Understanding ADAPT in the SDTM TS Domain: Adaptive vs Non-Adaptive Trials
Understanding ADAPT in the SDTM TS Domain: Adaptive vs Non-Adaptive Trials
Posted by Sarath Annapareddy | May 14, 2025
The Study Data Tabulation Model (SDTM) plays a critical role in organizing and submitting clinical trial data. One of the key parameters regulatory agencies look for in the Trial Summary (TS) domain is the ADAPT parameter (TSPARMCD=ADAPT), which identifies whether the trial uses an adaptive design.
In this post, we’ll break down what ADAPT really means, when to use TSVAL=Y, and why multiple protocol amendments alone do not make a study adaptive.
What is ADAPT in the TS Domain?
The ADAPT parameter tells regulators whether the clinical trial used an adaptive design (TSVAL=Y) or a non-adaptive (fixed) design (TSVAL=N).
"Adaptive clinical trials allow for changes in design or hypotheses during the study based on accumulating data, without undermining the validity or integrity of the trial." – FDA Guidance
Example 1: Non-Adaptive Trial (ADAPT = N)
Non-adaptive trials follow a fixed protocol. Most traditional randomized controlled trials (RCTs) fall into this category.
| STUDYID | TSPARMCD | TSVAL |
|---|---|---|
| ABC123 | ADAPT | N |
This study had no pre-planned adaptations. Even if it had amendments later, it’s still not adaptive unless those changes were defined from the beginning.
Example 2: Adaptive Trial (ADAPT = Y)
Adaptive trials allow changes based on interim analyses—like adjusting the sample size, changing dosage, or dropping treatment arms—based on pre-specified criteria in the protocol.
| STUDYID | TSPARMCD | TSVAL |
|---|---|---|
| DEF456 | ADAPT | Y |
This trial included predefined decision rules for interim analyses and therefore qualifies as adaptive.
Are Protocol Amendments the Same as Adaptive Design?
No. A trial with multiple protocol amendments is not automatically adaptive.
Amendments that arise from operational challenges (e.g., recruitment issues, eligibility tweaks) are reactive and do not constitute an adaptive design unless they were planned from the beginning and tied to accumulating trial data.
Example: A trial adds Protocol V2 and V3 to expand inclusion criteria and extend timelines. There’s no original plan for interim decisions or modifications. → TSVAL = N
When is TSVAL Set to "Y" for TSPARMCD=ADAPT?
You should only set TSVAL = Y if all the following are true:
- The protocol explicitly defines it as an adaptive trial
- Pre-specified adaptation rules are documented (e.g., sample size re-estimation, early stopping)
- Interim analyses are part of the design with proper statistical control
Common Adaptive Design Features
- Sample Size Re-estimation
- Early Stopping for Efficacy or Futility
- Adaptive Randomization
- Dose Adjustment
- Group Sequential Design
Quick Decision Table
| Scenario | Adaptive? | TSPARMCD | TSVAL |
|---|---|---|---|
| Protocol includes pre-specified interim analyses | Yes | ADAPT | Y |
| Protocol had multiple amendments but no pre-specified adaptations | No | ADAPT | N |
| Original protocol allows dose adjustments based on data | Yes | ADAPT | Y |
| Traditional fixed design with no planned changes | No | ADAPT | N |
Example TS Domain Representation
| TSPARMCD | TSPARM | TSVAL | Description |
|---|---|---|---|
| ADAPT | Adaptive Design | Y | Indicates that the study has an adaptive design approach |
Regulatory Considerations
Both the FDA and EMA emphasize the importance of planning adaptations before trial initiation. If changes are made after the fact, they are considered amendments—not adaptive features.
"Adaptations should be planned and documented in the protocol before trial initiation. Unplanned changes or protocol amendments made during the study are not considered adaptive design features." – FDA Adaptive Design Guidance
Conclusion
Understanding whether a trial is adaptive or non-adaptive is essential for accurate SDTM submission and regulatory transparency. Adaptive trials provide flexibility and efficiency—but only when designed that way from the start. If you’re working on the TS domain, be sure to verify the protocol and statistical analysis plan before setting TSVAL = Y for TSPARMCD = ADAPT.
References
- CDISC SDTM Documentation – cdisc.org
- FDA Guidance on Adaptive Design – fda.gov
- EMA Guidelines on Clinical Trial Design – ema.europa.eu
- SDTM Implementation Guide (SDTM IG)