jSurvival 15: Competing Survival Analysis

Introduction

Competing survival analysis is essential when patients can experience multiple types of events that prevent the occurrence of the primary event of interest. The competingsurvival function in jSurvival provides comprehensive tools for analyzing overall survival, cause-specific survival, and competing risks scenarios using advanced statistical methods.

Learning Objectives:

Understand the differences between overall, cause-specific, and competing risks survival analysis
Master the application of Fine-Gray subdistribution hazard models
Interpret cumulative incidence functions in competing risks scenarios
Apply competing survival methods to real-world clinical datasets
Create publication-ready competing risks plots and tables
Understand when to use each type of survival analysis

Module Overview

The competingsurvival function provides three complementary analysis approaches:

1. Overall Survival Analysis

Purpose: Analyze all-cause mortality regardless of specific cause
Event Definition: Any death (disease-related or other causes) vs. alive
Use Cases: Primary endpoint in many clinical trials, population health studies
Statistical Method: Standard Kaplan-Meier and Cox regression

2. Cause-Specific Survival Analysis

Purpose: Focus exclusively on disease-specific deaths
Event Definition: Disease death vs. censored (including other deaths)
Use Cases: When other causes of death are considered nuisance events
Limitation: May overestimate survival if competing risks are substantial

3. Competing Risks Analysis

Purpose: Account for multiple competing events simultaneously
Event Definition: Disease death vs. other death vs. alive outcomes
Statistical Method: Fine-Gray subdistribution hazard model with cumulative incidence functions
Advantage: Provides realistic probability estimates accounting for competition

library(ClinicoPath)
library(survival)
library(dplyr)

Getting Started

Sample Data Overview

We’ll use a comprehensive competing survival dataset designed specifically for this analysis type.

# Load competing survival data
data(competing_survival_data)

# Dataset overview
cat("Dataset dimensions:", nrow(competing_survival_data), "×", ncol(competing_survival_data), "\n")

# Key survival variables
survival_vars <- c("Overall_Time", "Outcome")
cat("Survival variables:", paste(survival_vars, collapse = ", "), "\n")

# Explanatory variables for stratification
explanatory_vars <- c("Treatment", "Tumor_Stage", "EGFR_Status", "Grade")
cat("Explanatory variables:", paste(explanatory_vars, collapse = ", "), "\n")

# Outcome distribution
outcome_summary <- competing_survival_data %>%
  group_by(Outcome) %>%
  summarise(
    n = n(),
    percentage = round(n() / nrow(competing_survival_data) * 100, 1),
    median_time = round(median(Overall_Time), 1),
    .groups = 'drop'
  )

cat("\nOutcome Distribution:\n")
print(outcome_summary)

# Treatment group summary
treatment_outcome <- competing_survival_data %>%
  group_by(Treatment, Outcome) %>%
  summarise(n = n(), .groups = 'drop') %>%
  tidyr::pivot_wider(names_from = Outcome, values_from = n, values_fill = 0)

cat("\nTreatment by Outcome:\n")
print(treatment_outcome)

Core Competing Survival Analysis

1. Overall Survival Analysis

Analyze all-cause mortality treating any death as an event.

# Overall Survival Analysis Example
# In jamovi: jsurvival > Competing Survival > Analysis Type: Overall

cat("Overall Survival Analysis\n")
cat("========================\n\n")

if(requireNamespace("survival", quietly = TRUE) && 
   requireNamespace("finalfit", quietly = TRUE)) {
  
  # Create overall survival indicator
  os_data <- competing_survival_data %>%
    mutate(
      death_overall = ifelse(Outcome %in% c("Dead_of_Disease", "Dead_of_Other"), 1, 0)
    )
  
  # Kaplan-Meier analysis by treatment
  surv_obj <- Surv(os_data$Overall_Time, os_data$death_overall)
  km_overall <- survfit(surv_obj ~ Treatment, data = os_data)
  
  cat("Overall Survival Summary by Treatment:\n")
  os_summary <- summary(km_overall, times = c(6, 12, 24, 36))$table
  print(os_summary)
  
  # Log-rank test
  logrank_test <- survdiff(surv_obj ~ Treatment, data = os_data)
  p_value <- 1 - pchisq(logrank_test$chisq, 1)
  
  cat("\nLog-rank Test Results:\n")
  cat("Chi-square statistic:", round(logrank_test$chisq, 3), "\n")
  cat("P-value:", round(p_value, 4), "\n")
  
  if(p_value < 0.05) {
    cat("Conclusion: Significant difference in overall survival by treatment\n")
  } else {
    cat("Conclusion: No significant difference in overall survival by treatment\n")
  }
  
