BMC Medical Research Methodology最新文献

{"title":"Estimating the causal effects of exposure mixtures: a generalized propensity score method.","authors":"Qian Gao, Ting Li, Guiming Zhu, Juping Wang, Kexin Qiu, Liangpo Liu, Xiujuan Yang, Tong Wang","doi":"10.1186/s12874-025-02673-4","DOIUrl":"10.1186/s12874-025-02673-4","url":null,"abstract":"<p><strong>Background: </strong>In environmental epidemiology and many other fields, estimating the causal effects of multiple concurrent exposures holds great promise for driving public health interventions and policy changes. Given the predominant reliance on observational data, confounding remains a key consideration, and generalized propensity score (GPS) methods are widely used as causal models to control measured confounders. However, current GPS methods for multiple continuous exposures remain scarce.</p><p><strong>Methods: </strong>We proposed a novel causal model for exposure mixtures, called nonparametric multivariate covariate balancing generalized propensity score (npmvCBGPS). A simulation study examined whether npmvCBGPS, an existing multivariate GPS (mvGPS) method, and a linear regression model for the outcome can accurately and precisely estimate the effects of exposure mixtures in a variety of common scenarios. An application study illustrated the analysis of the causal role of per- and polyfluoroalkyl substances (PFASs) on BMI.</p><p><strong>Results: </strong>The npmvCBGPS achieved acceptable covariate balance in all scenarios. The estimates were close to the true value as long as either the exposure or the outcome model was correctly specified, and the results were less impacted by correlations among mixture components. The accuracy and precision of mvGPS and the linear regression model relied on the correctly specified exposure model and outcome model, respectively. The npmvCBGPS outperformed mvGPS in all scenarios. The npmvCBGPS achieved better covariate balance than mvGPS and provided an overall inverse trend between the PFAS mixtures with BMI.</p><p><strong>Conclusions: </strong>In this study, we proposed npmvCBGPS to accurately estimate the causal effects of multiple exposure mixtures on health outcomes. Our approach is applicable across various domains, with a particular emphasis on environmental epidemiology.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"221"},"PeriodicalIF":3.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12482879/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145191048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A human-LLM collaborative annotation approach for screening articles on precision oncology randomized controlled trials.","authors":"Hui Chen, Jiale Zhao, Sheng Zheng, Xinyu Zhang, Huilong Duan, Xudong Lu","doi":"10.1186/s12874-025-02674-3","DOIUrl":"10.1186/s12874-025-02674-3","url":null,"abstract":"<p><strong>Background: </strong>Supervised learning can accelerate article screening in systematic reviews, but still requires labor-intensive manual annotation. While large language models (LLMs) like GPT-3.5 offer a rapid and convenient alternative, their reliability is challenging. This study aims to design an efficient and reliable annotation method for article screening.</p><p><strong>Methods: </strong>Given that relevant articles are typically a small subset of those retrieved articles during screening, we propose a human-LLM collaborative annotation method that focuses on verifying positive annotations made by the LLM. Initially, we optimized the prompt using a manually annotated standard dataset, refining it iteratively to achieve near-perfect recall for the LLM. Subsequently, the LLM, guided by the optimized prompt, annotated the articles, followed by human verification of the LLM-identified positive samples. This method was applied to screen articles on precision oncology randomized controlled trials, evaluating both its efficiency and reliability.</p><p><strong>Results: </strong>For prompt optimization, a standard dataset of 200 manually annotated articles was equally divided into a tuning set and a validation set (1:1 ratio). Through iterative prompt optimization, the LLM achieved near-perfect recall in the tuning and validation sets, with 100% and 85.71%, respectively. Using the optimized prompt, we conducted collaborative annotation. To evaluate its performance, we manually reviewed a random sample of 300 articles that had been annotated using the collaborative annotation method. The results showed that the collaborative annotation achieved an F1 score of 0.9583, reducing the annotation workload by approximately 80% compared to manual annotation alone. Additionally, we trained a BioBERT-based supervised model on the collaborative annotation data, which outperformed the model trained on data annotated solely by the LLM, further validating the reliability of the collaborative annotation method.</p><p><strong>Conclusions: </strong>The human-LLM collaborative annotation method demonstrates potential for enhancing the efficiency and reliability of article screening, offering valuable support for systematic reviews and meta-analyses.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"219"},"PeriodicalIF":3.