================================================================================
CRATE-AUGMENTED SOURCE BUNDLE (de novo fork)
================================================================================
Project: CHORUS
Document bundle: data/preprocessed/concatenated/CHORUS_preprocessed.txt
Crate package: data/ro-crate_packages/CHORUS
Crate manifest: data/ro-crate_packages/crate_manifest.yaml

This bundle is the document corpus plus RO-Crate evidence. Artifacts
that are already in D4D or datasheet form are deliberately withheld so
that this arm extracts rather than transcribes; see the exclusion list
below and notes/D4D_GENERATION_ARMS.md.

CRATE EVIDENCE INCLUDED
--------------------------------------------------------------------------------
  + CHORUS_crate_metadata_reduced.json — crate JSON-LD with file inventories collapsed; the substantive evidence (rai:* fields, ethics, access, provenance)
  + ai_ready_score.json — AI-readiness self-assessment

CRATE ARTIFACTS WITHHELD
--------------------------------------------------------------------------------
  - CHORUS_crate_d4d.yaml — already a schema-valid D4D record — this is the deterministic fork's output; including it would make the de novo arm copy-through
  - ro-crate-linkml.yaml — upstream's own D4D-shaped mapping; same copy-through risk
  - ro-crate-datasheet.html — upstream-authored datasheet rendering of the same content; a datasheet is the artifact being generated, so it is withheld as input

================================================================================
================================================================================
CONCATENATED DOCUMENT
================================================================================
Input Directory: data/preprocessed/individual/CHORUS
Total Files: 4
Extensions: ['.txt']
Recursive: False
Selection Manifest: data/preprocessed/source_manifest.yaml
================================================================================

TABLE OF CONTENTS
--------------------------------------------------------------------------------
  1. reporter_nih_gov_project-details-10472824_row7.txt
  2. bridge2ai-for-clinical-care-informational-webinar-cohort-2_row9.txt
  3. chorus4ai_org_row11.txt
  4. github_chorus_ai_overview_2025-11-14.txt
================================================================================

FILE: reporter_nih_gov_project-details-10472824_row7.txt
PATH: data/preprocessed/individual/CHORUS/reporter_nih_gov_project-details-10472824_row7.txt
SIZE: 5492 bytes
--------------------------------------------------------------------------------

SOURCE METADATA
Project: CHORUS
Source ID: nih_reporter_project
Source type: NIH project page
Source URL: https://reporter.nih.gov/project-details/10472824
Raw file: data/raw/CHORUS/reporter_nih_gov_project-details-10472824_row7.txt
--------------------------------------------------------------------------------
NIH RePORTER Project
Source: https://reporter.nih.gov/project-details/10472824
Application ID: 10472824
Project number: 1OT2OD032701-01
Core project number: OT2OD032701
Title:  Bridge2AI: Patient-Focused Collaborative Hospital Repository Uniting Standards (CHoRUS) for Equitable AI
Principal investigator: ROSENTHAL, ERIC S.
Organization: MASSACHUSETTS GENERAL HOSPITAL
Fiscal year: 2022
Award amount: 5880300
Project start: 2022-09-01T00:00:00
Project end: 2026-11-30T00:00:00

Abstract Text
There is an urgent need for infrastructure to support artificial intelligence and machine learning (AI/ML) in critical care. Developing high-resolution multi-center data sets is a critical first step towards actionable and trustworthy AI. As part of the NIH Common Fund’s Bridge2AI program, the CPatient-Focused Collaborative Hospital Repository Uniting Standards (CHoRUS) for Clinical Care AI data generation project will meet the need of generating data for ML/AI applications aimed at characterizing acute and critical care illness, predicting complications, and measuring treatment response among patients with acute or critical illness. Through 3 pillars, the CHoRUS data generation project will addresses multiple challenges relevant for acquiring an AI-ready data set from more than 100,000 critically ill patients: 1) Data (Standards, Tool Development and Optimization, and Data Acquistion) 2) Ethics (Ethical and Trustworthy AI) and 3) People (Team Science and Skill and Workforce Development). The project’s overarching goal is to develop a publicly available, AI-ready critical care dataset, while ensuring the methods promote privacy, accountability, and clinical benefit, while promoting a new generation of AI clinicians and scientists. The dataset will also provision a holdout test set, accessible for model external validation to aid marketplace adoption of AI-developed models for implementation in acute and critical care. Drawing expertise from a comprehensive set of disciplines such as team science, law, ethics, health services, biomedical science, engineering, and scientific journal publications, this project will A) establish a legal framework for collecting data at scale, sampling to ensure comprehensive sets of patient conditions and clinical treatment strategies; B) perform community-facing ethics focus groups to determine what data is appropriate for public sharing; C) ensure that data elements feature appropriate contextual factors such as geographic distance to the nearest hospital; D) develop capabilities across a multi-center network to acquire, standardize, tokenize, store, visualize, and label data such as structured electronic health record data, tokenized unstructured electronic health record data, telemetry and EEG waveforms, imaging, and social determinants of health; E) acquire data, standardize data to the OMOP Common Data Model, transform data using approaches that limit re-identification, and label data for ; and F) cultivate expertise in the lay and scientific community to improve AI literacy nd utilization through multimodal educational approaches. To accomplish this, the project will involve extensive collaboration between centers as well as through the NIH Bridge2AI program, the NIH Bridge2AI Bridge Center, external biomedical and clinical organizations, industry, and regulatory agencies.

