Drug Safety最新文献

{"title":"Authors' response to Tran et al.'s comment on \"The Use of Multiple Medications During Pregnancy Among an Ethnically Diverse Population in South-Eastern Melbourne: A Retrospective Analysis to Explore Potential Risks and Complications\".","authors":"Yitayeh Belsti, Aya Mousa, Hannah Jackson, Lisa J Moran, Kirsten R Palmer, Raja Ram Dhungana, Emily Callander, Daniel Lorber Rolnik, Helena Teede, Joanne Enticott","doi":"10.1007/s40264-025-01534-9","DOIUrl":"10.1007/s40264-025-01534-9","url":null,"abstract":"","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":"439-441"},"PeriodicalIF":4.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143540561","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Trends in Proton Pump Inhibitor Use in Sweden by Sex and Age: A Drug Utilisation Study.","authors":"Nele Brusselaers, Unnur Gudnadottir, Lars Engstrand, Helene E Lilja","doi":"10.1007/s40264-024-01502-9","DOIUrl":"10.1007/s40264-024-01502-9","url":null,"abstract":"<p><strong>Background: </strong>Proton pump inhibitors (PPIs) are among the most popular drugs worldwide. Yet, there are concerns on long-term safety and poor adherence to prescription guidelines. Off-label use in children and increasing maintenance use in older adults may be particularly worrisome.</p><p><strong>Objectives: </strong>To assess differences in PPI use by age, sex calendar year and PPI type, and to explore potential underlying indications (ulcerogenic drugs, and indications) in Sweden.</p><p><strong>Methods: </strong>Proton pump inhibitor drug utilisation study based on the Swedish nationwide prescribed drug (2006-2023) and patient registries (2006-2022).</p><p><strong>Results: </strong>Proton pump inhibitors were used by 14.4% (women) and 10.5% (men) of adults; and 1.0-1.5% of children and adolescents (aged < 20 years). Proton pump inhibitor use was higher in women in all age-groups except small children (aged < 5 years). Proton pump inhibitor use has increased in all age groups, especially in young children (aged < 10 years) and the oldest groups (aged > 65 years). Proton pump inhibitor users aged > 85 years filled most prescriptions with an annual average of 9.5 (men), 11.6 (women) prescriptions. Most prescriptions were for omeprazole and esomeprazole: 63.7% and 23.5% in adults; 23.5% and 44.7% in children (2023). Prescriptions for other drugs for peptic ulcers/reflux became rare, with 99% of prescriptions in this category being PPIs by 2023. Gastro-intestinal diagnoses were predominantly recorded in men, became less prevalent and only explained part of PPI use, while ulcerogenic drugs were common (particularly in women), suggesting PPIs are regularly used for gastroprotection.</p><p><strong>Conclusion: </strong>Proton pump inhibitor use has doubled in children and increased 50% in adults over the study period, in both sexes, while recorded gastrointestinal indications decreased. Alternative therapies were rarely prescribed in Sweden.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":"389-400"},"PeriodicalIF":4.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11903566/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142791426","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"OpenPVSignal Knowledge Graph: Pharmacovigilance Signal Reports in a Computationally Exploitable FAIR Representation.","authors":"Achilleas Chytas, George Gavriilides, Anargyros Kapetanakis, Alix de Langlais, Marie-Christine Jaulent, Pantelis Natsiavas","doi":"10.1007/s40264-024-01503-8","DOIUrl":"10.1007/s40264-024-01503-8","url":null,"abstract":"<p><strong>Introduction: </strong>Pharmacovigilance signal report (PVSR) documents contain valuable condensed information published by drug monitoring organizations, typically in a free-text format. They provide initial insights into potential links between drugs and harmful effects. Still, their unstructured format prevents this valuable information from being integrated into data-processing pipelines (e.g., to support either the investigation of drug safety signals or decision-making in the clinical context).</p><p><strong>Objective: </strong>OpenPVSignal is a data model designed specifically to publish PVSRs via a computationally exploitable format, compliant with the FAIR (Findable, Accessible, Interoperable, Reusable) principles to promote ease of access and reusability of these valuable data.</p><p><strong>Methods: </strong>This paper outlines the procedure for converting pharmacovigilance signals published by the World Health Organization Uppsala Monitoring Centre (WHO-UMC) into the OpenPVSignal data model, resulting in a Knowledge Graph (KG). It details each step of the process, including the technical validation by KG engineers and the qualitative verification by medical and pharmacovigilance experts, leading to the finalized KG.