TrialCarta
How to check that the trial dates in your memo are still right
A diligence memo, forecast or investment thesis is right on the day it's written. Then sponsors move readout dates, trials stop enrolling and FDA extends reviews. Here is what goes stale first, where to check each fact by hand, and a check that does it for a whole document on your own computer.
Five facts that go stale
1. Readout timing
"Topline data in mid-2026" rests on the trial's registered primary completion date. Sponsors update it, often by several months, and rarely announce the change.
2. Trial status
A trial "currently enrolling" may now be active, not recruiting, completed, suspended or terminated.
3. Enrollment
Targets get raised or cut, and the anticipated number turns into an actual one when enrollment closes.
4. Primary endpoint
Endpoints are sometimes reworded, swapped with a secondary endpoint, or changed outright mid-trial.
5. FDA decision dates
A PDUFA date can move, most often by three months when FDA treats new data as a major amendment. Companies report the new date in their SEC filings.
How to check by hand
- Open each trial's record on ClinicalTrials.gov. Search by NCT ID and compare the status, the enrollment (and whether it says estimated or actual), the primary outcome and the primary completion date with what the memo says.
- Look at the record's history. The history of changes shows each version the sponsor registered, so you can see when a date moved and from what.
- Search SEC filings for FDA dates. In EDGAR full-text search, search the drug's name with "PDUFA" and read the newest 8-K, 10-Q or 10-K. Older filings can give a date that has since moved.
- Check for an FDA advisory committee. FDA announces each meeting in the Federal Register, usually four to six weeks ahead, naming the application and the drug.
- Re-read the memo for knock-on sentences. If a readout moved, a later sentence about the financing runway or the next milestone may be wrong too.
That works for a few trials. A landscape memo can cite twenty, and the check has to be repeated every time the memo goes out again.
An example
A short test memo on hereditary angioedema, checked on November 10, 2026. Four of its statements, with what the sources said that day:
| Verdict | The memo says | Now |
|---|---|---|
| Changed | “Topline data are expected in mid-2026.” (NCT06842823) | Primary completion registered for Dec 2026 |
| Changed | “…donidalorsen (NCT05392114), which is currently enrolling.” | Active, not recruiting |
| Changed | “PDUFA target action date of February 28, 2027” (lonvo-z) | March 10, 2027, quoted from Intellia's 8-K of Nov 5, 2026 |
| Holds | “…enroll approximately 150 patients.” (NCT06842823) | 157 participants (actual) |
The memo was written for testing. The trial and filing data are real, as of the date shown, and may have changed since.
What TrialCarta's document check does
Check My Document reads a memo, forecast or thesis (PDF, Word, RTF, text or Markdown) and finds the statements that depend on public facts: a trial's readout timing, enrollment, primary endpoint or status, and FDA decision dates. It checks each one against ClinicalTrials.gov and the newest SEC filing that gives a date, and marks it:
- Changed, with what the source says now and a link to it
- Holds, when the source still agrees, allowing for "approximately" and for ranges like "mid-2026" or "second half"
- Can't check, when a statement can't be tied to one trial, or an endpoint is worded too differently to compare
Statements are tied to trials by an NCT ID in the sentence or the one before it, by a trial name or drug the document cites elsewhere, or, marked for you to confirm, by the one open trial matching a drug code.
Every verdict comes from fixed rules, not an AI model, and the document isn't changed. Afterwards, an AI model on your own computer drafts new wording for the changed statements in the memo's own style, with every quote checked against the document. You can also keep the memo on watch, so you hear when another statement stops holding.
Only trial IDs and drug names are searched for. The memo itself never leaves your machine.
What it can't do
It doesn't read Excel or PowerPoint yet. FDA dates come only from companies that file with the SEC. Readout timing is checked against the registered primary completion date, which is when the last patient's primary measurement is due, not the day results are announced. And a statement with no trial ID, trial name or drug the document ties to a trial is marked Can't check rather than guessed.
Common questions
Is the primary completion date the same as the readout?
No. It's when the last participant's primary outcome is measured. Topline results usually follow within weeks to a few months, so a primary completion date that moves later usually means the readout moves too.
Why would a PDUFA date change?
Most often because the company submitted substantial new information during the review and FDA extended the goal date, typically by three months. FDA can also act before the date.
Does the memo get uploaded anywhere?
No. TrialCarta runs on your own computer, or on your team's own server. It sends only trial IDs and drug names to ClinicalTrials.gov and SEC EDGAR.
Check a memo before it goes out again
The document check is part of TrialCarta, alongside the catalyst calendar and the financing screen. We'll run it on a test memo like the one above.
Ask for a walkthrough