Enhanced Survival Analysis Power & Sample Size

Comprehensive power analysis and sample size calculation for survival studies using Lachin-Foulkes and Schoenfeld methods from the gsDesign package. Enhanced with visualizations, export functionality, and clinical examples.

Usage

survivalpower(
  calculation_type = c("sample_size", "power", "events", "hazard_ratio"),
  method = c("lachin_foulkes", "schoenfeld"), 
  hazard_control = 0.083,
  hazard_treatment = 0.042,
  hazard_ratio = 0.6,
  study_duration = 24,
  accrual_duration = 12,
  alpha = 0.025,
  beta = 0.1,
  power = 0.9,
  allocation_ratio = 1,
  sample_size_input = 100,
  events_input = 50,
  sided = c("1", "2"),
  entry_type = c("unif", "expo"),
  gamma = 0,
  dropout_rate = 0,
  show_summary = TRUE,
  show_formulas = FALSE,
  show_interpretation = TRUE,
  show_power_plot = FALSE,
  show_timeline_plot = FALSE,
  power_plot_range = "auto",
  export_results = FALSE,
  export_power_curve = FALSE
)

Arguments

calculation_type: Type of calculation: "sample_size", "power", "events", or "hazard_ratio"
method: Calculation method: "lachin_foulkes" (full study design) or "schoenfeld" (events-based)
hazard_control: Event hazard rate for control group (events per time unit)
hazard_treatment: Event hazard rate for treatment group (events per time unit)
hazard_ratio: Hazard ratio (treatment/control) for Schoenfeld method
study_duration: Maximum study duration (months or years)
accrual_duration: Patient accrual (recruitment) duration
alpha: Type I error rate (significance level)
beta: Type II error rate (1 - power)
power: Statistical power (1 - beta)
allocation_ratio: Randomization ratio (treatment:control)
sample_size_input: Sample size for power calculation
events_input: Number of events for power calculation
sided: Test type: "1" for one-sided, "2" for two-sided
entry_type: Patient entry pattern: "unif" for uniform, "expo" for exponential
gamma: Rate parameter for exponential entry
dropout_rate: Equal dropout hazard rate for both groups
show_summary: Whether to display comprehensive study design summary
show_formulas: Whether to display mathematical formulas
show_interpretation: Whether to include clinical interpretation
show_power_plot: Whether to display power curve visualization
show_timeline_plot: Whether to display study timeline (Lachin-Foulkes only)
power_plot_range: Sample size range for power plots ("auto" or "min,max")
export_results: Whether to export detailed results
export_power_curve: Whether to export power curve data

Details

This enhanced version of survivalpower includes several major improvements:

Calculation Methods:

Lachin-Foulkes Method: Full study design approach accounting for accrual duration, follow-up period, dropout rates, and patient entry patterns
Schoenfeld Method: Events-based approximation using asymptotic normal distribution of log-rank statistic for quick estimates

Calculation Types:

Sample Size: Calculate required sample size for desired power
Power Analysis: Calculate power given sample size and effect size
Events Calculation: Determine number of events needed
Hazard Ratio Detection: Minimum detectable effect size

Enhanced Visualizations:

Power Curves: Interactive plots showing power vs sample size relationships
Study Timeline: Visual representation of accrual and follow-up periods
Effect Detection: Visualization of detectable effect sizes

Export Functionality:

Results Export: Comprehensive results suitable for protocol documents
Power Curve Export: Data points for external analysis and custom plotting
CSV/Excel Compatible: Ready for regulatory submissions

Clinical Examples: See survivalpower_scenarios dataset for pre-configured clinical trial scenarios including oncology, cardiology, neurology, and rare disease studies.

