FDA Study-Data Standards in 2026: Selecting Compatible Versions

TL;DR

Takeaway: For a hypothetical clinical study starting on 1 June 2026, FDA Data Standards Catalog v11.0 retains both SDTMIG 3.3 and 3.4 within their applicable windows. ADaMIG 1.1, 1.2 and 1.3, and Define-XML 2.0 and 2.1, also remain listed without established end dates. Older versions such as SDTMIG 3.2 and ADaMIG 1.0 have passed their listed ends for new starts. FDA encourages the latest supported version while allowing selection among supported alternatives. [1] [2]

The two CDISC terminology rows accept "06/10/2011 or later" and name no quarter. [1] A current-file snapshot of the NCI Enterprise Vocabulary Services CDISC tree on 26 August 2026 stamped SDTM and SEND as Q1 2026, release date 27 March 2026, modified 30 March 2026. [3] The same public SDTM, SEND, ADaM, Define-XML, DDF, and CDASH stamp files on 4 October 2026 read Q3 2026, release date 25 September 2026. [4] The June 2026 Study Data Technical Conformance Guide says sponsors should use the most recent dictionary version available at the start of the study and may use a newer version that appears later. [5]

That guide still tells the sponsor to put in SDTIGVER the exact term in column F, with the example 3.2, and in SDTMVER the exact term in column E, with the example 1.4, of a tab it still calls "Submission Data Exchange Stds." [5] In Catalog v11.0 the standards tab is Submission Data Standards, column E is SDO, column F is Property, and the 3.2 property cell reads SDTMIGv3.2. [1]

The three worked dates below are hypothetical study starts. They test the catalog's logic for a clinical study, an earlier transition period and a nonclinical study with different application-type dates. The resulting tables support a version-selection decision, rather than measuring industry adoption or submission success.

A study start picks a catalog row

The file at FDA's catalog download on 4 October 2026 is still Data Standards Catalog v11.0. Change History dates version 11.0 to 20 March 2025 and says the order of columns on the standards tab was adjusted. [1] The guidance page that hosts the download was last modified 24 March 2025 and still points at that file. [6] The Study Data Standards Resources page, content current as of 18 September 2026, still lists the catalog. [7] The June 2026 conformance guide's plan to post updated versions in March and October, sooner if needed, is a plan for the guide. [5]

Section II.A of FDA's June 2021 guidance covers standardized electronic study data for certain INDs, NDAs, ANDAs and BLAs, including relevant amendments, supplements and reports. Noncommercial INDs are exempt; the guidance defines these by the absence of intended commercial distribution and includes investigator-sponsored and expanded-access INDs. Voluntary standardized submissions remain possible. Specified CBER device submissions under section 745A(b) sit outside this guidance's scope. [2]

Section II.C identifies the catalog as the list of standards FDA can process, review and archive. Failure to conform to required standards can lead FDA to refuse to file an NDA or BLA, or refuse to receive an ANDA. This is why a version-selection record belongs in study planning, alongside CRF design, data management and statistical analysis planning. [2]

Section II.E keys the version to study start. "The requirement to submit using a particular standard is dependent on its support by FDA as listed in the Catalog at the time of study start." [2] For a clinical study that date is the earliest informed-consent date among enrolled subjects, Trial Summary parameter SSTDTC. For a nonclinical study it is the date the study director signs the protocol or plan, STSTDTC. [2] "Sponsors or applicants are encouraged to use the latest version listed in the Catalog. However, when there are multiple versions of a standard listed, sponsors or applicants can select the version to use for their study." [2] The catalog instructions say the same thing: where more than one standard or terminology is supported for a use, the sponsor may select one or discuss it with the review division, and a standard that is not listed should be discussed with the Agency in advance. [1]

The initial timetable has passed. Catalog standards were required for NDA, ANDA, and certain BLA studies with a start date 24 months after 18 December 2014, that is 18 December 2016, and for certain IND studies at 18 December 2017. [2] A separate EClinCloud article already states that gate for a first trial. [8] Later updates use the Federal Register timetable in section II.E. A start on a requirement-begin date or a support-end date stays unclassified below.

