Analysis code for diagnostic evaluation of alternative specimens for Mpox PCR

This repository contains the analysis code used to evaluate the diagnostic performance of alternative clinical specimens for Mpox PCR testing. The Jupyter Notebook reproduces all statistical analyses described in the Methods section of the accompanying manuscript, including descriptive statistics, Wilson confidence intervals for diagnostic accuracy, agreement metrics (Cohen’s kappa and Gwet’s AC1), exact McNemar tests with Holm correction, paired Ct comparisons, Bland–Altman analyses, equivalence testing using two one-sided tests (TOST) with paired bootstrap resampling, and generalized estimating equation (GEE) regression models accounting for within-participant correlation. No participant-level data are included. Access to the underlying Mpox dataset is subject to ethical and regulatory restrictions and may be requested from the corresponding author.