Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences

This work focuses on the global and partial identification problem for fractional differential equations. We provide a general numerical procedure based on global and local optimization algorithms with two refinements for biological systems that ensure solution positivity and homogeneous parameter units. The method is applied to a new fractional model of Dengue outbreak called the Fractional Homogeneous Nishiura (FHN) model, calibrated using data of newly infected people in Cape Verde. We show that our identification method yields a better fit between data and model solutions than previous approaches and that our FHN model captures the dynamics of Dengue more closely than existing systems.

Background: Traditional diagnostic models lack explainability, while multimodal language models prone to hallucination remain unsafe for medical education. An interactive, risk-free artificial intelligence framework is required to serve as a reliable clinical mentor for radiology trainees. Methods: We propose a multi-agent architecture decoupling deterministic image analysis from generative consultation. Specialized computer vision models perform anatomical localization and pathological segmentation. These quantitative outputs are synthesized into a structured payload, which grounds a locally hosted large language model (LLaVA 7B) using strict prompt guardrails and prerequisite protocols. Results: The system effectively eliminates visual hallucinations by intercepting unanchored queries. The artificial intelligence tutor successfully contextualizes spatial anomalies and baseline metrics, generating accurate conversational explanations and formally structured radiology reports while strictly enforcing medical safety disclaimers. Discussion and Conclusion: By anchoring language generation exclusively to verified algorithmic realities, this framework transforms opaque diagnostic models into safe, interactive educational simulators. This establishes a highly reliable paradigm for integrating explainable artificial intelligence into medical training.

Artificial intelligence (AI) tools have been rapidly adopted by medical researchers, yet whether early career researchers in low and middle income countries possess the awareness and habits needed to use these tools safely remains poorly documented. This study characterized AI adoption patterns, hallucination awareness, and verification and disclosure practices among early career medical researchers in Pakistan. A cross sectional anonymous online survey was conducted among medical students, house officers, residents, physicians, and faculty involved in research or academic work across Pakistan (May 2026). Descriptive statistics and chi square tests were applied to 373 eligible responses. AI use was near universal (99.7%), with 60.3% using AI tools daily. The most commonly reported tool in this sample was Claude (40.5%), followed by ChatGPT (29.2%) and Perplexity (26.0%), though this ranking likely reflects sampling characteristics. Despite high adoption, 59.2% typically did not verify AI outputs before use, and 40.2% had never heard that AI can generate fabricated scientific references. In behavioral vignettes, 36.5% assumed convincing AI generated references were authentic, and 54.2% would continue using remaining AI content after discovering one fabricated reference. Formal research training was strongly associated with consistent disclosure (51.7% vs. 17.1%; chi square=48.43, p less than 0.001). Role, daily use frequency, and research training were not significantly associated with verification behavior. Early career medical researchers in Pakistan demonstrate high AI adoption alongside incomplete hallucination awareness and infrequent verification, a pattern that may carry implications for research integrity. Formal training was the only factor significantly associated with consistent disclosure. Integration of AI literacy into medical curricula and institutional governance frameworks merits consideration.

Aim: Cross-sectional summaries of periodontitis based on clinical attachment loss (CAL) are, by definition, conditioned on surviving teeth. Because the most severely affected teeth are more likely to have been lost, these measures may underestimate cumulative disease burden and show an artificial flattening (attenuation) of severity with age. We hypothesised that measures more sensitive to severe attachment loss would show greater attenuation at older ages than measures defined across a broader range of sites. Materials and Methods: Using nationally representative data from adults aged 30+ years in NHANES 2009-2014, we examined age-specific trajectories across multiple continuous measures of periodontal severity and assessed whether divergence between measures followed the pattern predicted under severity-dependent tooth loss. Results: The proportion of observable sites declined from 93% at ages 30-34 to 68% at 80+ years, establishing the structural basis for the divergence observed across severity measures. All severity measures showed nonlinear attenuation with age, with distortion increasing with severity threshold. Higher-threshold measures exhibited the greatest attenuation, while lower-threshold measures showed more stable trajectories. Conclusions: Cross-sectional summaries of periodontitis reflect disease among surviving teeth rather than cumulative damage across teeth originally at risk. Attenuation at older ages is consistent with depletion of the most severely affected teeth rather than biological slowing. Distortion varies by measure, with higher-threshold and mean-based indices most affected, whereas the CAL 3+ mm threshold provides a more stable basis for age comparisons.

