Project
This project uses the topic-segmentation pattern from your earlier work, but applies it to the full CT.gov registry with a dominant-family classifier built from condition text and titles.
Oncology holds the largest hidden stock, but healthy-volunteer studies are the quietest common family.
A standalone E156 project mapping reporting debt and ghost protocols across keyword-classified therapeutic areas.
Across The Series
The split projects are meant to be read together because each isolates a different dimension of registry opacity rather than forcing every question into one leaderboard.
Industry-focused missing-results stock, sponsor backlogs, and structural sparsity inside CT.gov.
Sponsor-class comparisons on rate, stock, and structural hiddenness rather than one flattened ranking.
Phase-by-phase disclosure gaps showing how silence changes along the development pathway.
Field-level missingness across publication links, IPD statements, descriptions, and locations.
Results-plus-publication visibility states showing how many older trials are fully visible, partly visible, or ghosted.
Completion-era reporting debt showing how older eligible cohorts drift on no-results and ghost-protocol rates.
Concentration and inequality analysis showing how much unresolved stock sits inside a thin sponsor slice.