Datadory notebook
Clinical Trial Protocol Results Dataset Data: Dataset Structure and Field Coverage
Datadory delivers clinical trial protocol results dataset data covering comprehensive field definitions, entity mappings, and historical time series — structured for direct analytics and delivered on demand.
1,744 datasets. Pick your catch.
What is the clinical trial protocol results dataset?
The dataset behind the query is ClinicalTrials.gov, the U.S. National Library of Medicine's registry and results database, which reported 599,549 total studies through API v2 in August 2026. Datadory catalogs it as the anchor record of its biotechnology slice - quality score 10 out of 10, refreshed daily, free, and commercial delivery terms as a U.S. government work. Registration has run since 2000 and now spans studies from more than 200 countries.
One record covers one registered study end to end. The protocol section carries identification (nctId, titles, organization), status, sponsors, conditions, design, arms and interventions, outcomes, eligibility, locations, oversight and references; a derived section adds MeSH condition browsing terms; and a hasResults flag marks studies that have posted results in the registry. That structure is what separates a protocol-and-results dataset from a news feed of trial announcements: every field is submitted, versioned and queryable.
Which fields carry the protocol and results detail?
API v2 organizes each study into named modules you can request individually with the fields parameter, so a pipeline pulls only what its model needs:
| Module | What it answers |
|---|---|
| identificationModule | NCT ID, brief and official title, sponsoring organization |
| sponsorCollaboratorsModule | Lead sponsor and collaborators by name |
| conditionsModule | Conditions or diseases under study |
| designModule | Study type (INTERVENTIONAL, OBSERVATIONAL, EXPANDED_ACCESS), phases, enrollment count with ACTUAL or ANTICIPATED type |
| armsInterventionsModule | Arms plus interventions with name, type (Drug, Device, Behavioral) and description |
| outcomesModule | Primary and secondary outcome measures with time frames |
| eligibilityModule | Sex, age bounds, healthy-volunteers flag, full criteria text |
| contactsLocationsModule | Sites with facility, city, state, country and geo-point coordinates where provided |
| derivedSection.conditionBrowseModule | MeSH terms with id, term and ancestor tree |
Sample records show how far the granularity goes. NCT00787735 is a completed Johns Hopkins University behavioral study enrolling 380 participants that started in January 2004; NCT00429442 is a withdrawn Phase 3 simvastatin-plus-Copaxone trial in relapsing multiple sclerosis. Both keep their full protocol sections even though neither produced results - which is exactly why competitive-intelligence teams track withdrawals rather than discarding them.
How do you pull protocol and results records into a pipeline?
Five steps take you from zero to a working extract:
Who works with this dataset day to day?
Data scientists treat the registry as labeled training corpus material alongside NCBI GEO's roughly 250,000 Series of functional genomics data, and developers wrap API v2 directly because the JSON schema is stable and documented per module. Whichever seat you sit in, the entry points stay the same: the biotechnology data hub indexes all 24 primary records, the dataset profile for ClinicalTrials.gov documents every module field, and the head-to-head comparison with ChEMBL shows where registry data ends and bioactivity data begins.
| Field path | What it tells you |
|---|---|
| protocolSection.identificationModule.nctId | Unique NCT registry identifier for the study |
| identificationModule.organization | Sponsoring organization responsible for the study record |
| statusModule.overallStatus | Current recruitment status: RECRUITING, COMPLETED, WITHDRAWN, SUSPENDED |
| sponsorCollaboratorsModule.leadSponsor.name | Name of the lead sponsor entity |
| conditionsModule.conditions | Array of conditions or diseases being studied |
| designModule.phases | Phase codes such as PHASE1, PHASE2, PHASE3, PHASE4, or NA for observational studies |
| designModule.enrollmentInfo.count | Enrollment count with type ACTUAL or ANTICIPATED |
| designModule.studyType | INTERVENTIONAL, OBSERVATIONAL, or EXPANDED_ACCESS |
| armsInterventionsModule.interventions | Arms plus interventions with name, type and description |
| outcomesModule | Primary and secondary outcome measures with time frames |
| contactsLocationsModule.locations | Sites with facility, city, state, country and geo-point coordinates |
| derivedSection.conditionBrowseModule.meshes | MeSH terms with id, term and ancestor tree |
| hasResults | True when the study has posted results data in the registry |
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
ClinicalTrials.gov
openFDA — FDA Regulatory Data on Drugs, Devices and Food
openfda
ChEMBL
GWAS Catalog
Google Patents
NCBI GEO
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sampleQuestions worth asking
Does ClinicalTrials.gov require results posting for every trial?
No. Each record carries a hasResults flag precisely because posting is not universal, and withdrawn or suspended trials retain their protocol sections without results. Filter on filter.advanced or read the overallStatus enum (RECRUITING, COMPLETED, WITHDRAWN, SUSPENDED) to separate completed-with-results studies from the rest of the 599,549 records.