Drug Safety

Objectives. To present DrugSet, a validated R Shiny application supporting the construction medicinal products codelists based on the Anatomical Therapeutic Chemical (ATC) system and their mapping to Clinical Practice Research Datalink (CPRD) Aurum prodcodes within a single interactive workflow. Materials and Methods. DrugSet comprises four modules: data preparation, ATC-based hierarchical code selection, string-based CPRD Aurum prodcodes mapping, and codelist export. Validation was conducted against World Health Organization (WHO) ATC reference codelists and manually curated prodcodes mappings across three drug classes: metformin, beta-blocking agents, and topical salicylic acid. Sensitivity, specificity, and Positive Predictive Values (PPV) were calculated for ATC codelist generation. Agreement proportions (overlapping against total identified codes) were calculated for prodcodes mapping. Time needed for codelist construction using DrugSet was recorded and compared to manual approaches. Results. DrugSet ATC codelist generation against WHO manual reference achieved 100% sensitivity, specificity, and PPV across all medicinal products. Prodcodes mapping agreement ranged from 89.2% to 98.3% with discrepancies due to missing data in the prodcodes input vocabulary. DrugSet completed codelist construction in 9 minutes compared to 3 hours and 10 minutes manually, across all medicinal products classes. Discussion. DrugSet provides a unified workflow that runs directly on ATC and source CPRD Aurum vocabulary files. The reduction in codelist construction time and export of the generated codelists supports reproducibility in pharmacoepidemiologic studies where codelist creation can represent a significant proportion of study setup time. Conclusion. DrugSet is an open-source, validated tool that improves accuracy, and efficiency of codelist construction for medicinal products based on ATC codes towards CPRD Aurum prodcodes.

Introduction: Adverse drug withdrawal events (ADWEs) are a key safety concern with deprescribing but are infrequently reported in trials. Although pharmacovigilance systems have advanced our understanding of medication-related harms, it is unclear how extensively these systems have been used for ADWEs. Objectives: To examine the reporting patterns of ADWEs for all drugs recorded in United States and European pharmacovigilance databases between 2004 and 2023. Methods: A retrospective study was conducted using two pharmacovigilance databases, the publicly available FDA-FAERS dataset and EMA-EV Level 2A (individual-level) dataset. ADWE cases were identified using relevant MedDRA preferred terms. Data on patient characteristics, reporter type, drugs, indication, ADWE outcomes, dechallenge/rechallenge, seriousness criteria, time to onset, duration, and causality were summarised. Results: A total of 158,505 ADWE reports were analysed (FDA-FAERS: 145,514; EMA-EV: 12,987), with mean ages of 46.1 (FDA; 55.3% female) and 45.5 years (EMA; 57.1% female). The frequently reported drug classes were opioids (FDA: oxycodone, 29.8%; EMA: buprenorphine, 19%), antidepressants (FDA: duloxetine, 32%; EMA: venlafaxine, 25.9%) and gabapentinoids (FDA: pregabalin, 6.7%; EMA: pregabalin, 6.0%). The most common adverse outcomes were other serious medical conditions (FDA=63.9%; EMA=46.0%), hospitalisation (FDA=15.9%; EMA=28.3%), and disability (FDA=13.3%; EMA=6.2%) and these outcomes varied significantly based on sex and age group (p<0.05). Conclusions: This study provides novel evidence of reporting patterns and characteristics of ADWEs across drugs in pharmacovigilance data. These findings emphasise that adverse drug reaction reporting systems need to accommodate ADWEs (i.e., clarity on terminologies, dechallenge/rechallenge, causality assessment) to effectively capture ADWE-related data to support evidence-based deprescribing practices for better patient safety