INTRODUCTION: Alzheimers disease (AD) is a multifactorial disorder, yet current research largely focuses on downstream biomarkers with limited attention to environmental contributors. Experimental studies suggest that per and polyfluoroalkyl substances (PFAS) may contribute to neuroimmune and neurodegenerative pathways relevant to AD. OBJECTIVE: To examine associations between PFAS exposure and neuroimmune and AD related plasma biomarkers in cognitively unimpaired rural adults. METHODS: In a cross sectional pilot study (n=48), serum concentrations of 33 PFAS were measured, including four legacy compounds (PFOS, PFHxS, PFOA, PFNA). Plasma neuroimmune related (ITGB2, SMOC1, TREM2, GFAP) and AD related biomarkers (Ab42/40, ptau217) were detected using proteomic analysis. RESULTS: PFOS showed moderate associations with ITGB2, SMOC1, and Ab42/40 in unadjusted analyses, which attenuated after adjustment for age. PFOA and PFNA demonstrated consistent inverse associations with TREM2 before and after adjustment. DISCUSSION: Findings suggest possible compound specific PFAS associations with immune and amyloid related biomarkers, supporting further investigation in longitudinal and PFAS mixture based studies.

Introduction Air pollution is a leading driver of cardiovascular disease with a growing body of literature implicating this in worse glucose homeostasis. Increases in fine particulate matter air pollution (PM2.5) are associated with increased blood glucose and hemoglobin A1c across the glycemic spectrum from normoglycemia to prediabetes to all forms of diabetes. Despite strong evidence for positive associations of PM2.5 with dysglycemia, it remains unknown if reducing air pollution exposure through air filtration can effect improvements in glucose. This study aims to test the hypothesis that short-term, in-home air pollution reduction using high efficiency particulate air (HEPA) filtration will improve blood sugar in adults with prediabetes. Methods and analysis This trial is a randomized, double-blind, sham-controlled trial of the effects of lowering air pollution exposure using HEPA filtration on cardiometabolic health in adults with prediabetes living in the New York City area. Participants will be randomly assigned to use bedroom air cleaners, or sham air cleaners, while measuring PM2.5 continuously for 1 month. The primary outcomes will be continuous glucose monitoring metrics measured before and after HEPA air filtration. Exploratory outcomes will include insulin resistance measures, serum biomarkers and transcriptomics measured before and after HEPA intervention. We will quantify effects of HEPA filtration with models using treatment arm (true versus sham filtration) as the independent variable. Secondary analyses will model continuous measures of PM2.5 as the independent variable. Ethics and Dissemination This study has undergone peer review; and the work was supported by Grant 2023-0214 from the Doris Duke Foundation, who had no other role in study design or implementation. The study was registered in ClinicalTrials.gov (NCT05994937) prior to recruitment. Clinical Trials Clinical Trials NCT05994937; https://clinicaltrials.gov/study/NCT05994937

Our study evaluated whether a deep learning auto segmentation model combined with machine learning triage can streamline radiotherapy clinical trial quality assurance (QA). We analyzed 107 stereotactic ablative radiotherapy (SABR) cases from a multi-institutional phase II clinical trial of neurovascular sparing prostate SABR, focusing on physician contours of the internal pudendal artery (IPA) as a novel organ-at-risk with substantial interobserver variability. Contours were scored by the trial principal investigator as Per-Protocol or Minor Deviation/Unacceptable. We applied a deep learning model for IPA auto-segmentation. Agreement between human and AI contours was then quantified using 14 overlap, distance, and surface metrics, and a supervised classifier was trained on these metrics to flag clinical trial protocol deviations. While AI segmentation achieved only modest geometric accuracy with mean Dice similarity coefficient of 0.446 and 95th percentile Hausdorff distance of 14.23, when incorporating all 14 metrics, a machine learning classifier yielded AUROC of 0.836, flagging all Minor Deviation/Unacceptable cases with 100% sensitivity on the 27 case hold-out set with 6 false positives and no false negatives. AI segmentation combined with metrics-based machine learning can triage protocol deviations within a multi-institution radiotherapy clinical trial, supporting prospective evaluation of AI-assisted trial QA.

