How to Check Whether a Clinical Trial Was Published
Find a clinical trial’s published results using its NCT number, registry results, PubMed and publisher records, with a worked STEP 1 example.
Key takeaways
- Completed does not mean published.
- Check registry results and journal reports separately.
- Match identifiers, populations and primary outcomes.
In this article
Check both the trial registry and publication databases: posted registry results and a published journal article are different forms of reporting. “Completed” describes study status, not proof that results were published. An empty publications section is also not proof that no paper exists. NLM recommends checking ClinicalTrials.gov results and searching PubMed for the study or identifier.1
This walkthrough starts with a known trial. To find relevant studies first, use our clinical-trial evidence guide. The evidence center explains what registration can and cannot establish.
Start with the exact NCT identifier
An NCT number identifies a ClinicalTrials.gov record. Copy the complete identifier, including the eight digits after “NCT.” Search the number directly and confirm the trial’s intervention, sponsor, population and primary outcomes. A familiar acronym alone may be ambiguous.
Record the date you checked. Registry records are supplied by sponsors or investigators and can change; registration is not government endorsement of every study’s science or safety.
Check three places for reporting
| Place to look | What it can provide | What it does not establish |
|---|---|---|
| Registry Results Posted tab | Structured summary outcomes and adverse events, when submitted and posted | A peer-reviewed journal article |
| Registry Publications section | Listed or linked reports, when available | A complete inventory of every paper |
| PubMed and publisher | Journal report, protocol, follow-up or secondary analysis | That the report answers the primary trial question |
ClinicalTrials.gov’s modernized display uses a Results Posted tab when summary results are available. Its Publications section may be absent when no publications are listed or linked. Interface labels can change; follow the record’s current navigation.2
Worked example: STEP 1
The STEP 1 trial’s identifier is NCT03548935. Its original journal report is Wilding and colleagues’ 2021 semaglutide trial, PMID 33567185, DOI 10.1056/NEJMoa2032183.3,4
1. Open the registry record
Use NCT03548935. Confirm that it concerns semaglutide in adults with overweight or obesity, rather than another STEP study. Compare its planned primary outcomes with the report.
2. Search PubMed independently
Paste NCT03548935 into PubMed. If the identifier search misses a report, try the trial acronym plus intervention, author or publication title. NLM’s PubMed guide describes citation searching and associated-data fields. Not every paper has an indexed trial number.5
3. Match the paper to the trial
The primary report randomized 1,961 adults without diabetes, compared semaglutide with placebo alongside lifestyle intervention, and measured weight outcomes at 68 weeks. Those features help distinguish it from a follow-up, subgroup or different trial.4
4. Identify what the paper actually reports
Read the title, abstract, methods and registered identifier. A protocol describes planned methods; it is not a results paper. A secondary analysis may address body composition or a subgroup while the primary report answers a different question. Finding either does not justify saying all trial outcomes have been fully reported.
5. Check the current version
Follow the publisher link and look for corrections, retraction notices or updated supplementary material. Retain the PMID, DOI, NCT number and access date in your notes so someone else can repeat the search.
What “no results found” should mean
After checking the registry and searching the identifier, acronym, intervention and investigators, describe the search precisely: “I did not find a matching results report in these sources as of this date.” Avoid converting that into “the trial failed” or “nothing was published anywhere.”
Possible explanations include incomplete linking, publication delay, a paper indexed without the NCT number, a terminated study, or genuinely unreported results. Which explanation applies requires evidence. Missing reporting reduces what readers can assess; it does not reveal the outcome by itself.
Reporting deadlines require context
U.S. reporting rules apply to defined categories of trials and allow specified exceptions or extensions. Not every registered study has an identical legal deadline. A reader should not infer noncompliance simply from “completed” plus an empty results tab. Consult the official reporting-requirements page for the study’s situation.6
A posted summary and a journal paper can complement each other. If the denominators or outcomes differ, investigate whether the papers use different analysis populations, follow-up periods or estimands. Our results-table guide explains those distinctions.
References
- NLM: How do I find the results of a clinical trial?.
- NLM: study-record displays in modernized ClinicalTrials.gov, 2024.
- ClinicalTrials.gov: STEP 1, NCT03548935. Official record checked 10 October 2026.
- Wilding JPH et al. Once-weekly semaglutide in adults with overweight or obesity, 2021. PMID 33567185; DOI 10.1056/NEJMoa2032183.
- NLM: PubMed User Guide.
- ClinicalTrials.gov: reporting requirements.
Sources and navigation checked 10 October 2026. No claim of exhaustive searching or independent medical review.
