Use Cases › Overview Group Comparison & ANOVA ↳ Two-Group Comparison ↳ One-Way ANOVA ↳ Two-Way ANOVA ↳ Repeated Measures Regression & Curve Fitting ↳ Linear Regression ↳ Logistic Regression & AUC ↳ Dose-Response & IC50 ↳ Mixed Models (GLMM) Survival Analysis ↳ Kaplan-Meier Workflow ↳ Cox Regression Workflow Normality & Power ↳ Pre-Study Power Calculation ↳ Normality Testing Categorical & Association ↳ Chi-Square Test ↳ Fisher's Exact Test ↳ Odds Ratio & Risk Life Sciences Applications ↳ Flow Cytometry ↳ Gene Expression / Omics ↳ Animal Study Workflow ↳ Clinical Trial Endpoints

Kaplan-Meier Workflow

The Kaplan-Meier curve is the standard tool for visualizing survival data. Licklider generates the curve with a mandatory at-risk table, calculates median survival time with 95% CI, runs the log-rank test, and enforces censoring disclosure — all in a single publication-ready export.

STEP 1 — The Scenario

Researcher Profile

A clinical oncologist analyzing overall survival (OS) data from a Phase II trial comparing a targeted therapy (Arm A, n=44) vs. standard chemotherapy (Arm B, n=44) in advanced non-small cell lung cancer. Follow-up is 24 months.

Research Question

"Is there a significant difference in overall survival between the two treatment arms? What is the median OS and its 95% CI for each arm?"

Dataset

  • 88 patients: Arm A (n=44), Arm B (n=44)
  • Columns: PatientID, Arm, OS_months (time-to-event), Event (1=death, 0=censored)
  • At 24 months: 62 deaths (74 events recorded), 26 censored

STEP 2 — Licklider Analysis Path

1. Data Upload & Event Column Detection

Licklider detects two numeric columns: one time column (OS_months) and one binary event indicator (Event: 0/1). It asks: "Which value in the Event column represents the event of interest (death/relapse)?" → 1=event, 0=censored confirmed.

2. Data Contract Declaration

Observation unit: one patient. Event: death. Censoring: administrative (end of 24-month follow-up) and loss to follow-up. Analysis purpose: confirmatory (pre-specified primary endpoint in the trial protocol).

3. Censoring Summary

Licklider reports: total n, number of events, number censored (and %) per arm. This is a required disclosure before KM estimation proceeds.

4. Kaplan-Meier Estimation

KM survival curves are calculated for each arm. Median OS (time at which survival = 0.5) and its 95% CI (Brookmeyer-Crowley method) are reported. If median is not reached in one arm (survival stays above 50%), this is disclosed explicitly.

5. Log-Rank Test

The log-rank test (Mantel-Cox) is run to compare the survival functions. χ², degrees of freedom, and p-value are reported. If the event rate is low in early time periods, the Gehan-Breslow-Wilcoxon test (weights early events more heavily) is offered as an alternative.

6. Figure Generation

KM curves for both arms with: 95% CI bands (or ± SE whiskers), censoring tick marks, and a mandatory at-risk table below the x-axis. Log-rank p-value annotated on the figure. Colors are colorblind-safe (accessible palette).

STEP 3 — Guards & Disclosures Activated

At-Risk Table Enforcement

Any KM curve exported without an at-risk table is blocked. The at-risk table shows the number of patients still under observation at regular time intervals. This is a mandatory disclosure for all major clinical journals (NEJM, Lancet, JCO).

Censoring Disclosure

The censoring mechanism (administrative censoring at 24 months vs. loss to follow-up) must be declared and disclosed in the Methods text. The number and percentage censored in each arm are required in the figure legend or caption.

Median Survival CI Enforcement

Median OS without 95% CI is not permitted in the export. If median is not reached, the export includes "Median not reached (NR)" with the lower bound of the CI.

Effect Size + CI + N Mandatory (3.1.2)

The log-rank p-value alone is not sufficient. Licklider requires the HR (from a univariate Cox model, if applicable) or the median OS comparison with CIs as the primary effect measure.

N Disclosure (1.2.1)

Total n enrolled, n per arm, n events per arm, and n censored per arm are all disclosed in the figure legend and the Methods text.

STEP 4 — Export Package

  • Figure: KM curves with at-risk table, 95% CI bands, censoring marks, log-rank p annotation — 300 dpi TIFF/PDF
  • Statistical Report: KM table (survival probability at key time points); median OS with 95% CI per arm; log-rank χ², df, p; censoring summary per arm; n events/censored
  • Methods Text: "Overall survival was estimated using the Kaplan-Meier method and compared between arms using the log-rank test (Mantel-Cox). Median OS is reported with 95% CI (Brookmeyer-Crowley method). Censoring was administrative at 24 months and for loss to follow-up. An at-risk table is presented below the survival curve. n=88 (Arm A: 44, Arm B: 44); events: Arm A [n], Arm B [n]."
  • Figure Legend: Arm color coding, median OS ± CI, log-rank p, total n, events n, censored n per arm
  • Preprocessing Audit Log: Event/censoring coding verification; input N = analysis N = 88