Comprehensive Survival Analysis with ClinicoPath

1. Introduction

The survival function in ClinicoPath provides a comprehensive suite of survival analysis tools designed for clinicopathological research. It combines Kaplan-Meier estimation, Cox proportional hazards regression, and a wide range of advanced methods into a single, unified analysis.

What it produces by default (no optional flags needed):

Median survival time with 95% confidence intervals for each group
Cox proportional hazards regression with hazard ratios
Survival probability table at user-specified time points (default: 12, 36, 60 months)

What you can add with a single checkbox:

Kaplan-Meier plots, cumulative events/hazard plots, KMunicate-style plots
Pairwise group comparisons with multiple testing correction
Proportional hazards assumption testing
Restricted Mean Survival Time (RMST)
Person-time analysis with incidence rates
Weighted log-rank tests (Fleming-Harrington family)
Bootstrap internal validation (optimism-corrected C-index)
Calibration curves for Cox model assessment
Non-linearity assessment via restricted cubic splines
Age-adjusted, age-stratified, and age-as-time-scale models
Competing risks and cause-specific survival
REMARK reporting checklist
Natural language summaries and clinical interpretation

Clinical safety features:

Analysis is blocked when fewer than 10 events are detected
Warnings are issued for low event counts (10-19, 20-49)
Competing risk mode automatically skips analyses that are not valid in that context

2. Loading the Data

The survival_test dataset bundled with ClinicoPath contains 200 simulated patients with survival time, binary outcome, treatment group, stage, grade, age, sex, performance status, and a continuous biomarker value. It is designed to demonstrate all features of the survival function.

library(ClinicoPath)
data(survival_test, package = "ClinicoPath")
str(survival_test)

The key columns are:

Column	Description	Type
`elapsedtime`	Survival time in months	numeric
`outcome`	Event indicator (1 = event, 0 = censored)	numeric
`treatment`	Treatment group (Control, Treatment A, Treatment B)	factor
`stage`	Cancer stage (I, II, III, IV)	factor
`grade`	Tumor grade (1, 2, 3)	factor
`age`	Patient age at diagnosis	numeric
`sex`	Patient sex (Female, Male)	factor
`biomarker_value`	Continuous biomarker measurement	numeric

3. Basic Survival Analysis

A minimal call requires three variables: elapsedtime, outcome, and explanatory. The function automatically computes median survival, Cox regression, and survival probability tables. Note that dod, dooc, awd, and awod must be supplied (as empty strings) even when not used.

result <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = ""
)

This produces:

Median Survival Table – median survival time, restricted mean, number of events, and 95% CI for each level of the explanatory variable.
Cox Regression Table – hazard ratios (univariable) with 95% CI and p-values for each level compared to the reference.
Survival Probability Table – survival probabilities at 12, 36, and 60 months (customizable via cutp).

4. Survival Plots

ClinicoPath provides five distinct plot types. Enable each with a single boolean option.

4.1 Kaplan-Meier Survival Curve

result_plots <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  sc = TRUE,
  ci95 = TRUE,
  risktable = TRUE,
  censored = TRUE,
  pplot = TRUE,
  medianline = "hv",
  endplot = 60,
  byplot = 12,
  ybegin_plot = 0,
  yend_plot = 1
)

Plot appearance options:

ci95 – shaded 95% confidence band around each curve
risktable – number-at-risk table below the x-axis
censored – tick marks where censoring occurred
pplot – log-rank p-value annotation
medianline – horizontal (“h”), vertical (“v”), or both (“hv”) median line
endplot, byplot – control x-axis range and tick spacing
ybegin_plot, yend_plot – control y-axis range

4.2 Cumulative Events

Shows the cumulative proportion of events over time (1 minus survival).

result_ce <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  ce = TRUE,
  endplot = 60,
  byplot = 12
)

4.3 Cumulative Hazard

Displays the Nelson-Aalen cumulative hazard estimate.

result_ch <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  ch = TRUE,
  endplot = 60,
  byplot = 12
)

4.4 KMunicate-Style Plot

A publication-ready survival plot following the KMunicate recommendations for improved readability of Kaplan-Meier curves.

result_kmunicate <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  kmunicate = TRUE,
  endplot = 60,
  byplot = 12
)