BackgroundPost-market pharmacovigilance relies predominantly on single-database disproportionality analysis of spontaneous adverse event reports, which lacks corroboration across independent evidence streams and cannot integrate randomised trial evidence. No publicly accessible platform has previously combined European national pharmacovigilance registries, the US FDA Adverse Event Reporting System (FAERS), and clinical trial meta-analyses into a unified, continuously scored signal detection framework. MethodsWe describe the Signal Consensus Index (SCI), a composite 0-100 pharmacovigilance signal score integrating disproportionality evidence from the Danish National Pharmacovigilance Database, the UK MHRA Yellow Card scheme, and FAERS, with DerSimonian-Laird meta-analytic risk ratios from ClinicalTrials.gov, across 6,905,874 drug-adverse event pairs. Each source contributes a continuous score derived from the lower bounds of three complementary disproportionality metrics (ROR, PRR, IC025) for spontaneous reporting sources, and from the pooled risk ratio lower confidence bound for clinical trials. The SCI is publicly accessible via the Adverse Event Atlas (aeatlas.com). We report reference set validation against the EU-ADR reference standard, a single-source comparison with discordance characterisation, temporal stability analysis across eight cumulative data windows (2015-2023), and a weight sensitivity analysis across seven pre-specified weighting schemes. ResultsThe SCI generated 129,176 Moderate-or-Strong signals (SCI [&ge;] 50, confidence [&ge;] 50) and 7,290 Strong signals (SCI [&ge;] 70, confidence [&ge;] 70). Reference set validation against 88 classifiable drug-event pairs (44 positive controls, 44 negative controls) yielded 18 true positives, 0 false positives, 44 true negatives, and 26 false negatives (sensitivity 40.9%, specificity 100.0%, PPV 100.0%, NPV 62.9%). Zero false positives were observed across all 44 classifiable negative controls, with five false negatives attributable to the confidence gate correctly suppressing single-source signals pending multi-source corroboration. Single-source comparison demonstrated that FAERS alone generated 1,438,246 disproportionality signals, of which 94.8% were not confirmed by the SCI, while 54,184 SCI-detected signals were absent from FAERS, of which 8.3% involved drugs absent from the US reporting system. Discordance analysis showed that 99.8% of Danish non-confirmation reflected data availability constraints. Temporal stability was high: 98.5% of pairs received identical classifications across all seven weight scenarios, and 57.0% of final Strong signals were already detectable as Moderate or Strong in the earliest data window (2015-2016). Strong classifications were stable across weight scenarios (94.0% of Strong observations remaining Strong). ConclusionsThe SCI provides a transparent, openly accessible framework for cross-source pharmacovigilance signal prioritisation with 100% specificity and PPV against an established reference standard and stable classifications across weighting schemes. Progressive signal emergence through the Moderate tier supports its use as an early detection layer. The platform is available at aeatlas.com.

Background: Retrieval of adverse event reports based on coded drug-event co-occurrence enables large-scale pharmacovigilance analyses, but yields candidate reports rather than validated cases, risking misinterpretation if used alone. Aim: To develop and apply a framework for identification and characterization of clinically meaningful case series in pharmacovigilance. Methods: We conducted two case studies. The first developed and refined the framework in an information-rich setting, focusing on drug-induced impulsivity across selected drugs; the second tested its applicability in a more routine, information-poor setting, focusing on drug-induced suicidality. Results: In Case 1, non-relevant reports were frequent for drugs with uncertain evidence and negative controls ({approx}20-40%) compared to drugs with established causal roles (4%). The emerging framework assessed relevance based on exposure, event, drug-event relationship, and population. For suspected adverse drug reactions, relevant reports were further characterized by reporter suspicion and evidentiary qualifiers supporting or refuting causality; higher suspicion was associated with more supportive qualifiers. Applied to Case 2, the framework ruled out 69% of reports as non-relevant but highlighted substantial non-assessability (17%). Conclusions: In pharmacovigilance, retrieval is not equivalent to case identification. Relevance is question-specific and shaped by how reports are captured, processed, and retrieved. This can be especially critical for emerging or bias-prone safety questions. Transparent and reproducible case definition and adjudication are essential for interpretable analyses.

Background: This study aimed to evaluate real-world adverse event (AE) signals of EV to provide evidence-based guidance for its safe clinical application. Methods: Data from the FDA Adverse Event Reporting System (FAERS) database from the period of 2019 Q1-2025 Q3 were analyzed. Disproportionality analysis algorithms, including the reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network (BCPNN), and empirical Bayes geometric mean (EBGM), were utilized to mine safety signals.The time to onset (TTO) was evaluated using the Weibull distribution model. Results: Among 11,697,906 reports, 4,177 EV-treated patients experienced 14,511 AEs. The most common System Organ Classes (SOCs) were skin and subcutaneous tissue disorders (18.23%), general disorders and administration site conditions (13.17%).Multi-algorithm consensus identified 179 positive signals. Alongside known toxicities (rash, peripheral neuropathy, hyperglycemia), potential new signals emerged, including dysgeusia, atypical skin lesions, and myelosuppression. Median TTO was 14 days, with the Weibull {beta} of 0.736, confirming an "early failure" profile. Subgroup analysis revealed toxicity heterogeneity: patients aged [&ge;]65 and females exhibited stronger signals for fatal severe cutaneous adverse reactions, while patients aged < 65 and males showed higher susceptibility to neurological and metabolic toxicities. Conclusions: The real-world safety profile of EV confirms known toxicities, reveals new risks (e.g., dysgeusia), and shows toxicity concentrated in the first treatment cycle. Clinical practice requires proactive monitoring during the first two weeks using demographic-specific strategies: vigilance for fatal skin toxicity in elderly and female patients, and close follow-up of neurological and metabolic indicators in younger and male populations.

