CT.gov Condition Hiddenness Map

Abstract

When older closed interventional trials on ClinicalTrials.gov are grouped by condition, do some fields show up as quieter than others? We carried out a full-registry search of 249,507 older completed studies from a March 29, 2026 snapshot and organised them into dominant keyword-based families using registry condition strings. Our outcomes were two-year no-results rates, ghost-protocol rates, and results-and-publication visibility across common families. The biggest family was oncology, with 42,344 eligible older studies, making it the largest area of hidden evidence. Healthy-volunteer studies had the highest ghost-protocol rate at 63.5 percent, but gastrointestinal and metabolic areas were also obscured. Infectious-disease studies were more visible, with a 20.6 percent fully visible rate despite substantial non-reporting. Our classification is, however, keyword-based and single-label, so multi-topic trials can be compressed into one family and some records may remain in an ‘other’ bucket.