ADEXSUM

Variable	Parameter	Comment
ADTHFL		DTHFL is a SDTM variable in DM, which includes deaths after data cutoff date. Should ADSL.ADTHFL be here to replace DTHFL?.
AEVLINT		AEVLINT value of 'by cycle' should be optional; only needed if multiple doses given in the same infusion cycle and summary of dosing by cycle is required. 'By cycle' not applicable for oral dosing.
AVAL	CUMACDOS TOTACDOS	It is hard to tell the difference of the AVAL derived between these two PARAMCDs. Are they equal for the overall?.
	CUMACDOS TOTACDOS CUMPLDOS TOTPLDOS	What is the difference between CUMACDOS vs. TOTACDOS and CUMPLDOS vs. TOTPLDOS? Suggest wording on the notes to be 'Cumulative planned dose for each treatment up to each evaluation interval. This way we can distinguish it from the definition of TOTPLDOS/TOTACDOS.
	DOSINT NDOSINT	What is considered a dose interruption? Is this an interruption within one dose/infusion. Or does this pertain to skipped doses? Would be good to clarify.
	DURDLAY	Regarding DURDLAY… same subject may have multiple dose delays of different duration and reason for delay. We may want to summarise frequency of delays for each duration.
	General	Is there PARAMCD to represent number of doses for oral dosing?. Not clear which of these parameters apply to only oral, only infusions or both.
		For delays, reductions, interruptions should we include records with value '0' for subjects who did not have any delays, reductions, interruptions, respectively?. From the flags it appears that only records with 'Y' should be included in which case flag is redundant. In notes section clarify that these parameters are created by subject and EXTRT.
		Are dose interruption parameters intended to summarise infusions or oral dosing missed doses?.
		Are all parameters permissible based on study designs/needs?.
	NADMIN	Can we confirm that this is one observation per treatment per subject?.
		Does it make sense to count number of administrations across all treatments?.
		Why is PARAMCD for planned no. of administrations (corresponding to NADMIN) not included?.
		Will the Agency provide specific algorithms for this calculation? Specifically, how to count and handle situations such as infusion interruptions and oral medication administered multiple times per day (e.g., would an infusion that is interrupted then restarted count as one or two administrations?).
	NADMIN NUMCYC NUMPCYC	It is more useful to create NADMIN, NUMCYC and NUMPCYC per EXTRT? If for combination drugs.
	NDOSDCTS	What is considered a dose discontinuation? Can a subject discontinue the same treatment more than once? Would that not be counting interruptions then instead of discontinuations? Or is this an interruption within one dose/infusion. Would be good to clarify.
	NDOSDCTS	NDOSDCTS - please clarify whether it can be multiple dose discontinuation per subject and EXTRT.
	NUMPCYC CUMPLDOS TOTPLDOS	Is this the number of cycles/planned dose per protocol or should drug discontinuation also be taken into account?.
	NUMPCYC TRTPDURD CUMPLDOS TOTPLDOS	Why do we need this planned days?. Do we use this in our analysis?. We always use Actual Days for the analysis. If so why do we need the Planned Days Parameters?.
	TRTDCT	Why is TRTDCT needed?. Flag for permanent discontinuation is available in ADSL.
	TRTPDURD	Is PARAMCD="TRTPDURD" (treatment duration planned in days) required? Clarification is needed for the derivation of this value. Will this be the length of the planned treatment phase as specified in the protocol? For example, TRTPDURD = (Total treatment cycles * Days on treatment per cycle). Or, will this be the length of the planned treatment phase to be completed on the data cutoff date? For example, TRTPDURD = (Total treatment cycles completed by the data cutoff date * Days on treatment per cycle).
AVALC	DELAY	Please clarify if this is included only if collected on the EX CRF. Or is this a derived field? If this is a derived field, will a standard algorithm be defined and provided by the Agency?.
	DOSCTS	Please clarify if this is restricted to temporary dose discontinuations.
	DOSINT	Please clarify if this is calculated only for infusions and, if not, how does this differ from temporary dose discontinuation.
	TRTDCT	This information will be captured on EOT disposition page and will be in ADSL. Is there any reason for keeping same information in ADEXSUM?.
		Is the expectation that this flag is identical for all study drugs or would this be per study drug.
		The evaluation interval is 'Overall' for this parameter and PARQUAL='ALL', so is this referring to 'Any Dose Discontinuations' for each value of EXTRT?.
		Why is this only of interest for ALL treatments, rather than individual treatments, as well?.
AVALU / AEVLINT		Similar to the PARQUAL concern, does the use of AVALU and AEVLINT encourage the creation of PARAMs/PARAMCDs that are not unambiguously describing the contents in AVAL/AVALC? For example, if we needed to summariSe duration of treatment in both days and months, we would create 2 distinct PARAMs/PARAMCDs, 1 for days and 1 for months.
General		We presume all PARAMCD are not mandatory and their inclusion is dependent on study design, e.g. TRTPDURD, NUMPCYC, CUMPLDOS, TOTPLDOS, etc may not be applicable if exposure schedules do not have pre-specified number of doses or cycles. List of PARAMCD depends on the analysis planned in the study. If it is used in the analysis, we'll use the parameter name provided; if the parameter is not used, we don't need to derive and keep it in the dataset.
General		Is ADEXSUM datsest required? For some t-cell therapies which have a single infusion, there is little to no summaries of exposure that are of interest.
PARAMCD		Can the parameter list be expanded to include 'number of dose escalations' for DE study?.
PARQUAL		Since PARQUAL contains values from EXTRT or 'All', consider using variable ATRT instead?.
		This variable was not included in the final version of ADaM IG v1.2. However, conceptually this (or equivalent) is required for this dataset to enable the same parameter to be summarised for different study medications.
		Including treatment in PARQUAL and not in PARAM goes against ADaM principle that PARAM should uniquely identify the content of AVAL. But it does make things easier to for ADEXSUM, so I am not against it, but just wondering if this needs to be explained a bit more.
		Current descriptions are not clear, can more clarity OR example be provided? In the VLM, value of 'ALL' - is this across all the treatments?.