  # Cox regression for hazard ratio
  cox_model <- coxph(surv_obj ~ Treatment, data = os_data)
  hr <- exp(cox_model$coefficients)
  hr_ci <- exp(confint(cox_model))
  
  cat("\nHazard Ratio Analysis:\n")
  cat("Hazard Ratio:", round(hr, 2), "\n")
  cat("95% Confidence Interval:", round(hr_ci[1], 2), "-", round(hr_ci[2], 2), "\n")
  
  treatment_levels <- levels(os_data$Treatment)
  cat("Interpretation:", treatment_levels[2], "vs", treatment_levels[1], "\n")
}

2. Cause-Specific Survival Analysis

Focus exclusively on disease-specific deaths, treating other deaths as censored.

# Cause-Specific Survival Analysis Example
# In jamovi: jsurvival > Competing Survival > Analysis Type: Cause Specific

cat("\nCause-Specific Survival Analysis\n")
cat("=================================\n\n")

if(requireNamespace("survival", quietly = TRUE) && 
   requireNamespace("finalfit", quietly = TRUE)) {
  
  # Create cause-specific survival indicator
  css_data <- competing_survival_data %>%
    mutate(
      death_disease = ifelse(Outcome == "Dead_of_Disease", 1, 0)
    )
  
  # Kaplan-Meier analysis for disease-specific survival
  surv_css <- Surv(css_data$Overall_Time, css_data$death_disease)
  km_css <- survfit(surv_css ~ Treatment, data = css_data)
  
  cat("Cause-Specific Survival Summary by Treatment:\n")
  css_summary <- summary(km_css, times = c(6, 12, 24, 36))$table
  print(css_summary)
  
  # Log-rank test for cause-specific survival
  logrank_css <- survdiff(surv_css ~ Treatment, data = css_data)
  p_css <- 1 - pchisq(logrank_css$chisq, 1)
  
  cat("\nCause-Specific Log-rank Test:\n")
  cat("Chi-square statistic:", round(logrank_css$chisq, 3), "\n")
  cat("P-value:", round(p_css, 4), "\n")
  
  # Cox regression for cause-specific hazard ratio
  cox_css <- coxph(surv_css ~ Treatment, data = css_data)
  hr_css <- exp(cox_css$coefficients)
  hr_ci_css <- exp(confint(cox_css))
  
  cat("\nCause-Specific Hazard Ratio:\n")
  cat("Hazard Ratio:", round(hr_css, 2), "\n")
  cat("95% Confidence Interval:", round(hr_ci_css[1], 2), "-", round(hr_ci_css[2], 2), "\n")
  
  # Compare with overall survival
  cat("\nComparison with Overall Survival:\n")
  cat("Disease-specific deaths:", sum(css_data$death_disease), "of", nrow(css_data), 
      "(", round(mean(css_data$death_disease) * 100, 1), "%)\n")
  
  other_deaths <- sum(competing_survival_data$Outcome == "Dead_of_Other")
  cat("Other deaths:", other_deaths, "(", 
      round(other_deaths / nrow(css_data) * 100, 1), "%)\n")
  
  if(other_deaths / nrow(css_data) > 0.1) {
    cat("Note: Substantial competing mortality present - consider competing risks analysis\n")
  }
}

3. Competing Risks Analysis

Account for multiple competing events using Fine-Gray subdistribution hazard models.

# Competing Risks Analysis Example  
# In jamovi: jsurvival > Competing Survival > Analysis Type: Competing Risk

cat("\nCompeting Risks Analysis\n")
cat("=======================\n\n")

if(requireNamespace("survival", quietly = TRUE) && 
   requireNamespace("finalfit", quietly = TRUE) &&
   requireNamespace("cmprsk", quietly = TRUE)) {
  
  # Create competing risks status variable
  crr_data <- competing_survival_data %>%
    mutate(
      status_crr = case_when(
        Outcome %in% c("Alive_w_Disease", "Alive_wo_Disease") ~ 0,  # censored
        Outcome == "Dead_of_Disease" ~ 1,                          # disease death
        Outcome == "Dead_of_Other" ~ 2,                            # competing death
        TRUE ~ 0
      ),
      Treatment_numeric = as.numeric(Treatment) - 1  # for cmprsk
    )
  
  cat("Competing Events Summary:\n")
  event_summary <- crr_data %>%
    group_by(status_crr) %>%
    summarise(
      n = n(),
      percentage = round(n() / nrow(crr_data) * 100, 1),
      .groups = 'drop'
    ) %>%
    mutate(
      event_type = case_when(
        status_crr == 0 ~ "Censored/Alive",
        status_crr == 1 ~ "Disease Death", 
        status_crr == 2 ~ "Other Death"
      )
    )
  print(event_summary[, c("event_type", "n", "percentage")])
  
