BMC Medical Research Methodology最新文献

{"title":"Conducting two evidence syntheses in six weeks – experiences with and evaluation of a pilot project","authors":"Heather Melanie R. Ames, Hege Kornør, Line Holtet Evensen, Ingeborg Beate Lidal, Elisabet Hafstad, Christine Hillestad Hestevik, Patricia Sofia Jacobsen Jardim, Gyri Hval","doi":"10.1186/s12874-024-02334-y","DOIUrl":"https://doi.org/10.1186/s12874-024-02334-y","url":null,"abstract":"Evidence synthesis organisations are trying to meet commissioners’ needs for rapid responses to their evidence synthesis commissions. In this project we piloted an intensive process, working to complete evidence syntheses within six-weeks, rather than the standard lead time of 4–6 months. Our objective was to explore how researchers experience working intensively, identify barriers and facilitators, and determine how a more intensive approach to evidence synthesis could be more systematically introduced in the future. In a pre-planning phase, an intensive work group was established, and two commissions were selected for this pilot project. The evidence synthesis process was divided into two phases: planning and intensive. The planning phase, involved scheduling the intensive phase, exploring new digital tools, and identifying peer reviewers. The intensive phase encompassed the entire evidence synthesis process. Two review teams were formed, each with a team lead supported by a process lead and leadership contact point. Throughout the project, teams engaged in reflective meetings to evaluate and adjust processes as needed. During the planning phase, teams identified significant uncertainties regarding scopes, research questions, and inclusion criteria. To address this, they engaged with commissioners earlier than originally planned, clarified these aspects, and prepared protocols. Despite some minor deviations from the original plan, both reviews were completed on schedule, with one team expanding their scope due to the absence of eligible studies. Teams operated flexibly, held regular meetings, and found the process seamless due to fewer interruptions. Machine learning tools facilitated rapid study selection. The process lead role, created to guide and evaluate the project, proved beneficial, providing structure and support, although clearer role delineation with the leadership contact point could have improved efficiency. Overall, the intensive process fostered focus and productivity, allowing teams to manage short-term deliverables effectively. The researchers preferred working intensively with one evidence synthesis over being involved with many projects at the same time. They felt that time use was more effective, and they were able to complete the tasks in a focused way. However, there are several implications that should be considered before implementing an intensive approach in future evidence syntheses.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142248520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Response is increased using postal rather than electronic questionnaires – new results from an updated Cochrane Systematic Review","authors":"Phil Edwards, Chloe Perkins","doi":"10.1186/s12874-024-02332-0","DOIUrl":"https://doi.org/10.1186/s12874-024-02332-0","url":null,"abstract":"A decade ago paper questionnaires were more common in epidemiology than those administered online, but increasing Internet access may have changed this. Researchers planning to use a self-administered questionnaire should know whether response rates to questionnaires administered electronically differ to those of questionnaires administered by post. We analysed trials included in a recently updated Cochrane Review to answer this question. We exported data of randomised controlled trials included in three comparisons in the Cochrane Review that had evaluated hypotheses relevant to our research objective and imported them into Stata for a series of meta-analyses not conducted in the Cochrane review. We pooled odds ratios for response using random effects meta-analyses. We explored causes of heterogeneity among study results using subgroups. We assessed evidence for reporting bias using Harbord’s modified test for small-study effects. Twenty-seven trials (66,118 participants) evaluated the effect on response of an electronic questionnaire compared with postal. Results were heterogeneous (I-squared = 98%). There was evidence for biased (greater) effect estimates in studies at high risk of bias; A synthesis of studies at low risk of bias indicates that response was increased (OR = 1.43; 95% CI 1.08–1.89) using postal questionnaires. Ten trials (39,523 participants) evaluated the effect of providing a choice of mode (postal or electronic) compared to an electronic questionnaire only. Response was increased with a choice of mode (OR = 1.63; 95% CI 1.18–2.26). Eight trials (20,909 participants) evaluated the effect of a choice of mode (electronic or postal) compared to a postal questionnaire only. There was no evidence for an effect on response of a choice of mode compared with postal only (OR = 0.94; 95% CI 0.86–1.02). Postal questionnaires should be used in preference to, or offered in addition to, electronic modes.