I am fitting a K.M survival model on a data and the summary for model seems fine but whenever I plot it's reversed as it 0 at day 0 and 1 at day 2500 which should be the other way around?
kaplan <- survfit(Surv(time_diff, death)~1, data = new_data)
summary(kaplan)
| time | n.risk | n.event | P((s0)) | P(1) |
|---|---|---|---|---|
| 6 | 84 | 1 | 0.9881 | 0.0119 |
| 7 | 83 | 1 | 0.9762 | 0.0238 |
| 11 | 79 | 1 | 0.9638 | 0.0362 |
| 13 | 78 | 1 | 0.9515 | 0.0485 |
| 18 | 74 | 1 | 0.9386 | 0.0614 |
| 19 | 73 | 1 | 0.9258 | 0.0742 |
| 20 | 72 | 1 | 0.9129 | 0.0871 |
| 21 | 71 | 1 | 0.9000 | 0.1000 |
| 24 | 68 | 1 | 0.8868 | 0.1132 |
| 26 | 67 | 2 | 0.8603 | 0.1397 |
| 29 | 64 | 1 | 0.8469 | 0.1531 |
| 33 | 62 | 1 | 0.8332 | 0.1668 |
| 47 | 57 | 1 | 0.8186 | 0.1814 |
| 56 | 54 | 1 | 0.8035 | 0.1965 |
| 63 | 53 | 1 | 0.7883 | 0.2117 |
| 64 | 52 | 1 | 0.7731 | 0.2269 |
| 75 | 50 | 1 | 0.7577 | 0.2423 |
| 85 | 47 | 1 | 0.7416 | 0.2584 |
| 96 | 46 | 1 | 0.7254 | 0.2746 |
| 117 | 45 | 1 | 0.7093 | 0.2907 |
| 133 | 44 | 1 | 0.6932 | 0.3068 |
| 138 | 42 | 1 | 0.6767 | 0.3233 |
| 139 | 41 | 1 | 0.6602 | 0.3398 |
| 144 | 40 | 1 | 0.6437 | 0.3563 |
| 170 | 37 | 1 | 0.6263 | 0.3737 |
| 180 | 36 | 1 | 0.6089 | 0.3911 |
| 189 | 35 | 1 | 0.5915 | 0.4085 |
| 192 | 34 | 1 | 0.5741 | 0.4259 |
| 210 | 33 | 1 | 0.5567 | 0.4433 |
| 225 | 32 | 1 | 0.5393 | 0.4607 |
| 253 | 30 | 1 | 0.5213 | 0.4787 |
| 258 | 29 | 1 | 0.5033 | 0.4967 |
| 269 | 28 | 1 | 0.4854 | 0.5146 |
| 548 | 19 | 1 | 0.4598 | 0.5402 |
| 603 | 17 | 1 | 0.4328 | 0.5672 |
| 1078 | 11 | 1 | 0.3934 | 0.6066 |
| 1263 | 10 | 1 | 0.3541 | 0.6459 |
| 1388 | 8 | 1 | 0.3098 | 0.6902 |
| 1466 | 7 | 1 | 0.2656 | 0.7344 |
| 2060 | 5 | 1 | 0.2125 | 0.7875 |
| 2097 | 4 | 1 | 0.1593 | 0.8407 |
| 2281 | 3 | 1 | 0.1062 | 0.8938 |
| 2460 | 2 | 1 | 0.0531 | 0.9469 |
| 2828 | 1 | 1 | 0.0000 | 1.0000 |
plot(kaplan)
Sorry if the question is simple as I am new R!
Thanks a lot!
Probably your status variable death is a factor. It shouldn't be. It should be numeric.
Run a toy example below. Here, status is a factor, so plot will look like yours:
time<-1:10
status<-factor(rep(1:0, 5)) # factor!
fit<-survfit(Surv(time, status)~1)
plot(fit)
Now, status is numeric and plot looks like you wish it to look like.
time<-1:10
status<-rep(1:0, 5) # numeric!
fit<-survfit(Surv(time, status)~1)
plot(fit)