Updates to survfitms models in {survival} v 3.6.4 (#208)

* updates

* Update NEWS.md

* updates

* Update ggcuminc.R

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: ggsurvfit
 Title: Flexible Time-to-Event Figures
-Version: 1.0.0.9001
+Version: 1.0.0.9002
 Authors@R: c(
     person("Daniel D.", "Sjoberg", , "danield.sjoberg@gmail.com", role = c("aut", "cre", "cph"),
            comment = c(ORCID = "0000-0003-0862-2018")),
@@ -33,7 +33,7 @@ Imports:
     gtable,
     patchwork (>= 1.1.0),
     rlang (>= 1.0.0),
-    survival (>= 3.4-0),
+    survival (>= 3.6-4),
     tidyr (>= 1.0.0)
 Suggests: 
     covr,

NEWS.md

@@ -2,12 +2,14 @@
 
 * For multi-state models created with `survfit()`, the y-axis label is now "Probability in State". (#205)
 
-* Updated legend position syntax to account for changes in ggplot v3.5.0.
+* Updated legend position syntax to account for changes in {ggplot2} v3.5.0.
 
 * The `tidy_survfit()` (and subsequently `ggsurvfit()`) now honor the `survfit(start.time)` if specified. (#192)
 
 * We now allow for negative follow-up times in `tidy_survfit()` (and subsequently `ggsurvfit()`). When negative follow-up times are present users should specify `survfit(start.time)` and we print a note to this effect when not set. (#192) 
 
+* As of {survival} v3.6-4, the number censored are now returned as a matrix for multi-state models (i.e. competing risks models). The `tidy_survfit()` function has been updated to account for this new structure and the minimum version of {survival} has been increased to the latest version. (#199)
+
 # ggsurvfit 1.0.0
 
 * By default, a model plot created with `ggsurvfit()` or `ggcuminc()` uses the color aesthetic to plot curves by the stratifying variable(s), and further, `ggcuminc()` uses the linetype aesthetic for plots that contain multiple outcomes (i.e. competing events). We now introduce the global option `"ggsurvfit.switch-color-linetype"` to switch these defaults, giving users more flexibility over the output figures. Furthermore, when the `linetype_aes=` argument is called in a situation when it does not apply, it will be silently ignored (previously, an error message _may_ have been thrown). (#166)

R/ggcuminc.R

@@ -1,5 +1,6 @@
 #' Plot Cumulative Incidence
 #'
+#' @description
 #' Plot a cumulative incidence object created with `tidycmprsk::cuminc()`
 #' or a multi-state object created with `survfit2()`.
 #' Read more on multi-state models [here](https://cran.r-project.org/package=survival/vignettes/compete.pdf).

R/tidy_survfit.R

@@ -57,12 +57,12 @@ tidy_survfit <- function(x,
     df_tidy <-
       df_tidy %>%
       dplyr::filter(!.data$state %in% "(s0)") %>%
-      dplyr::select(-dplyr::all_of("n.risk")) %>%
+      dplyr::select(-dplyr::all_of(c("n.risk", "n.censor"))) %>%
       dplyr::left_join(
-        df_tidy %>% dplyr::filter(.data$state %in% "(s0)") %>% dplyr::select(dplyr::any_of(c("strata", "time", "n.risk"))),
+        df_tidy %>% dplyr::filter(.data$state %in% "(s0)") %>% dplyr::select(dplyr::any_of(c("strata", "time", "n.risk", "n.censor"))),
         by = intersect(c("strata", "time"), names(df_tidy))
       ) %>%
-      dplyr::relocate("n.risk", .after = "time") %>%
+      dplyr::relocate(dplyr::any_of(c("n.risk", "n.censor")), .after = "time") %>%
       dplyr::rename(outcome = "state")
   }
 

inst/WORDLIST

@@ -17,6 +17,7 @@ adjuvant
 behaviour
 cowplot
 cuminc
+geom's
 geoms
 geom’
 ggcuminc
@@ -39,6 +40,7 @@ survfit
 tablemab
 tibble
 tidycmprsk
+unmapped
 unremarkably
 unstratified
 vismab