Mahmood Ahmad
Tahir Heart Institute
author@example.com

MetaSprint DTA: Automated Open-Access Discovery for Diagnostic Test Accuracy Meta-Analysis

Can automated extraction from open-access abstracts produce pooled diagnostic accuracy estimates consistent with published meta-analyses across diverse specialties? MetaSprint DTA integrates a four-source discovery pipeline searching ClinicalTrials.gov, Europe PMC, OpenAlex, and PubMed with a bivariate GLMM and HSROC engine in a single browser application requiring no installation. The pipeline extracts sensitivity, specificity, and sample sizes using over 30 regex patterns with Unicode preprocessing, back-calculates two-by-two tables, and pools estimates within one session. Across 70 published DTA meta-analyses spanning 13 specialties, all pooled estimates fell within 15 percent of published values with study counts frequently exceeding published reviews, and R cross-validation achieved 33 of 33 parity against mada and metafor. Advanced diagnostics include Cook distance, DFBETAS, Copas selection model, profile-likelihood confidence intervals, and bootstrap BCa intervals. The platform bridges the months-long gap between clinical question and pooled diagnostic accuracy estimate. However, a limitation is that abstract-only extraction cannot capture studies reporting accuracy metrics solely in full-text tables.

Outside Notes

Type: methods
Primary estimand: Pooled sensitivity and specificity
App: MetaSprint DTA v1.0
Data: 70 published DTA meta-analyses across 13 specialties
Code: https://github.com/mahmood726-cyber/metasprint-dta
Version: 1.0
Validation: DRAFT

References

1. Reitsma JB, Glas AS, Rutjes AW, et al. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol. 2005;58(10):982-990.
2. Macaskill P, Gatsonis C, Deeks JJ, Harbord RM, Takwoingi Y. Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy. Cochrane; 2023.
3. Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to Meta-Analysis. 2nd ed. Wiley; 2021.