  # Fine-Gray competing risks regression
  library(cmprsk)
  cuminc_result <- cuminc(
    ftime = crr_data$Overall_Time,
    fstatus = crr_data$status_crr,
    group = crr_data$Treatment
  )
  
  cat("\nCumulative Incidence Function Results:\n")
  
  # Extract cumulative incidence estimates at key time points
  time_points <- c(12, 24, 36)
  
  for (tp in time_points) {
    cat(paste0("\nAt ", tp, " months:\n"))
    
    # Find closest time point in results
    for (i in 1:length(cuminc_result)) {
      group_name <- names(cuminc_result)[i]
      times <- cuminc_result[[i]]$time
      est <- cuminc_result[[i]]$est
      
      closest_idx <- which.min(abs(times - tp))
      if (length(closest_idx) > 0) {
        cat(paste0(group_name, ": ", round(est[closest_idx] * 100, 1), "%\n"))
      }
    }
  }
  
  # Fine-Gray regression for treatment effect
  crr_fit <- crr(
    ftime = crr_data$Overall_Time,
    fstatus = crr_data$status_crr, 
    cov1 = crr_data$Treatment_numeric,
    failcode = 1  # disease death
  )
  
  cat("\nFine-Gray Regression Results (Disease Death):\n")
  cat("Subdistribution Hazard Ratio:", round(exp(crr_fit$coef), 2), "\n")
  cat("95% Confidence Interval:", round(exp(crr_fit$coef - 1.96*sqrt(crr_fit$var)), 2), 
      "-", round(exp(crr_fit$coef + 1.96*sqrt(crr_fit$var)), 2), "\n")
  cat("P-value:", round(2 * (1 - pnorm(abs(crr_fit$coef / sqrt(crr_fit$var)))), 4), "\n")
  
  # Clinical interpretation
  cat("\nClinical Interpretation:\n")
  if (exp(crr_fit$coef) > 1) {
    cat("Experimental therapy associated with higher subdistribution hazard of disease death\n")
  } else {
    cat("Experimental therapy associated with lower subdistribution hazard of disease death\n")
  }
}

Advanced Applications

4. Stratified Analysis by Disease Stage

Evaluate competing survival patterns across different disease stages.

# Stratified Competing Risks Analysis by Tumor Stage
cat("\nStratified Analysis by Tumor Stage\n")
cat("==================================\n\n")

# Stage distribution
stage_dist <- competing_survival_data %>%
  group_by(Tumor_Stage) %>%
  summarise(
    n = n(),
    disease_deaths = sum(Outcome == "Dead_of_Disease"),
    other_deaths = sum(Outcome == "Dead_of_Other"),
    alive = sum(Outcome %in% c("Alive_w_Disease", "Alive_wo_Disease")),
    .groups = 'drop'
  ) %>%
  mutate(
    disease_death_rate = round(disease_deaths / n * 100, 1),
    other_death_rate = round(other_deaths / n * 100, 1)
  )

cat("Disease and Other Death Rates by Stage:\n")
print(stage_dist[, c("Tumor_Stage", "n", "disease_death_rate", "other_death_rate")])

# Competing risks analysis by stage
if(requireNamespace("cmprsk", quietly = TRUE)) {
  stage_data <- competing_survival_data %>%
    mutate(
      status_crr = case_when(
        Outcome %in% c("Alive_w_Disease", "Alive_wo_Disease") ~ 0,
        Outcome == "Dead_of_Disease" ~ 1,
        Outcome == "Dead_of_Other" ~ 2,
        TRUE ~ 0
      )
    )
  
  # Cumulative incidence by stage
  stage_cuminc <- cuminc(
    ftime = stage_data$Overall_Time,
    fstatus = stage_data$status_crr,
    group = stage_data$Tumor_Stage
  )
  
  cat("\nCumulative Incidence at 24 months by Stage:\n")
  for (i in 1:length(stage_cuminc)) {
    group_name <- names(stage_cuminc)[i]
    times <- stage_cuminc[[i]]$time
    est <- stage_cuminc[[i]]$est
    
