ssutil (version 1.0.0)

sim_power_ni_normal: Empirical Power for Non-Inferiority (Normal Outcomes)

Description

Estimates empirical power to declare non-inferiority between two groups across multiple outcomes using t-tests. Simulates normally distributed data under the null (no difference) and applies non-inferiority rules based on user-defined required and optional tests.

Usage

sim_power_ni_normal(
  nsim,
  npergroup,
  ntest,
  ni_limit,
  test_req,
  test_opt,
  sd,
  corr = 0,
  t_level = 0.95,
  conf.level = 0.95
)

Value

An S3 object of class empirical_power_result, which contains the estimated empirical power and its confidence interval. The object can be printed, formatted, or further processed using associated S3 methods. See also empirical_power_result.

Arguments

nsim

Integer. Number of simulations to perform.

npergroup

Integer. Number of observations per group.

ntest

Integer. Number of tests (outcomes) to compare.

ni_limit

Numeric. Limit to declare non-inferiority. Can be a scalar or vector of length ntest.

test_req

Integer. Number of required tests that must show non-inferiority (first test_req tests).

test_opt

Integer. Number of optional tests that must also show non-inferiority from the remaining tests.

sd

Numeric. Standard deviation(s) of the outcomes. Scalar or vector of length ntest.

corr

Numeric. Correlation between the tests. Scalar (common correlation), or vector of length ntest*(ntest-1)/2.

t_level

Numeric. Confidence level used for the t-tests (e.g., 0.95 for 95% CI).scalar or vector of length ntest.

conf.level

Numeric. Confidence level for the empirical power estimate

Details

A test is considered non-inferior if the lower bound of its confidence interval is greater than the specified non-inferiority limit. Overall non-inferiority is declared if all test_req and at least test_opt of the remaining tests are non-inferior.

See Also

empirical_power_result

Examples

sim_power_ni_normal(
  nsim = 1000,
  npergroup = 250,
  ntest = 7,
  ni_limit = log10(2/3),
  test_req = 2,
  test_opt = 3,
  sd = 0.4,
  corr = 0,
  t_level = 0.05
)