</p><p><strong>Results: </strong>A total of 101 PVSRs from 2011 to 2019 were incorporated into the openly available KG.</p><p><strong>Conclusion: </strong>The presented KG could be useful in various data-processing pipelines, including systems that support drug safety activities.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":"425-436"},"PeriodicalIF":4.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11903642/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143373834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Major Adverse Cardiovascular Events Related to JAK Inhibitors: A Disproportionality Analysis Using the WHO Global Individual Case Safety Database.","authors":"Raffaella Di Napoli, Christophe Richez, Cristina Scavone, Allison Singier, Maxime Demourgues, Annamaria Mascolo, Annalisa Capuano, Francesco Salvo","doi":"10.1007/s40264-025-01535-8","DOIUrl":"10.1007/s40264-025-01535-8","url":null,"abstract":"<p><strong>Background: </strong>Rheumatoid arthritis (RA) is commonly treated with Janus kinase inhibitors (JAKis) and anti-tumor necrosis factor-α (anti-TNFα), but the cardiovascular safety profiles of these drugs remain unclear.</p><p><strong>Objective: </strong>The aim of this study was to describe the individual case safety reports of major adverse cardiac events (MACE) or stroke and to determine whether there was a difference in the frequency of reporting of cardiovascular events between JAKis and anti-TNFα used in RA.</p><p><strong>Methods: </strong>A case/non-case study was conducted using the WHO VigiBase^® database. Descriptive analysis was performed, the time to onset (TTO) of MACE was calculated, and the reporting odds ratio (ROR) was used to estimate the frequency of MACE reports associated with JAKis versus anti-TNFα in RA.</p><p><strong>Results: </strong>A total of 18,099 cases of MACE were identified, of which 2543 (14%) were associated with JAKis, predominantly in women (65.4%) and in patients aged ≥65 years (49.9%). The median time to onset was 210 days (IQR 60-510) for JAKis and 690 days (210-1460) for anti-TNFα. JAKis were associated with higher odds of reporting MACE (ROR 1.38 [95% CI 1.32-1.44]), mainly due to non-fatal stroke (1.65 [1.55-1.75]). Stroke as a whole showed similar results (1.62 [1.53-1.72]). The ROR of MACE was also slightly increased in patients aged <65 years treated with JAKis (1.29 [1.21-1.39]).</p><p><strong>Conclusions: </strong>Compared with anti-TNFα, JAKis were more associated with MACE, especially stroke, and with a shorter time to onset. These data support the hypothesis of a different cardiovascular reporting frequency between JAKis and anti-TNFα. In patients with identified cardiovascular risk, anti-TNFα should be preferred to JAKis until more definitive results are available.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143691565","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mitigation and Management of Common Toxicities Associated with the Administration of CAR-T Therapies in Oncology Patients.","authors":"Jonathan Renninger, Lisa Kurz, Heather Stein","doi":"10.1007/s40264-025-01538-5","DOIUrl":"https://doi.org/10.1007/s40264-025-01538-5","url":null,"abstract":"<p><p>Chimeric antigen receptor T-cell (CAR-T) therapies are one of the main approaches among targeted cellular therapies. Despite the potential benefit and durable responses observed in some patients receiving CAR-T therapies, serious and potentially fatal toxicities remain a major challenge. The most common CAR-T-associated toxicities include cytokine release syndrome (CRS), neurotoxicity, cytopenias, and infections. While CRS and neurotoxicity are generally managed with tocilizumab and corticosteroids, respectively, high-grade toxicities can be life-threatening. Close postinfusion monitoring and assessment of clinical laboratory parameters, patient-related and clinical risk factors (e.g., age, tumor burden, comorbidities, baseline laboratory parameters, and underlying abnormalities), and therapy-related risk factors (e.g., CAR-T type, dose, and CAR-T-induced toxicity) are effective strategies to mitigate the toxicities. Clinical laboratory parameters, including various cytokines, have been identified for CRS (interleukin [IL]-1, IL-2, IL-5, IL-6, IL-8, IL-10, C-reactive protein [CRP], interferon [IFN]-γ, ferritin, granulocyte-macrophage colony-stimulating factor [GM-CSF], and monocyte chemoattractant protein-1), neurotoxicity (IL-1, IL-2, IL-6, IL-15, tumor necrosis factor [TNF]-α, GM-CSF, and IFN-γ), cytopenias (IL-2, IL-4, IL-6, IL-10, IFN-γ, ferritin, and CRP), and infections (IL-8, IL-1β, CRP, IFN-γ, and procalcitonin). CAR-T-associated toxicities can be monitored and treated to mitigate the risk to patients. Assessment of alterations in clinical laboratory parameter values that are correlated with CAR-T-associated toxicities may predict development and/or severity of a given toxicity, which can improve patient management strategies and ultimately enable the patients to better tolerate these therapies.