Background: Dihydropyridine calcium channel blockers (DHP-CCB) are widely prescribed antihypertensives whose adverse effects may trigger unnecessary prescribing of additional medications, termed prescribing cascades (PC). We aimed to identify potential DHP-CCB-induced PCs using high-throughput sequence symmetry analysis (HTSSA). Methods: Using Medicare claims data (2011-2020), we identified new users aged [&ge;]66 years with continuous enrollment [&ge;]360 days before and [&ge;]180 days after DHP-CCB initiation. We screened for initiation of 446 "marker" drug classes within {+/-}90 days of DHP-CCB initiation. Sequence ratios compared marker drug initiation after versus before DHP-CCB initiation. Adjusted sequence ratios (aSR), accounting for prescribing trends over time, were calculated with 95% CIs >1 considered statistically significant. Clinical experts classified statistically significant signals as potential PCs through consensus. Results: Among 388,862 DHP-CCB initiators (mean age 76.6 {+/-} 7.5 years; 62.5% women, 92.3% with hypertension), 82 of 446 marker drug classes had significantly elevated aSRs, of which 24 were classified as potential PCs. Strongest signals ranked by highest aSR included other systemic hemostatics (aSR 2.99; 95% CI, 1.10-8.16), other nasal preparations (aSR 1.99; 95% CI, 1.47-2.70), and drugs used in erectile dysfunction (aSR 1.85; 95% CI, 1.27-2.70). Other clinically relevant signals, ranked by number needed to harm (lowest to highest), included sulfonamides (NNTH 104; 95% CI, 98-111), electrolyte solutions (NNTH 216; 95% CI, 196-241), and osmotically acting laxatives (NNTH 710; 95% CI, 540-1056). Conclusion: Potential PCs identified in this Medicare cohort reflected known and underrecognized adverse effects of DHP-CCBs. Further studies are needed to evaluate the clinical consequences of these PCs.

In the context of global health, the ability of frontline primary health providers to identify potential Drug-Drug Interactions (DDIs) is a critical component of patient safety. This is particularly true in settings like Tanzania, where drug dispensers often serve as the primary point of contact for patients. In this study, we establish a baseline for drug decision-making capabilities across multiple cadres of healthcare providers in Kibaha, Tanzania. We specifically distinguish between the ability to recognize safe drug combinations versus harmful ones. The findings reveal a critical asymmetry in provider performance: while professional training improves the recognition of safe combinations, it provides no advantage over lay intuition (and in some cases, a significant disadvantage) in detecting potentially harmful interactions.

Background Large Language Models (LLMs) are increasingly explored for pharmacovigilance tasks, including information extraction, case documentation, and single-case causality assessment. However, their ability to support causality assessment at the case series level -- a complex, time-intensive task requiring clinical reasoning across multiple reports -- remains unexplored. Objective To investigate how a large-scale general-purpose LLM can support pharmacovigilance professionals in assessing causality in a case series, and to explore how prompt design influences the quality of the model's reasoning. Methods GPT-4o was used to assess causality for five drug - adverse event combinations, using an adaptation of the Bradford Hill viewpoints for case series assessment. The combinations represented varying drugs and vaccines, adverse events, and case series sizes (5-402 reports). One combination served as a negative control. Structured prompts were iteratively developed and refined using one combination, then applied to all combinations. LLM-generated assessments for each viewpoint were qualitatively evaluated by human annotators for accuracy (precision), and the LLM's coverage of key aspects from the original signal text was assessed for one combination (recall). Results Across all five combinations, annotators agreed with 79-92% of the LLM's output sentences. Full disagreement was consistently low (3-7%), with errors typically involving misinterpretation of complex report details rather than outright fabrication. Prompt design substantially influenced output quality; providing Bradford Hill viewpoint descriptions, including case series data, and adding explicit anti-hallucination instructions improved specificity and grounding. For the recall assessment, 15 of 23 key segments from the original signal text were reflected in the LLM output. The overall summary assessments demonstrated balanced reasoning, correctly distinguishing between positive safety signals and the negative control, and provided a coherent synthesis suitable as a starting point for human assessors. Conclusions LLMs have the potential to generate contextually nuanced and largely accurate preliminary causality assessments of case series aligned with the Bradford Hill viewpoints, with a low but non-zero hallucination rate. These findings support LLMs as a tool to augment, not replace, expert judgment in signal assessment. Future work should address larger and more diverse signal sets, improved evaluation frameworks for generative output, and the integration of pre-computed summary statistics to reduce errors.

