ClinicalTrials.gov study records: AllPublicXML snapshots ======================================================== This archive holds a series of AllPublicXML snapshots downloaded from ClinicalTrials.gov, the clinical trials registry run by the US National Library of Medicine (NLM) at the NIH. Each snapshot is a separate point-in-time capture of every publicly available study record on the site at the date it was taken. Directory layout ---------------- The parent directory contains one folder per snapshot, named by the date the snapshot was published, for example: 26th July 2017/ 8th March 2018/ 30th June 2019/ 6th July 2020/ !README.txt (this file) Each dated folder contains a single .zip file, the original AllPublicXML download for that date. Snapshots in this archive ------------------------- (The folder name gives the published date; the study count is in each zip's Contents.txt) Folder Studies ---------------- -------- 26th July 2017 250,607 8th March 2018 268,157 30th June 2019 309,645 6th July 2020 344,473 What each zip contains ---------------------- The "AllPublicXML" bulk download was a single zip file that ClinicalTrials.gov offered containing all public records in its legacy XML format. Inside each zip you will find: - Contents.txt, a short manifest (see below). - One XML file per study, named by its ClinicalTrials.gov identifier (the NCT number), e.g. NCT00000102.xml, grouped into subdirectories by the leading part of the NCT number so that no single folder holds the full set. This grouping is purely for manageability and carries no other meaning. Each XML record holds the protocol registration information for one study and, where the sponsor had submitted it, the summary results information. Files validate against the ClinicalTrials.gov public XML schema (public.xsd), which described the structure of records on the classic site. Fields cover the title, sponsor, brief and detailed descriptions, status, phase, study design, conditions, interventions, eligibility criteria, locations, and key dates. MeSH terms, where present, are assigned automatically by the registry and the records themselves flag them as imperfect. The Contents.txt manifest ------------------------- Each zip includes a Contents.txt with three lines giving the study count, the grouping rule, and the published date, for example: All 250,607 studies in XML format Studies grouped by leading part of NCT Id Studies published on July 26, 2017 Dating a snapshot ----------------- Use the "published" date from the zip's Contents.txt as the canonical date of the snapshot; this is also what the folder names here use. Note that the individual records carry their own element recording when ClinicalTrials.gov processed the data, and this can differ from the published date by a day. For the July 2017 snapshot, for instance, Contents.txt gives 26 July 2017 while the records report processing on 27 July 2017. All files within a single zip share the same processing date. Status of this data format -------------------------- This is a legacy format that is no longer produced. NLM launched a modernised ClinicalTrials.gov site in June 2023 and retired the classic site, its API, and the associated bulk downloads on 25 June 2024. The current registry is served through a REST API returning JSON (API v2), with a revised data model that does not map one-to-one onto the XML here. These snapshots therefore preserve both the historical content of the registry and the discontinued XML representation of it. Caveats ------- - Each snapshot is a point in time, not a live feed. Records may since have been updated, completed, corrected, or withdrawn on the registry. Always cite the snapshot date. - Registry data reflects what sponsors submitted. Completeness and accuracy vary between records and over time. - Public records only. Anything not public on ClinicalTrials.gov at the snapshot date is absent. Provenance and licence ---------------------- ClinicalTrials.gov data are made available by the US National Library of Medicine. For terms of use and citation guidance, see NLM's current policies at https://clinicaltrials.gov. This archive redistributes the data unmodified for preservation and research reuse. Archived by: Ben Meghreblian Contact: benmeg@benmeg.com / https://benmeg.com