Background: Hospital incident risk scoring has long relied on two- or three-dimensional frameworks (Severity Assessment Codes or Risk Priority Numbers),even though root cause analysis standards recognize that clinical risk is multi-factorial. The obstacle has been mainly cognitive: human reviewers cannotreliably score many dimensions across high incident volumes, so richer assessmenthas not been operationalized at scale.Objective: To extend the traditional three-dimensional FMEA to an eight-dimensional patient-safety risk feature framework, to establish a multi-modellarge language model (LLM) extraction pipeline that scores these dimensionsautomatically, and to demonstrate a variance-aware integer optimization (mean-variance integer programming, MV-IP) that provides a reproducible tie-breakingrule for incident prioritization under extraction uncertainty, rather than improvedrisk coverage.Methods: An 8-dimensional framework covering harm severity, potential harm,frequency, detectability, systemic impact, vulnerable populations, regulatoryrelevance, and economic impact was applied to 213 synthetic and 196 realcurated incident narratives. Three independent LLMs (GPT-5.4, Gemini 3.1 Pro, Grok-4.1 Fast) from different provider families extracted structured risk scores.Inter-model consistency was assessed via ICC(A,1). Among coverage-equivalentselections, MV-IP minimized inter-model variance to give a reproducible prioriti-zation rule. An English-language sensitivity analysis was conducted on 31 AHRQPSNet WebM&M cases.Results: On real cases, seven of eight dimensions reached Fair or betterinter-model reliability (ICC(A,1) 0.53 to 0.83); D5 (Systemic Impact) was theexception at Poor reliability (0.275), driven by little between-case variation ratherthan by wide model disagreement. Reliability was not uniform: two dimensionswere Excellent (D1 actual harm 0.834, D8 economic impact 0.782), two Good,and three only Fair, so some dimensions are more readily extractable than others.The same anchors gave broadly similar results on English-language narratives.When deterministic top-K selection returned several equal-coverage solutions(11 on real cases, total inter-model variance 0.205 to 1.274), MV-IP selected theminimum-disagreement set, replacing ad hoc tie-breaking with an explicit rulewithout improving coverage. Bootstrap resampling found 74% to 90% of per-casevariance estimates stable despite the three-model panel.Conclusions: The eight-dimensional framework operationalizes patient-safetyrisk features that quality teams have considered only implicitly, and three inde-pendent LLM families produced reproducible scores on most dimensions ofcurated narratives. Inter-model agreement, however, measures reproducibilityrather than clinical correctness, and high agreement does not by itself establishthat a score is right; the dimensions that are reliably extractable today (notablyD6 and D8) differ from those that are not yet (D5, and to a lesser degree D4 andD7), which has direct implications for incident-reporting form design. MV-IP con-tributes a reproducible, variance-aware tie-breaking rule rather than improvedcoverage. Validation against expert-prioritized RCA lists and deployment on rawinstitutional incident reports remain the next steps toward clinical use.

Subjective auditory-perceptual evaluation and uninterpretable deep learning models limit the clinical assessment of voice disorders. This study proposes a two-phase zero-shot framework to evaluate voice pathology. First, an Audio Spectrogram Transformer is fine-tuned on the Perceptual Voice Quality Database to generate an acoustic latent space. Second, Orthogonal Procrustes analysis maps these acoustic embeddings directly onto the semantic space of a pre-trained Sentence Transformer. The geometric alignment produced continuous semantic axes that outperformed a supervised machine learning baseline in regressing clinician-rated GRBAS (Grade, Roughness, Breathiness, Asthenia, and Strain) severity scales. Furthermore, these axes correlate with traditional acoustic measures, including Harmonics-to-Noise Ratio and local jitter, while remaining robust when applied to aperiodic signals by not requiring fundamental frequency extraction. Most importantly, the model achieved zero-shot semantic expansion, successfully evaluating voices using an untrained, natural clinical vocabulary beyond the GRBAS scale. External validation on the Voice ICarus Database confirmed cross-corpus stability and demonstrated the capacity for zero-shot differential phenotyping of specific etiologies, such as hypokinetic dysphonia and reflux laryngitis. By bridging acoustic and semantic latent spaces, this framework offers an objective, continuous, and transparent metric for evaluating voice quality using voice descriptive vocabulary.