4.5 Log-Log Plot

Used for visual assessment of the proportional hazards assumption. If the log-log curves are approximately parallel, the proportional hazards assumption is reasonable.

result_loglog <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  loglog = TRUE
)

5. Survival Probability Tables

The survival probability table reports estimates at user-specified time points. Use the cutp argument to define these cutpoints as a comma-separated string.

result_table <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  cutp = "6, 12, 24, 36, 48, 60"
)

The output table includes, for each stratum and time point:

Number at risk
Number of events
Survival probability (with 95% CI)

6. Pairwise Comparisons

When the explanatory variable has more than two levels, pairwise log-rank tests identify which specific groups differ significantly. Multiple testing correction is applied via the padjustmethod option.

result_pw <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  pw = TRUE,
  padjustmethod = "bonferroni"
)

Available adjustment methods: holm (default), hochberg, hommel, bonferroni, BH, BY, fdr, none.

7. Proportional Hazards Assumption Testing

The Cox model assumes proportional hazards – that the hazard ratio between groups is constant over time. This can be tested formally using Schoenfeld residuals and visually via the Schoenfeld residual plot.

result_ph <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  ph_cox = TRUE
)

The output includes:

Formal test – cox.zph() test with chi-square statistic and p-value. A significant p-value (< 0.05) suggests violation.
Interpretation – automated plain-language assessment with recommendations
Schoenfeld residual plot – scaled residuals vs. time; a flat trend supports proportional hazards

8. RMST Analysis

Restricted Mean Survival Time represents the area under the survival curve up to a specified time horizon (tau). It is particularly useful when:

The survival curves cross (non-proportional hazards)
Median survival cannot be estimated (too few events)
A clinically meaningful time horizon exists (e.g., 5-year RMST)

result_rmst <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  rmst_analysis = TRUE,
  rmst_tau = 48
)

If rmst_tau is set to 0 (default), the function automatically uses the 75th percentile of observed follow-up time. A table note indicates the chosen tau.

The RMST table reports, for each group:

RMST estimate (average survival time up to tau)
Standard error
95% confidence interval
The tau value used

9. Person-Time Analysis

Person-time analysis calculates incidence rates that properly account for varying follow-up durations. This is essential for epidemiological studies where participants have unequal observation periods.

result_pt <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  person_time = TRUE,
  time_intervals = "12, 24, 36, 48, 60",
  rate_multiplier = 100
)

The person-time table shows, for each time interval:

Number of events
Total person-time accumulated
Incidence rate (events per rate_multiplier person-time units)
95% confidence interval for the rate

The rate_multiplier controls the denominator: 100 gives rates per 100 person-months, 1000 gives rates per 1000 person-months.

10. Weighted Log-Rank Tests

The standard log-rank test weights all time points equally. Weighted variants from the Fleming-Harrington family emphasize different parts of the survival curve, which is important when hazards are non-proportional.

result_wlr <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  weightedLogRank = TRUE,
  survivalTestType = "fleming_harrington"
)

When enabled, the function runs all four weighted tests simultaneously:

Test	rho	Emphasizes
Log-Rank (standard)	0	All time points equally
Gehan-Breslow-Wilcoxon	–	Early differences
Tarone-Ware	–	Moderate early weighting
Peto-Peto	–	Early-to-mid differences

The survivalTestType option also controls which test is used for pairwise comparisons (via survminer::pairwise_survdiff(rho=...)).

11. Bootstrap Internal Validation

Bootstrap validation provides optimism-corrected estimates of model discriminative ability. It resamples the data, fits the Cox model on each bootstrap sample, and evaluates on both the bootstrap and original data to estimate the optimism bias.

result_boot <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  bootstrapValidation = TRUE,
  bootstrapValN = 200
)

The validation table reports:

Metric	Apparent	Optimism	Corrected
C-index	Training performance	Overfitting bias	Honest estimate
Somers’ Dxy	2 * (C-index - 0.5)	…	…
Calibration slope	Should be near 1	…	…

Important notes:

survival::concordance() treats higher linear predictor as better prognosis by default. The function uses reverse = TRUE because the Cox linear predictor has higher values = worse prognosis.
Minimum 50 resamples, maximum 1000. Default 200 provides a reasonable bias-variance trade-off.
This analysis can be computationally intensive for large datasets.