Abstract Introduction The sharing of individual patient data is essential for advancing pharmacometrics but is strictly limited by privacy regulations (e.g., GDPR). While synthetic data generation offers a legally compliant alternative, its structural impact on complex nonlinear mixed-effects (NLME) modelling remains largely unexplored. This study aimed to benchmark five generative artificial intelligence algorithms by evaluating the balance between data privacy and the preservation of structural PK properties and clinical dosing guidance. Material & methods A daptomycin two-compartment PopPK model was used to simulate a reference cohort of 500 patients. Five generative algorithms (Modified AVATAR, Gaussian Copula, Synthpop, TVAE, and CTGAN) produced 100 independent synthetic datasets each. A two-stage evaluation framework was applied: first, a statistical indistinguishability test based on logistic regression (AUC ROC) was used as a macroscopic pre-selection criterion to determine algorithm eligibility for NLME modelling and privacy risk assessment. Privacy risk was independently quantified using the Anonymeter framework (Singling Out and Linkability attacks). Eligible algorithms were further evaluated on PK parameter recovery bias and clinical dosing simulations. Results Deep learning architectures (TVAE, CTGAN) were excluded at the pre-selection stage due to both biologically implausible covariate generation and high macroscopic detectability (mean AUC ROC = 0.837 and 0.986, respectively). Synthpop, AVATAR, and Gaussian Copula all passed the indistinguishability threshold (AUC ROC = 0.475 +- 0.033, 0.490 +- 0.013, and 0.619 +- 0.031, respectively) and proceeded to NLME evaluation. However, attack-based privacy assessment revealed that Synthpop carried an unacceptable singling-out risk (0.035), disqualifying it from privacy-preserving data sharing. AVATAR and Gaussian Copula demonstrated acceptable privacy profiles (singling-out = 0.004 and 0.001; linkability = 0.010 and 0.003, respectively). At the structural level, Gaussian Copula injected stochastic noise inflating residual error (+157.0%) and V1; (+25.9%), blunting predicted Cmax and predisposing to empirical dose escalation and risk of toxicity. AVATAR acted aSs a smoothing filter, deflating V2; (-48.3%) and underestimating CL (-12.9%). Forward clinical simulations confirmed directionally opposed prediction errors: Gaussian Copula consistently underestimated Cmax across standard and renally impaired profiles (-14.5% and -16.0%, respectively), predisposing to empirical dose escalation, whereas AVATAR- and Synthpop-derived models overestimated Cmax and Cmin in the obese infected patient (+14.7% and +8.2%, respectively), compounding the accumulation risk already present in this profile. Conclusion While no generative algorithm currently offers a perfect solution, AVATAR and Gaussian Copula represent the most viable candidates, being the only methods to satisfy both macroscopic indistinguishability and attack-based privacy criteria. These findings highlight the necessity of a structured, two-stage validation framework and suggest that, when coupled with therapeutic drug monitoring, synthetic datasets could significantly enhance multicentre collaboration while maintaining strict regulatory compliance

Gray-market retatrutide use is increasing, but patient safety experiences remain poorly characterized. This cross-sectional analysis examined Reddit posts and comments from retatrutide-specific and broader peptide or weight-management communities through December 2025. A validated large language model classified self-reported retatrutide use and extracted author-attributed symptoms mapped to MedDRA Preferred Terms. Among 13,589 users reporting current use, 7,823 had at least one mapped symptom after exclusions. Unlike phase 2 trial findings dominated by gastrointestinal events, Reddit reports most often described appetite increase, fatigue, increased energy, nausea, food craving, insomnia, and elevated heart rate. Findings are hypothesis-generating and warrant pharmacovigilance attention.