The technical conformance guide contains nonbinding recommendations. The study-data guidance explains that its format and exemption provisions under section 745A(a) have binding effect. Keep these different legal functions distinct when recording a preference, a supported option or a requirement. [5] [2]

Our reading: Start with the applicable FDA catalog window, then check the compatible model and implementation guide. The newest CDISC publication and the latest FDA-supported version can have different dates.

What the four date columns and footnotes [1], [2], and [12] say

The standards sheet has four date columns. The Column Descriptions tab defines them, and an empty end or requirement cell is a value that tab is built to hold.

Date Support Begins is "the date when FDA is able to accept submissions of data formatted in the indicated property." The same cell defines the word that replaces a date on the oldest rows: "Ongoing means that FDA supports the listed standard for electronic submissions as of the initial Catalog release date, 06/13/2011." [1] In the tables below, Ongoing is 13 June 2011.

Date Support Ends is "the date when FDA will no longer accept submissions of data formatted in the indicated property." An empty field "means that a 'Date Support Ends' date has not yet been established." [1]

Date Requirement Begins is "the date when the requirement to submit data formatted in the indicated property becomes effective." An empty field "means that a 'Date Requirement Begins' date has not yet been established." [1]

Date Requirement Ends is "the date when the requirement to submit data formatted in the indicated property is no longer effective." An empty field "means that a 'Date Requirement Ends' date has not yet been established." [1]

Footnote [1] means "For NDAs, ANDAs, and certain BLAs," and points at section II.A. Footnote [2] means "For certain INDs," and points at the same section. [1] For the study-data rows assessed here, a single date without an application-type split applies to both timelines. A cell whose footnotes are something else was left out of these study-data clocks. That includes the eCTD specification dates marked [5] and [6], which point at the eCTD specifications guidance, and the SPL dates marked [3] and [4]. [1]

Footnote [12] is printed in full on the standards sheet:

If a study started before the support ends date, the study can continue to use this specified version of the standard. If a study starts after the support ends date, the study can NO longer use this specified version of the standard. Sponsors cannot use this version of the standard or terminology after this date. Generally, a waiver process may be available. Sponsors and applicants should consult with their review division. An empty field in this column means that a support end date has not been established. [1]

Read footnote [12] as a whole. Its first sentence expressly preserves use for a study that began before support ended; its second excludes later starts. Its more general third sentence should therefore be read with that explicit study-start distinction. Section II.E of the guidance also keys the version to study start. An earlier study may retain its eligible version; a proposed unsupported version for a later start requires a separate waiver discussion. [1] [2]

The [12] marker points to a footnote. On SDTM v2.0 and SDTMIG 3.4 it appears next to support beginning on 13 December 2023 and a requirement beginning on 15 March 2025. Their end cells remain empty. Interpret the cell's date by its column heading, then apply the footnote. [1]

Support includes a start on or after Date Support Begins and before an established Date Support Ends. Requirement includes a start after Date Requirement Begins and before an established Date Requirement Ends. That "after" follows the study-data examples and the SEND table header, "Studies started after the following dates." [2] [5] Section II.E.2 also says a version update is required for studies with a start date "no earlier than 12 months after the transition date," which on its face includes the anniversary day. [2] Those two wordings were not reconciled here, so a start on 15 March 2025 was not classified. None of the three hypothetical starts falls on a boundary.

