Paper
Trial scale matters, but size alone does not solve the registry's hiddenness problem.
How much of ClinicalTrials.gov hiddenness tracks trial enrollment size once older closed interventional studies are grouped into comparable size buckets? We analysed 249,507 eligible older closed interventional studies from the March 29, 2026 full-registry snapshot and binned them by recorded enrollment. The project compares two-year no-results rates, ghost-protocol rates, full visibility, and sponsor-class contrasts across enrollment buckets from 1-50 through 5,001+ participants. Studies enrolling 1 to 50 participants showed a 73.2 percent no-results rate and a 47.6 percent ghost-protocol rate. Studies enrolling 1,001 to 5,000 participants fell to 62.4 percent no results and 18.7 percent ghost protocols, while large OTHER-sponsored studies still remained highly obscured. Trial scale therefore matters, but size alone does not erase sponsor-driven reporting debt within the public registry surface. That pattern persists across tiny studies and surprisingly large nonindustry backlogs alike. Enrollment is registry-recorded and can be missing, estimated, or misclassified, so these buckets describe visible scale rather than adjudicated participant counts.