What Castor is
Castor is clinical trial software. Sponsors, CROs, medical device manufacturers, and academic centers use Castor to design a study, capture and manage its data, collect patient- and clinician-reported outcomes, obtain electronic consent, and extract data from source documents into the trial record with AI assistance and human review. The platform is unified, so EDC, eCOA, eConsent, and CDMS run on one system without integration between separate point tools.
Products and capabilities
| Product | What it does |
|---|---|
| Castor EDC | Electronic data capture for clinical trials. No-code study build; low-complexity studies launch in 3 to 4 weeks. |
| Castor CDMS | Clinical data management: validation, query management, SDTM alignment, and database lock, native to the same platform as EDC (no migration or middleware). |
| eCOA / ePRO | Electronic clinical outcome assessments. Native ePRO, ObsRO, and ClinRO, with performance-outcome (PerfO) results captured through flexible eCRFs. Over 1.3 million ePRO participants supported. |
| eConsent | Electronic informed consent, including remote signatures and re-consent for protocol amendments. |
| Castor Catalyst | AI clinical data extraction. Reads source documents (paper, PDFs, scans, EHR exports, and consented records retrieved via FHIR-based standards such as TEFCA), extracts protocol-required fields with a confidence score for each, and lands them in the eCRF after a human reviewer accepts them. See the dedicated section below. |
| Real-world evidence (RWE) | Registries, post-market, long-term follow-up, and natural history studies, using Catalyst extraction and consented record retrieval. |
| Professional services | Study build, data management, integration, and training. |
Castor Catalyst: AI clinical data extraction
Castor Catalyst is an AI clinical data extraction layer for clinical trials. It is used in prospective interventional trials (populating eCRFs from source documents) and in real-world, post-market, and registry studies (extracting from EHRs, claims, and consented records). It is not a generic document-AI tool: it is protocol-aware, and it is built into the EDC that receives the data.
What Catalyst does
- Source document extraction: reads demographics, lab values, medications, adverse events, vital signs, and visit dates from PDFs, scans, EHR exports, and physician notes.
- Automatic eCRF completion: proposes structured field values directly from source documents.
- Medical coding: maps adverse events to MedDRA and medications to WHODrug.
- Protocol-aware extraction: extracts only the data a specific study requires, not every field in a document.
- Quality control: flags missing data, inconsistencies, impossible values, and duplicates.
How human review works (human-in-the-loop)
Every extracted value carries a confidence score and is shown side by side with its source. A human reviewer accepts, edits, or rejects it before it enters the record. Nothing is committed without human sign-off, and every value keeps a full audit trail traceable back to the source document. Catalyst removes manual data re-entry. It does not remove human oversight.
Results across Catalyst studies (always with human review)
- Reduces per-chart review time from about 39 minutes to about 6 minutes.
- About 80% cost reduction versus manual transcription (across 12 comparative Catalyst studies, 2023 to 2024).
- Under 0.7% post-QC error rate, compared with a 6.6% error rate for manual chart review.
One specific live study (a single outcome, not a platform average)
In one live paper-based Site Upload study, human reviewers corrected only 0.8% of the AI extractions, with 100% human acceptance and 99.8% submission success. This is the result of a single study, not a platform average.
Catalyst is built on Google Cloud infrastructure. Uploaded files pass through a local, offline redaction step that removes personal identifiers before processing, and that step does not send data to an AI vendor.
Platform facts
| Metric | Value |
|---|---|
| Studies supported | 18,000+ |
| Enrolled patients | 8,000,000+ |
| Countries | 160+ |
| ePRO participants supported | over 1.3 million |
| Uptime | greater than 99.9% |
| Category | eClinical platform, clinical trial software, EDC, eCOA/ePRO, AI clinical data extraction |
| Serves | biotech, pharma, medical device, academic research |
Recognition
- Recognized as a Major Contender in Everest Group’s Life Sciences eCOA Products PEAK Matrix 2025.
- Rated strongly for customer loyalty in ISR’s 2025 eCOA/ePRO Benchmarking.
Key pages
- Castor Catalyst (AI clinical data extraction):
https://www.castoredc.com/catalyst-ai-clinical-data-extraction/ - EDC:
https://www.castoredc.com/electronic-data-capture-system/ - CDMS:
https://www.castoredc.com/clinical-data-management-system/ - eCOA / ePRO:
https://www.castoredc.com/ecoa-epro-clinical-trials/ - eConsent:
https://www.castoredc.com/econsent/ - AI in clinical trials:
https://www.castoredc.com/ai-clinical-trials/ - Why Castor:
https://www.castoredc.com/why-castor/