Value

A survivalpowerResults object containing:

power_results: HTML-formatted comprehensive results
formulas: Mathematical formulas and references (optional)
interpretation: Clinical interpretation and recommendations (optional)
power_plot: Power curve visualization (optional)
timeline_plot: Study timeline visualization (optional)
exported_results: Detailed results data frame (optional)
exported_power_curve: Power curve data points (optional)
export_summary: Export summary and usage instructions (optional)

References

Lachin JM, Foulkes MA (1986). Evaluation of sample size and power for analyses of survival with allowance for nonuniform patient entry, losses to follow-up, noncompliance, and stratification. Biometrics 42:507-519.
Schoenfeld D (1981). The asymptotic properties of nonparametric tests for comparing survival distributions. Biometrika 68:316-319.
Jennison C, Turnbull BW (2000). Group Sequential Methods with Applications to Clinical Trials. Chapman & Hall/CRC.
Anderson KM, Seibold H (2019). gsDesign: An R Package for Group Sequential Design. R package version 3.1.1.

Author

ClinicoPath Development Team

Examples

# Example 1: Oncology Phase III Trial - Sample Size Calculation
oncology_trial <- survivalpower(
  calculation_type = "sample_size",
  method = "lachin_foulkes",
  hazard_control = 0.083,      # ~12 month median survival
  hazard_treatment = 0.042,    # ~24 month median survival
  study_duration = 36,         # 3 year study
  accrual_duration = 24,       # 2 year accrual
  alpha = 0.025,              # One-sided 2.5%
  beta = 0.1,                 # 90% power
  show_interpretation = TRUE
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'

# Example 2: Cardiology Trial - Events-Based Quick Calculation
cardio_trial <- survivalpower(
  calculation_type = "events",
  method = "schoenfeld",
  hazard_ratio = 0.75,        # 25% risk reduction
  alpha = 0.025,
  beta = 0.2,                 # 80% power
  show_summary = TRUE
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'

# Example 3: Power Analysis with Visualization
power_analysis <- survivalpower(
  calculation_type = "power",
  method = "schoenfeld", 
  hazard_ratio = 0.6,
  events_input = 150,
  show_power_plot = TRUE,     # Show power curve
  power_plot_range = "50,400"
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'

# Example 4: Complex Study Design with Timeline
complex_study <- survivalpower(
  calculation_type = "sample_size",
  method = "lachin_foulkes",
  hazard_control = 0.05,
  hazard_treatment = 0.03,
  study_duration = 48,
  accrual_duration = 30,
  entry_type = "expo",        # Exponential accrual
  gamma = 1,
  dropout_rate = 0.02,
  allocation_ratio = 2,       # 2:1 randomization
  show_timeline_plot = TRUE,  # Show study phases
  show_formulas = TRUE
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'

# Example 5: Export for External Analysis
export_analysis <- survivalpower(
  calculation_type = "sample_size",
  method = "lachin_foulkes",
  hazard_control = 0.083,
  hazard_treatment = 0.042,
  study_duration = 36,
  accrual_duration = 24,
  export_results = TRUE,      # Export main results
  export_power_curve = TRUE,  # Export power curve
  power_plot_range = "200,600"
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'

# Example 6: Using Pre-configured Scenarios
data(survivalpower_scenarios)
oncology_scenario <- survivalpower_scenarios[1, ]  # Oncology Phase III

scenario_analysis <- survivalpower(
  calculation_type = "sample_size",
  method = "lachin_foulkes",
  hazard_control = oncology_scenario$Control_Hazard_Rate,
  hazard_treatment = oncology_scenario$Treatment_Hazard_Rate,
  study_duration = oncology_scenario$Study_Duration_Months,
  accrual_duration = oncology_scenario$Accrual_Duration_Months,
  alpha = oncology_scenario$Alpha_Level,
  power = oncology_scenario$Target_Power,
  allocation_ratio = oncology_scenario$Allocation_Ratio,
  dropout_rate = oncology_scenario$Dropout_Rate,
  show_interpretation = TRUE
)
#> Error : Could not evaluate 'calculation_type && method'
#>     invalid 'x' type in 'x && y'
#> Error: Could not resolve 'calculation_type && method'