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142248518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modelling wound area in studies of wound healing interventions","authors":"Samuel I. Watson, Eleni Gkini, Jon Bishop, Katie Scandrett, Indra Napit, Richard J. Lilford","doi":"10.1186/s12874-024-02326-y","DOIUrl":"https://doi.org/10.1186/s12874-024-02326-y","url":null,"abstract":"Experimental studies of wound healing often use survival analysis and time to event outcomes or differences in wound area at a specific time point. However, these methods do not use a potentially large number of observations made over the course of a trial and may be inefficient. A model-based approach can leverage all trial data, but there is little guidance on appropriate models and functional forms to describe wound healing. We derive a general statistical model and review a wide range of plausible mathematical models to describe wound healing. We identify a range of possible derived estimands and their derivation from the models. Using data from a trial of an intervention to promote ulcer healing in patients affected by leprosy that included three measurement methods repeated across the course of the study, we compare the goodness-of-fit of the models using a range of methods and estimate treatment effects and healing rate functions with the best-fitting models. Overall, we included 5,581 ulcer measurements of 1,578 unique images from 130 patients. We examined the performance of a range of models. The square root, log square root, and log quadratic models were the best fitting models across all outcome measurement methods. The estimated treatment effects magnitude and sign varied by time post-randomisation, model type, and outcome type, but across all models there was little evidence of effectiveness. The estimated effects were significantly more precise than non-parametric alternatives. For example, estimated differences from the three outcome measurements at 42-days post-randomisation were − 0.01 cm2 (-0.77, 0.74), -0.44 cm2 (-1.64, 0.76), and 0.11 cm2 (-0.87, 1.08) using a non-parametric method versus − 0.03 cm2 (-0.14, 0.06), 0.06 cm2 (-0.05, 0.17), and 0.03 cm2 (-0.07, 0.17) using a square-root model. Model-based analyses can dramatically improve the precision of estimates but care must be taken to carefully compare and select the best fitting models. The (log) square-root model is strongly recommended reflecting advice from a century ago.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142248521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The impact of different censoring methods for analyzing survival using real-world data with linked mortality information: a simulation study","authors":"Wei-Chun Hsu, Aaron Crowley, Craig S. Parzynski","doi":"10.1186/s12874-024-02313-3","DOIUrl":"https://doi.org/10.1186/s12874-024-02313-3","url":null,"abstract":"Evaluating outcome reliability is critical in real-world evidence studies. Overall survival is a common outcome in these studies; however, its capture in real-world data (RWD) sources is often incomplete and supplemented with linked mortality information from external sources. Conflicting recommendations exist for censoring overall survival in real-world evidence studies. This simulation study aimed to understand the impact of different censoring methods on estimating median survival and log hazard ratios when external mortality information is partially captured. We used Monte Carlo simulation to emulate a non-randomized comparative effectiveness study of two treatments with RWD from electronic health records and linked external mortality data. We simulated the time to death, the time to last database activity, and the time to data cutoff. Death events after the last database activity were attributed to linked external mortality data and randomly set to missing to reflect the sensitivity of contemporary real-world data sources. Two censoring schemes were evaluated: (1) censoring at the last activity date and (2) censoring at the end of data availability (data cutoff) without an observed death. We assessed the performance of each method in estimating median survival and log hazard ratios using bias, coverage, variance, and rejection rate under varying amounts of incomplete mortality information and varying treatment effects, length of follow-up, and sample size. When mortality information was fully captured, median survival estimates were unbiased when censoring at data cutoff and underestimated when censoring at the last activity. When linked mortality information was missing, censoring at the last activity date underestimated the median survival, while censoring at the data cutoff overestimated it. As missing linked mortality information increased, bias decreased when censoring at the last activity date and increased when censoring at data cutoff. Researchers should consider the completeness of linked external mortality information when choosing how to censor the analysis of overall survival using RWD. Substantial bias in median survival estimates can occur if an inappropriate censoring scheme is selected. We advocate for RWD providers to perform validation studies of their mortality data and publish their findings to inform methodological decisions better.