Create derive_var_bvcacritxfl function

[manciniedoardo]

Background Information

Create function derive_var_bcvacritxfl which can create criterion flags for CBVA gain/loss of x letters. Functionality could be as follows:

derive_var_bcvacritxfl <- function(
  ...
  gain_values,
  lose_values,
  ....
){
<function code>
}

which could be called like this:

adbcva <- adbcva %>% 
    derive_var_bcvacritxfl(
        gain_values = c(10,15),
        lose_values = c(5,10)
     )

to generate an updated ADBCVA dataset containing criterion flags CRIT1FL, CRIT2FL, CRIT3FL, CRIT4FL for gaining 10 letters, gaining 15 letters, losing 5 letter and losing 10 letters respectively, along with associated variables CRIT1, CRIT2, CRIT3, CRIT4 encoding the criterion for the flag (eg, CRIT1 = "10 letter BCVA gain" for BCVA records and blank otherwise).

Proposed convention:

• CRITxFL with 1<=x<=10 used for BCVA gain categories
• CRITxFL with 11<=x<20 used for BCVA loss categories

This will be quite a central function in {admiralophtha}, so we should aim to provide:

• many examples in the function roxygen code
• vignette
• example in the ADBCVA template.

---

Note: we should also add avoidloss_values and avoidgain_values corresponding to avoiding a loss/gain of x letters. An appropriate convention can also be considered like the one suggested above.

[ukey1025]

Sorry for my delay coding. I missed push my branch so copy my code here.

I have some questions.
i) Do gain_values and loss_values always have 2 vectors? If not, I need to modify code which.
ii) Sorry for I am not famillier with ADBCVA, how to define gain_values or loss _values. Do we usulally use minus numbers?

I will confirm with our in-house ophthalmologist tomorrow, and will modify.

/////

derived_var_bvcacritxfl <- function(
    adbcva_1,
    x,
    gain_values,
    lose_values
){
  
  adbcva_1 %>%
    mutate(CRIT1FL = ifelse((!!as.name(x) >= gain_values[1] & !!as.name(x) < gain_values[2]),
                            paste(gain_values[1], " letter BCVA gain"), "" )) %>%
    mutate(CRIT2FL = ifelse((!!as.name(x) >= gain_values[2]),
                            paste(gain_values[2], " letter BCVA gain"), "" )) %>%  
    mutate(CRIT3FL = ifelse((!!as.name(x) <= -(lose_values[1] & !!as.name(x) < -(lose_values[2])),
                            paste(lose_values[1], " letter BCVA loss"), "" )) %>%
    mutate(CRIT4FL = ifelse((!!as.name(x) <= -(lose_values[2])),
                            paste(lose_values[2], " letter BCVA loss"), "" ))
}

[manciniedoardo]

Thanks for this Yuki. A couple of notes:

• We really should be working in a separate branch, so I will make a new branch and copy the code across.
• We should be following the programming strategy listed here, which has info about variable names and formatting.
• I think it would make sense to loop over gain_values and lose_values rather than having many mutate statements.

I am happy to implement the above myself and then circle back with you to show you what I've done and how I've done it 😄