SDTM Baseline Flag v ADaM Baseline Flag

For one of my current studies, the derivation of the SDTM Baseline Flags was implemented before the derivation of analysis baseline was fully known. Therefore, the baseline flags do not always match between SDTM and ADaM.

In SDTM, we take the last nonmissing value prior to first dose. This could include unscheduled assessments.

In ADaM, we prioritise the scheduled Week 1 value as long as it is prior to first dose, otherwise if its not prior to first dose or its not available then the last nonmissing value prior to first dose is used.

There is one dataset where we have to create an extra record in ADaM (ADVS) in order to derive baseline correctly (in fact the extra record is created for every visit, to take the average value out of 3 readings). Otherwise, we don't have to do any creation of extra records.

With this in mind, should the SDTM Baseline flag match the ADaM Baseline flag? The client is insistent that as we knew the derivation of baseline, we should have changed the derivation in SDTM once we knew it, so that it would match ADaM. But our SDTM baseline derivation is the usual implementation that I have always used, other clients have accepted this before and sent submissions based upon this, even when the ADaM baseline flags are not necessarily the same.

Thanks

Lindsey Thomas