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203380","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of acute onset of chronic cor pulmonale: comparative analysis of Holt-Winters exponential smoothing and ARIMA model","authors":"Nan Wang, Weiyi Zhuang, Zhen Ran, Pinxi Wan, Jian Fu","doi":"10.1186/s12874-024-02325-z","DOIUrl":"https://doi.org/10.1186/s12874-024-02325-z","url":null,"abstract":"The aim of this study is to analyze the trend of acute onset of chronic cor pulmonale at Chenggong Hospital of Kunming Yan’an Hospital between January 2018 and December 2022.Additionally, the study will compare the application of the ARIMA model and Holt-Winters model in predicting the number of chronic cor pulmonale cases. The data on chronic cor pulmonale cases from 2018 to 2022 were collected from the electronic medical records system of Chenggong Hospital of Kunming Yan’an Hospital. The ARIMA and Holt-Winters models were constructed using monthly case numbers from January 2018 to December 2022 as training data. The performance of the model was tested using the monthly number of cases from January 2023 to December 2023 as the test set. The number of acute onset of chronic cor pulmonale in Chenggong Hospital of Kunming Yan’an Hospital exhibited a downward trend overall from 2018 to 2022. There were more cases in winter and spring, with peaks observed in November to December and January of the following year. The optimal ARIMA model was determined to be ARIMA (0,1,1) (0,1,1)12, while for the Holt-Winters model, the optimal choice was the Holt-Winters multiplicative model. It was found that the Holt-Winters multiplicative model yielded the lowest error. The Holt-Winters multiplicative model predicts better accuracy. The diagnosis of acute onset of chronic cor pulmonale is related to many risk factors, therefore, when using temporal models to fit and predict the data, we must consider such factors’ influence and try to incorporate them into the models.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203379","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Is this really Empowerment? Enhancing our understanding of empowerment in patient and public involvement within clinical research","authors":"Imke Schilling, Ansgar Gerhardus","doi":"10.1186/s12874-024-02323-1","DOIUrl":"https://doi.org/10.1186/s12874-024-02323-1","url":null,"abstract":"There has been a growing push to involve patients in clinical research, shifting from conducting research on, about, or for them to conducting it with them. Two arguments advocate for this approach, known as Patient and Public Involvement (PPI): to improve research quality, appropriateness, relevance, and credibility by including patients’ diverse perspectives, and to use PPI to empower patients and democratize research for more equity in research and healthcare. However, while empowerment is a core objective, it is often not clear what is meant by empowerment in the context of PPI in clinical research. This vacancy can lead to insecurities for both patients and researchers and a disconnect between the rhetoric of empowerment in PPI and the reality of its practice in clinical trials. Thus, clarifying the understanding of empowerment within PPI in clinical research is essential to ensure that involvement does not become tokenistic and depletes patients’ capacity to advocate for their rights and needs. We explored the historical roots of empowerment, primarily emerging from mid-20th century social movements like feminism and civil rights and reflected the conceptual roots of empowerment from diverse fields to better understand the (potential) role of empowerment in PPI in clinical research including its possibilities and limitations. Common themes of empowerment in PPI and other fields are participation, challenging power structures, valuing diverse perspectives, and promoting collaboration. On the other hand, themes such as contextual differences in the empowerment objectives, the relationship between empowerment and scientific demands, research expertise, and power asymmetries mark a clear distinction from empowerment in other fields. PPI offers potential for patient empowerment in clinical trials, even when its primary goal may be research quality. Elements like participation, sharing opinions, and active engagement can contribute to patient empowerment. Nonetheless, some expectations tied to empowerment might not be met within the constraints of clinical research. To empower patients, stakeholders must be explicit about what empowerment means in their research, engage in transparent communication about its realistic scope, and continuously reflect on how empowerment can be fostered and sustained within the research process.