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143662865","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Pregnancy and Infant Outcomes in Women with Multiple Sclerosis Exposed to Glatiramer Acetate Therapy: An Extended 4-Year Safety Update.","authors":"Sigal Kaplan, Andra Ghimpeteanu, Claudia Florentina Dragut","doi":"10.1007/s40264-025-01523-y","DOIUrl":"https://doi.org/10.1007/s40264-025-01523-y","url":null,"abstract":"<p><strong>Background and objectives: </strong>While glatiramer acetate (GA) is generally considered safe during pregnancy and breastfeeding, long-term data, particularly for the 40 mg/mL dose, are limited. Previous research found GA exposure rates and pregnancy outcomes comparable to the general population. This study evaluates pregnancy, fetal, and infant outcomes following maternal exposure to GA 20 and 40 mg/mL to provide a cumulative four-year update.</p><p><strong>Methods: </strong>Post-marketing pregnancy data reported between April 1, 2019 to March 31, 2023 were searched in Teva's Global Safety database and supplemented with 1- and 12-month post-delivery questionnaires. Prospective pregnancy data, collected prior to known pregnancy outcomes or congenital malformations, were used to estimate pregnancy and infant outcomes for GA 20 and 40 mg/mL exposure. Rates of major congenital malformations (MCM) and other pregnancy and infant outcomes were estimated.</p><p><strong>Results: </strong>Among 3514 pregnancies, multiple sclerosis (MS) was the primary indication (62.4%), with most exposure to GA 40 mg/mL (72.2%), in the first trimester (94.9%). Of these, 2455 (69.9%) had known pregnancy outcomes. Of 1211 prospective pregnancies (1239 fetuses) with known outcomes, 1138 (91.8%) resulted in live births. Fetal loss occurred in 101 cases (8.2%), including spontaneous abortion (6.7%), elective termination (0.8%), ectopic pregnancy (0.3%), stillbirth (0.2%), and other (0.2%). The prevalence of MCM was 1.5% overall (95% CI, 0.9-2.4) and 1.9% during organogenesis (95% CI, 1.1-3.1), comparable to background rates. Minor congenital malformations were less frequent (0.7%). Prospective pregnancies with completed questionnaires (n = 539) reported preterm birth (9.8%), low/very low birth weight (7.3%), neonatal intensive care unit (NICU) admission (8.8%), and adverse events (17.4%). Infant growth remained within normal ranges. Of 384 women completing the 12-month questionnaire, 146 reported breastfeeding with GA (average 8 months). Among these, 14/125 (11.2%) respondents reported infant hospitalization. Growth parameters for 55 breastfed infants were within normal limits. Overall, pregnancy and infant outcomes were similar across GA doses.</p><p><strong>Discussion: </strong>Despite limitations of post-marketing data, this four-year study found no increased risk of adverse pregnancy, fetal, or infant outcomes associated with GA exposure. The MCM rates aligned with the general population, and infant outcomes during breastfeeding were within normal ranges. These findings support the safety of both 20 and 40 mg/mL GA during pregnancy and breastfeeding.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143662877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"QT Interval, Antipsychotics and Correlates Among Patients with Schizophrenia: Cross-Sectional Data from the Multicentric Real-World FACE-SZ.","authors":"Mona Gerentes, Mohamed Lajnef, Andrei Szöke, Bruno Aouizerate, Fabrice Berna, Maud Cléry, Isabelle Chéreau, Nathalie Coulon, Julia Clauss-Kobayashi, Eric Fakra, Jean-Michel Dorey, Caroline Dubertret, Guillaume Fond, Ophélia Godin, Tudi Goze, Christophe Lançon, Marion Leboyer, Sylvain Leignier, Pierre-Michel Llorca, Jasmina Mallet, David Misdrahi, Nicolas Oriol, Romain Rey, Paul Roux, Benoit Schorr, Mathieu Urbach, Etienne Véry, Franck Schürhoff, Baptiste Pignon","doi":"10.1007/s40264-025-01526-9","DOIUrl":"https://doi.org/10.1007/s40264-025-01526-9","url":null,"abstract":"<p><strong>Background: </strong>The life expectancy of patients with schizophrenia is reduced, partly due to cardiovascular diseases. Antipsychotics are associated with QT interval prolongation, which is a risk factor for arrhythmia and cardiac arrest. The differences between antipsychotic with regard to QT interval prolongation are not well understood.</p><p><strong>Objective: </strong>The aim was to compare the QT values associated with different antipsychotics within a real-world population of subjects with clinically stable forms of schizophrenia.