G-Computation Risk-Ratio Inference for KK Designs with Binary Responses

Fits the same all-subject logistic working model as InferenceIncidKKGCompRiskDiff for a KK incidence outcome using treatment and, optionally, all recorded covariates, then estimates the marginal (standardized, g-computation) risk ratio \(\hat\theta = \left(n^{-1}\sum_i \hat p(1, X_i)\right) / \left(n^{-1}\sum_i \hat p(0, X_i)\right)\) by averaging fitted-model predicted risks under all-treated and all-control assignments over the empirical covariate distribution (Robins 1986). Matched pairs are treated as clusters and reservoir subjects are treated as singletons when computing the sandwich covariance; the delta method is applied on the log-risk-ratio scale to keep the reported ratio and its confidence interval positive, then back-transformed for reporting.

Details

This estimator has likelihood_tier = "none". If the working model fails to converge, has no treatment-arm variation, or the all-control standardized risk is zero (undefined ratio), the estimate is marked non-estimable via is_nonestimable().

References

Robins, J. (1986). A new approach to causal inference in mortality studies with a sustained exposure period. Mathematical Modelling, 7(9-12), 1393-1512. doi:10.1016/0270-0255(86)90088-6

Super class

Inference -> InferenceIncidKKGCompRiskRatio

Methods

Public methods

• InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_function()

• InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_cpp()

• InferenceIncidKKGCompRiskRatio$approximate_randomization_distribution_beta_hat_T()

• InferenceIncidKKGCompRiskRatio$supports_rand_pval_for_incidence()

• InferenceIncidKKGCompRiskRatio$clone()

+ inherited public methods from Inference

• Inference$capabilities()
• Inference$compute_asymp_confidence_interval()
• Inference$compute_asymp_two_sided_pval()
• Inference$compute_estimate()
• Inference$compute_exact_confidence_interval()
• Inference$compute_exact_two_sided_pval_for_treatment_effect()
• Inference$duplicate()
• Inference$get_analysis_data()
• Inference$get_covariates()
• Inference$get_design_object()
• Inference$get_model_formula()
• Inference$get_nonestimable_reason()
• Inference$get_nonestimable_stage()
• Inference$get_optimization_alg()
• Inference$get_response()
• Inference$get_response_type()
• Inference$get_treatment()
• Inference$initialize()
• Inference$is_nonestimable()
• Inference$set_optimization_alg()
• Inference$set_seed()
• Inference$supports()

---

InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_function()

Usage

InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_function(
  custom_randomization_statistic_function
)

---

InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_cpp()

Usage

InferenceIncidKKGCompRiskRatio$set_custom_randomization_statistic_cpp(fn)

---

InferenceIncidKKGCompRiskRatio$approximate_randomization_distribution_beta_hat_T()

Usage

InferenceIncidKKGCompRiskRatio$approximate_randomization_distribution_beta_hat_T(
  r = 501,
  delta = 0,
  transform_responses = "none",
  show_progress = TRUE,
  permutations = NULL,
  debug = FALSE,
  zero_one_logit_clamp = .Machine$double.eps
)

---

InferenceIncidKKGCompRiskRatio$supports_rand_pval_for_incidence()

Usage

InferenceIncidKKGCompRiskRatio$supports_rand_pval_for_incidence()

---

InferenceIncidKKGCompRiskRatio$clone()

The objects of this class are cloneable with this method.

Usage

InferenceIncidKKGCompRiskRatio$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# \donttest{
seq_des = DesignSeqOneByOneKK14$new(n = 10, response_type = 'incidence')
for (i in 1:10) {
  seq_des$add_one_subject_to_experiment_and_assign(data.frame(x1 = rnorm(1), x2 = rnorm(1)))
}
seq_des$add_all_subject_responses(rbinom(10, 1, 0.5))
inf = InferenceIncidKKGCompRiskRatio$new(seq_des)
inf$compute_estimate()
#> [1] 1.490008
# }