The short version
- A reviewer's guide is the orientation document for a set of study data: what is in the package, how it is organized and what is unusual about it.
- FDA's Technical Conformance Guide recommends one for tabulation data and one for analysis data, and points to the PHUSE templates.
- The most useful guide is specific to the study, explains every remaining validation finding and is drafted early, not assembled the week before submission.
What the guides are
There are two for most clinical studies. The clinical study data reviewer's guide (cSDRG) describes the tabulation data: the SDTM version, the domains, how the annotated CRF maps to them, sponsor-defined choices and any data issues. The analysis data reviewer's guide (ADRG) describes the analysis data: the ADaM datasets, how they support the analyses in the statistical analysis plan, key derivations, and how a reviewer can reproduce the primary results.
FDA's Study Data Technical Conformance Guide recommends both and refers sponsors to the templates developed by PHUSE [1]. The PHUSE packages include a template, completion guidelines and worked examples [2, 3]. Using them is not about formatting: the reviewer has seen hundreds of guides in that structure and knows where to look.
Read it as a reviewer would
A reviewer has the guide, the define.xml and the datasets open together. Their questions are practical. Which dataset holds the primary endpoint? Which population flag is used for the primary analysis? Why does this domain contain records with no visit? Why does a validation rule fire 300 times? If the guide answers these in a sentence each, the review moves on. If it does not, the reviewer writes a question, and the clock starts on a response.
The Technical Conformance Guide is explicit about one of these: conformance issues that are not corrected should be explained in the reviewer's guide [1]. The best guides list each remaining validation finding with the rule, the number of records, the reason and why it was not fixed. "Data collected as such" is a reason only if it says what was collected and why.
Common faults
| Fault | What the reviewer sees | Better |
|---|---|---|
| Template text left in | Instructions or examples from another study | Every section written for this study |
| Findings list without reasons | A copy of the validation report | Rule, count, reason, why not fixed |
| No route to the primary result | Dataset list only | The datasets, variables and flags behind the primary analysis, in order |
| Versions disagree | SDTM or terminology version differs from define.xml | One source for versions, checked before release |
| Written last | Gaps and contradictions with the define | Drafted at the first transfer, updated at each one |
A synthetic example
The fictional ZEPHYR-101 team drafts its ADRG at the first dry run instead of at submission. Writing the section on the primary analysis exposes that two programmers derived the analysis visit window differently in two datasets. Fixing it costs a day in October. The same discovery during a regulatory review would have cost a response letter and a re-run.
Do this today
- Open the PHUSE ADRG and cSDRG templates and compare their section headings with your current drafts.
- Write the paragraph that leads a reviewer from the primary endpoint to the dataset, variables and population flag behind it.
- Turn your latest validation report into a table of remaining findings with a reason for each.
Check yourself
One question
Show the answer
Answer 2. FDA's Technical Conformance Guide asks sponsors to explain conformance issues that remain in the reviewer's guide. An unexplained finding costs the reviewer time and invites a question; changing collected data to satisfy a rule breaks traceability.
Apply it
Turn it into a plan
The brief says what to check. The Navigator and the Handbook turn it into dated, owned actions for your milestone, in your browser, with no study data.
Sources
References
- U.S. Food and Drug Administration. Study Data Technical Conformance Guide: Technical Specifications Document. www.fda.gov ↗
- PHUSE. Clinical Study Data Reviewer's Guide (cSDRG) Package: template, completion guidelines and examples. advance.hub.phuse.global ↗
- PHUSE. Analysis Data Reviewer's Guide (ADRG) Package: template, completion guidelines and examples. advance.hub.phuse.global ↗
Standards and guidance change. Confirm the version in force for your study, sponsor and region before relying on any reference. Examples are synthetic and illustrative. This brief is generalized professional guidance, not regulatory advice.
Next brief · Thursday 10:00 ET · Capacity Math