A Windows package, a browser platform
MedCalc is a respected Windows program for medical statistics, licensed per seat (about 599 USD one-time or 189 USD per year as of 2026) and known best for its ROC analysis. If you work on a Mac, on Linux, or on a locked-down hospital machine where you cannot install software, the browser wins before the feature comparison even starts.
Diagnostic accuracy without the install
- ROC curves with AUC, its confidence interval, and a significance test against 0.5
- Sensitivity, specificity, predictive values, and likelihood ratios from a 2x2 table
- Intraclass correlation (ICC) for rater agreement
- Bland-Altman plots with bias and 95% limits of agreement
- Prevalence estimation with confidence intervals
The rest of the clinical toolkit
The comparison tests (t-tests, ANOVA, nonparametrics, chi-square, Fisher's exact, McNemar), correlation, and seven regression models including Cox proportional hazards and conditional logistic for matched designs. Kaplan-Meier survival with the log-rank test. A full meta-analysis suite validated against R's metafor. Every result shows its formula and a citation, and data comes in by CSV, Excel, or paste.
An honest note
MedCalc goes deeper on method comparison: it tests two paired ROC curves against each other (DeLong) and runs Passing-Bablok regression, which Covexe does not. If your paper hinges on comparing AUCs head to head, MedCalc has that tool. For the standard diagnostic and clinical analyses, on any machine, Covexe does them free.