 700 dental panoramic radiographs using comprehensive metrics, including BLEU, ROUGE, CIDEr, clinical precision/recall, and specialised hallucination analysis, comparing against both concatenation and coattention baselines.

RESULTS: MedFusionT5 demonstrated superior performance across all evaluated metrics. Compared to the coattention baseline, CIDEr increased by 122% (5.65 vs 2.54) and by 320% over simple concatenation. BLEU-4 reached 0.865, outperforming both baselines, while maintaining the lowest hallucination rate at 2.42% (39% reduction vs coattention, 46% vs concatenation). The model achieved an optimal balance between precision (0.982) and recall (0.923), with 90% of reports exhibiting near-zero hallucination. Notably, MedFusionT5 showed consistent quality independent of report length (r = -0.022), unlike coattention’s length-dependent performance (r = +0.795).

CONCLUSION: MedFusionT5 establishes a new state-of-the-art in automated dental report generation, demonstrating that unidirectional cross-modal alignment achieves superior semantic quality and clinical precision while minimising hallucinations. This work identifies unidirectional attention as the optimal alignment strategy for medical AI, providing a foundation for trustworthy clinical deployment where both accuracy and reliability are paramount.

PMID:41771189 | DOI:10.1016/j.identj.2025.109404

J Breath Res. 2026 Mar 2. doi: 10.1088/1752-7163/ae4bfd. Online ahead of print.

ABSTRACT

Melioidosis is a life-threatening infectious disease caused by Burkholderia pseudomallei (Bp). Rapid diagnosis and appropriate antimicrobial treatment are critical to reduce mortality, yet diagnosis is hindered by diverse clinical manifestations, mimicry with other diseases, and reliance on slow culture-based methods. Detecting volatile compounds offers a non-invasive approach for rapid infection detection. In this study, we aim to identify volatile compounds in patients’ breath that can aid in diagnosing melioidosis and indicating response to treatment.&#xD;Methods: Breath samples were collected from 17 patients with culture-confirmed melioidosis and eight patients with other febrile illnesses. Longitudinal samples were collected from five of the 17 melioidosis patients over approximately one month of antibiotic treatment. Breath samples were analyzed using comprehensive two-dimensional gas chromatography time-of-flight mass spectrometry. Data analysis involved statistical comparison and machine learning-based feature selection. &#xD;Results: We identified three breath markers -camphene, 1-butanol, and 3-methylheptyl acetate -that discriminated melioidosis (n=7) from febrile controls (n=6) with an area under the receiver operating characteristic curve of 1.00. These three markers correctly classified 11 additional samples from 11 melioidosis patients, with one febrile control misclassified. Separately, we selected four breath markers, three of which were hydrocarbons, that differentiated samples associated with a positive Bp culture from those with a negative Bp culture, with a random forest model developed upon these four markers showing a sensitivity of 98% and specificity of 95%. Moreover, we identified a set of 16 volatile compounds that significantly correlated (correlation coefficient > 0.6) with blood C-reactive protein levels. Lastly, a panel of 144 volatile compounds was identified that corresponded to treatment time, indicating that the breath profile may reflect treatment response or shifts in disease severity.&#xD;Conclusion: This pilot study reports candidate breath-based markers for diagnosing melioidosis and assessing treatment outcome, supporting further validation in larger studies. &#xD.

PMID:41771178 | DOI:10.1088/1752-7163/ae4bfd

J Radiol Prot. 2026 Mar 2. doi: 10.1088/1361-6498/ae4be9. Online ahead of print.

ABSTRACT

Ionizing radiation elicits complex cellular responses that are influenced by both total dose and delivery rate. Understanding how dose rate modulates molecular outcomes is important for accurate risk assessment. In this study, we apply an integrative multi-omics approach combining transcriptomic and proteomic profiling, adjusting for covariates, to investigate how differential dose rates alter gene and protein expression in human lymphocytes, with emphasis on alterations in key molecular pathways.&#xD;Methods: Peripheral blood from 14 healthy donors (8 males, 6 females) was irradiated ex vivo with X-rays at 0.05 Gy/minute (DR1) and 1.0 Gy/minute (DR2) across a dose range 0-6 Gy. Gene expression was assessed using TempO-Seq™, and relative protein abundance was determined by mass spectrometry. Differential expression analysis was conducted using edgeR and limma, adjusting for sex, age, and leukocyte counts (false discovery rate (FDR) < 0.05). Multi-omics integration was performed using regularized canonical correlation analysis (rCCA) implemented in mixOmics, followed by Reactome pathway enrichment analysis.&#xD;Results: We identified 2,477 and 2,612 differentially expressed genes at DR1 and DR2, respectively, and 368 and 386 differentially expressed proteins. Using canonical variates from rCCA, we show that covariate adjustment improved dose discrimination, particularly above 0.5 Gy. Using a correlation cut-off threshold of 0.5 in rCCA, 212 (DR1) and 276 (DR2) highly correlated gene-protein pairs were identified. DR2 exposure was associated with stronger enrichment of stress-related pathways, including unfolded protein response, senescence and oncogenic kinase signaling. In contrast, DR1 induced enrichment of pathways associated with immune engagement, including antigen presentation. At both dose rates, transcriptomic changes highlighted upstream regulatory processes (chromatin modeling) and proteomic changes captured downstream functional pathways such as immune activity and apoptosis. &#xD;Conclusion: Multi-omics approach with covariate adjustment revealed key radiation-responsive pathways and dose-rate-dependent molecular differences, highlighting the value of integrating transcriptomic and proteomic data to better understand radiation effects.