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12481989/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145191065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Handling missing outcomes in time-to-event analyses in randomised controlled trials: a scoping review with a focus on multiple imputation.","authors":"Saravanaraj Karuppasamy, Prasanna Samuel Premkumar, Venkata Raghava Mohan","doi":"10.1186/s12874-025-02676-1","DOIUrl":"10.1186/s12874-025-02676-1","url":null,"abstract":"<p><strong>Background: </strong>Randomised Controlled Trials (RCTs) are the gold standard for evaluating treatment effects. However, missing outcomes can threaten the validity of the results. Missing data pose a unique challenge in time-to-event analyses, where the event time may be censored rather than completely missing. Proper handling of missing event times is crucial to ensure unbiased and reliable conclusions in RCTs. This scoping review examines how missing outcomes in time-to-event studies have been addressed in high-impact medical journals and evaluates the implementation and reporting of multiple imputation (MI) techniques in RCTs.</p><p><strong>Method: </strong>This scoping review assessed methods for handling missing time-to-event outcomes in RCTs published between 2019 and 2024 in three high-impact medical journals: The New England Journal of Medicine, The Lancet, and The Journal of the American Medical Association. Studies were reviewed to identify whether missing outcome data were present and, if so, which methods were used to handle them. Studies that applied MI were examined in detail to assess how the MI approach was implemented and reported. The review also explored theoretical approaches for imputing censored event times.</p><p><strong>Results: </strong>A total of 834 articles were identified through a PubMed search. After screening, 383 RCTs underwent full-text review. Of these, 354 (92.4%) had no or < 10% missing outcomes without imputation. The remaining 29 studies (7.6%) addressed missing data using statistical approaches: 12 applied MI, 10 used complete case analysis, 6 conducted best-/worst-case sensitivity analyses, and 1 used a propensity score-based method. MI approaches varied, with some studies lacking detailed reporting.</p><p><strong>Conclusion: </strong>In RCTs with survival outcomes, properly handling missing event times is essential. This scoping review reveals that, despite the availability of statistical methods, the treatment of missing time-to-event outcomes remains underutilised and often poorly documented. While many studies reported non-administrative censoring, limited information was provided on whether such censoring was informative or non-informative. Additionally, the reporting of MI techniques is frequently insufficient. These findings highlight a critical gap in the handling and reporting of missing outcomes in survival analysis. Strengthening these practices will enhance the reliability and reproducibility of survival analyses in RCTs.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"220"},"PeriodicalIF":3.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12482435/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145191016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploration of Likert scale in terms of continuous variable with parametric statistical methods.","authors":"Iksoo Huh, Jungsoo Gim","doi":"10.1186/s12874-025-02668-1","DOIUrl":"10.1186/s12874-025-02668-1","url":null,"abstract":"","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"218"},"PeriodicalIF":3.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12482090/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145191067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development and validation of high-sensitivity and high-specificity pubmed search filters for systematic and non-systematic reviews.","authors":"Victor Nascimento Fontanive, Letícia Possebon Müller, Federico Riva, Mariana Seoane, Roger Keller Celeste","doi":"10.1186/s12874-025-02666-3","DOIUrl":"10.1186/s12874-025-02666-3","url":null,"abstract":"<p><strong>Objective: </strong>To develop and compare the accuracy of different PubMed search strategies (filters) for systematic and non-systematic reviews in dental journals.</p><p><strong>Methods: </strong>This validation study included articles published in 2019 in 15 dental journals. Two search filters were developed: (1) a high-sensitivity filter to retrieve all possible review articles, and (2) a high-specificity filter for systematic reviews. Two previously published filters were used as benchmarks. The gold standard method for identifying the study methodology was manual reading of the full text. Accuracy, sensitivity and specificity were calculated.</p><p><strong>Results: </strong>Among the 2246 articles published, 6.7% (n = 150) were systematic reviews and 5.9% (n = 132) were other types of reviews. The high-sensitivity filter retrieved 147 of 150 systematic reviews and showed a sensitivity of 98.0% (95%CI: 94.3-99.6) and specificity of 88.9% (95%CI: 87.5-90.2). The high-specificity filter had 96.7% (95%CI: 92.4-98.9) sensitivity and 99.1% (95%CI: 98.6-99.5) specificity for retrieving systematic reviews. The accuracy of this filter for systematic reviews was 97.9% (95%CI: 96.4-99.4), which was higher than the PubMed benchmark filter (p < 0.05) and similar to another longer filter.</p><p><strong>Conclusion: </strong>This study provides two new highly accurate search filters for PubMed that can be used by clinicians, researchers and policymakers.