</p><p><strong>Methods: </strong>The FACE-SZ cohort comprises subjects with psychotic disorders, referred to schizophrenia expert cents. QT interval was measured, as well as all treatments (psychotropic and others). The following maintenance treatment for schizophrenia was analysed cross-sectionally: aripiprazole, clozapine, haloperidol, amisulpride, olanzapine, quetiapine, risperidone. Age, sex, smoking status, body mass index, blood potassium levels, and the co-prescription of another QT-prolonging treatment were used as adjustment factors in multivariable linear regression analyses.</p><p><strong>Results: </strong>Among 792 patients, the mean corrected QT (QTc) interval in the sample of patients under monotherapy was 407 ms. The mean age was 31.7 years, and the majority were male (73.3 %). In comparison to the rest of the sample, clozapine was associated with a longer QTc interval (β = 0.012, 95% CI [0.006-0.018]), while aripiprazole was significantly associated with a shorter QTc interval (β = - 0.010, 95% CI [- 0.016 to - 0.005]). Other antipsychotics were not associated with significant variations of the QTc.</p><p><strong>Conclusions: </strong>The prescription of antipsychotics should always be accompanied by close monitoring of the QTc interval to prevent the risk of severe cardiac arrhythmia, particularly concerning clozapine.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143633824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Narrative Search Engine for Case Series Assessment Supported by Artificial Intelligence Query Suggestions.","authors":"Alem Zekarias, Eva-Lisa Meldau, Shachi Bista, Joana Félix China, Lovisa Sandberg","doi":"10.1007/s40264-025-01529-6","DOIUrl":"https://doi.org/10.1007/s40264-025-01529-6","url":null,"abstract":"<p><strong>Introduction: </strong>Manual identification of case narratives with specific relevant information can be challenging when working with large numbers of adverse event reports (case series). The process can be supported with a search engine, but building search queries often remains a manual task. Suggesting terms to add to the search query could support assessors in the identification of case narratives within a case series.</p><p><strong>Objective: </strong>The aim of this study is to explore the feasibility of identifying case narratives containing specific characteristics with a narrative search engine supported by artificial intelligence (AI) query suggestions.</p><p><strong>Methods: </strong>The narrative search engine uses Best Match 25 (BM25) and suggests additional query terms from two word embedding models providing English and biomedical words to a human in the loop. We calculated the percentage of relevant narratives retrieved by the system (recall) and the percentage of retrieved narratives relevant to the search (precision) on an evaluation dataset including narratives from VigiBase, the World Health Organization global database of adverse event reports for medicines and vaccines. Exact-match search and BM25 search with the Relevance Model (RM3), an alternative way to expand queries, were used as comparators.</p><p><strong>Results: </strong>The gold standard included 55/750 narratives labelled as relevant. Our narrative search engine retrieved on average 56.4% of the relevant narratives (recall), which is higher when compared with exact-match search (21.8%), without a significant drop in precision  (54.5% to 43.1%). The recall is also higher as compared with RM3 (34.4%).</p><p><strong>Conclusions: </strong>Our study demonstrates that a narrative search engine supported by AI query suggestions can be a viable alternative to an exact-match search and BM25 search with RM3, since it can facilitate the retrieval of additional relevant narratives during signal assessments.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143633822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evidence-Based Design of Prescription Medication Information: An Updated Scoping Review.","authors":"Andrea M Russell, Rebecca Lovett, Abigail Vogeley, Denise A Nunes, Carolyn McKelvie, Wayne Middleton, Michael Wolf","doi":"10.1007/s40264-025-01527-8","DOIUrl":"https://doi.org/10.1007/s40264-025-01527-8","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Well-designed prescription medication information (PMI), defined as materials which communicate the essential information needed for a patient to safely and accurately self-administer a medication at or near the time of prescribing, is important for patient education. A previous review identifying best practices in the design of PMI was last completed using studies published through 2015.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;The aim of this review was to present an updated description of studies comparing one or more types of PMI, including details of if or how patients were involved in PMI design, and to consolidate design elements associated with positive outcomes.