PMID:41771177 | DOI:10.1088/1361-6498/ae4be9

Thorac Cancer. 2026 Mar;17(5):e70255. doi: 10.1111/1759-7714.70255.

ABSTRACT

BACKGROUND: Pleural mesothelioma is a highly aggressive malignancy with a poor prognosis due to the limited efficacy of currently available therapies. Macroautophagy (hereafter “autophagy”) is a lysosome-mediated degradation pathway involved in cellular homeostasis that can either support or inhibit cancer progression depending on context. In this study, we investigated the effects of SAR405, an inhibitor of vacuolar protein-sorting 34 (VPS34), which is important for regulating the early stage of autophagy, on pleural mesothelioma.

METHODS: Human pleural mesothelioma cell lines H28, H2452, and 211H were cultured with SAR405. The effects of SAR405 on protein expression, cell viability, colony formation, cell invasion, and the cell cycle were investigated, as were its synergistic effects with cisplatin. Autophagy induction was evaluated in mesothelioma cells transfected with the pMRX-IP-GFP-LC3-RFP-LC3ΔG plasmid, which was developed for the quantitative and statistical estimation of autophagy.

RESULTS: SAR405 treatment alone significantly reduced cell viability, colony formation, and cell invasion, and increased G₂/M cell cycle arrest. In addition, SAR405 induced apoptosis in the H2452 cell line. Although cisplatin weakly induced autophagy in mesothelioma cells, its combination with SAR405 did not result in additive or synergistic effects on cell viability.

CONCLUSIONS: Based on these results, inhibition of VPS34 by SAR405 effectively suppressed cell viability in all mesothelioma cell lines and induced apoptosis in H2452 cells. The findings of this study indicate the potential for VPS34 inhibition as a new strategy for the treatment of mesothelioma and provide insights into the complex role of autophagy in this malignancy.

PMID:41771171 | DOI:10.1111/1759-7714.70255

Eur J Heart Fail. 2026 Jan 14:xuag009. doi: 10.1093/ejhf/xuag009. Online ahead of print.

ABSTRACT

AIMS: Cardiogenic shock (CS) is often treated with catecholamines titrated to an adequate target mean arterial pressure (MAP) while minimizing adverse effects. We aim to assess the optimal catecholamine dose/MAP balance in heart failure-associated CS (HF-CS).

METHODS: Patients with HF-CS were retrospectively enrolled from 16 tertiary centres in 5 European countries (2016-2021; NCT03313687). Dosage was quantified by inotropic scores (epinephrine, norepinephrine, and dobutamine). Associations of baseline and seven-day summarized dosage with intensive care unit (ICU) discharge (mixed-effects logistic regression) and 30-day mortality (Cox regression) were analysed. Potential catecholamine/MAP target ratios for optimized outcomes were assessed in models adjusted for age, sex, pH, lactate and prior resuscitation, stratified by centre.

RESULTS: N = 704 patients: median age 63 years, 74% male, 34% post-resuscitation, median lactate 5.2 mmol/l. Of these, 53% were discharged from ICU, 48% died within 30 days. Higher inotropic scores independently predicted a lower probability of ICU discharge (baseline score: OR 0.78 [95%-CI 0.69-0.88]; summarized score: OR 0.46 [0.38-0.56]; both P < .001) and higher risk of 30-day mortality (baseline score: HR 1.27 [1.15-1.40], summarized score HR 1.83 [1.60-2.09]; both P < .001). A score/MAP ratio <0.403 µg/kg/min/mmHg was associated with higher ICU discharge odds (ceiling effect); a < 0.426 µg/kg/min/mmHg with lower 30-day mortality hazards (no ceiling effect). Lowering catecholamine doses by accepting reduced MAP targets was linked to better outcomes.