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Identifying trials run in India that are registered in other clinical trial registries: a cross-sectional study","authors":"Rishima Borah, Anwesha Dhal Samanta, Jaishree Mendiratta, Manish Mishra, Gayatri Saberwal","doi":"10.1186/s12874-024-02336-w","DOIUrl":"https://doi.org/10.1186/s12874-024-02336-w","url":null,"abstract":"Clinical trials play a crucial role in biomedical research, and it is important to register them in public registries to ensure transparency and prevent research waste. In this study, we wished to determine what steps need to be taken to identify every clinical trial run in India that has been registered in any of the (non-Indian) World Health Organization-recognised primary registries. Of the 16 registries, we studied all except that of the European Union, which will be studied separately. Two methodologies were employed for each registry, except for four that did not facilitate one or the other method. Methodology A involved downloading all the records in a registry and querying them. Methodology B involved conducting a search via the registry website. Only four registries provided consistent results with both methodologies. Seven registries had different results from the two methodologies. Of these, in four cases, in Methodology A one field indicated that the study ran in India, while another indicated otherwise. The above-mentioned ambiguities should be addressed by the concerned registries. Overall, this study reinforces the need for improved data accuracy and transparency in clinical trial registries and emphasizes the importance of resolving complications faced by users while navigating the registries. Ensuring accurate and comprehensive registration of clinical trials is essential for meta-research and the use of such data by a variety of stakeholders.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203381","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptation and psychometric assessment of a sexual and reproductive empowerment scale in Arabic among refugee and non-refugee adolescent girls","authors":"Myriam Dagher, Sawsan Abdulrahim, Berthe Abi Zeid, Maia Sieverding","doi":"10.1186/s12874-024-02300-8","DOIUrl":"https://doi.org/10.1186/s12874-024-02300-8","url":null,"abstract":"Sexual and reproductive empowerment (SRE) is an important determinant of women’s and girls’ health yet measuring it is complex due to cultural and domain-specific variations. This study describes the process of adapting an SRE scale consisting of four domains (self-efficacy; future orientation; social support; and safety) and testing its psychometric properties among Arabic speaking adolescent girls in Lebanon. An SRE scale developed in a Western context was adapted in four steps: (1) reviewing the scale and selecting culturally appropriate domains for translation to standard Arabic; (2) conducting cognitive interviews with 30 11-17-year-old adolescent girls in Lebanon; (3) administering the scale to 339 refugee adolescent girls who participated in an early marriage intervention; and (4) conducting confirmatory factor analysis (CFA) on the data to assess the scale’s psychometric properties. The original model for the 13-item, four-domain adapted scale demonstrated poor fit in CFA. After iteratively removing two items, scale properties were improved, albeit were not optimal. The validity and reliability results for the self-efficacy domain were acceptable. Cognitive interview data revealed that Arab adolescent girls understood self-efficacy in relational terms, recognizing that autonomous decision-making is not necessarily favored but is influenced by parents and family. This study presents an effort to customize an SRE scale for use in studies on the health of adolescent girls in an Arab cultural context. Findings from cognitive interviews highlight the importance of taking into consideration relationality in adolescent sexual and reproductive decision-making. The self-efficacy domain in the adapted scale demonstrates acceptable psychometric properties and is recommended for use in health studies to capture SRE.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203383","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Recruiting people with selected citizenships for the health interview survey GEDA Fokus throughout Germany: evaluation of recruitment efforts and recommendations for future research","authors":"Carmen Koschollek, Beate Gaertner, Julia Geerlings, Ronny Kuhnert, Elvira Mauz, Claudia Hövener","doi":"10.1186/s12874-024-02328-w","DOIUrl":"https://doi.org/10.1186/s12874-024-02328-w","url":null,"abstract":"Germany is the second most common country of immigration after the US. However, people with own or familial history of migration are not represented proportionately to the population within public health monitoring and reporting. To bridge this data gap and enable differentiated analyses on migration and health, we conducted the health interview survey GEDA Fokus among adults with Croatian, Italian, Polish, Syrian, or Turkish citizenship living throughout Germany. The aim of this paper is to evaluate the effects of recruitment efforts regarding participation and sample composition. Data collection for this cross-sectional and