ObjectivesTo evaluate supply-chain vulnerabilities affecting medications essential for treating sexually transmitted infection in the United States and identify disruption mechanisms that may predispose these therapies to shortages. MethodsWe conducted a qualitative, structured supply-chain vulnerability assessment of first-line medications for five priority sexually transmitted pathogens recommended by the Centers for Disease Control and Prevention and the World Health Organization: azithromycin, doxycycline, ceftriaxone, benzathine penicillin G, metronidazole, tinidazole, acyclovir, and cefixime. Using a predefined framework derived from pharmaceutical supply-chain disruption literature, we evaluated 13 disruption categories spanning raw material sourcing, active pharmaceutical ingredient production, manufacturing, distribution, market dynamics, information systems, and post-distribution loss mechanisms. Each category was assessed using four binary indicators and classified as relevant when at least two criteria were satisfied. ResultsMultiple disruption domains applied across the drug set. Recurrent vulnerabilities included geographically concentrated active pharmaceutical ingredient production, limited manufacturing redundancy in low-margin generic markets, manufacturing constraints affecting sterile injectable products, reliance on consolidated distribution networks, and susceptibility to demand surges and information-system disruptions. All eight drugs exhibited at least one regulatory or market signal consistent with potential supply vulnerability, including documented shortages, product discontinuations, or limited manufacturer participation. ConclusionsSupply-chain vulnerabilities were identified across multiple first-line sexually transmitted infection therapies, indicating that disruption risk is not confined to a single drug. There is a need for policy interventions to strengthen supply-chain resilience, including diversification of active pharmaceutical ingredient sourcing and distribution networks, as well as incentives for sustainable generic production.

Aims and SettingIn the U.S., the emergence of new adulterants and novel psychoactive substances continues to complicate approaches to overdose, treatment, and public safety. Information about this changing drug supply is often gleaned from police drug seizures, but community drug checking services, which test the contents of a persons drug supply and share that data, provide another means to understand local drug supplies. However, it is unclear how seized drugs differ from those collected in the community, whether one approach is potentially more instructive, and what can be learned about local drug supplies from each source. We therefore compared drug samples tested from police departments (PDs) and community partner (CP) drug checking programs to examine what, if any, differences existed in sample content, form, submitter characteristics, and emerging substance presence. DesignWe conducted a retrospective cohort analysis of drug samples collected and tested between April 2018 and December 2025 by the Massachusetts Drug Supply DataStream derived from CPs and PDs operating in the same geographic area across eight locations. Bivariate analyses (Chi-square, Fishers exact) tested for differences in sample and submitter characteristics by source. FindingsThere were 2,430 unique samples submitted by CPs (68.1%) and PDs (31.9%) from the same location. Compared to CP samples, proportionally more PD samples showed fentanyl as primary substance (74.2% PD vs. 64% CP, p<.001) and less often contained additives (xylazine 15.0% PD vs. 27.4% CP; medetomidine 0.6% PD vs. 2.2% CP, both p<.001). PD samples were typically powders (73.2% vs. 37.9%) and pills (13.6% vs. 3.6%) while CP samples were more often residue (51.9% vs. 2.1%, p<.001). Submitter characteristics, when reported, differed by source: gender (n=528, male: 78.6% PD vs. 50.1% CP, p<.001), race/ethnicity (n=468, Black: 15.8% PD vs. 7.8% CP; Hispanic: 6.7% PD vs. 13.2% CP, p<.05), and associated overdose (n=242, fatal: 62.9% vs. 10.9%, p<.001). Emergent substances were detected a median of 249 days sooner in CP than co-located PD samples, though drugs exhibiting concerning patterns (e.g., unexpected fentanyl in stimulants) had similar, swift detection times. ConclusionDrug samples differ based on PD vs. CP source in significant ways that may introduce bias when drawing conclusions about drug supply trends but also offer unique insights for public health and responses to emerging drugs. Modern drug monitoring should include a broad range of sources to best prepare for changes the illicit supply may bring to overdose prevention, public safety, and health systems.

