Enhanced Survival Model Validation — survivalmodelvalidation • ClinicoPath

Comprehensive validation toolkit for survival models including calibration assessment, prediction performance metrics, bootstrap validation, and external validation frameworks. Essential for developing reliable prognostic models.

Usage

survivalmodelvalidation(
  data,
  time_var,
  status_var,
  risk_score,
  covariates,
  stratification_var,
  validation_type = "internal_bootstrap",
  model_type = "cox_ph",
  performance_metrics = "all_metrics",
  calibration_assessment = TRUE,
  calibration_method = "decile_based",
  bootstrap_samples = 1000,
  cv_folds = 10,
  prediction_times = "1,3,5,10",
  confidence_level = 0.95,
  optimism_correction = TRUE,
  shrinkage_estimation = TRUE,
  discrimination_plots = TRUE,
  calibration_plots = TRUE,
  roc_curves = TRUE,
  prediction_error_plots = TRUE,
  risk_distribution_plots = TRUE,
  model_comparison = FALSE,
  decision_curve_analysis = TRUE,
  net_benefit_range = "0.01,0.99",
  external_data_source = "",
  transportability_analysis = FALSE,
  subgroup_validation = TRUE,
  temporal_trends = FALSE,
  missing_data_sensitivity = FALSE,
  clinical_impact_metrics = TRUE,
  reporting_guidelines = "tripod",
  export_validation_report = FALSE
)

Arguments

data: The data as a data frame.
time_var: Time to event or censoring
status_var: Event indicator variable
risk_score: Model predictions to validate
covariates: Covariates for model fitting/validation
stratification_var: Stratification factor
validation_type: Validation methodology
model_type: Model class specification
performance_metrics: Metrics for model assessment
calibration_assessment: Enable calibration evaluation
calibration_method: Calibration evaluation approach
bootstrap_samples: Bootstrap iteration count
cv_folds: CV fold specification
prediction_times: Times for time-dependent metrics
confidence_level: CI level
optimism_correction: Correct for overfitting bias
shrinkage_estimation: Calculate shrinkage coefficient
discrimination_plots: Generate discrimination visualizations
calibration_plots: Generate calibration visualizations
roc_curves: Generate ROC visualizations
prediction_error_plots: Generate error curve plots
risk_distribution_plots: Generate risk distribution plots
model_comparison: Enable model comparison analysis
decision_curve_analysis: Evaluate clinical decision-making utility
net_benefit_range: Threshold range for net benefit
external_data_source: External validation data source
transportability_analysis: Evaluate population transportability
subgroup_validation: Enable subgroup-specific validation
temporal_trends: Assess time-based performance changes
missing_data_sensitivity: Evaluate missing data impact
clinical_impact_metrics: Assess clinical decision impact
reporting_guidelines: Structured reporting format
export_validation_report: Generate exportable report

Value

A results object containing:

`results$instructions`					Analysis instructions and overview
`results$modelSummary`					Summary of the survival model being validated
`results$performanceMetrics`					Comprehensive model performance metrics
`results$calibrationMetrics`					Model calibration metrics and tests
`results$validationResults`					Detailed validation results by method
`results$discriminationAnalysis`					Model discrimination assessment
`results$calibrationAnalysis`					Detailed calibration assessment
`results$subgroupAnalysis`					Performance across subgroups
`results$decisionCurveResults`					Clinical decision-making utility assessment
`results$clinicalImpact`					Clinical relevance and impact metrics
`results$discriminationPlot`					Discrimination visualization
`results$calibrationPlot`					Calibration assessment visualization
`results$rocCurvePlot`					Time-dependent ROC curve visualization
`results$predictionErrorPlot`					Prediction error assessment
`results$riskDistributionPlot`					Risk score distribution by outcome
`results$decisionCurvePlot`					Decision curve visualization
`results$validationReport`					Structured validation report
`results$methodologicalNotes`					Detailed methodology and interpretation guidance
`results$recommendations`					Clinical and methodological recommendations
`results$diagnostics`					Diagnostic information and warnings

Tables can be converted to data frames with asDF or as.data.frame. For example:

results$performanceMetrics$asDF

as.data.frame(results$performanceMetrics)

Examples

# \donttest{
# Example: Cox model validation
survivalmodelvalidation(
    data = clinical_data,
    time = followup_months,
    status = death_event,
    risk_score = predicted_risk,
    validation_type = "internal_bootstrap"
)
# }