multilingual survey took place between 11/2021 and 5/2022 utilizing a sequential mixed-mode design, including self-administered web- and paper-based questionnaires as well as face-to-face and telephone interviews. The gross sample (n = 33436; age range 18–79 years) was randomly drawn from the residents’ registers in 120 primary sampling units based on citizenship. Outcome rates according to the American Association for Public Opinion Research, the sample composition throughout the multistage recruitment process, utilization of survey modes, and questionnaire languages are presented. Overall, 6038 persons participated, which corresponded to a response rate of 18.4% (range: 13.8% for Turkish citizenship to 23.9% for Syrian citizenship). Home visits accounted for the largest single increase in response. During recruitment, more female, older, as well as participants with lower levels of education and income took part in the survey. People with physical health problems and less favourable health behaviour more often took part in the survey at a later stage, while participants with symptoms of depression or anxiety more often participated early. Utilization of survey modes and questionnaire languages differed by sociodemographic and migration-related characteristics, e.g. participants aged 50 years and above more often used paper- than web-based questionnaires and those with a shorter duration of residence more often used a translated questionnaire. Multiple contact attempts, including home visits and different survey languages, as well as offering different modes of survey administration, increased response rates and most likely reduced non-response bias. In order to adequately represent and include the diversifying population in public health monitoring, national public health institutes should tailor survey designs to meet the needs of different population groups considered hard to survey to enable their survey participation.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predictive etiological classification of acute ischemic stroke through interpretable machine learning algorithms: a multicenter, prospective cohort study","authors":"Siding Chen, Xiaomeng Yang, Hongqiu Gu, Yanzhao Wang, Zhe Xu, Yong Jiang, Yongjun Wang","doi":"10.1186/s12874-024-02331-1","DOIUrl":"https://doi.org/10.1186/s12874-024-02331-1","url":null,"abstract":"The prognosis, recurrence rates, and secondary prevention strategies varied significantly among different subtypes of acute ischemic stroke (AIS). Machine learning (ML) techniques can uncover intricate, non-linear relationships within medical data, enabling the identification of factors associated with etiological classification. However, there is currently a lack of research utilizing ML algorithms for predicting AIS etiology. We aimed to use interpretable ML algorithms to develop AIS etiology prediction models, identify critical factors in etiology classification, and enhance existing clinical categorization. This study involved patients with the Third China National Stroke Registry (CNSR-III). Nine models, which included Natural Gradient Boosting (NGBoost), Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Light Gradient Boosting Machine (LGBM), Gradient Boosting Decision Tree (GBDT), Adaptive Boosting (AdaBoost), Support Vector Machine (SVM), and logistic regression (LR), were employed to predict large artery atherosclerosis (LAA), small vessel occlusion (SVO), and cardioembolism (CE) using an 80:20 randomly split training and test set. We designed an SFS-XGB with 10-fold cross-validation for feature selection. The primary evaluation metrics for the models included the area under the receiver operating characteristic curve (AUC) for discrimination and the Brier score (or calibration plots) for calibration. A total of 5,213 patients were included, comprising 2,471 (47.4%) with LAA, 2,153 (41.3%) with SVO, and 589 (11.3%) with CE. In both LAA and SVO models, the AUC values of the ML models were significantly higher than that of the LR model (P < 0.001). The optimal model for predicting SVO (AUC [RF model] = 0.932) outperformed the optimal LAA model (AUC [NGB model] = 0.917) and the optimal CE model (AUC [LGBM model] = 0.846). Each model displayed relatively satisfactory calibration. Further analysis showed that the optimal CE model could identify potential CE patients in the undetermined etiology (SUE) group, accounting for 1,900 out of 4,156 (45.7%). The ML algorithm effectively classified patients with LAA, SVO, and CE, demonstrating superior classification performance compared to the LR model. The optimal ML model can identify potential CE patients among SUE patients. These newly identified predictive factors may complement the existing etiological classification system, enabling clinicians to promptly categorize stroke patients’ etiology and initiate optimal strategies for secondary prevention.","PeriodicalId":9114,"journal":{"name":"BMC Medical Research Methodology","volume":null,"pages":null},"PeriodicalIF":4.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142203395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}