Catalyst AI

Castor Catalyst AI is the only unified Real World Evidence (RWE) data platform designed to automate the generation of regulatory-grade RWE in clinical research studies. Built on Google Cloud's infrastructure and AI technologies, it combines AI-powered automation of EMR and Claims data with a complete, 21 CFR Part 11 compliant eClinical suite — including EDC/CDMS, ePRO/eCOA, and eConsent — in a single platform. Catalyst is purpose-built for life sciences organizations seeking to dramatically reduce the cost and time associated with observational research, post-marketing surveillance, and submission-ready RWE generation.

Traditional approaches to RWE generation force unacceptable compromises: site-based models consume 60–70% of study budgets on data management with average error rates of 6.57%, data aggregators lack patient-level traceability required by the FDA, and integrating point solutions creates compliance risks and data reconciliation challenges. Castor Catalyst addresses all of these gaps through a self-driving study architecture that delivers up to 80% cost reduction, 70% faster study completion, and 100% consented patient compliance.

Platform Architecture: Four Layers of Capability

• Compliant Infrastructure (Layer 1): A GxP and 21 CFR Part 11 compliant eClinical suite encompassing EDC/CDMS, ePRO/eCOA, eConsent, and APIs for data integration.
• Data Connectivity (Layer 2): Unified, patient-consented data acquisition with direct access to Direct-to-Patient (DTP) EMR via FHIR and HIPAA, complete medical claims, full pharmacy claims, and lab data — with zero tokenization costs.
• AI Orchestration (Layer 3): Castor Catalyst AI reads input data, maps it to the target study structure, transforms and processes data with confidence intervals, performs data review with mapping details, and automates workflow orchestration. Capabilities include data extraction, CDISC mapping, ePRO/eCOA triggering, and automated inclusion/exclusion screening (coming soon).
• Expert Services and Human Oversight (Layer 4): Full-service support by medically trained clinical staff covering recruitment, IRB submission, data management, and clinical data review.

End-to-End Automated Workflow

1. Multi-Channel Recruitment and eConsent: Patients are recruited through multiple channels and provide explicit digital consent with full audit trails via a HIPAA-compliant, 21 CFR Part 11 compliant eConsent workflow.
2. EMR and Claims Data Retrieval: Automated, direct-to-patient retrieval of structured EHR data and complete claims records using FHIR-based API integration.
3. AI Processing and Screening: Castor's CoPilot AI performs initial extraction, structuring, and screening against protocol criteria, with mandatory human oversight and validation by medically trained clinical staff.
4. Clinical Data Review and Push to EDC/CDMS: All AI-processed data undergoes clinical review before being committed to the EDC/CDMS, ensuring a 0% error rate through this hybrid approach.
5. Data Delivery via EDC/CDMS and API: Submission-ready datasets are delivered through the integrated EDC/CDMS and via API, supporting CDISC standards including SDTM, ADaM, and ODM.

Key Platform Capabilities and Differentiators

• Unified single-login platform for EDC/CDMS, eCOA, eConsent, and RWE generation — eliminating the need to integrate disparate vendor systems.
• AI-powered automation delivering an 85% reduction in manual chart review effort.
• Direct-to-patient EMR access and consented claims data acquisition without expensive tokenization, unlike data aggregators such as Flatiron or Komodo.
• Full patient-level audit trail and complete traceability from source documents to final datasets, aligned with FDA July 2024 RWD guidance.
• Rapid deployment in 4–8 weeks using pre-validated workflows and automated setup, compared to 12–16 weeks for traditional approaches.
• Global scale supporting 40+ languages and 190+ countries.
• Automated inclusion/exclusion screening capabilities available as an early access program.

Supported Study Types and Therapeutic Areas

• Prospective Registries and Natural History Studies: Long-term disease tracking with automated EMR refreshes and ePRO/eCOA collection.
• External Control Arms (ECAs): Precisely matched historical controls for single-arm trials.
• Post-Marketing Safety (PASS/PMR): FDA-compliant safety surveillance and reporting.
• Screen-to-Enroll Workflows: Building pre-screened patient cohorts by pulling medical records prior to clinical trial enrollment.
• Oncology: Cancer registries, real-world treatment patterns, and automated chart abstraction for progression, response rates, and survival outcomes.
• Central Nervous System: Natural history studies for neurodegenerative diseases with longitudinal cognitive assessments and disease progression tracking.
• Rare Disease: Patient registries for ultra-rare conditions with direct-to-patient enrollment and global natural history data collection.
• Respiratory and Allergy: Real-world effectiveness studies for asthma and COPD with automated extraction of exacerbations, hospitalizations, and treatment patterns.
• Infectious Disease: Vaccine effectiveness and outbreak surveillance registries with rapid deployment for real-world safety monitoring.
• Cardiovascular: GLP-1 registries and metabolic endpoint collection with automated data retrieval for cardiovascular and diabetes outcomes assessment.

Performance Claims and Methodology

• 80% cost reduction calculated by comparing traditional site-based chart review costs versus the direct-to-patient automated extraction model, based on analysis of 12 comparative studies. Savings are driven by elimination of site startup fees ($15K–$25K per site), 90% reduction in manual transcription time, and use of a single IRB versus multiple site IRBs.
• 70% faster study completion measured from protocol finalization to database lock. Traditional chart review averages 32–40 weeks; the Catalyst automated approach averages 12–18 weeks, based on retrospective analysis of 25+ studies.
• 100% consented patients achieved through HIPAA-compliant digital consent with patient-mediated data access permissions, full audit trails, and 21 CFR Part 11 compliant consent workflows.

Castor Catalyst is fully compliant with 21 CFR Part 11, HIPAA, GxP, and ALCOA+ data integrity principles. The platform supports CDISC data standards (SDTM, ADaM, ODM) for FDA submissions, aligns with the FDA's Real-World Evidence Program framework, and adheres to FDA guidance on the use of electronic health records and electronic source data in clinical investigations. It is deployed on Google Cloud infrastructure and designed to meet ICH E6(R3) requirements through functional oversight capabilities.