    # Find 24-month estimate
    closest_24m <- which.min(abs(times - 24))
    if (length(closest_24m) > 0) {
      cat(paste0(group_name, ": ", round(est[closest_24m] * 100, 1), "%\n"))
    }
  }
  
  # Trend analysis across stages
  stage_numeric <- as.numeric(factor(stage_data$Tumor_Stage, 
                                    levels = c("I", "II", "III", "IV")))
  
  stage_crr <- crr(
    ftime = stage_data$Overall_Time,
    fstatus = stage_data$status_crr,
    cov1 = stage_numeric,
    failcode = 1
  )
  
  cat("\nTrend Analysis Across Stages (per stage increase):\n")
  cat("Subdistribution HR:", round(exp(stage_crr$coef), 2), "\n")
  cat("95% CI:", round(exp(stage_crr$coef - 1.96*sqrt(stage_crr$var)), 2), 
      "-", round(exp(stage_crr$coef + 1.96*sqrt(stage_crr$var)), 2), "\n")
  
  trend_p <- 2 * (1 - pnorm(abs(stage_crr$coef / sqrt(stage_crr$var))))
  cat("P for trend:", round(trend_p, 4), "\n")
  
  if (trend_p < 0.05) {
    cat("Conclusion: Significant trend of increasing disease death risk with advancing stage\n")
  }
}

5. Biomarker-Based Competing Risks Analysis

Evaluate prognostic biomarkers in the context of competing risks.

# Biomarker Analysis in Competing Risks Context
cat("\nBiomarker Analysis: EGFR Status\n")
cat("===============================\n\n")

# EGFR status distribution
egfr_summary <- competing_survival_data %>%
  group_by(EGFR_Status, Outcome) %>%
  summarise(n = n(), .groups = 'drop') %>%
  group_by(EGFR_Status) %>%
  mutate(
    total = sum(n),
    percentage = round(n / total * 100, 1)
  )

cat("Outcome Distribution by EGFR Status:\n")
print(egfr_summary[, c("EGFR_Status", "Outcome", "n", "percentage")])

# Competing risks analysis by EGFR status
if(requireNamespace("cmprsk", quietly = TRUE)) {
  egfr_data <- competing_survival_data %>%
    mutate(
      status_crr = case_when(
        Outcome %in% c("Alive_w_Disease", "Alive_wo_Disease") ~ 0,
        Outcome == "Dead_of_Disease" ~ 1,
        Outcome == "Dead_of_Other" ~ 2,
        TRUE ~ 0
      ),
      EGFR_numeric = ifelse(EGFR_Status == "Positive", 1, 0)
    )
  
  # Cumulative incidence by EGFR status
  egfr_cuminc <- cuminc(
    ftime = egfr_data$Overall_Time,
    fstatus = egfr_data$status_crr,
    group = egfr_data$EGFR_Status
  )
  
  cat("\nCumulative Incidence Comparison (EGFR Positive vs Negative):\n")
  
  time_points <- c(12, 24, 36)
  for (tp in time_points) {
    cat(paste0("\nAt ", tp, " months:\n"))
    
    for (i in 1:length(egfr_cuminc)) {
      group_name <- names(egfr_cuminc)[i]
      times <- egfr_cuminc[[i]]$time
      est <- egfr_cuminc[[i]]$est
      
      closest_idx <- which.min(abs(times - tp))
      if (length(closest_idx) > 0) {
        cat(paste0("  ", group_name, ": ", round(est[closest_idx] * 100, 1), "%\n"))
      }
    }
  }
  
  # Fine-Gray regression for EGFR effect
  egfr_crr <- crr(
    ftime = egfr_data$Overall_Time,
    fstatus = egfr_data$status_crr,
    cov1 = egfr_data$EGFR_numeric,
    failcode = 1
  )
  
  cat("\nEGFR Status Effect on Disease Death (Positive vs Negative):\n")
  cat("Subdistribution HR:", round(exp(egfr_crr$coef), 2), "\n")
  cat("95% CI:", round(exp(egfr_crr$coef - 1.96*sqrt(egfr_crr$var)), 2), 
      "-", round(exp(egfr_crr$coef + 1.96*sqrt(egfr_crr$var)), 2), "\n")
  
  egfr_p <- 2 * (1 - pnorm(abs(egfr_crr$coef / sqrt(egfr_crr$var))))
  cat("P-value:", round(egfr_p, 4), "\n")
  
  if (egfr_p < 0.05) {
    if (exp(egfr_crr$coef) > 1) {
      cat("Interpretation: EGFR-positive tumors have significantly higher disease death risk\n")
    } else {
      cat("Interpretation: EGFR-positive tumors have significantly lower disease death risk\n")
    }
  } else {
    cat("Interpretation: No significant association between EGFR status and disease death risk\n")
  }
}

Statistical Considerations

Model Assumptions and Validation

Important considerations for competing risks analysis.