12. Calibration Curves

Calibration assesses whether predicted survival probabilities match observed outcomes. The function groups patients into risk quantiles, computes Kaplan-Meier survival within each group, and compares it to the model’s predicted survival.

result_cal <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  calibration_curves = TRUE,
  calibration_timepoint = 36,
  calibration_ngroups = 5
)

The output includes:

Calibration metrics table – calibration slope, calibration-in-the-large, with ideal values and interpretation
Calibration by risk group table – predicted vs. observed survival for each risk quantile
Calibration plot – predicted vs. observed with 45-degree perfect calibration reference line

If calibration_timepoint is 0, the median observed follow-up time is used. Each risk group should have at least 20-30 patients for reliable estimates.

13. Non-Linearity Assessment (Restricted Cubic Splines)

For continuous predictors like age, tumor size, or biomarker levels, the relationship with survival may not be linear on the log-hazard scale. RCS analysis tests for non-linearity and visualizes the dose-response curve.

result_rcs <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  rcs_analysis = TRUE,
  rcs_variable = "biomarker_value",
  rcs_knots = 4
)

The output includes:

Non-linearity test table – likelihood ratio test comparing the spline model (non-linear) against a linear model. A significant p-value indicates the relationship is non-linear.
Hazard ratio curve plot – HR as a function of the continuous predictor, with 95% confidence band. The reference value is typically the median.

Knot recommendations: 3 knots (simplest, 2 df), 4 knots (recommended default, 3 df), 5 knots (more flexible, 4 df). Knots are placed at Harrell-recommended percentiles.

14. Age-Adjusted Analysis

Age is the most important confounder in survival studies. ClinicoPath provides multiple approaches to handle age.

14.1 Age-Adjusted Cox Regression

Includes age as a covariate in the Cox model, producing both unadjusted and age-adjusted hazard ratios side by side.

result_age <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  age_adjustment = TRUE,
  age_variable = "age"
)

14.2 Age x Group Interaction

Tests whether the treatment effect varies by age.

result_age_int <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  age_adjustment = TRUE,
  age_variable = "age",
  age_interaction = TRUE
)

14.3 Age-Stratified KM Plots

Displays Kaplan-Meier curves within age groups defined by cutpoints.

result_age_km <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  age_adjustment = TRUE,
  age_variable = "age",
  age_stratified_km = TRUE,
  age_group_cutpoints = "50, 65, 75"
)

14.4 Age as Time Scale

Uses biological age rather than follow-up time as the time axis. This is the most rigorous approach for cancer epidemiology where age is the primary driver of risk.

result_age_ts <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  age_adjustment = TRUE,
  age_variable = "age",
  age_time_scale = TRUE
)

15. Competing Risks Analysis

When multiple causes of death exist, standard survival analysis can overestimate the probability of the event of interest. Competing risks analysis properly accounts for events that preclude observation of the primary outcome.

# Requires a dataset with multi-level outcome variable
# Example using histopathology dataset:
data(histopathology, package = "ClinicoPath")

result_compete <- survival(
  data = histopathology,
  elapsedtime = "OverallTime",
  outcome = "Outcome2",
  outcomeLevel = "",
  explanatory = "Grade_Level",
  multievent = TRUE,
  dod = "DOD",
  dooc = "DOOC",
  awd = "AWD",
  awod = "AWOD",
  analysistype = "compete",
  sc = TRUE
)

Analysis types:

overall – standard overall survival (death from any cause)
cause – cause-specific survival (censors deaths from other causes)
compete – competing risks (cumulative incidence functions)

Important: When analysistype = "compete", the following analyses are automatically skipped because they are not valid for competing risks: Cox regression, pairwise comparisons, RMST, calibration, RCS, bootstrap validation, and person-time analysis.

16. Landmark Analysis

Landmark analysis evaluates survival conditional on surviving past a specified time point. This is useful for assessing the effect of a time-varying treatment or exposure.

result_landmark <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  uselandmark = TRUE,
  landmark = 12
)

All patients who experienced an event or were censored before the landmark time are excluded. The analysis then proceeds on the remaining cohort.

17. Residual Diagnostics

Cox model residuals help identify influential observations and assess model fit.

result_resid <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  residual_diagnostics = TRUE
)

The output includes:

Residuals table – Martingale, deviance, score, and Schoenfeld residuals
Diagnostic plot – visual assessment of residual patterns

18. Clinical Interpretations and Explanations

Two complementary features help users understand their results.

