Registered clinical trials finish at ~46% and post results at only ~21% of completed trials — industry and NIH sponsors report results 2.5-3x as often as academic or state/local government sponsors

Summary

Of 386,834 clinical trials on ClinicalTrials.gov with any activity since 2020, 46.3% (178,992) have reached COMPLETED status. Of those completed trials, only 20.8% (37,254) have ever posted results to the registry's results database -- the rest finished but left the outcome unreported. Sponsor type explains most of the variation: industry sponsors report results for 39.4% of their completed trials and NIH for 38.7%, roughly 2.5 to 3 times the rate of academic/hospital sponsors (14.7%) and more than 10 times the rate of other state/local-government sponsors (3.0%). Completion rates themselves are far more uniform across sponsor types (44-56%) -- the real divergence is in whether a finished trial ever gets its results posted, not whether it finishes.

How this was measured

Query against health.clinical_trials (AskAmerica corpus, sourced from clinicaltrials.gov, restricted by table design to studies with a last-update-post-date on or after 2020-01-01). Completion is overall_status = 'COMPLETED'. Results reporting is the table's has_results flag, sourced from the API's top-level hasResults field -- it flags whether summary results were EVER posted to the results database, independent of whether that happened within the legally required window. Sponsor grouping used funder_type (INDUSTRY, NIH, FED, OTHER_GOV, NETWORK, OTHER -- OTHER is overwhelmingly academic medical centers, hospitals and individual-investigator-led trials) as the primary breakdown, with individual top lead_sponsor organizations shown for additional texture.

Individual large sponsors show the same split: Pfizer (65.6% completed, 59.1% with results), Novartis (60.2% completed, 35.7% with results) and the VA Office of Research and Development (57.2% completed, 51.5% with results) post results at multiples of Cairo University, Assiut University, Assistance Publique-Hopitaux de Paris and Riphah International University -- all large academic/hospital sponsors reporting results on well under 1% of their trials despite completing 16-62% of them.

How this compares to the published literature

Independent studies of FDAAA-mandated results reporting find a similar direction but higher absolute rates, because they use a narrower, legally-defined denominator (only trials the FDAAA Final Rule actually requires to report, generally later-phase interventional drug/device trials with US sites) rather than every registered study. A NEJM-style compliance analysis reported industry sponsors posting results for roughly 64.5% of FDAAA-applicable trials (about 50.3% within the mandated one-year deadline), versus roughly 31.4% for U.S. government-sponsored trials reported within one year -- both higher than this corpus's all-trials figures, as expected, since this table's denominator (all registered trials, including many not covered by the mandate: Phase 1s, non-US-only sites, observational studies) is broader and less compliance-selected than the FDAAA-applicable-only denominator the compliance literature uses. The direction is the same in both: government/academic sponsors lag well behind industry and NIH.

Caveats

• health.clinical_trials only carries studies with a last-update-post-date on or after 2020-01-01 -- this is a recent-activity slice of the registry, not its full ~30-year history, so completion and reporting shares here describe trials still being actively maintained in the database, not every trial ever registered.
• has_results measures whether results were EVER posted, with no deadline attached -- it is not the same measure the FDAAA-compliance literature uses (results posted within the legally mandated 1-year window), so the percentages here are not directly comparable in magnitude to published FDAAA compliance-rate studies, only in direction.
• enrollment_count is null for 100% of rows in this table, so trial size could not be used as an additional covariate.
• Sponsor-type groupings (funder_type) are ClinicalTrials.gov's own coarse categories; OTHER is a large, heterogeneous bucket dominated by academic medical centers and hospitals, not a single sponsor archetype.

Every query behind this report

4 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT overall_status, COUNT(*) c FROM health.clinical_trials GROUP BY overall_status ORDER BY c DESC

SELECT COUNT(*) total,
  SUM(CASE WHEN overall_status='COMPLETED' THEN 1 ELSE 0 END) completed,
  SUM(CASE WHEN has_results THEN 1 ELSE 0 END) has_results,
  SUM(CASE WHEN overall_status='COMPLETED' AND has_results THEN 1 ELSE 0 END) completed_with_results
FROM health.clinical_trials

SELECT funder_type, COUNT(*) total,
  SUM(CASE WHEN overall_status='COMPLETED' THEN 1 ELSE 0 END) completed,
  SUM(CASE WHEN has_results THEN 1 ELSE 0 END) has_results,
  SUM(CASE WHEN overall_status='COMPLETED' AND has_results THEN 1 ELSE 0 END) completed_with_results
FROM health.clinical_trials
GROUP BY funder_type ORDER BY total DESC

SELECT lead_sponsor, COUNT(*) total,
  SUM(CASE WHEN overall_status='COMPLETED' THEN 1 ELSE 0 END) completed,
  SUM(CASE WHEN has_results THEN 1 ELSE 0 END) has_results,
  ROUND(100.0*SUM(CASE WHEN has_results THEN 1 ELSE 0 END)/COUNT(*),1) pct_results,
  ROUND(100.0*SUM(CASE WHEN overall_status='COMPLETED' THEN 1 ELSE 0 END)/COUNT(*),1) pct_completed
FROM health.clinical_trials
GROUP BY lead_sponsor
HAVING COUNT(*) >= 300
ORDER BY total DESC
LIMIT 25