Clinical study folders tend to grow by accretion. An EDC export arrives, then lab files, then SDTM and ADaM datasets, define.xml, protocol versions, SAPs, and TFL outputs. If the folder structure is improvised, the file name becomes the only clue, and the latest cut may not be the one a reviewer opens. A better approach is source-first organization plus disciplined versioning and comparison. This guide walks through a practical method using a clinical trial data management platform as the shared workspace.
A clinical trial data management platform should not force the study team to invent a new filing logic. It should reflect the sources the study already has. Limora gives each study its own workspace and groups files by source such as EDC, SDTM, ADaM, lab, and documents. That structure is the foundation for everything that follows.
Step 1: Map each source to a home before uploading
Begin with a simple source map. EDC exports belong with CRF data. SDTM and ADaM datasets get their own areas. Lab data stays distinct from documents. Protocols, SAPs, annotated CRFs, define.xml, and TFL outputs belong with documents and outputs. The point is not to create a deep taxonomy for its own sake. The point is to ensure that when someone asks for the Week 12 lab review, no one has to guess whether the file sits beside the SAP, the raw lab transfer, or the analysis dataset.
Name the sources consistently across studies. If one study calls the source “EDC” and another calls it “clinical database,” searches and handovers become harder. The workspace should let you browse forms, visit schedules, datasets, and documents with subject filters, so the filing system supports both file-level and subject-level review. Keep the source map visible to the study team and apply it before the first upload. Retrofitting structure after several cuts is possible, but it takes longer and creates avoidable ambiguity.
Step 2: Treat every download, transfer, and rebuild as a separate cut
The most common versioning mistake is replacement. A new SDTM export arrives, someone drops it into the old folder, and the previous file disappears. The team then argues about which numbers were reviewed in the last meeting. Instead, make every upload a new revision with its own checksum. A checksum is a compact fingerprint of a file’s contents. If two files look alike but have different checksums, they are not identical. If a checksum matches, the file has not changed.
In Limora, every upload is a new revision with its own checksum. Nothing is overwritten, and each person chooses which cut they are looking at. That last point matters. Reviewers can open the cut they are responsible for without forcing everyone else into the same view. Keep cuts in order, label them clearly in your own standard, and avoid overwriting a cut with a corrected transfer. Upload the correction as the next cut. The revision history then tells the truth about what arrived and when.
This discipline also supports documents. Protocols, SAPs, and annotated CRFs change too. When a document is revised, keep the revision and make sure the data cut it describes remains identifiable. The workspace should not separate the data from the documents that define it.
Step 3: Compare cuts as a review activity, not a spot check
Once cuts are preserved, comparison becomes possible. Open a dataset, filter to the subjects you care about, and switch on compare mode to see what changed between cuts. Do not compare only row counts. Compare the records, fields, and subject-level values that matter to the question in front of you.
Use compare mode when an EDC transfer is corrected, when SDTM is rebuilt after a mapping update, or when an ADaM dataset is regenerated. Note the changes you expected, then inspect the changes you did not expect. If a subject’s value moves between cuts, attach a note or issue to the exact row or field. A comment that lives in a separate spreadsheet is easy to detach from the data; a note on the record keeps the discussion next to the evidence.
Document review works the same way. Search protocols, SAPs, and annotated CRFs page by page, and keep those pages beside the data they describe. If a reviewer questions a definition, the relevant page is within reach rather than buried in an email thread.
Step 4: Protect the context around sensitive files
File organization is not only about folders. It is also about who can see what and what happens when a question is raised. Blinding must be respected. In Limora, unblinded content needs its own grant; without it, those files, records, and search results stay hidden, whatever a person’s role. Build this into your source map from the beginning. Do not create a folder structure that quietly exposes unblinded material through document search or a shared view.
Access by study is equally important. People see only the studies they belong to, as a viewer, collaborator, or owner. Changes to membership, roles, unblinding, and records are written to an append-only audit trail. Records are retired or superseded, never removed, and every file revision is kept. These properties make the workspace useful during internal review and handover: the team can reconstruct what was seen, when it was seen, and which cut was current.
Step 5: Turn the method into a repeatable review routine
Before uploading a new cut, confirm the source, confirm the file name, and confirm whether the file replaces or supersedes an earlier item. After uploading, verify the checksum and record which cut is being used for the current review. Before a meeting, open the dataset, filter to the relevant subjects, and compare against the previous cut. During the meeting, capture issues on the exact form, field, row, or document page. After the meeting, check that open notes point to the correct cut and that superseded files remain visible as history.
A source-first workspace does not remove the need for judgment. It makes that judgment easier to audit. When files are grouped by source, cuts are kept as versions with checksums, and compare mode is used deliberately, the study team spends less time reconstructing what happened and more time reviewing what the data actually says.
Osaka quantum-physics postdoc now freelancing from Lisbon’s azulejo-lined alleys. Kaito unpacks quantum sensing gadgets, fado lyric meanings, and Japanese streetwear economics. He breakdances at sunrise on Praça do Comércio and road-tests productivity apps without mercy.