CHoRUS: Multi-Hospital Clinical Data for AI Research on Acute Illness

The Collaborative Hospital Repository Uniting Standards (CHoRUS) is a multi-hospital, multimodal clinical dataset developed under the NIH Bridge2AI program. It harmonizes electronic health record data, medical imaging, high-resolution physiologic signals, and tokenized clinical text across participating institutions. The repository is curated within a NIST 800-53–aligned secure enclave and was last updated in April 2026.

Use Cases

• Develop predictive models for patient recovery based on harmonized EHR data and physiologic signals.
• Validate multimodal AI algorithms using synchronized imaging, telemetry, and clinical text data.
• Benchmark AI model performance for clinical tasks across multiple hospital data sources.
• Conduct reproducible research on acute illness within a secure, auditable data enclave.

Strengths

• Data is harmonized across multiple hospitals using standardized data models and governance frameworks.
• Includes multiple high-resolution data modalities: EHRs, imaging, physiologic signals, and clinical text.
• Architected within a NIST 800-53–aligned secure enclave with role-based access controls and audit logging.
• Designed to support AI model development, validation, and benchmarking under NIH Bridge2AI program.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.
• Access is controlled and requires governance review, which may delay availability.

Provenance

Source

NIH Bridge2AI program, multiple participating hospitals.

Collection Method

Harmonized from institutional EHRs, imaging systems, and physiologic monitors using standardized data models.

Freshness

Last updated 2026-04-26 06:10:22; freshness should be verified.