Skip to contents

Rashnu

Rashnu is the Zoroastrian deity of truth and justice—the one who weighs souls on a golden scale.
This R package, developed by Zarathu, draws inspiration from Rashnu’s role as the divine judge to offer precision and fairness in sample size determination.

“Where truth is weighed, science begins.”

In clinical trials and research design, every decision matters.
Rashnu helps researchers define the right number of participants for: - Non-inferiority studies - Superiority comparisons (Lakatos method) - One-arm survival designs with transformation-based inference

This package brings clarity, rigor, and justice to your design process.

Installation

You can install the stable version of rashnu from CRAN with:

To access the latest development version, install it from GitHub with:

# install.packages("pak")
pak::pak("zarathucorp/rashnu")

Example

Rstudio Addins

Use rashnuBasic() or Click Interactive Sample Size Calculator Addin

Two sample survival non-inferiority 

twoSurvSampleSizeNI(
   syear = 12,
   yrsurv1 = 0.5,
   yrsurv2 = 0.5,
   alloc = 1,
   accrualTime = 24,
   followTime = 24,
   alpha = 0.025,
   power = 0.8,
   margin = 1.3
)
$Sample_size_of_standard_group
[1] 264

$Sample_size_of_test_group
[1] 264

$Total_sample_size
[1] 528

$Expected_event_numbers_of_standard_group
[1] 227.9

$Expected_event_numbers_of_test_group
[1] 227.9

$Total_expected_event_numbers
[1] 455.9

Two sample survival superiority 

lakatosSampleSize(
   syear = 12,
   yrsurv1 = 0.3,
   yrsurv2 = 0.5,
   alloc = 1,
   accrualTime = 24,
   followTime = 24,
   alpha = 0.05,
   power = 0.8,
   method = "logrank",
   side = "two.sided"
)
$Sample_size_of_standard_group
[1] 58

$Sample_size_of_test_group
[1] 58

$Total_sample_size
[1] 116

$Expected_event_numbers_of_standard_group
[1] 55.6

$Expected_event_numbers_of_test_group
[1] 49.7

$Total_expected_event_numbers
[1] 105.3

$Actual_power
[1] 0.803

One sample non-parametric survival 

oneSurvSampleSize(
   survTime = 12,
   p1 = 0.3,
   p2 = 0.4,
   accrualTime = 24,
   followTime = 24,
   alpha = 0.05,
   power = 0.8,
   side = "two.sided",
   method = "log-log"
)
SampleSize      Power 
   189.000      0.802