Extract column names and information for NONMEM $INPUT

[kylebaron]

Summary

As a user, I want to automatically extract information to population $INPUT in a NONMEM control stream

• Extract column names in order
• Add =DROP to columns that are character
• Append short name
• Include unit for continuous columns
• Add values and decodes for columns where this information is available in the spec object

Tests

• tests/testthat/test-nm_input.R
  • input text is formed from spec

[kylebaron]

Option 1

> nm_input(spec)
$INPUT
C         ; comment character 
NUM       ; record number 
ID        ; numeric subject identifier 
TIME      ; time after first dose (hour)
SEQ       ; data type 
CMT       ; compartment number 
EVID      ; event ID 
AMT       ; dose amount (mg)
DV        ; DV (ng/ml)
AGE       ; age (years)
WT        ; weight (kg)
CRCL      ; creatinine clearance (ml/min)
ALB       ; Albumin (g/dL)
BMI       ; BMI (m2/kg)
AAG       ; alpha-1-acid glycoprotein (mg/dL)
SCR       ; serum creatinine (mg/dL)
AST       ; aspartate aminotransferase 
ALT       ; alanine aminotransferase 
HT        ; height (cm)
CP        ; Child-Pugh score 
TAFD      ; time after first dose (hours)
TAD       ; time after dose (hours)
LDOS      ; last dose amount (mg)
MDV       ; MDV 
BLQ       ; BLQ 
PHASE     ; study phase indicator 
STUDY     ; study number 
SUBJ=DROP ; subject identifier 
RF=DROP   ; renal function stage 

; DECODE ---------------- 
; SEQ     : 0 = observation, 1 = dose
; EVID    : 0, 1
; CP      : 0 = normal, 1 = Pugh1, 2 = Pugh2, 3 = Pugh3
; MDV     : 0 = not missing, 1 = missing
; BLQ     : 1, 0
; PHASE   : 1
; STUDY   : 1 = 211-EXAMP-001, 2 = 211-EXAMP-011, 3 = 211-EXAMP-801, 4 = 211-EXAMP-802

[kylebaron]

Option 2

> nm_input2(spec)
$INPUT
C         ; comment character 
NUM       ; record number 
ID        ; numeric subject identifier 
TIME      ; time after first dose (hour)
SEQ       ; data type 
          ; 0 = observation, 1 = dose
CMT       ; compartment number 
EVID      ; event ID 
          ; values: 0, 1
AMT       ; dose amount (mg)
DV        ; DV (ng/ml)
AGE       ; age (years)
WT        ; weight (kg)
CRCL      ; creatinine clearance (ml/min)
ALB       ; Albumin (g/dL)
BMI       ; BMI (m2/kg)
AAG       ; alpha-1-acid glycoprotein (mg/dL)
SCR       ; serum creatinine (mg/dL)
AST       ; aspartate aminotransferase 
ALT       ; alanine aminotransferase 
HT        ; height (cm)
CP        ; Child-Pugh score 
          ; 0 = normal, 1 = Pugh1, 2 = Pugh2, 3 = Pugh3
TAFD      ; time after first dose (hours)
TAD       ; time after dose (hours)
LDOS      ; last dose amount (mg)
MDV       ; MDV 
          ; 0 = not missing, 1 = missing
BLQ       ; BLQ 
          ; values: 1, 0
PHASE     ; study phase indicator 
          ; values: 1
STUDY     ; study number 
          ; 1 = 211-EXAMP-001, 2 = 211-EXAMP-011, 3 = 211-EXAMP-801, 4 = 211-EXAMP-802
SUBJ=DROP ; subject identifier 
RF=DROP   ; renal function stage 

[curtisKJ]

@kylebaron I like having the decode with the descriptor; it seems easier to parse visually. So option 2 would be my preference, but maybe something like in yspec where there's a sub-descriptor for the decode would be nicer?

[KatherineKayMRG]

@kylebaron something like this would be great. I find the first option a little cleaner but if the column name were unfamiliar to me I would find the decode with the descriptor in option 2 to be quicker/simpler to interpret.

[kylebaron]

@KatherineKayMRG Yeah; I find it easier to digest option 2. But agree with @curtisKJ that option 1 seems more intuitive overall.

[kylebaron]

library(tidyverse)

library(yspec)

sp <- ys_help$spec()

yspec:::nm_input(sp) %>% cat(sep = "\n")
#> $INPUT
#> C         ; comment character 
#>           ; [. = analysis row, C = commented row]
#> NUM       ; record number
#> ID        ; subject identifier
#> SUBJ=DROP ; subject identifier
#> TIME      ; TIME
#> SEQ       ; SEQ 
#>           ; [0 = observation, 1 = dose]
#> CMT       ; compartment number
#> EVID      ; event ID 
#>           ; [values: 0, 1]
#> AMT       ; dose amount
#> DV        ; dependent variable
#> AGE       ; age
#> WT        ; weight
#> CRCL      ; CRCL
#> ALB       ; albumin
#> BMI       ; BMI
#> AAG       ; alpha-1-acid glycoprotein
#> SCR       ; serum creatinine
#> AST       ; aspartate aminotransferase
#> ALT       ; alanine aminotransferase
#> HT        ; height
#> CP        ; Child-Pugh score 
#>           ; [0 = normal, 1 = Pugh1, 2 = Pugh2, 3 = Pugh3]
#> TAFD      ; time after first dose
#> TAD       ; time after dose
#> LDOS      ; last dose amount
#> MDV       ; MDV 
#>           ; [values: 0, 1]
#> BLQ       ; below limit of quantification 
#>           ; [1 = above QL, 0 = below Q]
#> PHASE     ; study phase indicator 
#>           ; [values: 1]
#> STUDY     ; study number 
#>           ; [1 = SAD, 2 = MAD, 3 = Renal, 4 = Hepatic]
#> RF=DROP   ; renal function stage 
#>           ; [norm = Normal, mild = Mild, mod = Moderate,
#>           ;  sev = Severe]

^{Created on 2020-07-02 by the reprex package (v0.3.0)}