Clinical Evidence Digest – Published Trial Results

Which completed clinical trials have actually published results? Pulls recently completed Phase 2–4 trials from ClinicalTrials.gov, links each to its PubMed papers, extracts primary endpoints and posted results, classifies evidence status and writes a Markdown digest per therapeutic area.

Why use it

• Monthly evidence monitoring for medical affairs, journal clubs and teaching
• Find completed trials whose results are still unpublished (publication bias, follow-up)
• Background sections of grant proposals and reviews with verifiable NCT/PMID links
• Read the Markdown digest from the key-value store (key DIGEST) or feed the rows to an AI agent

No API keys, no scraping of HTML – the actor talks to the official public API, handles paging, rate limits and retries, and returns flat JSON you can export as CSV/Excel or call from code, Make, Zapier, n8n or AI agents (Apify MCP server).

Input

Field	Title	Description
therapeuticAreas	Therapeutic areas	Any of: obesity, addiction, cancer, hiv, infectious_diseases. Empty = all five.
customCondition	Custom condition	Optional extra area, e.g. 'MC4R deficiency' or 'Clostridioides difficile'.
intervention	Intervention filter	Optional, e.g. 'semaglutide', 'tirzepatide', 'lenacapavir'.
phases	Phases	Default: PHASE2, PHASE3, PHASE4.
completedWithinYears	Completed within (years)	Only trials completed in this many past years.
maxTrialsPerArea	Max trials per area	Sample size per therapeutic area.
publishedOnly	Only trials with published results	Save only trials linked to a PubMed publication.
email	Contact e-mail (recommended by NCBI)	Sent to NCBI as the 'email' parameter.
ncbiApiKey	NCBI API key	Optional, speeds up PubMed linking.

Example input:

{
  "therapeuticAreas": [
    "obesity",
    "hiv"
  ],
  "completedWithinYears": 2,
  "maxTrialsPerArea": 50,
  "publishedOnly": false
}

Output (one dataset item per result)

{
  "area": "obesity",
  "nctId": "NCT00000000",
  "title": "...",
  "phases": [
    "PHASE3"
  ],
  "enrollment": 1200,
  "interventions": [
    "Drug: ..."
  ],
  "completionDate": "2025-06-30",
  "primaryOutcomes": [
    "Percent change in body weight"
  ],
  "resultsPostedOnRegistry": true,
  "primaryOutcomeResults": [
    {
      "measure": "Percent change in body weight",
      "values": [
        {
          "group": "Drug",
          "value": "-14.9"
        }
      ],
      "analyses": [
        {
          "pValue": "<0.001"
        }
      ]
    }
  ],
  "publications": [
    {
      "pmid": "12345678",
      "journal": "N Engl J Med",
      "year": "2025",
      "doi": "10.1056/..."
    }
  ],
  "evidenceStatus": "results_published",
  "url": "https://clinicaltrials.gov/study/NCT00000000"
}

Status values

results_published (a results paper linked via trial references or NCT number; protocol/design papers, reviews and editorials do not count) · results_posted_registry_only · protocol_or_design_paper_only · completed_no_results_found · pubmed_lookup_failed (NCBI was temporarily unreachable for that trial; status unknown, rerun to retry). Every publication carries kind = results / protocol / other. The Markdown digest is stored in the run's key-value store under DIGEST.

Use with AI agents (Apify MCP server)

Add the Apify MCP server (https://mcp.apify.com) to Claude, Cursor or any MCP client and enable this actor (researchdata/clinical-evidence-digest). The agent can then call it directly: describe what you need, the agent fills the input fields above, runs the actor and reads the dataset. Because results are flat JSON with explicit nulls, an agent can use them without extra parsing. Set a maximum cost per run to keep agent calls capped.

Pricing

Pay per event: $0.01 per trial-summarized (trial summarized). You only pay for results you get; set a max cost per run to cap spending.

Cost example

At $0.01 per result: 50 results ≈ $0.50, 1,000 results ≈ $10.00. Set a maximum cost per run in the run options to cap spending.

Limits

• Data comes from the provider's public API; fields the provider does not publish are returned empty (null), never guessed.
• Results reflect the provider's database at run time.
• Not affiliated with the data provider.

Data source & terms

ClinicalTrials.gov API v2 (NLM) and NCBI E-utilities (PubMed). Registry/bibliographic summary only – not medical advice and no efficacy interpretation. Please respect the data provider's terms and cite the original resource in publications. This actor is an independent tool and is not affiliated with or endorsed by the data provider.