Fixed tidy method for Cumulative Incidence curves calcualted via surv…

[MarcusWalz]

Execution stopped when running the cumulative incidence (CI) example in the doc:

> fitCI <- survfit(Surv(stop, status * as.numeric(event), type = "mstate") ~ 1,
+                  data = mgus1, subset = (start == 0))
> td_multi <- tidy(fitCI)
Error in data.frame(..., check.names = FALSE) : 
  arguments imply differing number of rows: 237, 0, 711

Before the code was 1 - CI estimate, this isn't the correct behavior because survival is not the same as incidence. "Censorship" parts of the matrices outputted in the survfit object are omitted --- state == "". Ideally CI and KM would have their own functions in the survival library because the behavior is quite different. Current code executed and tested by comparing default plot method for a survfit with a ggplot method constructed with tidied survfit (with and without a "strata"). The example in the documentation executes as desired:

> fitCI <- survfit(Surv(stop, status * as.numeric(event), type = "mstate") ~ 1,
+                  data = mgus1, subset = (start == 0))
> td_multi <- tidy(fitCI)
> head(td_multi)
  time n.risk n.event n.censor estimate std.error conf.high conf.low state
1    6      0       0        0        0         0         0        0     2
2    7      0       0        0        0         0         0        0     2
3   31      0       0        0        0         0         0        0     2
4   32      0       0        0        0         0         0        0     2
5   39      0       0        0        0         0         0        0     2
6   60      0       0        0        0         0         0        0     2
> tail(td_multi)
     time n.risk n.event n.censor  estimate  std.error conf.high  conf.low state
469 12689      0       0        1 0.6735070 0.03126710 0.7293816 0.6060959     3
470 12931      0       1        0 0.6856106 0.03183376 0.7422024 0.6165956     3
471 13019      0       0        1 0.6856106 0.03183376 0.7422024 0.6165956     3
472 13152      0       1        0 0.7017487 0.03241725 0.7589737 0.6309371     3
473 14111      0       1        0 0.7178868 0.03136419 0.7731231 0.6492025     3
474 14325      0       0        1 0.7178868 0.03136419 0.7731231 0.6492025     3

[MarcusWalz]

Seems like a lot of changes are being made to multistate survfit in the survival package and backwards compatibility is likely broken. So the previous tidy code was likely broken by version 2.40 of survival which was released in late October and my patch probably won't work for the earlier version. travis-ci fails because it is using version 2.39.

[dgrtwo]

thanks!

[github-actions]

This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.