ClinicalTrials.gov results reporting: who, when, what and how
By Mubarak MutesasiraUpdated 7 min read
Short answer:
ClinicalTrials.gov results reporting means entering a trial's summary results into the Protocol Registration and Results System (PRS) in a fixed structure: participant flow, baseline characteristics, outcome measures with their statistical analyses, and adverse events, plus administrative sections. Under FDAAA 801 and 42 CFR Part 11, the responsible party for an applicable clinical trial must generally submit results no later than one year after the primary completion date. Prepare the data in the PRS structure from the start, check it against the registered outcomes, and expect review comments before posting.
This guide is for trial coordinators, investigators, statisticians and research offices preparing a results submission. It explains the rules in outline; it is not legal or regulatory advice. Read the current rule, the ClinicalTrials.gov data element definitions and your institution's policy, and ask your research office or PRS administrator when in doubt.
Who must report results
Under US law, the duty to report results applies to applicable clinical trials (ACTs). In outline, these are:
| Type | Broadly covers | Broadly excludes |
|---|---|---|
| Drug and biological product trials | Controlled interventional studies of an FDA-regulated drug or biologic | Phase 1 trials |
| Device trials | Prospective studies of health outcomes comparing an FDA-regulated device with a control, and pediatric postmarket surveillance | Small feasibility studies |
| Both | Trials with at least one US site, or run under an IND or IDE, or of a product made in the US and exported for research |
The 2017 final rule, 42 CFR Part 11, set out the details: which trials are covered, the data elements, the deadlines and the results modules.
Other policies go further. NIH-funded trials are expected to register and post results under the NIH policy, including trials that are not ACTs, such as some phase 1 and behavioral studies. Some funders, journals and institutions have their own requirements, and trials in the EU also report to the EU Clinical Trials Information System. Check every policy that applies to your trial, not only FDAAA.
The responsible party submits: the sponsor, or a principal investigator the sponsor designates who meets the conditions in the rule.
When results are due
- Standard deadline: no later than one year after the primary completion date, which is the date the final participant was examined or received an intervention for the primary outcome.
- Delayed submission: a certification can delay results when the product is not yet approved and approval is being sought, or a new use is being sought, within the limits in the rule.
- Extension for good cause: may be requested before the deadline; it is not automatic.
- Secondary outcomes not complete by the primary completion date are added when they are.
Keep the primary completion date in the record accurate. Updating it to "actual" when it happens starts the clock, and an outdated estimate is one of the first things that makes a record look overdue. Penalties for non-compliance can include civil money penalties and, for federally funded trials, loss of grant funds.
The results modules, one by one
The PRS results section has a fixed order. Each module has required data elements; the table summarizes them.
| Module | What goes in it | Watch for |
|---|---|---|
| Participant Flow | Recruitment details, pre-assignment details, arms or groups, periods, and for each the numbers started, completed and not completed, with reasons | Numbers must add up across periods; "started" in the first period should equal enrollment after pre-assignment |
| Baseline Characteristics | Arms or groups, the overall number of baseline participants, age, sex or gender, race and ethnicity if collected, and other prespecified measures | The baseline number should agree with the participant flow, or the difference should be explained |
| Outcome Measures | Every registered primary and secondary outcome: title, description, time frame, unit of measure, measure type, measure of dispersion, number analyzed, and data for each arm | The title and time frame must match the registration; scales need their name, range and direction |
| Statistical Analyses (within each outcome) | Groups compared, type of test (superiority, non-inferiority, equivalence), the method, P value, and the estimate with its confidence interval | Say exactly what was compared and how; a P value alone is weak |
| Adverse Event Information | Time frame, how adverse events were defined and collected, all-cause mortality, serious adverse events, and other adverse events above a frequency threshold, by arm | The threshold for "other" adverse events cannot be above 5 percent ; all-cause mortality is required |
| Limitations and Caveats | Problems that limit interpretation, such as early termination or measurement issues | Leave empty only if none apply |
| Certain Agreements | Whether an agreement restricts the PI's right to discuss or publish results | Check with the contracts office |
| Results Point of Contact | Name or title, organization, phone, email | Use a role mailbox that will outlive staff changes |
| Documents | The protocol and statistical analysis plan, and the informed consent form where required, as PDF/A | Redact only what the rule allows |
How the review works
After you submit, ClinicalTrials.gov staff run a quality control review of the results. They do not judge the science; they look for apparent errors, deficiencies and inconsistencies. If they find some, they return review comments in the PRS. You correct the record and resubmit, and the cycle repeats until the record passes. Respond promptly: the rule sets a period within which corrections are expected.
