Most people meet a clinical trial through a press release. The trial itself lives somewhere far less exciting: a public database entry with date fields, a status line and an enrolment count. Anyone on earth can open it in a browser. Almost nobody does, which is how one percentage from a sponsor announcement ends up carrying more weight in public argument than the 89-week study design sitting behind it.
That gap is worth closing, because reading a registry record is a skill that takes about five minutes to learn and pays off every time a new compound starts making noise.
What a registry entry is, and what it is not
ClinicalTrials.gov is operated by the US National Library of Medicine, part of the National Institutes of Health. Sponsors register a study at or near its start and update the entry as the study runs, closes and reports. For a large share of trials conducted under US jurisdiction, registration and the later posting of summary results are legal obligations rather than courtesies.
What the database does not do is judge. A record is not an approval. It is not a recommendation, not a safety finding, and not evidence that anything worked. A pharmaceutical company, a university lab and a single hospital department can all register studies in the same system using the same fields. The entry describes what a sponsor said it would do, and later, what it says it did.
Once you hold that distinction, the fields start to mean something.
One record, read field by field
Take the TRIUMPH-1 record on ClinicalTrials.gov, registered under the identifier NCT05929066. It is a Phase 3 master protocol run by Eli Lilly and Company, testing retatrutide against placebo in adults with obesity or overweight, including subsets of participants who also had knee osteoarthritis or obstructive sleep apnea.
The registry lists an actual study start of 10 July 2023, an actual primary completion date of 6 April 2026, and an actual study completion date of 30 April 2026. Overall status reads Completed. Actual enrolment was 2,335 participants. The entry was last updated on 21 August 2026. The design pairs an 80-week placebo-controlled treatment period with a 24-week extension, running roughly 89 weeks in total across as many as 24 study visits.
Nine words of that paragraph do the heavy lifting: actual, actual, actual, Completed, 2,335. They tell you the study is finished rather than planned, that the sample was large rather than exploratory, and that the dates in the record are observed events rather than projections. A reader who takes nothing else away from a registry entry should still take those.
Estimated and actual are different words for a reason
Registry entries mark almost every date and count as either estimated or actual. Estimated means the sponsor expects it. Actual means it happened, and the sponsor has come back to the record to say so.
This is the single fastest way to separate a running trial from a finished one without reading a word of prose. A Phase 3 entry showing an estimated primary completion two years out is a study in progress, whatever anyone claims about its findings. The same is true in reverse: a record that still says Recruiting long after its estimated completion date is telling you something about the trial’s pace, or about how carefully the sponsor maintains the entry.
Check the last update date while you are there. A record verified eight months ago and a record verified last week carry different levels of confidence about the status line sitting at the top.
Four things a record will never tell you
The first is whether findings have been through peer review. Sponsors announce topline results before publication as a matter of routine. Lilly reported positive 80-week topline results from TRIUMPH-1 on 21 May 2026, well ahead of full peer-reviewed reporting. A topline announcement is a company’s summary of its own data. It is useful and it is not the same object as a journal paper with methods, confidence intervals and an editor who pushed back.
The second is regulatory status. Nothing in a registry entry says a compound is approved, approvable or close to a filing. Retatrutide remains an investigational compound, and a completed Phase 3 record is a data point in that process rather than the end of it.
The third is what sits in a vial that someone else is selling. Material used in a registered trial is manufactured, characterised and released under pharmaceutical controls, with chain of custody documented from batch to participant. A registry identifier describes that supply chain and no other.
The fourth is context. The record gives you a primary outcome measure and the week it was assessed. It does not tell you how that endpoint compares with the endpoint another trial used, which is where most confused public comparisons actually begin.
Why an English-only database matters in Costa Rica
ClinicalTrials.gov publishes in English. There is no Spanish interface and no translated record. For a researcher or a curious reader in San José, the primary source on a compound like this one is written in a second language, while the marketing around it arrives in the first. That asymmetry is how bad information spreads in regional markets, and it is not the reader’s fault.
The honest fix is boring and effective: Spanish-language resources that repeat the registry’s framing instead of contradicting it. Costa Rican sites covering this compound sit at exactly that point. The FAQ on retatrutida.cr, for example, states in plain Spanish that retatrutide is an investigational peptide acting on the GLP-1, GIP and glucagon receptors, that it remains under scientific study, and that it is not approved for human use. That is the same status the registry implies, written in the language the reader thinks in.
The catalogue pages at Retatrutida Costa Rica follow the same line, describing each presentation as material for laboratory research rather than as anything a person is meant to take. Whether a reader agrees with that market existing at all, consistency between what a registry says and what a local page says is the minimum test any Spanish-language source should pass.
Five minutes, six fields
If you only ever check six things in a registry entry, check these:
- Overall status, and the date the status was last verified
- Phase, which tells you what question the study was built to answer
- Enrolment, marked actual or estimated
- Primary outcome measure, with the week it was assessed
- Allocation, masking and the comparator used
- Sponsor, and whether results have been posted to the record
Those six fields will settle most arguments you encounter online about whether a compound has been studied, how seriously, and whether anything has actually concluded.
None of this requires a scientific background. It requires opening the record instead of the headline about it. The TRIUMPH-1 entry has been publicly readable since July 2023, through every round of speculation about the compound it describes, and it was never behind a paywall for a single day of that time.