Introduction Air pollution is the largest environmental risk to human health. Children are disproportionately affected by air pollution and their exposure is amplified during physical activity. Observed concentrations of nitrogen dioxide in 1 in 4 London school playground exceeds the European limit, but the health impacts of air pollution exposure in London school playgrounds remain unexplored. Our study aims to assess and compare the acute changes in lung function and airway inflammation of primary school-aged children exercising in school playgrounds. Methods and analysis 330 children aged 8 to 11 years from ten London schools will be recruited to complete 90 minutes of physical activity and 90 minutes of rest in their school playground in a randomised crossover design. Pre-, post-, and 24-hour post-exposure oscillometry measurements will be performed with airway resistance at 5 Hz (R5) the primary physiological outcome. Nasal lavage samples will be collected pre-exposure and 24-hour post-exposure for analysis of inflammatory, oxidative, and vascular biomarkers, with IL-6 as the primary biological outcome. Mixed-effects regression models will examine associations between estimated pollutant exposures, exercise and physiological responses.

Teledermatology expands access to dermatologic expertise in rural settings, yet diagnostic uncertainty persists in low-resource primary care. This retrospective study evaluated MedGemma-4B-IT, a compact multimodal vision-language model, as adjunctive clinical decision support for challenging diagnostic cases. We analyzed 77 zero-concordance cases (360 clinical photographs) from a Dermatology Extension for Community Healthcare Outcomes (ECHO) tele-mentoring program (2016-2021). Zero-concordance cases showed no overlap between primary clinician provisional diagnosis and dermatologist-confirmed diagnosis. The model was prompted using dermatologist-style format to generate ranked differential diagnoses. Performance was assessed using strict case-level top-k exact-match accuracy and relaxed matching criteria based on fuzzy string similarity. MedGemma achieved 0.0% strict top-1 accuracy, 1.3% top-3 accuracy, 3.9% top-5 accuracy, and 3.9% top-10 accuracy. Relaxed concept-level matching achieved 28.6% top-1, 63.6% top-5, and 67.5% top-10 accuracy. Image-level accuracy was 44.2% (159/360, 95% CI 39.0-49.5%). The model surfaced the correct diagnosis within differential lists in 45.5% of cases despite no exact top-1 matches, suggesting utility for differential expansion rather than definitive diagnosis. Performance varied across diagnostic categories, with highest accuracy in Other categories (54.5%) and lowest in neoplastic conditions (0.0%). Common errors included confusion between inflammatory and other diagnostic groupings. These findings characterize MedGemma performance on real-world teledermatology cases and inform safe, clinician-in-the-loop integration into teledermatology workflows where specialist oversight remains essential.

Background Childhood overweight and obesity represent major public health challenges, shaped by socio-economic and environmental factors. This study investigates the mediating and moderating role of urban environmental exposures in socio-economic disparities in childhood excess weight. Methods Data was drawn from a population-based sample of children (2-9 years) and adolescents (10-17 years) living in Geneva, Switzerland. Parents reported household financial situation and children's height and weight, from which excess weight (i.e. overweight or obesity) was derived. Residential exposures to air pollution (PM2.5, NO2), noise (daytime, nighttime), and neighborhood greenness (green areas, canopy coverage) were estimated based on geocoded residential addresses. The association between household financial situation and excess weight was evaluated, as well as the mediating and moderating roles of urban environmental exposures. Results The analysis included 1006 children and 1154 adolescents. Among children, an average-to-poor household financial situation was associated with higher odds of excess weight in children (adjusted odds ratio [aOR]: 1.79, 95% confidence interval [CI]: 1.13; 2.84). Higher noise exposure was associated with excess weight (daytime: aOR: 1.40, 95% CI: 1.10; 1.77, nighttime: aOR: 1.37, 95% CI: 1.08; 1.74), while the association with PM2.5 appeared stronger among socio-economically disadvantaged children, though the interaction did not reach statistical significance (financial situation x PM2.5 interaction: aOR: 1.59, 95% CI: 0.98; 2.59). No significant associations were observed among adolescents. Conclusion These findings highlight the joint influence of social and environmental inequalities on childhood excess weight and stress the need to address these interconnected determinants to design equitable, targeted public health interventions.

