Public Health
2026-04-1711 min read

Survival Analysis in R: survminer, Cox PH, and Competing Risks

A practical guide to survival analysis in R — from Surv objects to Cox PH to competing risks — for public health and biostatistics researchers.

Sytra Team
Research Engineering Team, Sytra AI

R’s survival analysis ecosystem is arguably the best in any language. The survival package (maintained by Terry Therneau, one of the original Cox model implementers) is the gold standard. survminer adds publication-ready visualizations. Together, they handle everything from Kaplan-Meier curves to competing risks.

Setting Up Survival Data

library(survival)
library(survminer)
 
# Create Surv object (R’s equivalent of Stata’s stset)
surv_obj <- Surv(time = df$time, event = df$status)

The Surv() object encodes both the time and censoring information. Once created, it plugs into all survival functions as the response variable.

Kaplan-Meier Curves

km_fit <- survfit(surv_obj ~ treatment, data = df)
 
# Publication-ready KM plot
ggsurvplot(km_fit,
pval = TRUE,
conf.int = TRUE,
risk.table = TRUE,
xlab = "Time (months)",
ylab = "Survival probability")

ggsurvplot() from survminer produces journal-quality KM curves with risk tables, confidence intervals, and log-rank p-values in a single call. This would take ~30 lines of manual ggplot2 code.

Stop fighting with syntax.

Sytra is an AI research assistant built specifically for statistical computing. No more copy-pasting code into ChatGPT.

Get Early Access

Cox Proportional Hazards

cox_model <- coxph(surv_obj ~ treatment + age + sex,
data = df)
 
# PH assumption test (critical!)
ph_test <- cox.zph(cox_model)
print(ph_test)
ggcoxzph(ph_test)
 
# Forest plot of hazard ratios
ggforest(cox_model)

The cox.zph() function tests the proportional hazards assumption for each covariate and globally. ggcoxzph() visualizes the Schoenfeld residuals. ggforest() produces a forest plot of hazard ratios — the standard visualization for Cox model results in medical journals.

Competing Risks

library(cmprsk)
 
# Cumulative incidence function
cif <- cuminc(ftime = df$time,
fstatus = df$event_type,
group = df$treatment)
plot(cif)
 
# Fine-Gray model
library(tidycmprsk)
fg <- crr(Surv(time, event_type) ~ treatment + age,
data = df, failcode = 1)

Comparison with Stata

R’s survival analysis has two clear advantages: (1) survminer’s visualization tools are significantly better than Stata’s sts graph, and (2) the competing risks ecosystem (cmprsk, tidycmprsk) is more mature and better documented. Stata’s advantage is the unified st prefix: stset, stcox, sts, streg — a consistent interface. R requires learning multiple packages with different APIs.

#Survival Analysis#R#Public Health#Biostatistics

Enjoyed this article?

Related Guides

Public Health11 min read

Survival Analysis in Stata: A Guide for Epidemiologists

A practical guide to survival analysis in Stata — from stset to Cox PH to competing risks. Written for public health researchers and epidemiologists.

R + AI9 min read

Why ChatGPT Produces Invalid R Code for Statistical Analysis

ChatGPT knows R syntax better than Stata. But it still produces statistically invalid code — wrong standard errors, missing diagnostics, and unvalidated inference.

Methodology12 min read

Difference-in-Differences in R: didimputation, did, and fixest

A complete guide to modern DiD estimation in R — using fixest, did, and didimputation — with parallel trends testing and event study plots.

View all articles