Start Here
Not sure which test to use? The Survival Analysis Wizard asks a few questions and gives you a direct recommendation.
Open the Survival Analysis Wizard →Survival Analysis Hub
Analyzing time-to-event data -- survival probabilities, censoring, and adjusted hazard ratios.
Related Calculators
Related Articles
Medical Research Examples
Kaplan-Meier Survival Analysis: An oncologist follows 40 NSCLC patients on immunotherapy (22 deaths, 18 censored): 12-month survival=62%, 24-month survival=31%, median survival=14 months.
Cox Proportional Hazards: 120 post-MI patients followed 3 years: Age HR=1.08 (95% CI 1.03-1.14, p=0.002); concordance c=0.78 (good discrimination).
Hazard Ratio Calculator: A published trial reports HR=0.65 (95% CI 0.48-0.88) for cardiovascular death with a new therapy; the calculator confirms p=0.005, supporting a statistically significant protective effect.
Common Mistakes
- Excluding censored subjects from the analysis instead of using a method (Kaplan-Meier, Cox) designed to handle censoring.
- Comparing survival curves visually without a formal log-rank test or hazard ratio.
- Reporting a Cox hazard ratio without checking the proportional hazards assumption.
- Confusing hazard ratio with relative risk -- they answer related but distinct questions and are not numerically interchangeable.
- Treating a single reported HR as valid across the whole follow-up period when the effect clearly changes over time.
- Not reporting median survival time and its 95% CI alongside the survival curve.
Reporting Recommendations
Journals expect more than a p-value. For each test, report the effect estimate and its 95% confidence interval alongside the p-value:
Kaplan-Meier Survival AnalysisSurvival curve, median survival time, 95% CI, log-rank p-value for group comparison
Hazard Ratio CalculatorHR, 95% CI, p-value
Recommended Learning Order
- Kaplan-Meier Survival Analysis — start with unadjusted survival curves and the log-rank test
- Hazard Ratio Calculator — interpret a reported HR + CI from a published study
- Cox Proportional Hazards Regression — most advanced: adjust hazard ratios for multiple covariates
Frequently Compared Tests
Kaplan-Meier vs. Cox Regression vs. Hazard Ratio
Kaplan-Meier estimates unadjusted survival curves; Cox regression adjusts hazard ratios for multiple covariates (see the Regression Hub); the Hazard Ratio Calculator converts a reported HR/CI or raw event counts into a p-value.
Frequently Asked Questions
What is censoring and why does it matter?
Censoring means a subject left the study (lost to follow-up, withdrew, or the study ended) before their event occurred. Kaplan-Meier and Cox regression correctly account for censored subjects instead of excluding them.
When do I use Kaplan-Meier vs. Cox regression?
Use Kaplan-Meier to estimate and visualize unadjusted survival curves, optionally comparing 2+ groups with a log-rank test. Use Cox regression when you need to adjust the comparison for multiple covariates at once.
What does a hazard ratio (HR) mean?
HR > 1 means increased risk (hazard) relative to the reference group; HR < 1 means a protective effect. An HR of 2.0 means roughly double the instantaneous event rate at any given moment.
What is the proportional hazards assumption?
Cox regression assumes the hazard ratio is constant over the entire follow-up period. Check it with Schoenfeld residuals; if violated, consider time-varying coefficients or a stratified Cox model.
How do I compare survival between two groups?
Use the log-rank test (built into the Kaplan-Meier calculator) for an unadjusted comparison, or include a group covariate in a Cox model for an adjusted hazard ratio.
What is the concordance index (c-statistic)?
It is the probability that, for a random pair of subjects, the one with the higher predicted risk from the model actually has the shorter survival time. c > 0.7 indicates acceptable discrimination.