Unanchored Binary Analysis

Loading R packages

# install.packages("maicplus")
library(maicplus)

Additional R packages for this vignette:

library(dplyr)

Illustration using example data

This example reads in centered_ipd_sat data that was created in calculating_weights vignette and uses adrs_sat dataset to run binary outcome analysis using the maic_unanchored function by specifying endpoint_type = "binary".

Note that parameters ipd and pseudo_ipd in the maic_unanchored function for binary data analysis needs to have the following columns: USUBJID, ARM, RESPONSE. USUBJID in ipd needs to match USUBJID in weights_object. pseudo_ipd for binary data can be conveniently generated using get_pseudo_ipd_binary function.

Robust standard error for the adjusted result are calculated by sandwich variance estimator in sandwich package with the function vcovHC. Default type of variance estimator (specified by parameter binary_robust_cov_type) is HC3, but other types can be specified. For more information on different types, see vcovHC.

data(centered_ipd_sat)
data(adrs_sat)

centered_colnames <- c("AGE", "AGE_SQUARED", "SEX_MALE", "ECOG0", "SMOKE", "N_PR_THER_MEDIAN")
centered_colnames <- paste0(centered_colnames, "_CENTERED")

weighted_data <- estimate_weights(
  data = centered_ipd_sat,
  centered_colnames = centered_colnames
)

# get dummy binary pseudo IPD
pseudo_adrs <- get_pseudo_ipd_binary(
  binary_agd = data.frame(
    ARM = "B",
    RESPONSE = c("YES", "NO"),
    COUNT = c(280, 120)
  ),
  format = "stacked"
)

result <- maic_unanchored(
  weights_object = weighted_data,
  ipd = adrs_sat,
  pseudo_ipd = pseudo_adrs,
  trt_ipd = "A",
  trt_agd = "B",
  normalize_weight = FALSE,
  endpoint_type = "binary",
  endpoint_name = "Binary Endpoint",
  eff_measure = "OR",
  # binary specific args
  binary_robust_cov_type = "HC3"
)

There are two summaries available in the result: descriptive and inferential. In the descriptive section, we have summaries of events.

result$descriptive
## $summary
##   trt_ind treatment            type   n   events events_pct
## 1       B         B Before matching 400 280.0000   70.00000
## 2       A         A Before matching 500 390.0000   78.00000
## 3       B         B  After matching 400 280.0000   70.00000
## 4       A         A  After matching 500 142.8968   28.57935

In the inferential section, we have the fitted models stored (i.e. logistic regression) and the results from the glm models (i.e. odds ratios and CI). If other effect measures are needed besides odds ratios, we have an option to fit risk ratios or risk differences via eff_measure parameter. Here is the overall summary.

result$inferential$summary
##          case       OR       LCL      UCL        pval
## 1          AB 1.519481 1.1247154 2.052805 0.006417064
## 2 adjusted_AB 1.083350 0.7268601 1.614683 0.694183560

Here are model and results before adjustment.

result$inferential$fit$model_before
## 
## Call:  glm(formula = RESPONSE ~ ARM, family = glm_link, data = dat)
## 
## Coefficients:
## (Intercept)         ARMA  
##      0.8473       0.4184  
## 
## Degrees of Freedom: 899 Total (i.e. Null);  898 Residual
## Null Deviance:       1023 
## Residual Deviance: 1016  AIC: 1020
result$inferential$fit$res_AB_unadj
## $est
## [1] 1.519481
## 
## $se
## [1] 0.2373883
## 
## $ci_l
## [1] 1.124715
## 
## $ci_u
## [1] 2.052805
## 
## $pval
## [1] 0.006417064

Here are model and results after adjustment.

result$inferential$fit$model_after
## 
## Call:  glm(formula = RESPONSE ~ ARM, family = glm_link, data = dat, 
##     weights = weights)
## 
## Coefficients:
## (Intercept)         ARMA  
##     0.84730      0.08006  
## 
## Degrees of Freedom: 899 Total (i.e. Null);  898 Residual
## Null Deviance:       726.7 
## Residual Deviance: 726.5     AIC: 712.5
result$inferential$fit$res_AB
## $est
## [1] 1.08335
## 
## $se
## [1] 0.2275624
## 
## $ci_l
## [1] 0.7268601
## 
## $ci_u
## [1] 1.614683
## 
## $pval
## [1] 0.6941836

Using bootstrap to calculate standard errors

If bootstrap standard errors are preferred, we need to specify the number of bootstrap iteration (n_boot_iteration) in estimate_weights function and proceed fitting maic_unanchored function. Then, the outputs include bootstrapped CI. Different types of bootstrap CI can be found by using parameter boot_ci_type. See boot.ci in boot package for more details.

weighted_data2 <- estimate_weights(
  data = centered_ipd_sat,
  centered_colnames = centered_colnames,
  n_boot_iteration = 100,
  set_seed_boot = 1234
)

result_boot <- maic_unanchored(
  weights_object = weighted_data2,
  ipd = adrs_sat,
  pseudo_ipd = pseudo_adrs,
  trt_ipd = "A",
  trt_agd = "B",
  normalize_weight = FALSE,
  endpoint_type = "binary",
  endpoint_name = "Binary Endpoint",
  eff_measure = "OR",
  boot_ci_type = "perc",
  # binary specific args
  binary_robust_cov_type = "HC3"
)

result_boot$inferential$fit$boot_res_AB
## $est
## [1] 1.08335
## 
## $se
## [1] NA
## 
## $ci_l
## [1] 0.846157
## 
## $ci_u
## [1] 1.796863
## 
## $pval
## [1] NA