Diagnostic accuracy software for clinical test evaluation ROC curves with DeLong AUC comparison, sensitivity, specificity, likelihood ratios, predictive values, and optimal threshold determination — for quantitative and qualitative tests, with the confidence intervals and statistical tests reviewers expect to see.
Try free for 15 days — no credit card required
Trusted by 75,000 researchers, analysts, and scientists at leading universities, hospitals, and companies worldwide — for over 30 years.
Compare up to 10 tests rigorously
DeLong AUC comparison for paired and independent designs — with confidence intervals on the difference and formal tests for equality, equivalence, or non-inferiority. A statistically rigorous foundation for claiming one test outperforms another in a publication.
Defensible at peer review
Every calculation runs in Analyse-it's own validated engine — no Excel formulas, no third-party functions. Validated before every release so the diagnostic performance statistics in your manuscript can be defended under reviewer scrutiny.
Patient data stays local
Analyse-it runs entirely within Excel on your PC. No cloud processing, no data transmission — your patient data stays under institutional data governance and ethics approval controls.
Producing a ROC curve and reporting AUC is the easy part. The harder questions are the ones that matter clinically: at what threshold does this test perform best in your patient population, where the prevalence of disease and the consequences of misclassification are specific to your context? A test with high AUC can still have poor positive predictive value in a low-prevalence population. A threshold that maximises the Youden index may not be the right choice when the costs of false negatives and false positives are asymmetric. And a new biomarker that looks promising on a curve needs to be compared formally against the existing test before a publication can claim it performs better.
Analyse-it handles all of this. ROC curves for up to 10 paired or independent quantitative tests, with AUC estimation using DeLong confidence intervals and formal comparison tests — equality, equivalence, and non-inferiority. The full set of diagnostic accuracy measures at every threshold: sensitivity, specificity, positive and negative likelihood ratios, predictive values, odds ratio, and Youden index. Optimal threshold determination by Youden index, cost-based optimisation that accounts for the relative consequences of each type of misclassification, or fixed constraints on minimum sensitivity or specificity. An interactive decision plot across all thresholds showing how every measure shifts as the cut-point moves. For qualitative tests, sensitivity and specificity with exact and score confidence intervals, paired and independent test comparison, and McNemar-Mosteller and Fisher exact tests.
Diagnostic accuracy evaluation is rarely the only analysis in a clinical study. The same dataset may need survival analysis for time-to-event outcomes, logistic regression for a prediction model, or Bland-Altman for method agreement. Analyse-it covers all of these in the Medical edition — the analysis stays in one workbook, with one consistent interface, rather than being spread across separate tools with separate file formats.
75,000 researchers, analysts, and scientists at leading universities, hospitals, and research institutions have relied on Analyse-it for over 30 years.
What's included
-
Establish diagnostic accuracy with ROC curves and AUC
ROC curves for up to 10 paired or independent quantitative tests. Wilcoxon-Mann-Whitney AUC with DeLong confidence intervals. Z test of AUC against chance. Bi-histogram and dot-plot of positive and negative outcome distributions.
-
Report the full set of diagnostic measures at any threshold
Sensitivity and specificity with confidence intervals. Positive and negative likelihood ratios. Positive and negative predictive values. Odds ratio and Youden index. True positives, true negatives, false positives, and false negatives. All the measures journals and reviewers expect, in a single output.
-
Find the optimal decision threshold for your clinical context
Youden index for maximum overall accuracy. Cost-based optimisation when false positives and false negatives carry different consequences. Fixed constraints on minimum sensitivity or specificity. Full decision plot showing how every measure shifts as the threshold moves.
-
Compare diagnostic tests with statistical rigour
DeLong difference in AUC with confidence intervals. Formal tests for equality, equivalence, or non-inferiority of paired or independent tests. The statistical foundation you need before claiming one test outperforms another.
-
Evaluate qualitative test performance
Sensitivity and specificity with Clopper-Pearson exact or Wilson score confidence intervals. Paired and independent test comparison with Newcombe or Tango score confidence intervals on differences. McNemar-Mosteller exact, Fisher exact, and Score Z tests for equality of sensitivity and specificity.
-
Assess performance across patient subgroups
Evaluate diagnostic accuracy across different populations, disease severity strata, or demographic subgroups. Determine whether the test performs consistently across the groups it will be used in clinically — important for both publication and clinical implementation.
-
Part of the Medical edition
Diagnostic accuracy is part of the Medical edition, which also covers survival analysis, logistic regression, Bland-Altman agreement analysis, reference intervals, contingency tables, and the full range of standard hypothesis tests, ANOVA, and regression. One tool for your entire clinical research workflow.
Trusted by 75,000 researchers, analysts, and scientists at leading universities, hospitals, and companies worldwide — for over 30 years.
Try free for 15 days — no credit card required
