Astraea / Technical FAQ
The evidence behind the workflow.
Practical answers on data standards, traceability, human review, and deployment inside your environment.
How does Astraea support SDTM mapping?
Astraea proposes mappings from collected study data to SDTM domains. Programmers review the mapping rationale, controlled terminology, and study-specific interpretation before accepting the output. SDTM defines how tabulation data are organized; the applicable implementation guide and submission requirements must also be selected.
How are ADaM datasets connected to their sources?
Astraea connects analysis datasets to SDTM inputs and the statistical analysis plan, keeping derivations and source-to-result lineage available for review. CDISC’s ADaM standards support analysis reproducibility and traceability; study-specific analysis decisions still require statistical review.
What should reviewers check in Define-XML?
Review dataset and variable metadata, terminology, origins, derivations, and links to supporting documents against the delivered data. Applicable versions and requirements should be checked against the relevant regulator’s current data standards catalog and study context; generation alone is not proof of submission readiness.
What does source-to-result traceability mean?
A reviewer should be able to identify the source records, transformations, analysis population, derivation logic, and program version that produced an output. Astraea keeps those relationships available in the workflow so a result can be investigated in context.
Who approves the clinical outputs?
Qualified statisticians and programmers retain responsibility for reviewing methods, checking outputs, and approving deliverables. Automation supports execution; expert judgment stays with the study team. For the distinction between software execution and outsourced trial delivery, compare Astraea with a CRO.
What happens when an automated proposal is uncertain?
Reviewers evaluate ambiguous mappings, derivations, or analysis choices against the protocol, analysis plan, standards, and source data. A proposal should remain a review decision until its interpretation and supporting evidence have been checked.
Does an air gap establish HIPAA or Part 11 compliance?
No. An air gap can restrict network connectivity, but compliance also depends on the intended use, agreements, safeguards, validation, access controls, records, and operating procedures. Those responsibilities must be assessed for the deployed system and the client’s use case.
How should 21 CFR Part 11 be evaluated?
For applicable electronic records and signatures, evaluate validation, authorized access, record protection, audit trails, and signature controls in the context of the underlying regulatory requirements. Part 11 is a regulation, not a product certification. Astraea’s controls must be evaluated alongside the client’s procedures and validation approach.
How should a team verify an analysis can be reproduced?
Retain the approved inputs, program and dependency versions, parameters, population definitions, and execution context. Then rerun a defined example and compare it with the reviewed result. Traceability explains where a result came from; a controlled rerun tests whether it can be reproduced.
What belongs in a practical quality-control review?
Check conformity to the chosen standards, consistency with the statistical analysis plan, derivation logic, population and denominator definitions, expected values, and independent checks appropriate to the output. Review validation findings and their disposition rather than relying on a single automated pass indicator.
How is the client environment evaluated?
Technical evaluation establishes infrastructure, access and identity controls, model and network configuration, logging, retention, backup responsibilities, and change management. Deployment requirements and operational ownership should be documented before a production workflow is approved.