The review criteria are published, and reading them before you start saves rounds.
Common errors and how to avoid them
- Vague outcome titles. "Efficacy" or "Quality of life" says nothing. Write "Change from baseline in HbA1c at week 24".
- Scales without context. Name the scale, its range, and whether higher means better or worse: "SF-36 physical component score, 0 to 100; higher scores mean better function" (or the scale you used).
- Measure type and unit mismatch. If the measure type is "Count of participants", the unit is "Participants", not "percentage" or "mmHg".
- Numbers that do not agree. The number analyzed in an outcome exceeds the number who started, or baseline totals differ from participant flow with no explanation.
- Time frame drift. The registered time frame is 12 weeks, the results say 24. Results must report the outcomes as registered; changes need an explanation.
- Missing dispersion. A mean without a standard deviation or a confidence interval.
- Thin statistical analyses. No statement of what was compared, no estimate, or a confidence interval without its level.
- Adverse event gaps. No time frame, no description of how adverse events were collected, no all-cause mortality table, or serious events missing from one arm.
- Stale protocol section. Overall status, actual enrollment and actual primary completion date not updated before results are entered.
For how to report the statistics themselves (effect sizes, confidence intervals, exact P values), the SAMPL guidelines apply as much to a results record as to a paper.
A worked example (invented)
Example (invented for illustration). A two-arm trial of a sleep intervention, 120 randomized, primary outcome "Change in Insomnia Severity Index score from baseline to week 8". The first submission lists the outcome as "Insomnia severity", measure type "Mean", dispersion "Standard deviation", 118 analyzed. Review comments ask for the scale's range and direction, the time frame, and why 118 were analyzed when the participant flow shows 112 completed. The team rewrites the title to match the registration, adds "ISI, range 0 to 28; higher scores mean more severe insomnia", gives the time frame as "Baseline and week 8", and explains in the population description that analysis used multiple imputation for six participants with missing week 8 data. The next round passes.
Checklist before you submit
- The protocol section is current: overall status, actual enrollment, actual primary completion date.
- You know the deadline and whether a delay certification or extension applies.
- Participant flow numbers add up across periods and match enrollment.
- Baseline numbers match participant flow, or the difference is explained.
- Every registered primary and secondary outcome is reported, with titles and time frames that match the registration.
- Each scale has its name, range and direction.
- Measure type, unit and dispersion fit the data.
- Statistical analyses say what was compared, the method, the estimate with its confidence interval, and the P value.
- Adverse events: time frame, collection method, all-cause mortality, serious events and other events above the threshold, by arm.
- Limitations, certain agreements and the point of contact are filled.
- Protocol and statistical analysis plan uploaded as PDF/A.
- Someone who did not prepare the record has checked it against the review criteria.
Doing it in AI for Research
AI for Research has a ClinicalTrials.gov results checklist among its Study checklists in Proposal Studio (/proposals?module=study-checklist, group "Checklists"). The blank template "Checklist: ClinicalTrials.gov results" is free in the editor and follows the PRS sections in order, from "Before you start" to "Review". The tailored version costs 3 credits: describe your trial (or upload the protocol synopsis and the registered outcomes) and the checklist is fitted to your arms, outcomes and adverse event collection.
It does not submit anything to the PRS, it does not decide whether your trial is an applicable clinical trial, and it is not legal advice. Your research office and the responsible party confirm what applies. To look up registered trials, Official Sources includes ClinicalTrials.gov at no credit cost. The Free plan has 45 credits a month; see pricing. More: clinical research checklists.
Learn more
- Site initiation visit checklist, the pillar of this series
- SAMPL guidelines for reporting statistics
- SAGER guidelines for reporting sex and gender
- Statistical power and sample size
- Clinical research checklists and AI for medicine