A date attached to a medical discovery feels concrete. “Trials in 2027” or “available by 2030” can sound more useful than a list of uncertainties. But the usefulness of a date depends on what it describes and who supplied it.
This is our editorial method for reading those dates. It is not a model that predicts approvals.
Give every date a label
Record the event, not just the year. An announcement can concern a planned study, permission to proceed, the first person treated, expected final follow-up, publication of results or a proposed commercial launch. Do not combine those into a single “release date.”
Then record the speaker. “The developer expects” means something different from “the regulator decided.” Both can be newsworthy, but only the latter describes a regulator's completed decision when supported by the actual decision document.
Separate expectation from observation
Use two fields: expected event and observed event. Leave the observed field blank until there is attributable evidence that it occurred. If the expected date passes without a located update, write “no confirming update located in the sources checked,” with a check date.
Do not silently convert that absence into “canceled” or “failed.” It may justify further reporting. It does not identify the reason for the missing update.
The TRG035 announcement report demonstrates why an administrative step and human results need separate entries. The question is not whether the program is exciting; it is whether the source answers the question being asked.
Approval is one question; practical access is another
A regulator's decision needs its own record, including jurisdiction and use. Access needs additional reporting about actual delivery, eligibility, services and payment arrangements. Do not infer all of those from a single approval headline.
For example, an approval letter can establish that a specified product was authorized for a specified indication. It cannot tell a reader, without further evidence, whether a particular hospital has appointments or whether an insurer will cover their circumstances. Our Otarmeni report links the decision rather than inventing those answers. 1
A timeline readers can trust
A useful entry has a source URL, event date, date checked, source type and one sentence explaining what changed. A prediction should remain visibly a prediction even after being repeated by several publications.
If a date changes, preserve the earlier estimate and explain the change when a source gives a reason. Quietly replacing the old date erases the information readers need to understand uncertainty.
The answer may be “not established”
“Not established in the evidence reviewed” is more informative than a confident invented year. It tells the reader what the page can support and what remains an open reporting task.
We do not provide individual treatment timetables. We provide a way to distinguish an observed milestone from an expectation. A trustworthy timeline is a record of developments, not a countdown clock added to make a story feel closer to a sale.