BackgroundPharmacogenomic testing (PGx) can optimise drug efficacy and minimise toxicity, but the extent of prescriber adherence to PGx recommendations remains unclear. We aimed to quantify clinician adherence to international genotype-guided prescribing recommendations in a cohort of paediatric oncology patients. MethodsWe reviewed files of children enrolled in the MARVEL-PIC (NCT05667766) randomised control trial, who had PGx recommendations available. Patients were included if 12 weeks had passed since their PGx report was released to clinicians. Prescribing events were identified for actionable PGx recommendations, and classified as "explicitly followed", "inadvertently followed", or "not followed". Adherence was assessed by patient, drug, and recommendation. Results2,063 PGx recommendations were available for 216 patients. 64 (3.1%) recommendations were actionable for 44 patients and 10 drugs within the 12-week study period. Recommendations were explicitly followed in 57/288 (19.8%) of prescribing events, inadvertently followed in 145 (50.3%), and not followed in 86 (29.9%). Mercaptopurine demonstrated the highest rate of explicit adherence (87.5%). No significant associations were observed between adherence and age group, cancer type, drug type, or strength of recommendation. ConclusionAdherence to pharmacogenomic recommendations was very low, highlighting the need to understand barriers to PGx implementation, and consideration of clinical decision supports to facilitate adherence. Plain Language SummaryPharmacogenomic medicine (PGx) looks at how our genes affect our response to drugs, including their effectiveness and toxicity. Through genetic analysis we can create recommendations for drug dosing, avoidance, and monitoring. The MARVEL-PIC study aims to understand if having PGx recommendations decreases the rate of adverse events in children with cancer. We aimed to understand how often prescribers follow PGx recommendations after they are made available, in the MARVEL-PIC trial. To do this, we reviewed medical records and identified relevant prescribing events. We marked these as "recommendation explicitly followed", "recommendation not followed", or "recommendation inadvertently followed" (where the recommendation was followed, but it wasnt clear if this due to PGx). We found that when recommendations were available, they were only explicitly followed in around 20% of cases. In 50% of cases, they were followed but it was unclear whether this was due to PGx. In the remaining 30%, they were not followed. We also found that alerts on our electronic system were fired in about 80% of events where the recommendation was not followed, but did not change the outcome. These findings show that prescriber adherence to PGx recommendations is low. We need to better understand why this is the case and implement more specific tools to assist prescribers in following recommendations. Article HighlightsO_LIPharmacogenomic (PGx) testing can reduce adverse drug reactions by guiding drug choice, dosing, and monitoring. C_LIO_LI!Prescriber to PGx recommendation adherence has not been widely investigated. C_LIO_LIRetrospective analysis showed that explicit adherence to recommendations occurred in only 19.8% of relevant prescribing events. C_LIO_LIIn 50.1% of prescribing events, recommendations were followed, but there was no clear reference to PGx. C_LIO_LIMercaptopurine had the highest explicit adherence (87.5%) from the drugs analysed. C_LIO_LIThere were no statistically significant associations between adherence and age group, cancer type, drug type, or recommendation strength. C_LIO_LIRecommendations were explicitly followed in 29% of events where an interruptive alert was fired, and inadvertently followed in 8%. C_LIO_LITailored interruptive alerts have been shown to increase adherence in other studies, suggesting that the specific design of interruptive alerts may influence adherence. C_LIO_LIWe concluded that explicit prescriber adherence to PGx recommendations is very low (19.8%), and further research needs to be done to understand barriers to implementation. C_LI

Adverse drug events (ADEs) are a significant source of preventable patient harm, yet many ADE signals remain buried in free-text clinical notes. Clinical notes often describe adverse events (AEs) in relation to drugs in two ways: whether a drug causes the AE (the AE is an ADE) or a drug is given to treat an AE (it is considered the Reason for drug treatment). In the N2C2 2018 benchmark, ADEs and Reasons are annotated as separate entity types, despite often being similar in both wording and clinical meaning. This shared similarity makes them difficult to distinguish during entity extraction, leading to errors in relation classification. Therefore, we propose a two-stage framework that first detects AEs as a unified event category and then classifies drug-event pairs into Drug-ADE, Drug-Reason, or No-Relation. In the end-to-end evaluation on the N2C2 2018 benchmark, our system achieves F1 scores of 0.93 for Drug-ADE and 0.94 for Drug-Reason, improving over previously reported end-to-end benchmarks of 0.48 for Drug-ADE and 0.59 for Drug-Reason. Overall, these results support a more precise task formulation in which AEs are detected broadly first, and the ADE vs Reason distinction is resolved at the relation layer. Furthermore, they motivate the development of AE-focused datasets annotated independently of drug linkage to enable more reliable end-to-end pharmacovigilance systems.

