Performs survival analysis with time-dependent variables using the Simon-Makuch method. This analysis is specifically designed for situations where patient characteristics or treatments change during follow-up, such as biomarker changes, treatment switches, or progression events. The Simon-Makuch approach properly accounts for these time-varying exposures in survival estimation and comparison. Key features include Simon-Makuch plots (modified Kaplan-Meier curves), landmark analysis, time-dependent Cox regression, and proper handling of immortal time bias.
Usage
simonmakuch(
data,
survivalTime,
event,
eventLevel,
timeDepVariable,
timeDepTime,
timeDepStatus,
exposedLevel,
enableMultipleTimeDep = FALSE,
additionalTimeDepVars,
analysisType = "comprehensive",
confidenceLevel = 0.95,
performLandmarkAnalysis = FALSE,
landmarkTimes = "6, 12, 24",
landmarkWindow = 1,
performTimeDependentCox = FALSE,
timeDependentCovariates,
testTimeVaryingEffect = FALSE,
assessImmortalTimeBias = FALSE,
naiveComparison = FALSE,
showSimonMakuchPlot = TRUE,
showLandmarkPlots = FALSE,
showCumulativeIncidencePlot = FALSE,
plotTimeRange = "auto",
showConfidenceIntervals = TRUE,
showRiskTables = TRUE,
plotExposureStatus = FALSE,
performLogRankTest = TRUE,
performMantelByarTest = FALSE,
performTimeDepLRTest = FALSE,
handleTieBreaking = "efron",
robustVariance = FALSE,
clusterVariable,
showSurvivalEstimates = TRUE,
showHazardRatios = TRUE,
showLandmarkResults = FALSE,
showExposurePatterns = FALSE,
showModelDiagnostics = FALSE,
performBootstrapValidation = FALSE,
bootstrapSamples = 500,
performSensitivityAnalysis = FALSE,
showExplanations = TRUE,
showMethodologyNotes = FALSE,
includeClinicalGuidance = FALSE
)Arguments
- data
The dataset containing survival and time-dependent variable information.
- survivalTime
Time to event or censoring in consistent units (months recommended). This should be the total follow-up time from study entry.
- event
Event indicator (1 = event occurred, 0 = censored) or factor with event levels. For overall survival, this represents death from any cause.
- eventLevel
The level indicating event occurrence when using factor variables.
- timeDepVariable
The main time-dependent variable of interest (e.g., treatment status, biomarker status, disease progression). This variable can change during follow-up.
- timeDepTime
Time when the time-dependent variable changes status. Use 0 for baseline status, and actual time for when the variable changes (e.g., time of treatment initiation, biomarker conversion).
- timeDepStatus
Status of the time-dependent variable at each time point. For example: "Unexposed", "Exposed", or "Pre-treatment", "Post-treatment".
- exposedLevel
The level that represents the "exposed" or "active" status of the time-dependent variable.
- enableMultipleTimeDep
Enable analysis of multiple time-dependent variables simultaneously. Useful for complex exposure patterns or multiple biomarkers.
- additionalTimeDepVars
Additional time-dependent variables to include in the analysis. Each should have corresponding time and status variables.
- analysisType
Type of Simon-Makuch analysis to perform. Comprehensive includes all methods for thorough time-dependent survival analysis.
- confidenceLevel
Confidence level for survival estimates and statistical tests.
- performLandmarkAnalysis
Perform landmark analysis at specified time points. This method analyzes survival from specific landmark times, avoiding immortal time bias when studying time-dependent exposures.
- landmarkTimes
Comma-separated list of landmark time points for analysis. Patients must survive to each landmark time to be included in that analysis.
- landmarkWindow
Time window around landmark time for determining exposure status. Exposure status is determined within this window before the landmark.
- performTimeDependentCox
Perform Cox regression with time-dependent covariates. This provides hazard ratios that properly account for changing exposure status.
- timeDependentCovariates
Additional covariates to include in time-dependent Cox regression. These can be baseline variables or other time-dependent variables.
- testTimeVaryingEffect
Test whether the effect of the time-dependent variable changes over time (non-proportional hazards for time-dependent covariates).
- assessImmortalTimeBias
Assess and correct for immortal time bias in the analysis. Compares naive analysis (without proper time-dependent handling) with corrected Simon-Makuch analysis.
