128 aet02 template

[Meixh2021]

close insightsengineering/chevron-tasks#42

Hi @wangh107 @barnett11, thanks to the @BFalquet, I believed the current version of chevron can help sort the AEBODSYS correctly.

For the No Coding Available, I just added additional dm_explicit_na(na_level = "No Coding available") to fix that issue.

I corrected a sorting problem when the argument lbl_overall = NULL, the sorting will be based on the last column but not the whole column in aet02_1().

You might check the issue again with the following codes:

dpath <- "root/clinical_studies/ro_testing/cdp_testing/spa_testing/data_analysis/spaenv_testing/qa/outdata_vad"
adsl <- rice::rice_read(file.path(dpath,"adsl.sas7bdat"))
adae <- rice::rice_read(file.path(dpath,"adae.sas7bdat"))

library(dm)
library(chevron)

dm.data <- dm(adsl, adae) %>%
  dm_add_pk(adae, AESEQ) %>%
  dm_add_pk(adsl, c(USUBJID, STUDYID)) %>%
  dm_add_fk(adae, c(USUBJID, STUDYID), adsl) %>% 
  dm_explicit_na(na_level = "No Coding available") # Missing data

### Option to see the label format
dm.data <- dm.data %>%
  dm_zoom_to(adae) %>%
  mutate(AEBODSYS = formatters::with_label(AEBODSYS, "MedDRA System Organ Class")) %>%
  mutate(AEDECOD = formatters::with_label(AEDECOD, "MedDRA Preferred Term")) %>%
  dm_update_zoomed()

dm.data.1 <- dm.data %>%
  aet02_1_pre()
aet02_1(adam_db = dm.data.1, .study = list(actualarm = "TRT01A",
                                         lbl_overall = NULL))
dm.data.3 <- dm.data %>%
  aet02_3_pre()
aet02_3(adam_db = dm.data.3, .study = list(actualarm = "TRT01A",
                                         lbl_overall = NULL))

Thanks!

[github-actions]

Code Coverage Summary

Filename                           Stmts    Miss  Cover    Missing
-------------------------------  -------  ------  -------  ---------
R/aet01.R                            252     252  0.00%    38-455
R/aet02.R                            189     189  0.00%    47-456
R/aet03.R                             66      66  0.00%    41-157
R/aet04.R                             86      86  0.00%    49-182
R/assertions.R                        23      23  0.00%    21-46
R/cmt01a.R                           156     156  0.00%    50-427
R/cmt02_pt.R                          37      37  0.00%    34-108
R/dm_explicit_na.R                    44      44  0.00%    46-123
R/dmt01.R                             29      29  0.00%    44-125
R/dst01.R                            231     231  0.00%    5-609
R/dtht01.R                            77      77  0.00%    44-192
R/egt01.R                             44      44  0.00%    45-137
R/egt02.R                             56      56  0.00%    42-218
R/ext01.R                             69      69  0.00%    39-243
R/gen_args.R                           1       1  0.00%    18
R/lbt01.R                             39      39  0.00%    43-127
R/mht01.R                             57      57  0.00%    39-139
R/sample_study_object.R               14      14  0.00%    14-35
R/standard_data_preprocessing.R      221     221  0.00%    52-508
R/utils.R                            144     144  0.00%    11-377
R/vst01.R                             45      45  0.00%    46-140
R/vst02.R                             92      92  0.00%    42-254
TOTAL                               1972    1972  0.00%

Results for commit: b52fd42c9c2697357d23d4fae236d7bd4e6c8d75

Minimum allowed coverage is 80%

♻️ This comment has been updated with latest results

[wangh107]

Thanks @Meixh2021

Tim @barnett11 , what do you think if we include the following steps in the corresponding preprocess functions aet02_x_pre:

1. Replacing missing with "No Code available" for variables AEBODSYS, AEHLT, AEDECOD (depending on which variables are used)
2. Assign correct label format in table header
   e.g. mutate(AEBODSYS = formatters::with_label(AEBODSYS, "MedDRA System Organ Class"))

[barnett11]

Thanks @Meixh2021

Tim @barnett11 , what do you think if we include the following steps in the corresponding preprocess functions aet02_x_pre:

1. Replacing missing with "No Code available" for variables AEBODSYS, AEHLT, AEDECOD (depending on which variables are used)
2. Assign correct label format in table header
   e.g. mutate(AEBODSYS = formatters::with_label(AEBODSYS, "MedDRA System Organ Class"))

Yes that should do it thanks @wangh107

[Meixh2021]

Thanks @Meixh2021
Tim @barnett11 , what do you think if we include the following steps in the corresponding preprocess functions aet02_x_pre:

1. Replacing missing with "No Code available" for variables AEBODSYS, AEHLT, AEDECOD (depending on which variables are used)
2. Assign correct label format in table header
   e.g. mutate(AEBODSYS = formatters::with_label(AEBODSYS, "MedDRA System Organ Class"))

Yes that should do it thanks @wangh107

@wangh107 @barnett11 for item 2, currently the label of output is based on the label of the data. My question is shall we force the labels? For example, do all outputs generated by aet02_x() have the same label "MedDRA System Organ Class" but not its original label for AEBODSYS?

[barnett11]

Thanks @Meixh2021
Tim @barnett11 , what do you think if we include the following steps in the corresponding preprocess functions aet02_x_pre:

1. Replacing missing with "No Code available" for variables AEBODSYS, AEHLT, AEDECOD (depending on which variables are used)
2. Assign correct label format in table header
   e.g. mutate(AEBODSYS = formatters::with_label(AEBODSYS, "MedDRA System Organ Class"))

Yes that should do it thanks @wangh107

@wangh107 @barnett11 for item 2, currently the label of output is based on the label of the data. My question is shall we force the labels? For example, do all outputs generated by aet02_x() have the same label "MedDRA System Organ Class" but not its original label for AEBODSYS?

Yes, we should use the template-defined column name regardless of the CDISC attributed label thanks

[wangh107]

Another question for Tim @barnett11
How should we treat the No Coding Available when sorting?
Do we treat it as the regular category and sort by frequency?
OR
Should the No Coding Available category always be placed in the last?

[barnett11]

Another question for Tim @barnett11 How should we treat the No Coding Available when sorting? Do we treat it as the regular category and sort by frequency? OR Should the No Coding Available category always be placed in the last?

There is no specific guidance @Meixh2021 , but the general practice of just including in the frequency list is sufficient thanks

[BFalquet]

Thank you for the good work. Please see my comments inside.

you can have a look at insightsengineering/tern#577 if that helps for dtht01.

cheers

[Meixh2021]

Thank you for the good work. Please see my comments inside.

you can have a look at insightsengineering/tern#577 if that helps for dtht01.

cheers

Thanks @BFalquet, now I believe that this pull request can also help solve some issues in insightsengineering/chevron-tasks#47. 😄

[wangh107]

@Meixh2021
Suggest to include a data process in aet02_x_pre to code missing in AEBODSYS, AEHLT, AEDECOD as "No Coding Available" (only do it in the 3 mentioned variables in adae dataset, not the whole dm object.)

[Meixh2021]

@barnett11 Hi Tim, if you are available, would you please review the current template can meet the TLG criteria? Thank you!

[BFalquet]

looks good to me, thank you for the work. Beware that important changes are coming to chevron soon.

[barnett11]

Thanks @Meixh2021
Please note this pre-processing of Not Coded does not work if dm_explicit_na has been used on the data prior to passing of function. The terms for AEBODSYS and AEDECOD then appear as <MISSING>.
FYI @wangh107 , @BFalquet