Best method for transforming ADaM data sets into NONMEM format?


I have several ADaM formatted csv files that I would like to join and format for use in constructing a pop PK dataset that includes patient demographics as well as reported adverse events but I don’t know the best method to get this done.

To be more specific, I have serum conc. vs time data listed in adpc, and dosing times listed in adex, subject demographics in adsl and adverse event reoprting in adae. I want to combine the dosing times, plasma conc-time data in chronological order that accounts for deviations in PK sampling time and additional sort columns that contain subject demographics and reported AE’s.

I am trying to avoid any manual manipulation of the data set such as defining the dosing events in chronological order per subject. I was hoping someone else has already encountered this and could share the method they use?

Kind regards,

Dan Hines

Hi @daniel.hines,

This is a perennial problem that I don’t know of a great solution for, either. I’m working on an Rsdtm package ( with some collaborators hopefully to make some parts of this better. If the only issue is to put doses and concentrations into chronological order, and you have an ADTC column, then you should be able to simply sort on STUDYID, USUBJID, ADTC with something similar to:


d_prepared <-
  bind_rows(d_adpc, d_adex) %>%

After that, there are usually additional standardizations required, but that is hopefully a good start.

I’m all ears to hear what others do, too.