When building an SDTM dataset, the challenge lies in converting raw trial data into the standardized format. The variables ARM, ACTARM, and ARMNRS are particularly complex, as they are not always captured as straightforward data elements and often require assumptions to be made. Properly determining the values for these variables is crucial since they define the pooled population used in subsequent analyses.
In this blog, we’ll explore the considerations that need to be taken into account when assigning values to these variables, with a particular focus on Screen Failures.
Should Screen Failures Be Included in SDTM?
The first question that arises is whether Screen Failures should be included in the SDTM dataset. The answer is typically outlined in the study protocol. If Screen Failures are excluded, only subjects who pass enrollment and are assigned an ARM will be included in the SDTM data.
This blog will focus on the scenario where Screen Failures are included in SDTM.
Typical Study Flow and ARM Assignments
Consider the following simplified study flow:
The ARM, ACTARM and ARMNRS assignment would then be the following –
In this scenario, the ARM, ACTARM, and ARMNRS assignment would follow this sequence:
- Subjects pass Informed Consent and proceed to Inclusion/Exclusion Criteria.
- Subjects who meet the Eligibility Criteria are enrolled and assigned to an ARM.
- ARM and ACTARM are populated based on the treatment assigned to the subject during enrollment.
- ARMNRS reflects the Screen Failure status for those who do not meet the eligibility criteria.
This flow applies to both Double-Blinded and Open Label studies. The difference lies in the fact that in a Double-Blinded study, the ARM and ACTARM variables remain null until after database lock, or are populated with a placeholder ARM via a fictive randomization file. In contrast, in Open Label studies, the ARM and ACTARM are populated with the actual treatment at the enrollment stage.
Screen Failures: When Do They Apply?
In the case of a subject who has completed Informed Consent and Inclusion/Exclusion Criteria, they will be considered a Screen Failure if they do not pass the Additional Eligibility Criteria. However, if they meet the Additional Eligibility Criteria, they are no longer considered a Screen Failure but rather as a subject with no assigned ARM.
Protocol Review is Key:
It’s essential to carefully review the study protocol to determine how Screen Failures are defined. There are cases for example where Screen Failures include all subjects who were not treated (including the ones that got enrolled and assigned with an Arm), which would then override the approach discussed here.
Handling Subjects Without an Additional Eligibility Criteria Stage
In studies without an Additional Eligibility Criteria stage, the best practice is still to consider subjects who completed the Inclusion/Exclusion Criteria stage as Screen Failures. However, as with all aspects of clinical trials, protocol definitions take precedence. If the protocol differentiates between subjects who fail at the Informed Consent stage versus those who fail the Inclusion/Exclusion Criteria, you can use the value ‘NOT ASSIGNED’ to distinguish between these two groups.
Considering Other Sources of Data – Termination CRF
The above examples are relatively simple to follow, but in some cases, Screen Failure information is captured in other forms, such as a Termination CRF. Consider the following Termination CRF –
As you can see the first reason listed in Screen Failure which means we have another source of data we need to consider when defining Screen Failures –
Now, imagine a subject who reaches the enrollment stage, has an ARM assigned, but is marked as a Screen Failure. In such a case, unless the protocol doesn’t allow such a scenario (in which case it will be handled as a data issue), the ARM should be populated, but the ARMNRS field would be marked as ‘SCREEN FAILURE’. While this situation may not be common and typically is not allowed, it’s important to address it to accurately reflect the data.
Conclusions
In conclusion, Screen Failures can be represented in various ways across a clinical trial’s data. You may find yourself playing the role of a “data detective,” uncovering clues in different forms and data elements to determine whether a subject qualifies as a Screen Failure. A thorough review of the protocol and careful attention to data consistency are critical to making the right determinations.