Background: Periodontitis is defined by cumulative, irreversible tissue destruction, yet population-based measurement typically relies on cross-sectional indicators derived from retained teeth. Destruction that occurred earlier in life, particularly disease severe enough to result in tooth loss, is structurally excluded from these measures, potentially leading to systematic underestimation of lifetime periodontal burden. Objective: To develop and evaluate a measurement framework that estimates lifetime periodontal burden from cross-sectional data by explicitly incorporating informative tooth loss under etiological uncertainty. Methods: Data were drawn from 10,324 adults aged [≥]30 years participating in the 20090-2016 National Health and Nutrition Examination Survey (NHANES) who completed full-mouth periodontal examination and glycated hemoglobin (HbA1c) testing. Lifetime periodontal burden was estimated by combining observed clinical attachment loss in retained teeth with probabilistic contributions from missing teeth, using three alternative age-stratified attribution schedules derived from epidemiological studies of periodontal extraction. Performance was compared with conventional measures of periodontal severity and extent using distributional analyses, correlations with HbA1c, discrimination of diabetes status, and relative importance analysis. Age-adjusted models were treated as sensitivity analyses. Results: Estimated lifetime periodontal burden exhibited strong, monotonic age gradients across glycemic categories, in contrast to more attenuated patterns observed for severity and extent. Across attribution schedules, lifetime burden showed stronger correlations with HbA1c ({rho} = 0.30-0.32) than conventional measures. In multivariable models including all indices, lifetime burden retained an independent association with HbA1c, whereas severity and extent contributed little unique information. Discriminative performance for diabetes status was consistently higher for lifetime burden than for conventional measures and remained stable across attribution schedules. Conclusions: Lifetime periodontal burden can be estimated from cross-sectional data by explicitly modelling informative tooth loss rather than restricting measurement to retained teeth. Incorporating historical tissue loss under uncertainty yields a more coherent representation of cumulative periodontal destruction than snapshot-based measures and provides a methodological basis for life-course-oriented periodontal epidemiology.

Healthcare waste (HCW) management is a critical determinant of occupational safety, infection control, and environmental protection, particularly in low- and middle-income settings. Using the knowledge-attitude-practice (KAP) framework, this study assessed cognitive, behavioral, and institutional dimensions of HCW management among healthcare workers in urban Bangladesh. A cross-sectional survey was conducted among 342 cleaners and nurses in hospitals in the Chattogram Metropolitan Area (CMA) and Cumilla City Corporation (CuCC). Marked disparities were observed across professional groups. Training coverage was significantly lower among nurses than cleaners in CMA (22.5% vs. 48.7%; p = 0.002), whereas in CuCC nurses showed higher coverage (69.0% vs. 52.3%; p < 0.01). Knowledge of color-coded waste segregation was generally inadequate, with only 39.3% of CMA cleaners correctly identifying pharmaceutical waste bins compared with 60.0% of nurses (p < 0.01); CuCC nurses demonstrated substantially higher awareness (82.8%). Attitudinal indicators favored nurses, with strong hygiene and environmental risk awareness (95-100%) compared with cleaners (66-87.3%; p < 0.001). Despite this, compliance with segregation practices remained low across both sites (<30%). Several institutional support indicators were more favorable among nurses, particularly in CuCC. These findings indicate a significant knowledge-practice gap, emphasizing that effective HCW management requires not only training but also strengthened institutional structures and enforcement mechanisms to reduce public health and environmental risks.

