Analyze Kaplan Meier curves and log rank contrasts. Input interval events, censoring, and group labels. Export tables, inspect plots, and report corrected comparisons easily.
Use shared time points for all groups. Enter event counts and censored counts for each interval. Leave unused groups unchecked.
This sample matches the default values loaded in the calculator.
| Time | Treatment A Events | Treatment A Censored | Treatment B Events | Treatment B Censored | Treatment C Events | Treatment C Censored |
|---|---|---|---|---|---|---|
| 3 | 4 | 2 | 6 | 3 | 3 | 2 |
| 6 | 6 | 3 | 8 | 3 | 4 | 3 |
| 9 | 7 | 4 | 9 | 4 | 6 | 3 |
| 12 | 8 | 4 | 11 | 5 | 7 | 4 |
| 15 | 9 | 5 | 10 | 5 | 8 | 5 |
At each time point t, survival updates as:
S(t) = S(previous) × (1 − d / n)
where d is the number of events at that time and n is the number at risk just before that time.
For each pair of groups, expected events in Group A at each shared time are:
E(A) = d × nA / (nA + nB)
The comparison statistic uses the sum of observed minus expected events across all times:
Chi-square = [Σ(OA − EA)]² / ΣV
The variance at each time is:
V = nA × nB × d × (n − d) / [n² × (n − 1)]
where n = nA + nB and d is total events in the pair at that time.
Raw pairwise p-values are adjusted using Holm, Bonferroni, or Šidák correction. This reduces false positives when many pairwise tests are reviewed together.
RMST is the area under the estimated survival curve from time zero to the final listed time. It summarizes average survival time over the observed window.
It compares survival curves across several groups using reconstructed Kaplan Meier steps and pairwise log-rank style statistics from grouped interval counts.
Yes. This version assumes shared interval boundaries for every group so event totals and risk sets line up correctly during pairwise comparisons.
Yes. Enter interval event counts, censoring counts, and the starting sample size for each group. Individual subject records are not required here.
Holm is a strong default. It controls family-wise error and is usually less conservative than Bonferroni while keeping interpretation straightforward.
It means the estimated survival curve never dropped to 0.50 or below during the listed follow-up window, so a median cannot be read yet.
Use adjusted p-values for final decisions when many pairwise tests are run. They account for multiplicity and help reduce false positive findings.
Visible separation does not guarantee statistical significance. Small sample sizes, limited events, heavy censoring, or overlapping risk patterns can weaken evidence.
Yes. The page provides CSV export for tables and PDF export for the result panel so you can share calculations and charts easily.
Important Note: All the Calculators listed in this site are for educational purpose only and we do not guarentee the accuracy of results. Please do consult with other sources as well.