# Statistical Assumptions and Model Validation
cat("Statistical Considerations for Competing Risks Analysis\n")
cat("======================================================\n\n")

# Sample size considerations
total_events <- sum(competing_survival_data$Outcome %in% c("Dead_of_Disease", "Dead_of_Other"))
disease_events <- sum(competing_survival_data$Outcome == "Dead_of_Disease")
other_events <- sum(competing_survival_data$Outcome == "Dead_of_Other")

cat("Sample Size Assessment:\n")
cat("Total patients:", nrow(competing_survival_data), "\n")
cat("Total events:", total_events, "(", round(total_events/nrow(competing_survival_data)*100, 1), "%)\n")
cat("Disease deaths:", disease_events, "(", round(disease_events/nrow(competing_survival_data)*100, 1), "%)\n")
cat("Other deaths:", other_events, "(", round(other_events/nrow(competing_survival_data)*100, 1), "%)\n")

# Power considerations
min_events_per_group <- 10
groups <- length(unique(competing_survival_data$Treatment))
cat("\nPower Considerations:\n")
cat("Minimum events per group recommended:", min_events_per_group, "\n")
cat("Number of treatment groups:", groups, "\n")

disease_events_per_group <- disease_events / groups
cat("Disease events per group:", round(disease_events_per_group, 1), "\n")

if (disease_events_per_group >= min_events_per_group) {
  cat("Assessment: Adequate power for disease death analysis\n")
} else {
  cat("Assessment: Limited power - consider longer follow-up or larger sample size\n")
}

# Competing risk proportion
competing_proportion <- other_events / total_events
cat("\nCompeting Risk Assessment:\n")
cat("Proportion of deaths from other causes:", round(competing_proportion * 100, 1), "%\n")

if (competing_proportion >= 0.2) {
  cat("Recommendation: Competing risks analysis strongly recommended\n")
  cat("Rationale: Substantial competing mortality may bias cause-specific analysis\n")
} else if (competing_proportion >= 0.1) {
  cat("Recommendation: Consider competing risks analysis\n")
  cat("Rationale: Moderate competing mortality present\n")
} else {
  cat("Recommendation: Cause-specific analysis may be adequate\n")
  cat("Rationale: Limited competing mortality\n")
}

# Follow-up adequacy
median_followup <- median(competing_survival_data$Overall_Time)
max_followup <- max(competing_survival_data$Overall_Time)

cat("\nFollow-up Assessment:\n")
cat("Median follow-up:", round(median_followup, 1), "months\n")
cat("Maximum follow-up:", round(max_followup, 1), "months\n")

if (median_followup >= 24) {
  cat("Assessment: Adequate follow-up for survival analysis\n")
} else {
  cat("Assessment: Limited follow-up - interpret results cautiously\n")
}

Interpretation Guidelines

Framework for interpreting competing risks results.

# Interpretation Guidelines for Competing Risks Analysis
cat("\nInterpretation Guidelines\n")
cat("========================\n\n")

cat("1. Overall vs Cause-Specific vs Competing Risks:\n")
cat("   • Overall Survival: Most clinically relevant for patients\n")
cat("   • Cause-Specific: Useful for understanding disease-specific effects\n") 
cat("   • Competing Risks: Most appropriate when competing events are common\n\n")

cat("2. Subdistribution Hazard Ratio Interpretation:\n")
cat("   • HR > 1: Increased risk of the event of interest\n")
cat("   • HR < 1: Decreased risk of the event of interest\n")
cat("   • Accounts for competing events in the calculation\n\n")

cat("3. Cumulative Incidence Function (CIF):\n")
cat("   • Probability of experiencing the event by time t\n")
cat("   • Accounts for competing events\n")
cat("   • Sum of all CIFs ≤ 1 (remaining probability is survival)\n\n")

cat("4. Clinical Decision Making:\n")
cat("   • Use overall survival for prognosis discussions\n")
cat("   • Use cause-specific for treatment effect evaluation\n")
cat("   • Use competing risks for realistic risk prediction\n\n")

cat("5. Reporting Requirements:\n")
cat("   • Report number at risk and events for each time point\n")
cat("   • Include confidence intervals for all estimates\n")
cat("   • Specify the statistical method used\n")
cat("   • Justify choice between cause-specific and competing risks\n")

Integration with jamovi

Using competingsurvival in jamovi

Step-by-step guide for jamovi users.

