Search

The study aims to develop and validate a predictive model for effectively predicting in-hospital mortality in patients admitted to the intensive care unit due to Gastrointestinal bleeding (GIB). A retrospective cohort study was conducted, including data from patients in the Electronic Intensive Care Unit Collaborative Research Database (eICU-CRD) and Medical Information Mart for Intensive Care-IV Database (MIMIC-IV) with a diagnosis of GIB. Patients from the eICU-CRD were randomly allocated into both development and validation sets, and those from MIMIC-IV were assigned as an external validation group. Multivariate logistic regression was employed to create a predictive model, which was depicted as a nomogram. This study included a total of 2929 patients with GIB from the eICU-CRD and 718 patients from the MIMIC-IV. To access the dynamic nomogram, please use the following link: https://kangzou.shinyapps.io/DynNomapp%5fGIB/. The area under the receiver operating characteristic curve for the nomogram was 0.893 in the development set, 0.860 in the internal validation set, and 0.781 in the external validation set. The mortality rate was 25.7% in the high-risk group (nomogram scores > 101.974) and 2.8% in the low-risk group (nomogram scores ≤ 101.974). The nomogram exhibited excellent discrimination, calibration, and clinical utility in predicting in-hospital mortality among patients admitted to the intensive care unit with GIB. [ABSTRACT FROM AUTHOR]

Background: To evaluate the methodological and reporting quality of published meta-analyses (MAs) in four major gastrointestinal endoscopic journals, and identify the predicted factors for high quality. Methods: A systematic search was performed in PubMed to identify MAs from 1, January, 2016 to 31, December, 2020 in four major gastrointestinal endoscopic journals (including Digestive Endoscopy, Gastrointestinal Endoscopy, Surgical Endoscopy, and Endoscopy). We collected the characteristics of MAs after filtering unqualified articles, and assessed methodological and reporting qualities for eligible articles by AMSTAR tool and PRISMA checklist, respectively. Logistic regression was used for identifying predictive factors for high quality. Results: A total of 289 MAs were identified after screening by predefined inclusion and exclusion criteria. The scores (mean ± SD) of AMSTAR and PRISMA were 7.73 ± 1.11 and 22.90 ± 1.85, respectively. In PRISMA checklist, some items had less than 50% complete adherence, including item 2 (structured summary), items 5 (protocol and registration), items 12 and 19 (risk of bias in studies), item 27 (funding support). Item 1 (a priori design), item 4 (gray literature research), item 5 (list of included and excluded) were inferior to 50% adherence in AMSTAR tool. We found the predictive factors for high quality through logistic regression analysis: a priori design and funding support were associated with methodological quality. Protocol and registration influenced the methodological and reporting quality closely. Conclusion: In general, qualities on the methodology and the reporting of MAs published in the gastrointestinal endoscopic journals are good, but both of which still potentially need further improvement. [ABSTRACT FROM AUTHOR]