ADADAS resolve issues when PARAMCD=ACTOT, DTYPE=LOCF

[kaz462]

Used derive_locf_records() to derive PARAMCD=ACTOT, DTYPE=LOCF, differences with original adadas are due to below two issues:

1. derive_locf_records() from admiral 0.9.0:

Only variables specified in by_vars will be populated in the newly created records.

• There are variables (e.g., VISIT, VISITNUM, ADY, ADT, PARAM, PARAMN, QSSEQ) need to be populated based on the last valid observation, but shouldn't be included in the by_vars parameter. Those variables are blanks for DTYPE=LOCF if we use this function.
• I raised a feature request to add a new parameter keep_vars in derive_locf_records()
• Temporary solution: added keep_vars to derive_locf_records(), and saved it as R/derive_locf_records_(), adadas.R is using this updated function for now

[kaz462]

2. I think the remaining issues are from original adadas from CDISC pilot data

• CDISC SDTM/qs: 818 records for QSTESTCD=ACTOT
• CDISC ADaM/adadas: 1040 records for PARAMCD=ACTOT, 799 (directly from qs, should it be 818?) + 241 imputed records (DTYPE=LOCF)
• adadas generated by R: 1040 records for PARAMCD=ACTOT, 818 (directly from qs) + 222 imputed records (DTYPE=LOCF)

Take a detailed example USUBJID="01-701-1294" CDISC qs

> qs %>% filter(QSTESTCD=="ACTOT") %>% 
+   select(USUBJID, QSSEQ, VISIT, QSTESTCD, QSTEST,QSSTRESN) %>% 
+   filter(USUBJID=="01-701-1294")
# A tibble: 4 × 6
  USUBJID     QSSEQ VISIT     QSTESTCD QSTEST                QSSTRESN
  <chr>       <dbl> <chr>     <chr>    <chr>                    <dbl>
1 01-701-1294  5015 BASELINE  ACTOT    ADAS-COG(11) Subscore        9
2 01-701-1294  5030 WEEK 8    ACTOT    ADAS-COG(11) Subscore       14
3 01-701-1294  5045 WEEK 12   ACTOT    ADAS-COG(11) Subscore        6
4 01-701-1294  5060 RETRIEVAL ACTOT    ADAS-COG(11) Subscore        9

CDISC adadas For the record with QSSEQ=5045 and AVISIT=Week 8, DTYPE is populated as LOCF , but this record is directly from qs dataset not imputed.

> qc_adadas %>% filter(PARAMCD=="ACTOT") %>% 
+   select(USUBJID, QSSEQ, PARAMCD, AVISITN, AVISIT, VISIT, AVAL, DTYPE, ANL01FL, ADT, ADY) %>% 
+   arrange(USUBJID, AVISITN) %>% filter(USUBJID=="01-701-1294")
# A tibble: 5 × 11
  USUBJID     QSSEQ PARAMCD AVISITN AVISIT   VISIT      AVAL DTYPE  ANL01FL ADT          ADY
  <chr>       <dbl> <chr>     <dbl> <chr>    <chr>     <dbl> <chr>  <chr>   <date>     <dbl>
1 01-701-1294  5015 ACTOT         0 Baseline BASELINE      9 ""     "Y"     2013-03-24     1
2 01-701-1294  5030 ACTOT         8 Week 8   WEEK 8       14 ""     "Y"     2013-05-22    60
3 01-701-1294  5045 ACTOT         8 Week 8   WEEK 12      14 "LOCF" ""      2013-06-14    83
4 01-701-1294  5045 ACTOT        16 Week 16  WEEK 12      14 "LOCF" "Y"     2013-06-14    83
5 01-701-1294  5060 ACTOT        24 Week 24  RETRIEVAL     9 ""     "Y"     2013-10-08   199

adadas generated by R DTYPE is not LOCF for the record with QSSEQ=5045 and AVISIT=Week 8, which makes more sense to me.

> adadas %>% filter(PARAMCD=="ACTOT") %>% 
+   select(USUBJID, QSSEQ, PARAMCD, AVISITN, AVISIT, VISIT, AVAL, DTYPE, ANL01FL, ADT, ADY) %>% 
+   arrange(USUBJID, AVISITN) %>% filter(USUBJID=="01-701-1294")
# A tibble: 5 × 11
  USUBJID     QSSEQ PARAMCD AVISITN AVISIT   VISIT      AVAL DTYPE  ANL01FL ADT          ADY
  <chr>       <dbl> <chr>     <dbl> <chr>    <chr>     <dbl> <chr>  <chr>   <date>     <dbl>
1 01-701-1294  5015 ACTOT         0 Baseline BASELINE      9 ""     "Y"     2013-03-24     1
2 01-701-1294  5030 ACTOT         8 Week 8   WEEK 8       14 ""     "Y"     2013-05-22    60
3 01-701-1294  5045 ACTOT         8 Week 8   WEEK 12       6 ""     ""      2013-06-14    83
4 01-701-1294  5045 ACTOT        16 Week 16  WEEK 12       6 "LOCF" "Y"     2013-06-14    83
5 01-701-1294  5060 ACTOT        24 Week 24  RETRIEVAL     9 ""     "Y"     2013-10-08   199