as i read the various FDA guidlines the only two rules that need to be adhered to in order to avoid a technical rejection are 1734 and 1736. Is this still the case or have other rules been added?
To elaborate a bit further. Especially i TS the start date parameter must be in place. Let's for argue sake say that ACTUB was filled out incorrect. That would not lead to a technical rejection since it won't be checked by a machine.
Is my understanding correct that technical rejections only can occur when a machine finds and error and that only the two above mentioned rules are checked?
Forums: Validation Rule Suggestions
because this site already has some great presentations on the matter inidcating that this is the place to go to. + this is particular question was more of a deep dive into the details that i need to get on top of. I think though i have the answer.
But a great tip regarding contacting FDA directly. I wasn't aware of that one. Also the CDISC community link.
There are several validations of submitted study data at FDA. FDA technical rejection criteria are a part of eCTD validation. It is all about how folder structure and file names of your submission are compliant with eCTD specifications. Are your files located in pre-specified folders?
(yes, there are some additional requirements like correct file tags, but they are handled by Regulatory teams and generally not relevant to Biostat teams)
Define.xml represents documentation of your study data. DM and ADSL datasets are always expected to be in your study. For example, some studies may not collect Lab Test Results or Exposure data, but Demographics and Analysis Subject Level datasets are always collected and created in every study. So, rejection criteria ensure that you have 1. datasets (by a spot check for DM or ADSL) and 2. documentation (define.xml) in pre-specified folder like “NDA123456/0001/m5/datasets/study_xyz/tabulation/sdtm”
FDA request for TS dataset serves a different need. It allows automation of validation process itself. If you study started after certain date, your study data and documentation must be in CDISC-compliant format (SDTM, SEND, ADaM, Define-XML). TS domain can be a source of standardized machine-readable information about the study start date. Using this info, automated validation tools can decide if you study data must have datasets in “tabulation/stdm”, “tabulation/send”, “analysis/adam/datasets” folders or can still utilize non-standard “legacy” format? Therefore, in terms of eCTD rejection criteria, requirements for TS domain are limited to ensure that automated tools can read the study start date. It means that you need to 1. Have TS dataset (in SAS XPORT v5 format) 2. Have a record with TSPARMCD=’ SSTDTC’ in SDTM or TSPARMCD=’STSTDTC’ in SEND 3. For this record, TSVAL should have a non-missing date in ISO 8601 format. Everything else in TS domain is not relevant to automation process and formally not considered as rejection criteria.
In your example, missing or incorrect information about ‘ACTSUB’ (or misspelling of this parameter as ‘ACTUB’ in your example) will not be a part of FDA rejection criteria. All data quality and standard compliance issues will be reported to FDA reviewers and other FDA team members. So, they can make decision about potential impact of identified data issues to their intend use of study data and what corrective actions are needed? However, this is a part of data quality evaluation which is separate from eCTD validation process.
Thank you very much Sergiy for spelling this out to me. This is how interpreted it but couldn't get down 100%.
* Required Field
Technical rejection criteria
I wonder why you ask this question here. The FDA guidelines are owned by the FDA, not by ...
If you navigate to https://www.fda.gov/industry/fda-resources-data-standards/study-data-standards-resources
and then to the bottom under "stay connected", you will find a number of e-mail addresses where you can ask the FDA directly..
If however, you would like to ask the CDISC community about their experiences with this topic, another extremely forum is the LinkedIn "CDISC - SDTM Experts" forum at: https://www.linkedin.com/groups/2758697/