CONCLUSION: In HF-CS, higher catecholamine support independently associates with worse outcomes. Accepting lower blood pressure targets to reduce catecholamine dosage may improve outcomes. Validation in randomized controlled trials is urgently needed.

PMID:41771117 | DOI:10.1093/ejhf/xuag009

Eur J Heart Fail. 2026 Jan 12:xuaf025. doi: 10.1093/ejhf/xuaf025. Online ahead of print.

ABSTRACT

AIMS: In patients with heart failure (HF) influenza vaccination has shown beneficial effects in preventing cardiac decompensations. However, no conclusive results have been achieved in the few studies that have evaluated the impact of vaccination during episodes of acute HF (AHF) decompensation. We conducted a systematic review and meta-analysis to determine the possible effects of influenza vaccination on all-cause mortality in patients diagnosed with AHF.

METHODS: PubMed, Medline, Cochrane Central Register of Controlled Trials and Cochrane Database of Systematic Reviews databases were searched for longitudinal studies comparing patients with AHF vaccinated against influenza with unvaccinated patients. The primary outcome selected for meta-analysis was 1-year all-cause mortality, and secondary outcomes consisted of other outcomes reported in at least in two different studies. Statistical heterogeneity was determined by calculating the I² statistic. Individual adjusted results were pooled using a random effects model. Sensitivity analysis was run for the primary outcome by removing each individual study and then re-doing the meta-analysis.

RESULTS: Up to 30 June 2025, five observational cohort studies examining the effect of influenza vaccination on 1-year all-cause mortality in AHF patients had been published. Statistical heterogeneity was low (I2 = 33.7%), meaning that between-study results were consistent. Pooled analysis of confounder-adjusted hazard ratio (HR) for all-cause mortality in vaccinated patients was 0.89 (95% CI 0.83-0.96) compared with unvaccinated patients. All sensitivity analyses rendered very similar results. In-hospital and 90-day mortality were reported in three and two studies and showed similar reductions in risk, with an adjusted odds ratio of 0.85, 95% CI 0.70-1.01, and adjusted HR of 0.86, 95% CI 0.76-0.96; respectively. Isolated data from single studies suggest no effect on hospitalization following discharge after the AHF episode.

CONCLUSIONS: Influenza vaccination is associated with a lower short- and long-term all-cause mortality in patients with decompensated HF; however, as all the studies included in this meta-analysis were observational, these results could be subject to residual confounding and causality cannot be directly inferred from them.

PMID:41771112 | DOI:10.1093/ejhf/xuaf025

Eur J Heart Fail. 2026 Jan 8:xuaf001. doi: 10.1093/ejhf/xuaf001. Online ahead of print.

ABSTRACT

AIMS: To evaluate whether circulating N-terminal pro-B type natriuretic peptide (NT-proBNP) and high-sensitivity cardiac troponin T (hsTnT) can stage hypertensive heart disease (HHD), by assessing their association with adverse cardiac remodelling and cardiovascular outcomes in individuals with essential hypertension.

METHODS AND RESULTS: The REMODEL study prospectively enrolled 1054 asymptomatic individuals with essential hypertension and no prior cardiovascular diseases (59 ± 11 years old; systolic blood pressure 131 ± 14 mmHg; left ventricular ejection fraction 60 ± 7%). All participants underwent cardiovascular magnetic resonance (CMR) and blood sampling for NT-proBNP and hsTnT. The primary outcome was a composite of acute coronary syndromes, heart failure hospitalization, stroke and all-cause mortality. Median follow-up was 53 (23, 72) months. Maximal log-rank statistic identified thresholds of 152 pg/ml for NT-proBNP and 12.7 pg/ml for hsTnT. Individuals with elevations in both biomarkers (high-risk) were older, had the highest 24-h systolic blood pressure and more diabetes mellitus. They showed the most adverse CMR phenotype, with increased myocardial mass, greater diffuse and replacement fibrosis, impaired left ventricular strain and higher left atrial volumes. Event rates differed significantly across biomarker strata (log-rank P < .001). High-risk individuals had the greatest hazard of cardiovascular events [hazard ratio (HR) 17.11; 95% confidence interval (CI) 8.12-36.09), while intermediate-risk individuals showed intermediate risk (HR 3.44; 95% CI 1.71-6.94).

CONCLUSION: NT-proBNP and hsTnT are complementary biomarkers that not only predict cardiovascular outcomes and but also reflect the severity of cardiac remodelling in HHD. Their combined use enables effective staging of disease severity and may support stage-specific management strategies in patients with hypertension.

PMID:41771092 | DOI:10.1093/ejhf/xuaf001