I am trying to validate a SDTM package that has SUPPSV dataset.
Although the SD0095 rule clearly says that you can use SUPPQUAL only for general-observation-class datasets (Events, Findings, Interventions) and Demographics, this rule was not fired in the validation report.
I am using p21 community version 3.1.2, Validation Engine Version: 2010.1
Can you help me determine the reason.
As per latest FDA technical conformance guide it is stated that SUPPSV should NOT be used but data placed in the parent SV domain instead.
"It is also Agency preference that three non-standard variables (NSVs) for missed visits, --REASOC (Reason for Occur Value), --EPCHGI (Epi/Pandemic Related Change Indicator), and --CNTMOD (Contact Mode), outlined in the CDISC document “Guidance for Ongoing Studies Disrupted by COVID-19 Pandemic” be included within the SV domain and not within the supplemental SUPPSV domain or in other SDTM datasets. Submitting subject visits information in one single structured dataset allows both the human and technology consumer of this information to operate efficiently and with confidence that all visit data are considered during regulatory review."
Including NSVs in the parent domain is however a deviation from SDTM and would lead to other P21 output. How would you handle this, and would you foresee different rules in FDA Engine versus PMDA Engine?
Unfortunately, one sees over and over again that FDA requests/requires (see the wording "preference") things that violate the SDTM standard.
In this case probably because some reviewers are not capable to recombine SUPPxx datasets with the parent dataset, although that is pretty easy.
It was however exactly the FDA who wanted to have SUPPxx datasets in the past, as reviewers were not able to distinguish standard variables from non-standard variables.
This "back and forth" is very sad, and demonstrates the lack of coordination and alignment between FDA and CDISC.
IMO, the best way to handle this for the moment is that one uses two different sets of rules, depending on the use case, FDA-submission or non-FDA-submission. This can be accomplished by the "engine". Of course, the FDA set of rules should then not have entries that conflict with each other ("whatever one does, one always gets an error"). Essentially, that would mean that the "engine" is immediately updated each time a new conformance guide is published.
Catastrophic would however be that the rules also depend on the therapeutic area, e.g. different rules for COVID-19 and cancer studies, which seems to be the case here.
But better would of course be that CDISC and FDA, PMDA agree on one single set of rules, and that each party then also keeps to that set of rules without any exceptions...
* Required Field
Hi Saravanan, SD0095 rule…
SD0095 rule reports only incorrect usage of SUPPQUAL for Trial Design domains. It will be updated in the next release by SUPPCO and SUPPSE datasets. However, most likely SUPPSV datasets will be an exclusion for this rule similar to SUPPDM dataset. Recent CDISC and FDA guidance documentation for COVID-19 studies request additional information for Subjects Visits which is not covered by current SDTM model. Therefore, usage of SUPPSV dataset is expected by FDA guidance.