ObjectiveConventional pharmacodynamic (PD) modeling workflows require manual model selection, repeated equation rewriting, and empirical parameter adjustment, resulting in limited automation, high cross-scenario migration costs, and insufficient reproducibility. This study aims to develop PD Union, a unified, automated, and interpretable framework for mechanistic PD modeling. MethodsPD Union is built upon a unified continuous dynamical skeleton that organizes absorption and systemic exposure module, the receptor module, the drug input module, the first delay module, the primary pharmacodynamic function module, the primary pharmacodynamic state module, the downstream pharmacodynamic state module, the second delay module, the feedback module, the circadian modulation module, the biophase module, the direct effect module, the disease state module, the second PD axis first delay module, the second PD axis primary pharmacodynamic function module, the second PD axis primary pharmacodynamic state module, the second PD axis downstream pharmacodynamic state module, the second PD axis second delay module, and the second PD axis feedback module. A machine learning-based structure identification module is incorporated to recognize drug input modes and mechanism labels from population PK/PD time series, followed by constrained population parameter optimization, forming an integrated pipeline of structure identification, candidate generation, and parameter fitting. ResultsValidation was conducted at two levels. In standardized synthetic benchmarking across 14 representative single-endpoint scenarios, the structure identification model achieved an output mode accuracy(NRMSE) of 0.7600 and macro-average F1 of 0.6307; parameter fitting yielded an NRMSE mean of 0.146 and median of 0.117. In the unified reconstruction validation based on 15 population pharmacokinetics/pharmacodynamics (PK/PD) literature data, the mean NRMSE of PDUnion model for PD was 0.261, and the median was 0.228. Among the 15 studies, 14 performed better than the models provided in the original literature. ConclusionsPD Union demonstrates that interpretable mechanistic modularization combined with machine learning-assisted structure identification is feasible for automated PD modeling. The framework provides an executable methodological foundation for unified, reproducible, and extensible mechanistic PD modeling, with potential applicability to multi-endpoint and complex disease-state modeling scenarios.

Background: The impact of the Inflation Reduction Act (IRA) upon late-stage developments has been assumed to be limited. The Congressional Budget Office's IRA analysis excluded post-approval innovation, potentially overlooking substantial economic risks to drug developers and declines in the availability of treatments in areas of high unmet medical need such as oncology. Methods: A total of 1148 secondary trials from 364 FDA-approved medicines, published from 2018 to 2025, were obtained from Biomedtracker and clinicaltrials.gov. Using fractional multinomial logit, we model the share distribution of secondary indication studies across 19 disease groups and assess the change in this distribution post-IRA. We also assessed the number of secondary treatment studies pre- vs. post-IRA using multiple linear regression. Results: After the IRA's introduction, small molecule follow-on studies in oncology exhibited a statistically significant 35% decline (R2 = .48, p < 0.014) and lead indication small molecule oncology approvals exhibited a statistically significant 27% decline (R2 = .70, p < 0.002). We also find a statistically significant 14% decline in the share of orphan oncology studies pre- vs. post-IRA (p<0.001). Research Conclusions: This study's results refute claims that the IRA would have minimal negative effects on patient access or late-stage biopharmaceutical R&D. We hope this study reinvigorates debate about the law's unintended consequences and encourages thoughtful policy solutions, as the IRA manifestly creates disincentives that negatively impact patients seeking needed new medicines, particularly those requiring cures addressing metastatic late-stage cancers.

BackgroundLarge language models (LLMs) are increasingly used in medical education and clinical decision-making, but their reliability in high-risk medication dosing remains unclear. Opioid rotation is a common task requiring precise calculations where errors may result in overdose or inadequate pain relief. MethodsThirteen LLMs were tested using an API-based framework to ensure independent queries across trials. First, fictional clinical scenarios were tested to simulate real-world clinical situations involving opioid rotation; to test the effects of changes in wording, scenarios were revised into 4 "vignettes" showing the same clinical situation. Next, opioid pairs were tested with a random-dose paradigm across a clinically-pertinent range (5-120 mg daily morphine equivalents). LLM outputs were compared with expected values derived from reference standards. Accuracy was assessed using predefined safety thresholds: tight accuracy (0.85-1.15x expected dose) and broad accuracy (0.6-1.7x). We tested models naively and with prompts augmented with reference tables and unit explanations. ResultsNaive models generally exhibited low tight-range accuracy across opioid pairs. For any given opioid pair, each model would consistently produce similar incorrect conversion ratios despite wide variability across opioid pairs and language models. Vignette wording changes accounted for 76% of within-scenario response variance. Reference-based prompt augmentation significantly improved performance, with over half of models achieving high proportions of conversions within tight accuracy for morphine-equivalent conversions. ConclusionsWhile commercial LLMs demonstrated variable accuracy in the native state, prompt augmentation significantly improved their performance.

