Edit tidy.pyears so that strata are tidied.

[Dpananos]

I was working on some survival analyses which required the use of survival::pyears. In particular, I needed to compute person years and number of events within some strata. The code I used was something like

library(tidyverse)
library(survival)


x <- lung %>% 
     select(time, status, sex, age) %>% 
     mutate(
       age = factor(ntile(age, 5))
     )


fit <- pyears(Surv(time, status) ~ sex + age, data=x)

broom::tidy(fit)
#> # A tibble: 2 × 15
#>   pyears.1 pyears.2 pyears.3 pyears.4 pyears.5   n.1   n.2   n.3   n.4   n.5
#>      <dbl>    <dbl>    <dbl>    <dbl>    <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1     18.9     24.8     21.3     22.3     19.7    24    24    28    31    31
#> 2     21.8     17.0     16.1     16.2     12.4    22    22    18    14    14
#> # ℹ 5 more variables: event.1 <dbl>, event.2 <dbl>, event.3 <dbl>,
#> #   event.4 <dbl>, event.5 <dbl>

^{Created on 2025-06-28 with reprex v2.1.1}

You can see that there is one column for each strata, which to my understanding is not tidy!.

I've edited the tidy.pyears function so that the output looks like

library(tidyverse)
library(survival)


x <- lung %>% 
     select(time, status, sex, age, wt.loss) %>% 
     mutate(
       age = factor(ntile(age, 5)),
       wt.loss = factor(ntile(wt.loss, 5))
     )


fit <- pyears(Surv(time, status) ~ sex + age, data=x)

broom::tidy(fit)
#> Joining with `by = join_by(sex, age)`
#> Joining with `by = join_by(sex, age)`
#> # A tibble: 10 × 5
#>    sex   age   pyears     n event
#>    <fct> <fct>  <dbl> <dbl> <dbl>
#>  1 1     1       18.9    24    19
#>  2 2     1       21.8    22    11
#>  3 1     2       24.8    24    19
#>  4 2     2       17.0    22    14
#>  5 1     3       21.3    28    22
#>  6 2     3       16.1    18     8
#>  7 1     4       22.3    31    24
#>  8 2     4       16.2    14    11
#>  9 1     5       19.7    31    28
#> 10 2     5       12.4    14     9

^{Created on 2025-06-28 with reprex v2.1.1}

Now, there is one row per strata,l which to my understanding is tidy.