Dataset and code: Systematic review and meta-analysis of stroke blood biomarker data highlights need for more translational research methods.

This repository contains the raw data and essential code needed to replicate the analyses reported in “Systematic review and meta-analysis of stroke blood biomarker data highlights need for more translational research methods” authored by the Neurodiagnostics Lab at Case Western Reserve University (https://doi.org/10.1038/s41467-026-71329-8). The raw data aggregated from the primary literature are provided as a relational database comprised of multiple R data frames that can be downloaded as a single .RData file (relational_database.RData). A descriptive overview of both the data and the database structure is also provided (database_structure.pdf). The custom R script used to unify and model the raw diagnostic performance estimates reported in the primary literature is provided as a plain text file (unification_and_modeling.R), as is the custom R script used in simulation experiments carried out to characterize the potential sources and impact of ROC curve asymmetry on the accuracy of sAUC estimates (symmetry_tests.R). A descriptive overview of the code and how it can be run is also available (code_description.pdf).Reference:Christine G. Smothers, Shivam A. Gandhi, Julia H. Chang, Sadia M. Tahir, Jarrett Kelley, Emily Wang, Kareem Fareed, Hannah Mack, Cindy Dong, Madaline Witort, Bethany L. Armentrout, Darynaisha Crawford, Jing Wang, Kara Williams, Zea Budg, Christine Horvat-Davey, Chris Winkelman, Sarah R. Martha, Jon W. Schrock, Agnieszka A. Ardelt, Grant C. O’Connell. Systematic review and meta-analysis of stroke blood biomarker data highlights need for more translational research methods. Nature Communications (2026). https://doi.org/10.1038/s41467-026-71329-8