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"216"},"PeriodicalIF":3.4,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12482493/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145191052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A randomized Bayesian phase I-II dose optimization design for combination cancer therapies with progression-free survival end point.","authors":"Yingjie Qiu, Mingyue Li","doi":"10.1186/s12874-025-02665-4","DOIUrl":"10.1186/s12874-025-02665-4","url":null,"abstract":"<p><strong>Background: </strong>Combination therapies involving novel agents, such as immunotherapies and targeted therapies, offer significant antitumor benefits by increasing dose intensity, targeting multiple pathways, and benefiting a broader patient population. To further explore these advantages, the National Cancer Institute (NCI) has initiated Combination Therapy Platform Trial with Molecular Analysis for Therapy Choice (ComboMATCH) to evaluate the effectiveness of new drug combinations in treating both adults and children. However, designing dose optimization trials for these combination therapies presents substantial challenges due to the complex interactions and unique mechanisms of action.</p><p><strong>Methods: </strong>To address these challenges, we propose COMPACT, a Bayesian phase I-II randomized design for combination cancer therapies that uses progression-free survival (PFS) as the primary efficacy endpoint to identify the optimal dose combination (ODC) based on restricted mean survival time (RMST). The COMPACT design jointly evaluates both toxicity and PFS, with continuous toxicity monitoring throughout the trial. Toxicity probabilities are modeled using a partial ordering assumption without relying on complex parametric models, while PFS is modeled through a Bayesian Pareto proportional hazards model with gamma-shared frailty. The trial consists of two seamlessly connected stages. In the first stage, the dose space is explored primarily based on toxicity, while PFS data are concurrently collected. In the second stage, patients are adaptively randomized to safe and potentially promising dose combinations based on PFS, and the dose combination with the highest RMST among those deemed safe is selected as the ODC.</p><p><strong>Results: </strong>Simulation studies demonstrate that COMPACT has desirable operating characteristics and outperforms conventional designs in identifying the ODC, allocating more patients to ODC, while maintaining patient safety. Sensitivity analysis is performed to examine the robustness of the proposed design. A trial example is provided to facilitate the practical implementation of the proposed COMPACT design.</p><p><strong>Conclusions: </strong>The proposed COMPACT design offers a novel and robust framework for combination cancer therapies with progression-free survival end point.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"214"},"PeriodicalIF":3.4,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12466004/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145173685","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Semiautomatic mapping of a national drug terminology to standardised OMOP drug concepts using publicly available supplementary information.","authors":"Florian Katsch, Ágota Mészáros, Tibor Héja, Rada Hussein, Georg Duftschmid","doi":"10.1186/s12874-025-02669-0","DOIUrl":"10.1186/s12874-025-02669-0","url":null,"abstract":"<p><strong>Background: </strong>Mapping national drug terminologies to internationally recognized standards is essential for harmonising health data across regions and supporting secondary data use. In Austria, the national drug terminology lacks fine-granular mappings to RxNorm and RxNorm Extension (RxN/E), limiting its integration into the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). This study aims to semiautomatically map Austria's national drug terminology to RxN/E, to enable improved interoperability and data standardisation for secondary use.</p><p><strong>Methods: </strong>We implemented a semiautomated mapping approach using public supplementary data to bridge the gap between national drug concepts and RxN/E concepts. Probabilistic matching and hierarchical refinement techniques were applied to derive finer-grained and more meaningful mappings than previously available ingredient level mappings via the Anatomical Therapeutic Chemical (ATC) classification. We linked our mappings to other available European drug mappings for a validation of our results.</p><p><strong>Results: </strong>Our process successfully mapped 18,390 (95.42%) of Austria's 19,273 drug concepts to RxN/E, surpassing previous mappings that focused solely on ingredient-level relationships. Specifically, we mapped 73.65% of the concepts to more specific RxN/E targets, such as branded drug boxes and quantified clinical drugs. We identified multiple vocabulary inconsistencies, including duplications and erroneous relationships within RxN/E, which were documented for improvement. The results are disseminated as Usagi-formatted CSV files and HL7 FHIR ConceptMaps to encourage transparency, ease of use, and community-driven refinement.</p><p><strong>Conclusions: </strong>The presented mapping approach highlights the feasibility and utility of leveraging publicly available supplementary data to create mappings between national drug terminology and RxN/E. Our method yields fine-grained mappings, enabling precise and comprehensive drug data integration for secondary use.