18.1 Analysis Explanations

Provides detailed statistical methodology explanations for each output section (median survival, Cox regression, survival tables, plots, etc.).

result_explain <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  showExplanations = TRUE
)

18.2 Natural Language Summaries

Provides plain-language interpretations of the statistical results, including clinical interpretation, a glossary of terms, and copy-ready sentences for reports.

result_summary <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  showSummaries = TRUE
)

These features produce:

Clinical interpretation – context-aware summary of findings
Clinical glossary – definitions of survival analysis terminology
Copy-ready sentences – publication-ready text for results sections

19. REMARK Reporting Checklist

The REMARK (REporting recommendations for tumor MARKer prognostic studies) checklist helps ensure comprehensive reporting of biomarker-based survival studies.

result_remark <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  remark_checklist = TRUE
)

The checklist shows which REMARK items are addressed by the current analysis configuration and which require additional reporting.

20. Data Export

Survival estimates can be exported for use in external analyses or visualization tools.

result_export <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",
  export_survival_data = TRUE
)

21. Complete Analysis Example

Here is a comprehensive example that enables most features simultaneously.

result_full <- survival(
  data = survival_test,
  elapsedtime = "elapsedtime",
  outcome = "outcome",
  outcomeLevel = "1",
  explanatory = "treatment",
  dod = "",
  dooc = "",
  awd = "",
  awod = "",

  # Survival probability cutpoints

  cutp = "6, 12, 24, 36, 48, 60",

  # Plots
  sc = TRUE,
  ce = TRUE,
  ch = TRUE,
  kmunicate = TRUE,
  loglog = TRUE,
  ci95 = TRUE,
  risktable = TRUE,
  censored = TRUE,
  pplot = TRUE,
  medianline = "hv",
  endplot = 60,
  byplot = 12,

  # Statistical tests
  pw = TRUE,
  padjustmethod = "holm",
  ph_cox = TRUE,

  # Advanced analysis
  rmst_analysis = TRUE,
  rmst_tau = 48,
  weightedLogRank = TRUE,
  person_time = TRUE,
  time_intervals = "12, 24, 36, 48, 60",
  rate_multiplier = 100,

  # Model validation
  bootstrapValidation = TRUE,
  bootstrapValN = 200,
  calibration_curves = TRUE,
  calibration_timepoint = 36,
  calibration_ngroups = 5,

