ClinicalTrials.gov Study Monitor (new & status changes)

Search ClinicalTrials.gov studies and monitor new studies and changes to status, phases, enrollment, completion dates, results availability, and last update dates. The Actor reads the public documented API v2. It fetches current study data on every run and adds the ClinicalTrials.gov data processing timestamp to every study record.

Input examples

Fetch up to five recruiting studies, including studies returned by earlier runs.

{
  "condition": "type 2 diabetes",
  "statuses": ["RECRUITING"],
  "maxItems": 5,
  "onlyNew": false
}

Monitor new and changed phase 2 or phase 3 studies updated within the last 19 days.

{
  "term": "immunotherapy",
  "sponsor": "National Cancer Institute",
  "location": "United States",
  "phases": ["PHASE2", "PHASE3"],
  "updatedWithinDays": 19,
  "maxItems": 100,
  "onlyNew": true,
  "stateStoreName": "clinical-trials-monitor-state"
}

Input fields

Field	Default	Meaning
condition	Empty	API query.cond.
term	Empty	API query.term.
sponsor	Empty	API query.spons.
location	Empty	API query.locn.
statuses	[]	API filter.overallStatus, as a comma list.
phases	[]	Local phase filter, matched against any listed phase. NA also matches absent phases.
updatedWithinDays	Unset	Positive integer, up to 36500. Uses AREA[LastUpdatePostDate]RANGE[YYYY-MM-DD,MAX], starting that many UTC calendar days ago.
maxItems	100	Maximum study pushes, from 1 to 10000.
onlyNew	true	Emit only new or changed studies. Set false to include unchanged studies.
stateStoreName	clinical-trials-monitor-state	Named KV store containing the STUDIES history.

At least one nonempty condition, term, or sponsor is required. location alone is insufficient. Search strings use the source API's query syntax.

Supported statuses are RECRUITING, NOT_YET_RECRUITING, ACTIVE_NOT_RECRUITING, COMPLETED, TERMINATED, WITHDRAWN, SUSPENDED, ENROLLING_BY_INVITATION, and UNKNOWN.

Supported phases are EARLY_PHASE1, PHASE1, PHASE2, PHASE3, PHASE4, and NA.

Output sample

This illustrative record uses fictional study details.

{
  "nctId": "NCT01234567",
  "briefTitle": "Study of Drug A in Type 2 Diabetes",
  "officialTitle": "A Randomized Study of Drug A in Adults With Type 2 Diabetes",
  "overallStatus": "RECRUITING",
  "phases": [
    "PHASE2",
    "PHASE3"
  ],
  "studyType": "INTERVENTIONAL",
  "conditions": [
    "Type 2 Diabetes"
  ],
  "interventions": [
    {
      "name": "Drug A",
      "type": "DRUG"
    }
  ],
  "leadSponsor": "Example University",
  "collaborators": [
    "Example Research Institute"
  ],
  "enrollment": 120,
  "startDate": "2026-01-01",
  "primaryCompletionDate": "2027-01",
  "completionDate": "2027-06",
  "firstPostDate": "2025-12-01",
  "lastUpdatePostDate": "2026-10-19",
  "hasResults": false,
  "locationCountries": [
    "United States",
    "Canada"
  ],
  "locationsCount": 3,
  "studyUrl": "https://clinicaltrials.gov/study/NCT01234567",
  "source": "ClinicalTrials.gov",
  "dataProcessedAt": "2026-10-20T06:00:00",
  "checkedAt": "2026-10-20T12:30:00.000Z",
  "changeType": "new",
  "changedFields": []
}

Missing scalar fields are null. Missing lists are empty arrays, and missing hasResults is false. Dates retain the source precision, such as 2027-01.

The default dataset holds studies and uncharged error records. The overview view exposes study fields. The errors view projects recordType, error, source, and checkedAt for diagnostics. The default KV store holds RUN_SUMMARY, including start and finish times, counts of fetched, eligible, and pushed studies, charge-limit status, and any error. Fetched counts cover studies yielded after the local phase filter and before stopping.

Change detection

The named state store maps NCT IDs to SHA-256 fingerprints and the corresponding tracked field values. Fingerprints use stable JSON of overallStatus, phases, enrollment, primaryCompletionDate, completionDate, hasResults, and lastUpdatePostDate. Phase order does not affect the fingerprint.

