ahr split function with simple = TRUE / FALSE

[elong0527]

The output of a function should be the same structure.

The output wen simple = FALSE is not even generate AHR. Propose to split them into two functions:

ahr -> original one with simple = TRUE
ahr_hr -> report hr of each period.

>  ahr(enroll_rate = enroll_rate, fail_rate = fail_rate, total_duration = c(15, 30))
# A tibble: 2 × 5
   time   ahr event  info info0
  <dbl> <dbl> <dbl> <dbl> <dbl>
1    15 0.812  91.0  22.3  22.7
2    30 0.712 222.   54.4  55.4
> ahr(enroll_rate = enroll_rate, fail_rate = fail_rate, total_duration = c(15, 30), simple = FALSE)
# A tibble: 4 × 7
   time stratum     t    hr event  info info0
  <dbl> <chr>   <dbl> <dbl> <dbl> <dbl> <dbl>
1    15 All         0   1    53.9 13.5  13.5 
2    15 All         4   0.6  37.1  8.86  9.28
3    30 All         0   1    74.1 18.5  18.5 
4    30 All         4   0.6 147.  35.9  36.9 

[nanxstats]

I completely agree. More generally, "type-stable" functions are preferred: https://design.tidyverse.org/out-type-stability.html

[LittleBeannie]

Decision:
ahr -> original ahr function with simple = TRUE
pw_info -> original ahr function with simple = FALSE
ahr will call ``pw_info`

[elong0527]

Great! Maybe we can brainstorm a better function name.

If the function will become a general utility function, maybe piecewise_hazard_ratio

If the function only apply to AHR method, maybe starting from ahr_xxx e.g. ahr_piecewise.

@nanxstats, you must have better suggestion on name convention.

[LittleBeannie]

Great! Maybe we can brainstorm a better function name.

If the function will become a general utility function, maybe piecewise_hazard_ratio

If the function only apply to AHR method, maybe starting from ahr_xxx e.g. ahr_piecewise.

@nanxstats, you must have better suggestion on name convention.

Can we call it pw_info? Because it basically computes the statistical information for the piecewise model. Considering pw_info is also used in gs_info_wlr, let's not start it with ahr_xxx.