CDISC SDTM and ADaM Basics for Non-Statisticians: What Clinical Teams Need to Know

Why CDISC matters to Clinical Operations (even if you never build a dataset)

CDISC standards (notably SDTM and ADaM) shape how regulators review your study. Clinical teams influence CDISC success more than many realize because protocol design, CRF structure, and operational decisions determine whether data can be standardized cleanly.

This section is operational guidance only and not legal advice.

Inspection and submission reality

• Reviewers need to trace: protocol requirements → collected data → standardized datasets → analysis outputs.
• When traceability is weak (unclear visit definitions, inconsistent date handling, undocumented changes), teams spend time on remediation late in the study.
• Strong SDTM/ADaM traceability supports inspection readiness because it forces clarity on what happened and when.

SDTM basics: think “standardized way to describe what was collected”

SDTM is primarily about organizing collected observations into standard domains with controlled terminology and consistent metadata. For clinical teams, the key is understanding what operational decisions create SDTM complexity.

Common SDTM domains you touch indirectly

• DM (Demographics): subject identifiers, key dates, site info.
• SV (Subject Visits): planned vs actual visit dates—highly sensitive to protocol window rules.
• AE (Adverse Events): onset/stop dates, seriousness, action taken—connect to PV workflows (see PV & Safety Reporting).
• CM (Concomitant Medications) and MH (Medical History): baseline vs on-study distinctions.
• LB (Laboratory), VS (Vital Signs), EG (ECG): unit consistency and reference ranges depend on operational setup.
• EX (Exposure): dosing records and accountability processes.
• DS (Disposition): discontinuations, completion status, reasons.

SDTM pitfalls caused by operational ambiguity (and how to prevent them)

• Unclear visit definitions: If “Visit 3” can occur across a wide window without clear anchoring rules, SV and timepoint mapping become messy. Prevention: define anchor dates and window logic precisely.
• Inconsistent date capture: Partial dates or conflicting date sources create imputation burdens. Prevention: standardize source expectations and train sites; align with ALCOA+ principles (see ALCOA+ Data Integrity).
• Unscheduled visits not standardized: Unplanned assessments need consistent labeling and source documentation. Prevention: define unscheduled visit conventions and when forms should be used.
• Endpoint method drift: different devices or methods used across sites without documentation. Prevention: control equipment, training, and deviations; trend via RBM (see RBM That Works).

ADaM basics: analysis-ready data and traceability

ADaM datasets support statistical analyses. They include derived variables (e.g., baseline flags, analysis windows, change from baseline) and are designed to be traceable back to SDTM and ultimately to the source data.

Where ClinOps impacts ADaM quality

• Protocol window enforcement: analysis windows depend on documented visit timing and deviations.
• Deviation documentation: if deviations are poorly captured, analysis populations and sensitivity analyses become harder (see Protocol Deviations and CAPA).
• Treatment start/stop clarity: exposure dates and dose holds must be clear to support analysis.
• Blinding events: unblinding needs precise documentation and control.

Example: why a single missing date becomes an analysis problem

If an AE onset date is missing or inconsistent, the analysis may not be able to determine whether the event occurred treatment-emergent. That can cascade into safety outputs and narratives. Prevention is operational: clear source expectations, prompt query resolution, and reconciliation between safety and EDC systems.

What Clinical teams can do early to reduce CDISC pain later

Most CDISC remediation happens late because early protocol/CRF decisions weren’t mapped to standardized outcomes. Small steps early can save significant time and reduce inspection risk.

ClinOps checklist (early and ongoing)

• Participate in CRF design review with a “standardization lens” (visit naming, dates, units, controlled terms).
• Ensure protocol schedule and window rules are operationally executable and unambiguous.
• Define conventions for unscheduled visits, repeat assessments, and re-tests.
• Establish a query management cadence and escalation rules for critical data fields.
• Agree on where key records live and how they will be filed for retrieval (see TMF/eTMF Excellence).
• Use RBM to trend data quality risks that commonly drive CDISC issues (late data, high change rates, inconsistent units).

