Registration and death #26
Conversation
| as.Date(TPP_death_date) - as.Date(last_registration_start_date), | ||
| TRUE ~ as.difftime(NA_real_, units = "days") | ||
| ), | ||
| TPP_death_dereg = case_when( |
There was a problem hiding this comment.
there should be an option for if the last registration end date is missing - we should quantify how commonly that happens
There was a problem hiding this comment.
Added
| TPP_death_dereg = case_when( | |
| ons_death_dereg_group = case_when( | |
| is.na(last_registration_end_date) & !is.na(ons_death_date) ~ "missing_last_reg_end", |
| TPP_death_dereg = case_when( | |
| TPP_death_dereg_group = case_when( | |
| is.na(last_registration_end_date) & !is.na(TPP_death_date) ~ "missing_last_reg_end", |
| ons_death_reg == 0 ~ "last_reg_same_death", | ||
| ons_death_reg > 0 ~ "last_reg_before_death" | ||
| ), | ||
| ons_death_dereg_group = case_when( |
There was a problem hiding this comment.
It might be interesting to have more granularity - can you do this for individual days, say up to 30 days? you probably couldn't release that info, b/c of disclosure but might be more enlightening. You can keep what you have here (for release purposes).
There was a problem hiding this comment.
created and ONS and TPP new tables with values -31 < x < 31
There was a problem hiding this comment.
For this analysis I would do it with and without grace period. Perhaps as separate actions?
There was a problem hiding this comment.
First, I think you need to restrict to people who have a death during study period.
Also, I would do summary stats with / without grace period.
There was a problem hiding this comment.
I created summary stats
- raw
- inc criteria with grace period
- inc criteria without grace period (dereg_end_date > min_date_death)
No description provided.