Background: Patients with CKD and polypharmacy face high rates of drug-related problems, yet comprehensive medication review remains time-intensive and inconsistently performed. Large language models (LLMs) may augment this process, but existing benchmarks use multiple-choice formats that do not reflect open-ended, nephrology-specific review. We developed a trap-embedded synthetic CKD benchmark and evaluated five current-generation LLMs (GPT-5.4, Claude Sonnet 4.6, Gemini 3.1 Pro, Grok 4.1 Fast, DeepSeek R1; tested April-May 2026) for open-ended medication review. Methods: Fifty synthetic CKD cases across three complexity groups (G3a-G3b [n=20], G4 [n=15], G5/G5D/transplant [n=15]) with 8-12 medications and [&ge;]2 embedded clinical traps each were scored against nephrologist-adjudicated gold standards. Each model produced three independent responses per case (temperature 0; 750 total outputs). Primary endpoint was per-case macro F1; secondary endpoints were safety-critical omission rate, PI-adjudicated hallucination rate, and intra-model consistency. Blinded inter-rater reliability for gold-standard item detection was assessed on a 30% sample. Results: Consensus-level macro F1 ranged from 0.41 (Claude Sonnet 4.6) to 0.49 (Grok 4.1 Fast) (Friedman P < 0.001). Phosphate binder timing (11%) and hyperkalemia combinations (33%) were poorly detected across all models. Safety-critical omission rate ranged from 22% to 48% (P < 0.001); PI-adjudicated hallucination ranged from 0% (GPT-5.4) to 54% (DeepSeek R1), including fabricated dose caps and non-existent guideline citations. Blinded reliability for gold-standard item detection was high (kappa = 0.934, n = 92). Conclusions: This nephrology-specific benchmark exposes clinically important LLM blind spots that generic multiple-choice evaluations would not detect. Heterogeneous hallucination and omission rates indicate that model selection and domain-specific guardrails should precede any clinical deployment of LLM-assisted CKD medication review. Prospective validation with real patient data and human comparators is required before deployment recommendations can be made.

Rationale: Efficacious dose selection for anti-tuberculosis drugs has traditionally relied on achieving plasma exposures above the minimum inhibitory concentration, but this approach has not consistently aligned with clinical outcomes. Objectives: We sought to identify early pharmacokinetic-pharmacodynamic targets most predictive of clinical efficacious dose. Methods: We conducted a back-translational, pharmacokinetic-pharmacodynamic simulation-based analysis of 15 anti-tuberculosis drugs. Using pharmacokinetic data from multiple biological matrices and a range of pharmacodynamic metrics, we established candidate exposure-response targets for attainment. We systematically evaluated the predictive accuracy of each target pair against established clinical doses to formulate a decision-making framework linking key drug properties to the most predictive targets. Measurements and Main Results: Depending on the target used, projected clinical doses varied widely - both within and across compounds - highlighting the importance of target selection for dose projection and go/no-go decisions. In general, targeting cellular lesion-level drug exposures relative to in vivo preclinical potency provided an effective approach for early dose selection. However, for highly penetrating drugs, targeting site-of-action therapeutic exposures in the caseum was more predictive of clinical dose. Based on these findings, we developed a preliminary dose prediction tool that enables drug developers to estimate clinically relevant dose ranges of compounds using in vitro and early in vivo data. Conclusions: This work establishes and validates a simple, evidence-based framework to standardize early translational decision-making on dose selection of anti-tuberculosis candidates in development.

Pharmacogenomics (PGx) offers a pathway towards personalised medicine, which relies on health consumer involvement in making informed decisions. As consumers increasingly seek health information online, high-quality digital resources are essential to support informed consent and shared decision making. The complexity of PGx and widespread limitations in health literacy raise concerns about whether existing consumer-facing online PGx resources are understandable and sufficiently comprehensive. This study evaluates the readability, visual design, and informational quality of publicly available online written PGx health information. Twenty-three webpages met inclusion criteria. The mean readability corresponded to approximately 15 years of formal education (university level), substantially exceeding the Australian Government's recommended Year 7 reading level for public health materials. Informational quality was generally low, with most webpages being rated as poor or very poor. In contrast, visual design quality was relatively strong, with webpages achieving on average around three-quarters of the criteria. Although the visual presentation of PGx webpages is generally professional, their high reading difficulty and limited discussion of treatment choices and uncertainties reduce their usefulness for health consumer education. Improving readability, clearly communicating risks and limitations, and incorporating decision-support features may enhance the ability of online resources to support informed consent and shared decision making.