When clinical operations, data management, and biostatistics share a common set of definitions and escalation rules, SDTM/ADaM becomes less of a “final submission scramble” and more of a controlled output of your trial processes.

CDISC deliverable artifacts: what to expect and how to review them

Even if Clinical Operations does not author SDTM/ADaM deliverables, you can strengthen quality by knowing what the deliverables are and what “reviewable” looks like.

Common CDISC artifacts (high level)

• Annotated CRF (aCRF): shows where each CRF field maps into SDTM variables/domains.
• SDTM mapping specification: detailed mapping rules, controlled terminology, and derivations.
• ADaM specification: analysis dataset structure and derivation rules aligned to statistical analysis plan.
• Define.xml: metadata “map” of the submission datasets.
• Reviewer’s Guides: explains key decisions, conventions, and known limitations.

What ClinOps can review effectively

• Do visit labels and visit dates make sense relative to how the trial operated (SV domain logic)?
• Are protocol deviations that affect endpoints visible in the data story (e.g., missed visits, out-of-window visits)?
• Do key safety events (deaths, SAEs) reconcile between AE-like data and safety reporting records?
• Are units and reference ranges consistent with vendor setup (labs, devices)?

For inspection readiness, decide where these deliverable artifacts will be filed and how retrieval will work (see TMF/eTMF Excellence).

Visit structure and window rules: a small protocol choice with big SDTM impact

Visit-level ambiguity is one of the most common drivers of late mapping debates. The SV domain depends on consistent definitions of planned vs actual visits and how unscheduled assessments are represented.

Example: baseline anchored to randomization

Protocol rule: baseline labs must be within 7 days prior to randomization; dosing occurs on Day 1; follow-up visit windows are ±3 days from nominal day.

Operational questions you should resolve early:

• If screening labs are repeated due to an out-of-range value, is that a repeat screening visit or an unscheduled assessment?
• If the Day 15 visit occurs on Day 20 due to hospitalization, how is the event documented and how does it affect endpoint windows?
• What is the rule for “visit date” when procedures occur on multiple days (e.g., multi-day imaging)?

Clear answers reduce late re-derivations and reduce deviation/CAPA noise when windows are missed (see Protocol Deviations and CAPA).

Controlled terminology and coding: why “consistent words” are compliance controls

CDISC relies heavily on controlled terminology. Operational inconsistency—different site phrasing, mixed units, inconsistent “reason” selections—creates mapping complexity and can obscure safety signals.

Operational controls that help

• Use controlled picklists in EDC for key fields rather than free text when feasible
• Train sites on preferred terms and unit standards (labs, vitals)
• Align AE seriousness and outcomes capture with PV workflows to reduce reconciliation churn (see PV & Safety Reporting)
• Use RBM to detect sites with unusual coding patterns or high free-text usage (see RBM That Works)

Change control: protocol amendments and mid-study changes ripple into SDTM/ADaM

When protocol amendments change assessments, windows, or endpoints, the SDTM/ADaM mapping may need updating. ClinOps can reduce risk by ensuring amendments are implemented cleanly and documented, including:

• Clear effective dates by site and by subject (especially for re-consent and new procedures)
• EDC build change documentation and UAT evidence for critical changes (see CSV vs CSA)
• Consistent filing of amendment approvals and implementation communications in the TMF/eTMF

These controls improve both submission readiness and inspection readiness because they keep the “why” behind data changes visible and retrievable.

Inspection perspective: CDISC traceability is an operations story

When reviewers ask about SDTM/ADaM, they are often testing traceability: can your team explain how protocol requirements became collected data and then became standardized and analyzed outputs? ClinOps supports this traceability by ensuring the operational record is coherent.

