Input checking, minor rearrangement, and docs for coh_effectiveness()

R/coh_effectiveness.R

@@ -19,29 +19,34 @@
 #' @return summary: hazards ratio (CI95%), vaccine effectiveness (CI95%),
 #' and Schoenfeld test
 #' @examples
-#' \dontrun{
+#' # load example data from package
 #' data("cohortdata")
+#'
+#' # add immunization dates
 #' cohortdata$immunization_death <- get_immunization_date(
-#'   date = cohortdata,
+#'   data = cohortdata,
 #'   outcome_date_col = "death_date",
 #'   outcome_delay = 0,
-#'   immunisation_delay = 14,
+#'   immunization_delay = 14,
 #'   vacc_date_col = c("vaccine_date_1", "vaccine_date_2"),
 #'   end_cohort = "2021-12-31",
 #'   take_first = FALSE
 #' )
 #'
+#' # add vaccine status
 #' cohortdata$vaccine_status <- set_status(
 #'   data = cohortdata,
 #'   col_names = c("immunization_death"),
 #'   status = c("v", "u")
 #' )
 #'
+#' # add death status
 #' cohortdata$death_status <- set_status(
 #'   data = cohortdata,
 #'   col_names = c("death_date")
 #' )
 #'
+#' # add time to death
 #' cohortdata$time_to_death <- get_time_to_event(
 #'   data = cohortdata,
 #'   outcome_date_col = "death_date",
@@ -50,30 +55,51 @@
 #'   FALSE
 #' )
 #'
+#' # estimate vaccine effectiveness
 #' coh_eff_noconf(
 #'   cohortdata,
 #'   "death_status",
 #'   "time_to_death",
 #'   "vaccine_status"
 #' )
-#' }
 #' @export
 coh_eff_noconf <- function(data,
                            outcome_status_col,
                            time_to_event_col,
                            status_vacc_col,
                            p_thr = 0.05) {
-  cx <- survival::coxph(survival::Surv(
-    data[[time_to_event_col]],
-    data[[outcome_status_col]]
+
+  # input checking
+  checkmate::assert_data_frame(
+    data,
+    min.rows = 1L
+  )
+  checkmate::assert_names(
+    names(data),
+    must.include = c(outcome_status_col, time_to_event_col, status_vacc_col)
+  )
+  checkmate::assert_number(p_thr, lower = 0.0, upper = 1.0)
+
+  # cox regression
+  cx <- survival::coxph(
+    survival::Surv(
+      data[[time_to_event_col]],
+      data[[outcome_status_col]]
+    ) ~ data[[status_vacc_col]]
   )
-  ~ data[[status_vacc_col]])
+
+  # Test the Proportional Hazards Assumption
   test <- survival::cox.zph(cx)
-  hr <- c(round(exp(stats::coef(cx)), 4)[1])
-  ci025 <- c(round(exp(stats::confint(cx)), 4)[1])
-  ci975 <- c(round(exp(stats::confint(cx)), 4)[2])
-  p <- test$table[5]
-  p_value <- c(format(p, digits = 3))
+  hr <- round(exp(stats::coef(cx)), digits = 4)
+
+  # extract first and second element as limits
+  ci025 <- round(exp(stats::confint(cx)), 4)[1]
+  ci975 <- round(exp(stats::confint(cx)), 4)[2]
+
+  # extract from matrix by name
+  p <- test$table["GLOBAL", "p"]
+
+  p_value <- format(p, digits = 3)
   if (p < p_thr) {
     ph <- "reject"
   } else {