Rows that cover a clinical start on 1 June 2026

Takeaway: On 1 June 2026, either clock, SDTMIG 3.3, ADaMIG 1.1, and Define-XML 2.0 still sit beside the later rows, and SDTMIG 3.2, ADaMIG 1.0, and Define-XML 1.0 have ended. SDTMIG 3.3 remains supported alongside the requirement that begins for SDTMIG 3.4 on 15 March 2025, with empty end cells in its catalog row. The table covers the current alternatives and selected retired comparators; it omits several earlier retired legacy versions. [1] [2]

Both application-type timelines give the same result for this clinical start. SDTMIG 3.3's requirement began on 15 March 2022 for [1] and 15 March 2023 for [2]. SDTMIG 3.4 has one requirement-begin date, 15 March 2025, covering both. [1]

Source: FDA Data Standards Catalog v11.0, Submission Data Standards. Selected CBER/CDER clinical data, definition and transport rows. Dates reproduce the sheet values. An empty end cell means no end date has been established. Window includes the start means that the start follows the listed support and requirement beginnings and both ends remain open. [1] [2]

SDTM v1.1 sits outside the ended-guide pattern: support ended on 28 January 2015, and the row has no requirement date. [1] SDTM v1.4 and SDTMIG 3.2 end on 13 December 2023, the day SDTM v2.0 and SDTMIG 3.4 begin support. SDTMIG 3.3's end cells stay empty. [1]

The earlier outcome-instruments report already describes the coexistence of SDTMIG 3.3 and 3.4. [9] Here, the added comparison places those alternatives beside ADaM, Define-XML, transport and terminology decisions. Each component needs its own version entry. Nonclinical SEND rows are assessed separately below; an empty nonclinical CBER SDTM v1.5 row remains unclassified because its support-begin cell is also empty. [1]

Our reading: A later version's requirement-begin date leaves an earlier version's separate end-date fields unchanged. For the worked start, retain the still-supported alternatives in the decision record, then explain the chosen version. FDA's preference for the latest supported version provides a sensible starting point. [2]

A catalog window is only the first compatibility check. CDISC specifies that SDTMIG 3.4 is used with SDTM 2.0. Selecting a supported guide and a separately supported model does not make an arbitrary pairing valid. Record the intended model/guide pair, relevant errata, the terminology release and the Define-XML metadata convention in the same study specification. The catalog controls FDA support timing; the standards' own specifications control how their components work together. [10]

On 15 January 2024, support and requirement are different states

The same four columns, read on an earlier start, separate "FDA can accept it" from "the requirement is effective."

For a clinical start of 15 January 2024 on the footnote [1] clock, the SDTM rows are in three states. [1]

For a study starting on 15 January 2024, SDTMIG 3.2 has already reached its support end. SDTMIG 3.4 is supported from 13 December 2023, but its requirement date of 15 March 2025 is still in the future. Both application-type timelines give that result. This interval illustrates why support and requirement deserve separate fields in the planning record. [1] [2]

The same test puts ADaMIG 1.2 and 1.3 (shared catalog row) with SDTMIG 3.4: support began 18 July 2022 and the requirement begins 15 March 2024, so on this start the requirement has not begun. ADaMIG 1.1 includes the start. ADaMIG 1.0 and Define-XML 1.0 have ended. Define-XML 2.0, Define-XML 2.1, and CBER/CDER XPORT v5 include it. [1]

A supported version is available for consideration during the transition; the requirement window answers a separate timing question.

SENDIG-DART and SENDIG-AR still split by clock and center

A nonclinical start can produce different results by center and submission type. For CDER SENDIG-DART v1.1, a study-director signature on 1 September 2023 falls after the [1] requirement date and before the [2] requirement date. The catalog lists no CBER DART counterpart. SENDIG-AR has separate CDER and CBER rows, with different support and requirement dates. [1]

On 1 September 2023, the CDER Animal Rule row is within its requirement window for both application types. CBER support begins later, on 26 March 2024. Selecting by the property name alone would lose this center-specific distinction. [1]

The June 2026 guide maps study type in the SEND tables and says to prefer the catalog if the tables lag:

While these tables will be updated to reflect any changes, there may be a delay based on publication schedules; therefore, always refer to the FDA Data Standards Catalog for the most up to date information on data standards requirements. [5]