High-yield questions to prepare for

• How were visit windows managed and what proportion of key assessments were out of window?
• How were unscheduled assessments handled and represented in the data?
• How were protocol deviations captured, trended, and corrected?
• How were safety events reconciled between clinical and safety systems?

If you can show that these topics are controlled operationally (monitoring, central review, reconciliation, CAPA), the SDTM/ADaM story usually follows naturally (see Inspection Readiness). Consider maintaining a simple cross-reference sheet that lists where the aCRF, mapping specs, define.xml, and reviewer’s guides are filed, plus who can explain them. That small retrieval aid can save significant time during interviews and reduces the risk of producing partial or incorrect dataset documentation. When dataset conventions change (e.g., after an amendment), update the sheet and document the rationale so the “story” remains consistent and ownership remains clear across teams and vendors.

ClinOps checklist: decisions that make or break SDTM/ADaM downstream

Clinical Operations does not need to memorize CDISC variable names to influence CDISC success. What matters is agreeing (and documenting) operational conventions that reduce ambiguity and late rework. Use the checklist below during protocol finalization and again during CRF/UAT review.

1) Endpoint and assessment conventions

• What counts as the assessment? Define the exact method, device, and conditions (fasting, posture, time of day) where relevant.
• What if it’s repeated? Define which record is primary (first, best, average) and how repeats are documented.
• What if it’s missed? Define “not done” reasons and what follow-up is expected; ensure sites can operationalize it.
• Visit window rules: define anchor dates, allowable windows, and how “unscheduled” assessments are labeled.

2) Date/time and chronology rules (high impact)

Many reviewer questions revolve around chronology: did the AE occur before dosing? was the endpoint assessment within the window? was the consent signed before procedures? To support traceability:

• Standardize when time is required vs optional (e.g., dosing time, AE onset time, safety event reporting).
• Define how partial dates are handled and where the authoritative date comes from when sources differ.
• Train sites on contemporaneous documentation and the “reason for change” expectation for corrections (see ALCOA+).

3) Medical history and concomitant medications: baseline vs on-study consistency

Mapping complexity increases when baseline/on-study boundaries are unclear. Define operational guidance such as:

• What time frame is captured as medical history vs AE (and what the site should do when it’s unclear).
• How to capture ongoing medications and changes (start/stop, dose changes, route changes).
• How prohibited/concomitant medication checks are documented and trended.

4) Safety data alignment: reduce reconciliation churn

ClinOps can reduce SDTM/ADaM cleanup by ensuring AE documentation is consistent and timely. Practical alignment points include seriousness criteria capture, consistent onset/stop date rules, and clarity on how follow-up information is documented. Plan reconciliation between EDC and safety systems (see PV Workflows).

Timeline and governance: avoid the “submission scramble” by aligning early

Teams often discover CDISC problems late because decisions were never owned: units were inconsistent, visit windows were interpreted differently across regions, or CRFs lacked “not done” reasons. A lightweight governance model prevents this without slowing the study:

• Pre-build alignment session: ClinOps, DM, Stats, and key vendors agree on visit conventions, timepoint rules, and key derivations that affect endpoints.
• CRF/UAT review focus: ClinOps reviews CRFs for operational feasibility and ambiguity, not formatting. Identify fields that will be painful to standardize (free text, missing units, unclear date precision).
• Early trend review: once enrollment starts, use RBM/central monitoring to detect mapping-impacting patterns (late entries, high edit rates, inconsistent unit usage) and intervene quickly (see RBM).
• Amendment impact checks: treat amendments as change events that can alter datasets and derivations; document effective dates and update mapping decisions accordingly (see CSV vs CSA).

Finally, ensure CDISC-related artifacts are retrievable: aCRF, mapping/specs, define.xml, reviewer guides, and change history. Retrieval is part of inspection readiness (see Inspection Readiness) and should be supported by your TMF plan (see TMF/eTMF Excellence).