Saliva, Urine, and Anorectal Swabs as Viable Alternatives to Lesion Swabs for Molecular Diagnosis of Mpox.

Le résumé fourni par la source

This file contains de-identified participant-level source data and all statistical analysis outputs underlying the findings reported in the manuscript. It is structured as a nine-sheet workbook, with each sheet corresponding to a distinct component of the analysis. All sheets include column headers, data definitions, and footnotes describing methodology, abbreviations, and interpretation guidance. Sheet-by-Sheet Contents Cover — Table of contents listing all nine sheets with titles and descriptions. Includes a notes section specifying that all cycle values are reported as quantification cycle (Cq, formerly Ct), that PCR positivity is defined as Cq ≤ 40, that the lesion swab is the reference standard throughout, and that analyses were conducted in Python 3 using scipy, statsmodels, and numpy. S1 — Source Data — De-identified participant-level dataset (N = 120). Contains one row per participant with the following variables: participant ID, biological sex (self-reported), age in years, case category (Confirmed Positive / Alert), specimen type labels for each of the four collection sites, binary PCR result for each specimen (Positive / Negative), case status code, and Cq values for each specimen (recorded as "Not detected" where the specimen was PCR-negative). This sheet constitutes the primary source data for all downstream analyses. S2 — Normality Testing — Shapiro–Wilk normality test results for two variable groups: (1) the age distribution across all 120 participants; and (2) Cq value distributions among PCR-positive specimens for each of the four specimen types. Reports sample size, median, Q1, Q3, minimum, maximum, Shapiro–Wilk W statistic, p-value, and a binary normality classification (p ≥ 0.05 = Normal). Results justify the use of nonparametric statistical tests throughout the analysis. S3 — Diagnostic Performance — Full 2×2 contingency table counts (a, b, c, d) and derived diagnostic performance metrics for each alternative specimen type relative to the lesion swab reference standard. Metrics include sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), each with 95% Wilson score confidence intervals. Agreement metrics include Cohen's κ and Gwet's AC1. N = 120 for all comparisons. S4 — McNemar Discordance Tests — Exact McNemar test results for pairwise discordance between each alternative specimen and the lesion swab reference. Reports discordant pair counts (b and c cells), total discordant pairs, raw p-values, Holm-adjusted p-values (corrected across three simultaneous comparisons), significance classification (Holm p < 0.05), directional discordance odds ratios, and 95% confidence intervals for the odds ratios computed on the log scale. Significant results (urine, Holm-adjusted p = 0.035) are highlighted. S5 — Wilcoxon Cq Comparisons — Wilcoxon signed-rank test results for pairwise comparison of Cq values between each alternative specimen and the lesion swab, restricted to participants with positive PCR results on both specimens in each pair. Reports paired sample size, median Cq for each specimen, median difference, Wilcoxon W statistic, two-sided p-value, matched-pairs rank-biserial correlation effect size (r = |Z|/√N), and qualitative effect size interpretation (trivial / small / medium / large). All three comparisons returned trivial effect sizes (r < 0.08) and non-significant p-values (all p ≥ 0.586). S6 — Bland–Altman Analysis — Bland–Altman agreement analysis for Cq values between each alternative specimen and the lesion swab, restricted to co-positive pairs. Reports paired sample size, mean Cq difference (alternative minus lesion swab), standard deviation of differences, lower and upper limits of agreement (mean ± 1.96 SD), 95% confidence interval of the mean difference, regression slope for proportional bias assessment (Bland & Altman 1999), slope p-value, and a binary classification of proportional bias presence. A note flags mild proportional bias for urine (slope p = 0.032) with an interpretation caution. S7 — TOST Equivalence Testing — Bootstrap two one-sided tests (TOST) of diagnostic equivalence for each alternative specimen relative to the lesion swab reference, based on differences in PCR positivity proportions. Reports observed Δ (alternative minus lesion swab positivity proportion), 90% bootstrap confidence interval (20,000 paired iterations, random seed = 42), prespecified equivalence bounds (±0.10), and the equivalence conclusion. Saliva met equivalence criteria; anorectal swab and urine did not. Equivalent rows are highlighted in green. S8 — GEE Regression Results — Results of two generalised estimating equation (GEE) models using an exchangeable working correlation structure to account for within-participant correlation across four specimens. Model 1 (Binomial GEE): outcome is binary PCR positivity; results reported as adjusted odds ratios with 95% confidence intervals and p-values for specimen type (anorectal swab, urine, saliva, each versus lesion swab), case status (confirmed versus alert), and sex (male versus female). Model 2 (Gaussian GEE): outcome is Cq value among PCR-positive specimens; results reported as mean Cq differences. Reference categories are lesion swab and Alert. Significant results are highlighted.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.