The guide's SEND table header applies to studies starting after its listed dates. CDER lists embryo-fetal development studies under SENDIG-DART v1.1 at 15 March 2023 for NDAs/BLAs and 15 March 2024 for commercial INDs. Its Animal Rule natural-history and efficacy dates are 15 March 2022 and 15 March 2023. CBER lists Animal Rule studies at 15 March 2027 for NDAs, BLAs and commercial INDs. Match both the date and the study type; retain the catalog's scope qualifiers for certain BLAs and INDs. [5] [1]

Section 4.1.3.4.3 discusses which developmental and reproductive designs the guide reads as inside SENDIG-DART v1.1. [5] This comparison stops at the study types the table names. Bolded dates are "the start of the SEND requirement for that study type." [5] With the "started after" header, a start on 15 March 2024 stays unclassified. The worked start avoids that boundary.

ADaMIG 1.2 and 1.3 share a date window

Catalog cell F9 contains two lines, ADaMIG v1.2 and ADaMIG v1.3, with shared dates: support from 18 July 2022 and a requirement from 15 March 2024. Both end cells are empty. These are two implementation-guide versions sharing one catalog window, not a combined version identifier. Record the single version actually used in the study's implementation and metadata. The workbook supplies no intervening retirement of v1.2. [1]

The instructions say that from version 10.0 the catalog is being restructured so implementation guides can be listed independently of the underlying data models. [1] The sheet note above the header says model rows are formatted in bold with a blue fill, that dependent properties sit beneath, and that sorting may obscure those dependencies. [1] Read each property with its underlying model and related documentation; preserve the sheet's dependency information when exporting or sorting.

ADaM v2.1 is the model row. Its requirement dates are 17 December 2016 and 17 December 2017, with no established ends. ADaMIG 1.0 has its own ended window. Model support and implementation-guide support are distinct checks, so both belong in the selected configuration. [1]

The catalog names no NCI terminology quarter

Takeaway: The CDISC terminology rows accept "06/10/2011 or later" and name no quarter. The table is the 26 August 2026 Q1 stamp against the 4 October 2026 Q3 stamp for SDTM, SEND, ADaM, Define-XML, DDF, and CDASH. The guide's "most recent version available at study start" sentence stays a should. [1] [3] [4] [5]

The terminology sheet has two CDISC rows maintained by NCI-EVS, both for CBER and CDER. Both list the accepted version as "06/10/2011 or later," support begins 13 June 2011, requirement begins 17 December 2016 on clock [1] and 17 December 2017 on clock [2], and both ends are empty. The clinical information source is the index of CDISC SDTM terminology. The nonclinical information source is the index of CDISC SEND terminology. [1] The accepted-version column "specifies version(s)/release currently accepted by the Centers indicated." [1] On these two rows that value is a floor date. MedDRA, WHODrug/UMCDD, LOINC, SNOMED CT, and UNII are other rows, with other providers. [1]

The catalog's accepted-version floor establishes a regulatory range; NCI's current and archived files establish which actual terminology release was available. The August snapshot documents its capture date, while reconstruction of an earlier study's terminology should use a retained study-start package or the relevant archive. [3]

The guide's terminology section is a should, addressed to whatever dictionary the study uses:

Sponsors should use the most recent version of the dictionary available at the start of a clinical or nonclinical study. If a new version becomes available after the start of the study, sponsors may use the most current version of the dictionary for that clinical or nonclinical study. [5]

The CDISC subsection says sponsors should use the CDISC controlled terminology at NCI Enterprise Vocabulary Services, and that the version of an external dictionary should be stated in the define file and in the full Trial Summary domain. [5] Under the MedDRA heading, the guide says there is no requirement to recode earlier studies. [5] That sentence stays under MedDRA. It was not applied to CDISC terminology.

