Sample Size Calculation Using Lakatos Method for Survival Analysis

Computes the required sample size and expected event numbers for two-group survival analysis using Lakatos' method under exponential survival assumptions and varying weight functions (log-rank, Gehan, Tarone-Ware).

Usage

lakatosSampleSize(
  syear,
  yrsurv1,
  yrsurv2,
  alloc,
  accrualTime,
  followTime,
  alpha,
  power,
  method = c("logrank", "gehan", "tarone-ware"),
  side = c("two.sided", "one.sided"),
  b = 24
)

Arguments

syear: Survival time horizon in years.
yrsurv1: Survival probability of the standard group at syear.
yrsurv2: Survival probability of the test group at syear.
alloc: Allocation ratio (Test / Standard). For equal allocation, use 1.
accrualTime: Accrual period duration.
followTime: Additional follow-up time after last patient is accrued.
alpha: Significance level (e.g., 0.05 for two-sided tests).
power: Desired statistical power (e.g., 0.8).
method: Weighting method for test statistic. One of "logrank", "gehan", or "tarone-ware".
side: Type of test: "two.sided" or "one.sided".
b: Number of time divisions per year for numerical integration (default = 24).

Value

A list containing:

Sample_size_of_standard_group: Required sample size in the standard group.
Sample_size_of_test_group: Required sample size in the test group.
Total_sample_size: Total sample size.
Expected_event_numbers_of_standard_group: Expected number of events in the standard group.
Expected_event_numbers_of_test_group: Expected number of events in the test group.
Total_expected_event_numbers: Total number of expected events.
Actual_power: Achieved power given the calculated sample size.
error: (Optional) Error message when sample size cannot be calculated.

References

Lakatos E. (1988). Sample sizes based on the log-rank statistic in complex clinical trials. Biometrics, 44, 229–241.

Lakatos E, Lan KK. (1992). A comparison of sample size methods for the logrank statistic. Statistics in Medicine, 11(2), 179–191.

Web calculator (Superiority): https://nshi.jp/en/js/twosurvyr/

Examples

lakatosSampleSize(
  syear = 2,
  yrsurv1 = 0.7,
  yrsurv2 = 0.6,
  alloc = 1,
  accrualTime = 1,
  followTime = 1,
  alpha = 0.05,
  power = 0.8,
  method = "logrank",
  side = "two.sided"
)
#> $Sample_size_of_standard_group
#> [1] 446
#> 
#> $Sample_size_of_test_group
#> [1] 446
#> 
#> $Total_sample_size
#> [1] 892
#> 
#> $Expected_event_numbers_of_standard_group
#> [1] 104.7
#> 
#> $Expected_event_numbers_of_test_group
#> [1] 141.8
#> 
#> $Total_expected_event_numbers
#> [1] 246.5
#> 
#> $Actual_power
#> [1] 0.8
#>