Objective: Stigmatizing language in the electronic health record (EHR) has been associated with adverse patient experience in substance use disorder care, including opioid use disorder (OUD). This study evaluated a privacy-preserving, locally-deployed large language model as a method to detect stigmatizing language documentation in OUD patients with patient-directed discharge (PDD). Methods: A retrospective cohort study of 477 inpatient admissions from the MIMIC-IV database with a diagnosis of opioid use disorder were classified using a locally deployed Gemma-4-31b-it-bf16 model and predefined 140 term lexicon to identify stigmatizing language in clinical documentation. Results: Analysis of clinical documentation showed stigmatizing language was present in 84.1% (190/226) in the PDD cohort vs 62.2% (156/251) in the non-PDD cohort, with an unadjusted odds ratio of 3.21 (95% CI 2.07-4.98; p < 0.0001). After adjustment for age, sex, insurance status, marital status, and race, PDD discharge remained an independent predictor of stigmatizing documentation (aOR 2.24, 95% CI 1.40-3.59; p < 0.0001). Further analysis of stigma intensity showed higher stigmatizing markers in the PDD cohort vs the non-PDD cohort (2.85 {+/-} 2.39 vs 2.02 {+/-} 2.44; p < 0.0001). Discussion and Conclusion: Stigmatizing language is detected with increased frequency and prevalence in clinical documentation of OUD patients that initiate PDD compared to those that adhere to standard discharge processes. A locally deployed large language model (LLM) offers a scalable, privacy-preserving method to audit clinical documentation for stigmatizing language.

The availability of electronic health records has facilitated data-driven approaches to the understanding of multimorbidity, with clustering becoming a common tool for uncovering relevant groups of associated conditions. Previous studies, however, have found challenges in their reproducibility, with wide disparity in the reported clusters. At the core of this issue lays a vagueness of the definition of a cluster, leading to a lack of standards in their methods and evaluation, while implementation details are often not completely reported or explicit in their assumptions. We present a methodological pipeline that can be adapted to different cluster definitions (e.g. multiple cluster membership or clusters where all nodes are mutually associated) and a set of scores that can be composed into an evaluation metric that explicitly incorporates assumptions that align with the research aims. We apply our pipeline to a healthcare dataset of over 7 million patients in England and show how clusters may drastically differ when varying the parameter choices, exposing the risks of reporting a single clustering realisation. Our methodological pipeline, evaluation framework, and tools for analysis and network visualisation serve as a reference to transparently explore and align methodological decisions to the aims of multimorbidity clustering, contributing to overcome the reproducibility challenges of the field.

Background: Frontotemporal dementia (FTD) lacks widely accessible disease-specific biomarkers. Optical coherence tomography (OCT) and OCT angiography (OCTA) may provide non-invasive measures of retinal changes associated with neurodegeneration. We conducted a systematic review and meta-analysis evaluating retinal biomarkers in FTD compared with Alzheimer disease (AD) and controls. Methods: A systematic search of PubMed and Embase was conducted through April 25, 2026 according to PRISMA guidelines. Studies evaluating OCT/OCTA biomarkers in FTD with comparator groups were included. Inverse weighted random-effects models, publication bias assessments, and meta-regressions were performed. Results: Ten studies involving 139 individuals with FTD, 87 with AD, 29 with mild cognitive impairment, 14 with TDP-43 proteinopathy, 5 with tauopathy, and 255 controls were included in the systematic review; five studies were eligible for meta-analysis. Compared with AD, individuals with FTD demonstrated significantly thinner retinal nerve fiber layer (RNFL) thickness (SMD = -0.61, 95% CI -0.98, -0.24). Compared with controls, individuals with FTD exhibited significantly thinner ganglion cell layer-inner plexiform layer (GCL-IPL) thickness (SMD = -0.55, 95% CI -1.02, -0.08), whereas pooled analyses across multiple retinal biomarkers were non-significant (SMD = -0.19, 95% CI -0.52, 0.14). RNFL thickness correlated negatively with female % in FTD and positively with age in both AD and controls. Conclusions: Individuals with FTD exhibit lower RNFL thickness than AD and lower GCL-IPL thickness than controls, suggesting retinal alterations may reflect neurodegeneration. However, larger longitudinal studies with standardized OCT/OCTA protocols are needed to determine the diagnostic and prognostic utility of retinal biomarkers in FTD