On 4 October the SDTM, SEND, ADaM, Define-XML, DDF, and CDASH stamp files were byte-identical. Identical bytes are the same stamp text, and the package is the URL. In the August snapshot the CDASH stamp sat under the SDTM directory. [3] [4]

The catalog links the clinical and nonclinical CDISC rows to SDTM and SEND terminology. NCI distributes additional packages with their own release histories. A useful caution is Protocol terminology: CDISC announced its retirement with the 27 March 2026 release and replacement by ICH M11 terminology. An unchanged older Protocol stamp therefore needs that lifecycle context. A timestamp alone leaves retirement and ongoing maintenance indistinguishable. [1] [11]

The saved August snapshot and the October live retrieval establish two observed file states. A September release postdates a 1 June study start, so it can only be a later-version choice for that study. The conformance guide permits use of a newer dictionary after study start. Document the package actually adopted and any subsequent migration; use a dated archive or retained file to establish the original release. [3] [4] [5]

Our reading: Write the quarter, release date, and retrieval date from the stamp file for the package the study uses, next to the catalog floor. For a 1 June 2026 start, the 25 September 2026 release is after the start. For a later start, retrieve the then-current package rather than relying on the August snapshot.

Reconciling the guide's column references with the current workbook

Takeaway: The June 2026 guide says SDTIGVER should be the exact term in column F, example 3.2, and SDTMVER the exact term in column E, example 1.4. In Catalog v11.0, column E is SDO, column F is Property, and the property cell for the 3.2 guide is SDTMIGv3.2, not 3.2. [5] [1]

The Trial Summary instructions in the guide read:

SDTIGVER. SDTM IG Version. Should be the exact term listed in column F of the FDA Data Standards Catalog, Submission Data Exchange Stds tab. If multiple SDTM IG Versions are used for a study, then each version used should have a unique value for TSPARM. Ex. 3.2

SDTMVER. SDTM Version. Should be the exact term listed in column E of the FDA Data Standards Catalog, Submission Data Exchange Stds tab. If multiple SDTM Versions are used for a study, then each version used should have a unique value for TSPARM. Ex. 1.4 [5]

Catalog v11.0 has no tab named Submission Data Exchange Stds. The standards tab is Submission Data Standards. Change History for 20 March 2025 says the column order was adjusted. [1] Header row 5 reads, from column A: Use, Standard, FDA Center(s), Exchange Format, SDO, Property. Column E is SDO. Column F is Property. [1]

The parameter appendix repeats the old column references. This is a documentation mismatch, not evidence that a numeric version value such as 3.2 is intrinsically invalid. A study should report the version actually used, in the representation required by the applicable SDTM specification and current validation rules. Avoid copying the organization name CDISC from column E into a model-version field, or converting the Property label into a submission token by assumption. The historical examples 1.4 and 3.2 also need the study-start eligibility check above. [5] [1]

The eCTD check reads the study-start file, and XPORT v5 remains the transport row

The guide's technical-rejection checks use the study ID, the start date in ts.xpt, submission metadata and expected dataset files. They address package completeness and eligibility. They should be kept separate from the standards-conformance checks on the values inside each dataset. Appendix F of the same guide explains the study-ID and start-date linkage. [5]

For eCTD v3.2.2 the date comes from ts.xpt referenced in the study tagging file. The guide names error 1789 for a valid study tagging file, 1734 for a Trial Summary domain, 1736 for the required dataset files, and 1735 for the file tag within the study tagging file. [5] For eCTD 4.0 the date comes from ts.xpt associated with a Study Id Study Title keyword. The guide names US-eCTD4-516 for the Trial Summary domain, US-eCTD4-518 for the required dataset files, and US-eCTD4-517 for the document-type keyword. [5] Those numbers were read from the guide and were not re-run in a validator.

"when the SSD is after the established required deadlines, the study data must comply with the standards in the FDA Data Standards Catalog." [5] The study-data guidance says FDA relies on the submitter for the study-start date and the file type, and that the datasets generally needed include dm.xpt, adsl.xpt, and define.xml. [2]

The guide provides examples where SDTM and ADaM are inapplicable despite a start after 17 December 2016, including pilot and failed studies submitted to an ANDA. Its list is illustrative. Where those datasets are inapplicable, the simplified ts.xpt approach uses TSVALNF set to NA as specified by the guide. Establish the actual study type before applying that exception. [5] [1]

Catalog eCTD v4.0 support begins 16 September 2024, and the requirement-begins cell is empty, as are the requirement cells on the implementation-guide rows that begin support the same day. [1] That empty cell is separate from a validation check that runs when a sponsor submits in eCTD 4.0. The M2 eCTD specification row carries footnotes [5] and [6]. They point at the eCTD specifications guidance: NDAs, ANDAs, and BLAs on one date, and commercial INDs and master files on another. [1] That row was not assigned to clocks [1] and [2].