# jamovi Integration Guide for Competing Survival
cat("Using competingsurvival in jamovi\n")
cat("=================================\n\n")

cat("Step-by-Step Instructions:\n\n")

cat("1. Data Preparation:\n")
cat("   • Ensure survival time variable (continuous, numeric)\n")
cat("   • Create outcome variable with 4 levels:\n")
cat("     - Dead of Disease\n")
cat("     - Dead of Other Causes\n") 
cat("     - Alive with Disease\n")
cat("     - Alive without Disease\n")
cat("   • Prepare explanatory variable (categorical)\n\n")

cat("2. Access the Function:\n")
cat("   • Navigate to: jSurvival → Competing Survival\n")
cat("   • Or use: Analyses → ClinicoPath → jSurvival → Competing Survival\n\n")

cat("3. Variable Assignment:\n")
cat("   • Explanatory Variable: Select grouping variable\n")
cat("   • Overall Time (in months): Select survival time column\n")
cat("   • Outcome: Select outcome variable\n")
cat("   • Configure outcome levels:\n")
cat("     - Dead of Disease: Select appropriate level\n")
cat("     - Dead of Other: Select appropriate level\n")
cat("     - Alive w Disease: Select appropriate level\n")
cat("     - Alive w/o Disease: Select appropriate level\n\n")

cat("4. Analysis Selection:\n")
cat("   • Overall: All-cause mortality analysis\n")
cat("   • Cause Specific: Disease-specific mortality only\n")
cat("   • Competing Risk: Full competing risks analysis\n\n")

cat("5. Output Interpretation:\n")
cat("   • Review analysis summary for methods used\n")
cat("   • Examine survival table for hazard ratios\n")
cat("   • For competing risks: review cumulative incidence table\n")
cat("   • For competing risks: examine competing risks plot\n")
cat("   • Read clinical interpretation section\n\n")

cat("6. Troubleshooting:\n")
cat("   • Ensure no missing values in key variables\n")
cat("   • Verify outcome variable has exactly 4 levels\n")
cat("   • Check that follow-up times are positive\n")
cat("   • Ensure adequate sample sizes per group\n")
cat("   • Verify appropriate level assignments\n")

Conclusion

The competingsurvival function provides comprehensive tools for analyzing survival data when multiple competing events can occur. The choice between overall survival, cause-specific survival, and competing risks analysis depends on the research question, the proportion of competing events, and the clinical context.

Key Takeaways

Method Selection: Choose the appropriate analysis type based on research objectives and competing event frequency
Clinical Relevance: Overall survival is most clinically relevant; competing risks provides most realistic probabilities
Statistical Rigor: Use Fine-Gray models when competing events comprise >10-20% of total events
Interpretation: Subdistribution hazard ratios account for competing events in their calculation
Visualization: Cumulative incidence plots provide intuitive displays of competing risks

Best Practices

Always report the proportion of competing events
Justify the choice of analysis method
Present confidence intervals for all estimates
Consider sensitivity analyses using different methods
Validate findings with clinical domain experts

Next Steps

Explore multivariable competing risks models
Consider time-dependent covariates in competing risks
Integrate findings with other ClinicoPath survival modules
Develop prognostic models incorporating competing risks

This vignette demonstrates comprehensive competing survival analysis methodology using the competingsurvival function. Apply these principles to your own clinical datasets for robust survival analysis accounting for competing events.

ClinicoPath Development Team

2025-07-13

Introduction

Module Overview

1. Overall Survival Analysis

2. Cause-Specific Survival Analysis

3. Competing Risks Analysis

Getting Started

Sample Data Overview

Core Competing Survival Analysis

1. Overall Survival Analysis

2. Cause-Specific Survival Analysis

3. Competing Risks Analysis

Advanced Applications

4. Stratified Analysis by Disease Stage

5. Biomarker-Based Competing Risks Analysis

Statistical Considerations

Model Assumptions and Validation

Interpretation Guidelines

Integration with jamovi

Using competingsurvival in jamovi

Conclusion

Key Takeaways

Best Practices

Next Steps