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"213"},"PeriodicalIF":3.4,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12465475/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145173719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Towards robust electronic health record systems: integrating formal verification and process modeling techniques.","authors":"Saba Khan, Nadeem Akhtar, Muhammad Faheem Mushtaq, Nagwan Abdel Samee, Noha F Mahmoud, Imran Ashraf","doi":"10.1186/s12874-025-02637-8","DOIUrl":"10.1186/s12874-025-02637-8","url":null,"abstract":"<p><p>The integration of complex software systems such as clinical decision aid platforms and electronic health records (EHR) has substantially improved human healthcare support by enhancing diagnostic accuracy and optimizing medical workflows. Developing these systems requires meticulous processes including requirement specification, design, implementation, testing, and deployment. While numerous approaches exist for system development, formal verification through modeling has become crucial for ensuring system reliability and security. In this research, an approach based on Hierarchical Colored Petri-Nets (HCP-Nets) for process modeling and formal verification to develop an EHR system is proposed that is accurate, complete, and consistent. The World Health Organization (WHO) emphasizes integrating modern computing techniques such as data mining and machine learning into health systems in the context of viral disease outbreaks and the need for robust public health surveillance. There is a critical need for process mining and formal verification-based systems to improve data sharing and system integration in regions like Pakistan, where healthcare infrastructure and interoperability are limited. This research addresses key gaps in EHR systems including platform effectiveness, data storage consistency, data accuracy, completeness, and security against unauthorized access. The findings suggest that formal verification using HCP-Nets with model checking and process mining can greatly enhance the reliability and security of EHR systems with an accuracy of 80 5%, providing a strong foundation to advance health informatics and supporting better health outcomes in complex and diverse environments.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"215"},"PeriodicalIF":3.4,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12465584/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145173713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating the properties of the fragility index of meta-analyses.","authors":"Aiwen Xing, Xing Xing, Mohammad Hassan Murad, Lifeng Lin","doi":"10.1186/s12874-025-02648-5","DOIUrl":"10.1186/s12874-025-02648-5","url":null,"abstract":"<p><strong>Background: </strong>The fragility index (FI) has become an increasingly popular supplementary measure for evaluating the robustness of a study's conclusions. While initially developed for individual clinical trials, the FI has been extended to meta-analyses (MAs) of multiple studies. However, the existing literature provides limited insights into the properties of the FI in the context of MAs. This study aims to explore various statistical methods for MAs and assess the improvement in FI of MAs compared to the individual studies they comprise.</p><p><strong>Methods: </strong>We investigated the empirical distributions of FI and fragility quotient (FQ) using a large database of Cochrane MAs with binary outcomes. The FI of MAs was calculated under different statistical methods, including fixed-effect and random-effects models, with between-study variance estimators (restricted maximum-likelihood and DerSimonian-Laird), alongside Hartung-Knapp-Sidik-Jonkman (HKSJ) confidence interval adjustments. Subgroup analyses were performed to explore the impact of heterogeneity, sample size, and effect measures on fragility. Furthermore, we employed a metric to evaluate the improvement in fragility by comparing the FI of MAs with the FIs of the individual studies they included.</p><p><strong>Results: </strong>The median FI was 5 (IQR: 2-11) among 3,758 MAs analyzed, with 29% reporting statistically significant results. Notably, 15% of MAs had an FI of 1, and 54% had an FI ≤ 5. MAs with larger sample sizes or higher [Formula: see text] values, tended to exhibit greater robustness. HKSJ adjustments introduced more uncertainty, yielding more fragile results compared to analyses without these adjustments. Fragility improvement was higher in MAs with considerable heterogeneity.</p><p><strong>Conclusions: </strong>This study highlights the variability in fragility across MAs and underscores the influence of heterogeneity and statistical methods on FI. Further research is warranted to refine the assessment of fragility and incorporate clinical relevance into these evaluations.</p>","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"212"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12465983/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145147827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Moving towards accurate and transparent AMSTAR 2 ratings and correcting the record on Cochrane reviews.","authors":"Juan Victor Ariel Franco, Jae Hung Jung","doi":"10.1186/s12874-025-02682-3","DOIUrl":"10.1186/s12874-025-02682-3","url":null,"abstract":"","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":"25 1","pages":"211"},"PeriodicalIF":3.4,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12459029/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145136500","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}