Define-XML is the other file that check expects. "The Catalog lists the currently supported version(s) of Define-XML. It should be noted that Define-XML version 2.0 or later is strongly preferred." [5] "2.0 or later" includes 2.0. The 2.0 and 2.1 rows both have empty end cells, so both include a 1 June 2026 start. [1] "Strongly preferred" stays inside the nonbinding guide. [5]

The CBER/CDER transport row remains SAS Transport XPORT v5, with no established end and requirement dates of 17 December 2016 and 17 December 2017. FDA's guide identifies XPORT v5 for electronic datasets and says CPORT-produced transport files are unsuitable for its processing and review. The catalog has a separate centers row for CDRH, HFP and CVM. [1] [5]

FDA's Study Data Standards Resources page describes Dataset-JSON as undergoing further testing for possible use as a replacement for XPT v5. On the catalog version reviewed here, the study-data transport entry remains XPORT v5. [7]

A team planning Dataset-JSON should discuss that approach with FDA in advance under the catalog's instructions for unlisted standards. A testing announcement establishes development status, while the supported-format decision comes from the catalog and any applicable agency agreement. [1]

A one-study selection record

The record is the decision. Fill it from the catalog row and from the NCI stamp, then configure the version the record names. The example is a hypothetical clinical study whose first informed consent among enrolled subjects is 1 June 2026, aimed at an NDA reviewed by CDER. Clock [1] is the footnote for NDAs, ANDAs, and certain BLAs. [1] [2]

Two other starts change named fields and leave the form alone.

On 15 January 2024, clock [1], the SDTM lines follow the January table. SDTM v1.4 and SDTMIG 3.2 have ended. SDTM v1.7 and SDTMIG 3.3 include the start. SDTM v2.0 and SDTMIG 3.4 are supported, and the requirement is not yet effective. The shared ADaMIG 1.2/1.3 cell is in that same state: its requirement begins 15 March 2024. [1]

On 1 September 2023 the start definition switches to STSTDTC. CDER SENDIG-DART v1.1 includes the start on clock [1] and is before the requirement on clock [2]. CDER SENDIG-AR v1.0 includes the start on both clocks. CBER SENDIG-AR v1.0 begins support later, on 26 March 2024. There is no CBER DART row. [1]

Section II.D permits requests for a waiver from a specific version requirement, generally to a previously supported version. It excludes a general waiver to submit outside FDA-supported study-data standards. The guidance directs requests to cder-edata@fda.hhs.gov or cber-edata@fda.hhs.gov with a subject beginning Waiver Request, as early as planning and no later than protocol submission to the IND. FDA intends to respond within 30 days. An email request starts that process; an agency decision establishes the granted scope. [2]

Carry the selected versions into the study's data-standardization plan, mapping specifications, metadata and delivery tests. EDC collection, SDTM transformation and ADaM analysis are related stages with distinct specifications. EClinCloud describes study build and CRF design within its EDC offering; that capability claim establishes neither automatic submission conformance nor FDA endorsement of the platform.

Scope of the comparison

The comparison addresses the catalog's support and requirement windows for the covered FDA submissions. It leaves clinical appropriateness, study-specific mapping, statistical analysis quality and actual validation results to their respective review processes. The data support no estimate of rejection rates, migration costs or industry version uptake.

The worked dates deliberately avoid transition days. A start exactly on a listed boundary requires attention to the applicable notice, guidance and center advice. Historical terminology availability should be established from retained releases. No live submission validator was executed in this review.