A study absent from history has changeType: "new" and changedFields: []. A different fingerprint produces changeType: "changed" and the names of changed tracked fields. An equal fingerprint produces changeType: "unchanged" and an empty list. Changes to other fields do not trigger an output with onlyNew: true.

The Actor saves history only after a successful study push. Duplicate IDs within one run emit once. Charge limits and failed pushes leave unpushed studies eligible for a later run. With onlyNew: false, successful pushes also update history.

Dataset pushes and state writes are separate operations. A crash between them can produce duplicates on retry. Avoid overlapping runs sharing a state store. Use separate store names for independent histories. Reset history by removing the STUDIES record. History grows with the number of emitted NCT IDs.

Pricing

The PPE event is study, priced at $0.003 per emitted study, or $3 per 1,000 studies. There is no run fee. Unchanged studies emitted with onlyNew: false incur the same event charge. Skipped studies and error records are not charged. Error records use Actor.pushData without a charge event. Local runs do not bill PPE. Store pricing requires manual configuration as described in PUBLISHING.md.

Limits

• Requests start at most once per second per host. HTTP 429 and 5xx responses retry up to three times with exponential backoff and Retry-After support. Requests time out after 30 seconds.
• API requests use pageSize=100, JSON format, and nextPageToken pagination. The API supports larger pages, but this Actor uses 100.
• maxItems limits emitted studies. Phase filters and unchanged records may require reading more studies or all matching pages. A run stops at its output or charge limit. Each run starts a fresh search, so repeated capped runs can revisit earlier pages.
• Results depend on source availability and the source's latest processed dataset. They are not a complete record of every intermediate change. Removed studies and studies that no longer match the query do not emit deletion events.
• This tool provides study data, not medical advice, enrollment eligibility, or a guarantee that a study is still recruiting. Check the linked study page for current details.

Attribution and terms

Data source: ClinicalTrials.gov, a United States Government database. ClinicalTrials.gov provides the source data without charge. This Actor's price covers its collection and monitoring service. No proprietary rights are asserted over the database.

dataProcessedAt clearly displays the date the data were processed by ClinicalTrials.gov. It comes from the dataTimestamp in /api/v2/version, fetched once per run. checkedAt is the Actor's record processing time. Keep the ClinicalTrials.gov attribution and data processing date visible when distributing these records. Runs retrieve current data each time; refresh exported data to keep it current.

ClinicalTrials.gov data carry an international copyright outside the United States and its Territories or Possessions. Some data may be subject to third-party copyright. Use and redistribution remain subject to the ClinicalTrials.gov terms and conditions. The local TERMS.md records the terms check used for this implementation.

Modifications

The Actor selects the fields shown in the output sample and flattens nested API modules. It extracts sponsor and collaborator names, intervention names and types, enrollment counts, and date strings without date-type metadata. It reduces locations to a location count and unique country names. It omits contact objects entirely, including all contact emails and phone numbers. Other source fields are omitted.

It supplies null or empty defaults for missing fields, adds the study URL, attribution, source processing timestamp, check time, and change metadata, and applies optional local phase filters. It sorts and deduplicates phases only for fingerprint comparison. Study content is otherwise retained in the source's form. It does not output the full source database.

Disclaimer

This Actor is unofficial and is not affiliated with or endorsed by ClinicalTrials.gov, the U.S. National Library of Medicine, or the National Institutes of Health. It uses public documented APIs only, no logos, and no login.

Local development

Node 20 or newer is required. Tests use node:test, stubbed fetch, and local temporary storage. They do not need network access.

npm test

A normal npm start reads the sample input and calls the live API. For an offline local run, preload scripts/offline-fetch.mjs to replace fetch with small realistic API fixtures. Use an isolated storage directory with an INPUT.json under key_value_stores/default/.

mkdir -p storage/offline-verification/key_value_stores/default
cp storage/key_value_stores/default/INPUT.json storage/offline-verification/key_value_stores/default/INPUT.json
APIFY_TOKEN='' APIFY_IS_AT_HOME='' APIFY_PURGE_ON_START=false \
CRAWLEE_STORAGE_DIR=storage/offline-verification \
APIFY_LOCAL_STORAGE_DIR=storage/offline-verification \
NODE_OPTIONS='--import ./scripts/offline-fetch.mjs' npm start

Read datasets/default/ and key_value_stores/default/RUN_SUMMARY.json beneath that directory after the run. The preload is a verification tool and is never imported by production code.