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;Four databases (Ovid, Embase, CINAHL, and Cochrane) were searched for studies comparing one or more types of PMI using a specified literature search with follow-up citation searching of included articles. Eligible studies were (1) conducted in English-speaking countries, (2) randomized controlled trials, and (3) published in 2016 or later. Consistent findings from at least two well-conducted studies were deemed 'strong' evidence and inconsistency in study findings or quality were deemed 'moderate' evidence.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Thirty-two articles were included and most had some risk (n = 24) or high risk of bias (n = 4). Two-thirds of articles (n = 21) reported on the details of PMI development, and over half (n = 14) conducted formal pilot testing or obtained feedback from patients. Findings suggested benefits when patients were involved in PMI development. Twelve studies examined written medication information (e.g., leaflets), ten examined pharmacy-generated contained labelling (e.g., instructions printed on pill bottles), two examined supplemental information (e.g., medication regimen charts), and fourteen examined PMI delivered using technology-supported tools (e.g., text message instructions). The most prevalent assessed outcomes were knowledge (n = 19), behaviors (n = 17), attitudes/beliefs (n = 11), and clinical outcomes, such as blood pressure (n = 3). Several studies demonstrated positive outcomes when PMI was designed according to health literacy principles of plain language, typographic cues, actionable instructions, large font, and white space. Multiple trials of pictograms and illustrations alongside paired text and text messages to deliver PMI also had positive outcomes. Although there were several studies that examined interactive websites, audio, and video delivery of PMI, mixed findings resulted in moderate evidence. Novel methods of PMI delivery were identified: a plain language label for as-needed medications, strategic memory training, inclusion of patient photos and quotes, Quick Response codes, and electronic health record strategies.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;Overall, a high proportion of studies included patients in the development of PMI and focused on behavioral outcomes. How","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143613977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Off-the-Shelf Large Language Models for Causality Assessment of Individual Case Safety Reports: A Proof-of-Concept with COVID-19 Vaccines.","authors":"Andrea Abate, Elisa Poncato, Maria Antonietta Barbieri, Greg Powell, Andrea Rossi, Simay Peker, Anders Hviid, Andrew Bate, Maurizio Sessa","doi":"10.1007/s40264-025-01531-y","DOIUrl":"https://doi.org/10.1007/s40264-025-01531-y","url":null,"abstract":"<p><strong>Background: </strong>This study evaluated the feasibility of ChatGPT and Gemini, two off-the-shelf large language models (LLMs), to automate causality assessments, focusing on Adverse Events Following Immunizations (AEFIs) of myocarditis and pericarditis related to COVID-19 vaccines.</p><p><strong>Methods: </strong>We assessed 150 COVID-19-related cases of myocarditis and pericarditis reported to the Vaccine Adverse Event Reporting System (VAERS) in the United States of America (USA). Both LLMs and human experts conducted the World Health Organization (WHO) algorithm for vaccine causality assessments, and inter-rater agreement was measured using percentage agreement. Adherence to the WHO algorithm was evaluated by comparing LLM responses to the expected sequence of the algorithm. Statistical analyses, including descriptive statistics and Random Forest modeling, explored case complexity (e.g., string length measurements) and factors affecting LLM performance and adherence.</p><p><strong>Results: </strong>ChatGPT showed higher adherence to the WHO algorithm (34%) compared to Gemini (7%) and had moderate agreement (71%) with human experts, whereas Gemini had fair agreement (53%). Both LLMs often failed to recognize listed AEFIs, with ChatGPT and Gemini incorrectly identifying 6.7% and 13.3% of AEFIs, respectively. ChatGPT showed inconsistencies in 8.0% of cases and Gemini in 46.7%. For ChatGPT, adherence to the algorithm was associated with lower string complexity in prompt sections. The random forest analysis achieved an accuracy of 55% (95% confidence interval: 35.7-73.5) for predicting adherence to the WHO algorithm for ChatGPT.</p><p><strong>Conclusion: </strong>Notable limitations of ChatGPT and Gemini have been identified in their use for aiding causality assessments in vaccine safety. ChatGPT performed better, with higher adherence and agreement with human experts. In the investigated scenario, both models are better suited as complementary tools to human expertise.</p>","PeriodicalId":11382,"journal":{"name":"Drug Safety","volume":" ","pages":""},"PeriodicalIF":4.0,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143613979","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}