The selection record is the use of the sheet. Section II.E encourages the latest listed version. An earlier row stays on the record while its end cells are empty.

Sources

1. U.S. Food and Drug Administration. "FDA Data Standards Catalog," version 11.0. Instructions; Column Descriptions, the four date columns; Submission Data Standards header row 5, the property rows cited in the tables, and notes including footnotes [1], [2], [5], [6], [11], and [12]; Submission Data Terminologies, the clinical and nonclinical CDISC rows and the Accepted Version(s) column description; Change History entry for 20 March 2025. File retrieved 4 October 2026, 783,240 bytes, SHA-256 591ff69ac3efc84e128fd9fd9e4ff231c4bf3793fee4a9f9ca0a9396ca9de9c6. https://www.fda.gov/media/185927/download

2. U.S. Food and Drug Administration. "Providing Regulatory Submissions in Electronic Format — Standardized Study Data." Guidance for Industry, Revision 2, June 2021. Sections II.A through II.E, including the study-start footnote, the binding-effect paragraph for section 745A(a), and the version-update examples. https://www.fda.gov/media/82716/download

3. National Cancer Institute, Enterprise Vocabulary Services. CDISC terminology current directory, including ReadMe.txt dated 29 March 2024 and the Publication Date Stamp files captured as the current tree on 26 August 2026. SDTM and SEND stamps: Q1 2026, release 27 March 2026, modified 30 March 2026. https://evs.nci.nih.gov/ftp1/CDISC/

4. National Cancer Institute, Enterprise Vocabulary Services. CDISC Publication Date Stamp files retrieved 4 October 2026. SDTM, SEND, ADaM, Define-XML, DDF, and CDASH: Q3 2026, release 25 September 2026. Glossary: Q3 2026, release 25 September 2026. Protocol unchanged at Q3 2025, release 26 September 2025. TMF unchanged at Q3 2024, release 27 September 2024. Lead file: https://evs.nci.nih.gov/ftp1/CDISC/SDTM/SDTM%20Publication%20Date%20Stamp.txt

5. U.S. Food and Drug Administration. "Study Data Technical Conformance Guide," June 2026. Contains Nonbinding Recommendations. Document-control paragraph; Trial Summary parameters SDTIGVER and SDTMVER; parameter appendix columns E and F; section 3.1.1 on XPORT v5; section 4.1.3 SEND tables; section 4.1.4.5, Data Definition Files for SDTM, SEND, and ADaM; section 6.1.2 and the CDISC and MedDRA subsections; section 8.1.2 eCTD validation paragraphs. https://www.fda.gov/media/153632/download

6. U.S. Food and Drug Administration. "Data Standards Catalog" guidance page. Last modified 24 March 2025. Download link /media/185927/download. Accessed 4 October 2026. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/data-standards-catalog

7. U.S. Food and Drug Administration. "Study Data Standards Resources." Content current as of 18 September 2026. Catalog and conformance-guide links, and the Dataset JSON preliminary-testing paragraph. Accessed 4 October 2026. https://www.fda.gov/industry/fda-data-standards-advisory-board/study-data-standards-resources

8. EClinCloud Insights. "How Many Systems Does Your First Trial Need? A Registry-Derived eClinical Module Map for Small Sponsors (2026)." 18 September 2026. Related reading on the 18 December 2016 and 2017 timing and on submission type. https://www.eclincloud.com/insights/first-trial-eclinical-module-map-2026

9. EClinCloud Insights. "Clinical Outcome Instruments: From Registry Names to CDISC QRS and Study-Build Decisions." 6 September 2026. Already states the SDTMIG 3.4 requirement start of 15 March 2025 and the open SDTMIG 3.3 row. https://www.eclincloud.com/insights/2026-09-outcome-instruments-cdisc-cde-readiness

10. CDISC. SDTMIG v3.4. Published 29 November 2021; model pairing and errata. Accessed 4 October 2026.

11. CDISC. Controlled Terminology, release update 27 March 2026. Protocol terminology retirement and replacement by ICH M11 terminology.