Patient Journey Intelligence

Patient Journey Intelligence is a healthcare analytics platform that transforms raw clinical data into standardized, longitudinal patient pathways. It is designed for healthcare organizations seeking to support research, population health management, AI development, quality improvement, and regulatory reporting from a single, reusable data foundation.

The platform addresses a core challenge in secondary use of clinical data: information collected during routine patient care is typically captured for billing and documentation, leaving it fragmented across EHR systems, clinical notes, scanned PDFs, imaging reports, and lab systems in inconsistent formats. Patient Journey Intelligence continuously ingests, extracts, standardizes, and maintains this data so that downstream use cases — registries, cohort studies, predictive models, quality measures — can operate on the same curated dataset rather than rebuilding pipelines independently.

Core Platform Capabilities

• Data Integration: Connects to external data sources, ingests clinical documents in multiple formats, and transforms raw data into structured, standardized OMOP Common Data Model (CDM) format.
• Data Curation: Supports building and managing clinical ontologies, automating case finding, and maintaining expert-reviewed registries with full audit trails.
• De-identified OMOP: Automatically maintains synchronized identified and de-identified OMOP datasets in parallel, keeping research and operational data current together.
• AI Agents: Provides pre-built AI assistants for patient co-pilot functions, journey timelines, and cohort building through conversational interfaces.
• Patient Registries: Automates cancer, research, and custom registries with AI-powered extraction, NAACCR compliance, and regulatory reporting support.
• Governance: Includes AI governance controls, security best practices, de-identification standards, and comprehensive audit logging.

The Three Gaps the Platform Addresses

• Data Accuracy Gap: Up to 40% of critical diagnoses exist only in unstructured clinical notes and are never coded into structured fields. Treatment rationale, disease progression, and clinical reasoning documented as free text are invisible to traditional analytics without extraction.
• Data Engineering Gap: Each new use case — cancer registry, clinical trial cohort, quality measure, AI model — typically requires rebuilding similar data pipelines from scratch, consuming 10 or more FTE-years annually on redundant engineering across an organization.
• AI Governance Gap: Models trained on research datasets often fail in production because operational data uses inconsistent terminologies and different preprocessing. Maintaining separate identified and de-identified datasets doubles infrastructure costs, and audit trails, data lineage, provenance tracking, and PHI management require custom tooling that most teams lack.

How the Unified Data Pipeline Works

1. Ingest Any Data Source: Connects EHR systems, clinical notes, imaging (DICOM), PDFs, lab feeds, and external registries without extensive preprocessing, handling data as-is regardless of format or structure.
2. AI-Powered Extraction: John Snow Labs Medical Language Models extract structured facts from unstructured content, detect negations and temporal relationships, and resolve entities across documents, capturing clinical context trapped in free text.
3. Standardize to OMOP CDM: All data is mapped to OMOP Common Data Model v5.4 with standard terminologies including SNOMED CT, RxNorm, LOINC, and ICD-10, creating a unified schema compatible with OHDSI analytics tools, BI tools, and AI frameworks.
4. Continuously Update Patient Journeys: Longitudinal timelines combine all data about each patient — visits, conditions, medications, procedures, labs — and are automatically updated as new clinical documents arrive.

Living, Governed Data Assets

• Provenance: Every extracted fact includes source documents, extraction timestamps, and transformation lineage.
• Confidence Scores: ML model certainty scores support quality control and expert review workflows.
• Versioning: Time-stamped updates preserve historical states for reproducibility.
• Clinical Context: Complete temporal sequences show disease progression, treatment response, and care patterns.

Parallel Dataset Types Maintained Simultaneously

1. Identified OMOP (Operational Dataset): Full PHI for clinical operations, care coordination, point-of-care AI, and internal quality improvement.
2. De-Identified OMOP (Research Dataset): HIPAA Safe Harbor compliant with consistent pseudonyms, date-shifting, and PHI removal for research, external collaborations, and AI model training.
3. Patient Registry: NAACCR-compliant cancer registries with automated abstraction and full provenance information, covering all cancer sites.
4. Custom Registries: User-defined registries specifying target fields and patient cohorts, continuously extracted and maintained from incoming clinical documentation.

Reported Impact Metrics

• 40% more complete data by capturing diagnoses and clinical context from unstructured notes that structured EHR fields alone miss.
• 10+ FTE-years saved annually by eliminating redundant pipeline rebuilding for each registry, cohort, dashboard, or AI model.
• Weeks reduced to hours for patient timeline creation, cohort queries, and feature engineering.
• Built-in compliance covering de-identification, audit trails, data lineage tracking, and governance workflows without separate tools or custom development.

Supported Use Cases

• Clinical Research: Identify trial-eligible patients across all data sources, analyze real-world treatment effectiveness with complete medication and outcome timelines, and federate multi-site studies using standardized OMOP cohorts.
• AI and Predictive Analytics: Train models on de-identified research data and deploy on identified operational data with no feature drift; access pre-extracted features such as diagnosis timelines, medication adherence, and lab trends without custom NLP pipelines.
• Quality Improvement and Population Health: Measure outcomes against clinical guidelines using standardized terminologies, track chronic disease management across all touchpoints, and detect care gaps at scale.
• Regulatory Reporting and Registries: Automate cancer registry abstraction with NAACCR compliance, generate quality measure reports (HEDIS, CMS) without manual chart review, and maintain public health surveillance feeds with full data lineage.

Open Standards Architecture

• OMOP CDM v5.4: All patient data is standardized to OMOP CDM v5.4, enabling compatibility with OHDSI ecosystem tools (ATLAS, ACHILLES, CohortMethod), multi-institutional federated analytics, reproducible research, and AI model portability. Supported domains include Condition, Drug Exposure, Procedure, Measurement, Observation, Visit, Person, Provider, Device Exposure, Note, and Specimen.
• Standard Medical Terminologies: Clinical concepts are mapped to SNOMED CT, RxNorm, LOINC, ICD-10-CM, HPO (Human Phenotype Ontology), and UMLS Metathesaurus, enabling semantic interoperability across systems without manual mapping.
• Model Context Protocol (MCP): All platform capabilities — data extraction, cohort queries, patient timelines, registry abstraction — are exposed via MCP, an open standard for AI agent interoperability developed by Anthropic. This enables composable multi-step workflows, automatic tool discovery by agents, ecosystem integration with any MCP-compatible agent, and custom extensions using proprietary institutional logic.

The platform is architected around open specifications with no proprietary formats or licensing fees for the underlying standards, meaning OMOP data remains accessible and portable with any OMOP-compatible tool independent of the platform. Governance controls, access permissions, and transformations are applied per dataset type based on intended use, while the