  # Residual diagnostics
  residual_diagnostics = TRUE,

  # Interpretation aids
  showExplanations = TRUE,
  showSummaries = TRUE,
  remark_checklist = TRUE
)

22. Summary of All Options

Option	Type	Default	Description
`elapsedtime`	Variable	–	Survival time variable
`outcome`	Variable	–	Event indicator
`outcomeLevel`	Level	–	Which level = event
`explanatory`	Variable	–	Grouping variable
`cutp`	String	“12, 36, 60”	Cutpoints for survival table
`tint`	Bool	FALSE	Calculate time from dates
`dxdate`	Variable	–	Diagnosis date (if tint)
`fudate`	Variable	–	Follow-up date (if tint)
`timetypedata`	List	ymd	Date format in data
`timetypeoutput`	List	months	Output time unit
`uselandmark`	Bool	FALSE	Landmark analysis
`landmark`	Integer	3	Landmark time
`multievent`	Bool	FALSE	Multiple event levels
`dod`	Level	–	Dead of Disease level
`dooc`	Level	–	Dead of Other Causes level
`awd`	Level	–	Alive with Disease level
`awod`	Level	–	Alive without Disease level
`analysistype`	List	overall	overall/cause/compete
`sc`	Bool	FALSE	Survival curve plot
`kmunicate`	Bool	FALSE	KMunicate-style plot
`ce`	Bool	FALSE	Cumulative events plot
`ch`	Bool	FALSE	Cumulative hazard plot
`loglog`	Bool	FALSE	Log-log plot
`ci95`	Bool	FALSE	95% CI on plots
`risktable`	Bool	FALSE	Risk table on plots
`censored`	Bool	FALSE	Censoring marks
`pplot`	Bool	FALSE	P-value on plots
`medianline`	List	none	Median line style
`endplot`	Integer	60	Plot end time
`byplot`	Integer	12	Time interval for axis
`ybegin_plot`	Number	0.0	Y-axis start
`yend_plot`	Number	1.0	Y-axis end
`pw`	Bool	FALSE	Pairwise comparisons
`padjustmethod`	List	holm	P-value adjustment
`ph_cox`	Bool	FALSE	PH assumption test
`stratified_cox`	Bool	FALSE	Stratified Cox model
`strata_variable`	Variable	–	Stratification variable
`rmst_analysis`	Bool	FALSE	RMST analysis
`rmst_tau`	Number	0	RMST time horizon
`person_time`	Bool	FALSE	Person-time analysis
`time_intervals`	String	“12, 36, 60”	Person-time intervals
`rate_multiplier`	Integer	100	Rate per N person-time
`weightedLogRank`	Bool	FALSE	Weighted log-rank tests
`survivalTestType`	List	logrank	Test family selection
`bootstrapValidation`	Bool	FALSE	Bootstrap validation
`bootstrapValN`	Integer	200	Number of resamples
`calibration_curves`	Bool	FALSE	Calibration assessment
`calibration_timepoint`	Number	0	Timepoint (0=median)
`calibration_ngroups`	Integer	5	Risk group count
`rcs_analysis`	Bool	FALSE	Non-linearity test
`rcs_variable`	Variable	–	Continuous predictor
`rcs_knots`	Integer	4	Number of knots
`residual_diagnostics`	Bool	FALSE	Model diagnostics
`age_adjustment`	Bool	FALSE	Age-adjusted Cox
`age_variable`	Variable	–	Age variable
`age_interaction`	Bool	FALSE	Age x group test
`age_stratified_cox`	Bool	FALSE	Stratify Cox by age
`age_group_cutpoints`	String	“50, 65, 75”	Age group boundaries
`age_time_scale`	Bool	FALSE	Age as time axis
`age_standardization`	Bool	FALSE	SMR computation
`age_standardization_method`	List	indirect	SMR method
`age_stratified_km`	Bool	FALSE	Age-stratified KM
`adjusted_curves`	Bool	FALSE	Age-adjusted curves
`showExplanations`	Bool	FALSE	Method explanations
`showSummaries`	Bool	FALSE	Plain-language summaries
`remark_checklist`	Bool	FALSE	REMARK checklist
`export_survival_data`	Output	–	Export estimates

References

Klein JP, Moeschberger ML (2003). Survival Analysis: Techniques for Censored and Truncated Data. Springer.
Therneau TM, Grambsch PM (2000). Modeling Survival Data: Extending the Cox Model. Springer.
Royston P, Parmar MK (2013). Restricted mean survival time: an alternative to the hazard ratio. BMC Medical Research Methodology 13:152.
Morris TP et al. (2019). Proposals on Kaplan-Meier plots in medical research and a survey of stakeholder views: KMunicate. BMJ Open 9:e030215.
McShane LM et al. (2005). REporting recommendations for tumour MARKer prognostic studies (REMARK). British Journal of Cancer 93:387-391.

ClinicoPath Development Team

2026-03-22

1. Introduction

2. Loading the Data

3. Basic Survival Analysis

4. Survival Plots

4.1 Kaplan-Meier Survival Curve

4.2 Cumulative Events

4.3 Cumulative Hazard

4.4 KMunicate-Style Plot

4.5 Log-Log Plot

5. Survival Probability Tables

6. Pairwise Comparisons

7. Proportional Hazards Assumption Testing

8. RMST Analysis

9. Person-Time Analysis

10. Weighted Log-Rank Tests

11. Bootstrap Internal Validation

12. Calibration Curves

13. Non-Linearity Assessment (Restricted Cubic Splines)

14. Age-Adjusted Analysis

14.1 Age-Adjusted Cox Regression

14.2 Age x Group Interaction

14.3 Age-Stratified KM Plots

14.4 Age as Time Scale

15. Competing Risks Analysis

16. Landmark Analysis

17. Residual Diagnostics

18. Clinical Interpretations and Explanations

18.1 Analysis Explanations

18.2 Natural Language Summaries

19. REMARK Reporting Checklist

20. Data Export

21. Complete Analysis Example

22. Summary of All Options

References