- naiveComparison
Include comparison with naive analysis that ignores the time-dependent nature of the exposure. This demonstrates the importance of proper time-dependent analysis.
- showSimonMakuchPlot
Display Simon-Makuch survival curves that properly account for time-dependent exposure status changes during follow-up.
- showLandmarkPlots
Display survival curves from each landmark time point, showing how the effect of exposure varies with timing.
- showCumulativeIncidencePlot
Display cumulative incidence curves for time-dependent exposures, useful for understanding exposure patterns over time.
- plotTimeRange
Maximum time for survival plots. Use "auto" for automatic range or specify maximum months (e.g., "60" for 5-year follow-up).
- showConfidenceIntervals
Display confidence intervals around survival curves.
- showRiskTables
Display at-risk tables below survival curves, stratified by time-dependent exposure status.
- plotExposureStatus
Display plot showing how exposure status changes over time for the study population.
- performLogRankTest
Perform log-rank test comparing survival between exposure groups, properly accounting for time-dependent nature.
- performMantelByarTest
Perform Mantel-Byar test, which is specifically designed for comparing survival with time-dependent exposures.
- performTimeDepLRTest
Perform time-dependent log-rank test that allows for changing effects of exposure over time.
- handleTieBreaking
Method for handling tied event times in Cox regression.
- robustVariance
Use robust (sandwich) variance estimation for Cox regression to account for potential model misspecification.
- clusterVariable
Variable identifying clusters for robust variance estimation (e.g., hospital, physician, family).
- showSurvivalEstimates
Display table with survival estimates at key time points, stratified by time-dependent exposure status.
- showHazardRatios
Display hazard ratios from time-dependent Cox regression with confidence intervals and p-values.
- showLandmarkResults
Display detailed results from landmark analysis at each specified time point.
- showExposurePatterns
Display summary of exposure patterns, including time to exposure, exposure duration, and switching patterns.
- showModelDiagnostics
Display diagnostic plots and statistics for time-dependent Cox regression models.
- performBootstrapValidation
Perform bootstrap validation to assess stability of results and provide bias-corrected estimates.
- bootstrapSamples
Number of bootstrap samples for validation analysis.
- performSensitivityAnalysis
Perform sensitivity analysis with different assumptions about exposure timing and immortal time handling.
- showExplanations
Include detailed explanations for Simon-Makuch methodology and interpretation of results.
- showMethodologyNotes
Display detailed notes on the statistical methods used and their appropriate interpretation.
- includeClinicalGuidance
Include guidance for clinical interpretation of time-dependent survival analysis results.
Value
A results object containing:
results$welcomeMessage | a html | ||||
results$simonMakuchExplanation | a html | ||||
results$exposurePatterns | a table | ||||
results$survivalEstimates | a table | ||||
results$timeDependentCox | a table | ||||
results$hazardRatios | a table | ||||
results$statisticalTests | a table | ||||
results$landmarkResults | a table | ||||
results$immortalTimeBias | a table | ||||
results$timeVaryingEffects | a table | ||||
results$modelDiagnostics | a table | ||||
results$bootstrapValidation | a table | ||||
results$sensitivityAnalysis | a table | ||||
results$landmarkAnalysisExplanation | a html | ||||
results$immortalTimeBiasExplanation | a html | ||||
results$timeDependentCoxExplanation | a html | ||||
results$clinicalGuidance | a html | ||||
results$methodologyNotes | a html | ||||
results$simonMakuchPlotExplanation | a html | ||||
results$simonMakuchPlot | an image | ||||
results$landmarkPlots | an image | ||||
results$cumulativeIncidencePlot | an image | ||||
results$exposureStatusPlot | an image | ||||
results$diagnosticPlots | an image |
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$exposurePatterns$asDF
as.data.frame(results$exposurePatterns)
Examples
# Example: Analyzing survival with time-dependent treatment status
# simonmakuch(
# data = patient_data,
# survivalTime = "follow_up_months",
# event = "death_status",
# eventLevel = "Dead",
# timeDepVariable = "treatment_received",
# timeDepTime = "treatment_start_time",
# timeDepStatus = "treatment_status",
# analysisType = "comprehensive"
# )