Background: The global SARS-CoV-2 pandemic disrupted healthcare systems worldwide, raising concerns about its impact on clinical research. Early reports suggested reductions in participant enrollment, interruptions to ongoing trials, and challenges to protocol adherence, yet the magnitude and duration of these operational disruptions remain unclear. Methods: We conducted a registry-based analysis comparing clinical trials during the COVID-19 pandemic (December 2019 to November 2022) with a matched pre-pandemic cohort (December 2016 to November 2019). Studies were included if they reported any modifications to trial status, enrollment, or protocols during the study periods. Key variables included trial stoppage, enrollment changes, and adoption of remote or hybrid procedures. Results: The global SARS-CoV-2 pandemic resulted in widespread disruptions to trial operations with 13,323 clinical trials terminated, suspended or withdrawn over the course of the pandemic, a 38% increase compared to the 9,665 trials that stopped in the 3 years prior to the pandemic. Registries indicated a sharp decline in new participant enrollment across geographic regions and therapeutic areas, with partial recovery in later months. Review findings highlighted barriers including patient inaccessibility, staff redeployment, and supply chain interruptions. Conclusions: The pandemic caused system-wide operational shocks that compromised trial timelines and may have downstream methodological consequences. Recovery in enrollment does not imply restoration of pre-pandemic protocol fidelity or outcome ascertainment. Standardized reporting of disruptions, proactive contingency planning, and resilient trial designs are needed to maintain data integrity during large-scale disruptions and to support reliable evidence generation.

ObjectiveThe aim of this study was to develop and validate an automated, scalable framework to harmonise fragmented UK primary care prescription records into a research-ready dataset by mapping four diverse medical ontologies to a unified, historically comprehensive reference standard. Materials and MethodsWe used raw prescription records for consented participants in the UK Biobank, in which participants are uniquely characterized by multiple data modalities. Primary care data were preprocessed by selecting one drug code if multiple were recorded, cleaning codes to match reference presentations, expanding code granularity based on drug descriptions, and updating outdated codes to a single reference version. Harmonisation entailed mapping British National Formulary (BNF) and Read2 codes to dm+d, the universal NHS standard vocabulary for uniquely identifying and prescribing medicines. Harmonised dm+d records were then homogenised to a single concept granularity, the Virtual Medicinal Product (VMP). We validated our methods by creating medication profiles mapping contemporary drug prescribing patterns in 312 physical and mental health conditions. ResultsWe preprocessed 57,659,844 records (100%) from 221,868 participants (100%). Of those, 48,950 records were dropped due to lack of drug code. 7,357,572 records (13%) used multiple ontologies. Most (76%) records were encoded in BNF and most had the code granularity expanded via the drug description (N=28,034,282; 49%). 41,244,315 records (72%) were harmonised to dm+d and 99.98% of these were converted to VMP as a homogeneous dataset. Across 312 diseases, we identified 23,352 disease-drug associations with 237 medications (represented as BNF subparagraphs) that survived statistical correction of which most resembled drug - indication pairs. ConclusionOur methodology converts highly fragmented and raw prescription records with inconsistent data quality into a streamlined, enriched dataset at a single reference, version, and granularity of information. Harmonised prescription records can be easily utilised by researchers to perform large-scale analyses in research.

Drug overdose deaths in the United States reached record levels during the fentanyl era before recently declining. A plausible hypothesis is that a sudden drop in fentanyl purity beginning in 2023 caused the downturn in overdose mortality. We evaluated this hypothesis by replicating a published analysis with regional overdose data, using models that account for time trends and autocorrelation, and negative control indicators to test for spurious correlation. When fentanyl purity was rising, the national purity series did not track overdose increases in most regions and showed only a modest association in the West. When both purity and mortality later declined, the observed associations were also seen with unrelated macroeconomic indicators that shared the same time pattern. National fentanyl purity alone does not provide a sufficient explanation for recent overdose declines.