The Patient-Focused Collaborative Hospital Repository Uniting Standards (CHoRUS) for Equitable AI will develop the necessary network, tools, standards, data, and education to build machine-learning (ML) and artificial intelligence (AI)-derived models to improve the care and outcomes of patients requiring critical care.  By performing legal and ethical research, developing tools for data storage, labeling and analysis, acquiring and standardizing a dataset of unprecedented diversity and high resolution, sequestering holdout datasets for external validation, and enriching the community through education, this data generation project will catalyze the biomedical research and expertise necessary to promote the patient-focused deployment of AI in acute and critical care.

Preferred terms:
Accountability;Acute;Address;Adoption;Artificial Intelligence;Biomedical Research;Bridge to Artificial Intelligence;Clinical;Collaborations;Communities;Critical Care;Critical Illness;Data;Data Element;Data Set;Data Storage and Retrieval;Deterioration;Diagnosis;Discipline;Education;Electroencephalography;Electronic Health Record;Engineering;Ensure;Ethics;Event;Focus Groups;Funding;Generations;Goals;Health Services;Hospitals;Image;Industry;Infrastructure;Intelligence;Journals;Label;Laws;Legal;Machine Learning;Measures;Methods;Modeling;Patient-Focused Outcomes;Patients;Privacy;Publications;Research;Resolution;Sampling;Science;Scientist;Standardization;Telemetry;Testing;United States National Institutes of Health;Validation;Workforce Development;acute care;care delivery;care outcomes;data acquisition;data modeling;data standards;data tools;electronic structure;improved;literacy;multimodality;programs;repository;skill acquisition;social health determinants;tool;tool development;treatment response;trustworthiness


================================================================================

FILE: bridge2ai-for-clinical-care-informational-webinar-cohort-2_row9.txt
PATH: data/preprocessed/individual/CHORUS/bridge2ai-for-clinical-care-informational-webinar-cohort-2_row9.txt
SIZE: 16257 bytes
--------------------------------------------------------------------------------

SOURCE METADATA
Project: CHORUS
Source ID: cohort_2_webinar
Source type: tutorial
Source URL: https://www.aim-ahead.net/media/jnzdnid3/bridge2ai-for-clinical-care-informational-webinar-cohort-2.pdf
Raw file: data/raw/CHORUS/bridge2ai-for-clinical-care-informational-webinar-cohort-2_row9.pdf
--------------------------------------------------------------------------------
AIM-AHEAD Bridge2AI for Clinical Care
Training Program

Cohort 2

Informational Webinar

September 09, 2025, 2:00pm Central

JK


AIM-AHEAD Consortium

AIM-AHEAD  is  a  nationwide  network  of  institutions  and  organizations
designed  to  build  AI  talent  among  researchers  and  clinicians,  support
multidisciplinary  research  projects  that  harness  AI/ML  to  improve  the
health of Americans, and enhance the AI capabilities and infrastructure
of  communities  or  hospitals  that  otherwise  would  not  have  had  the
resources or the capacity to beneﬁt from the advancement of AI/ML.

JK


The AIM-AHEAD Coordinating Center

The A-CC consists of four cores, focused on various initiatives to achieve
AIM-AHEAD’s mission.

Leadership Core

Data Science Training Core

Lead, recruit, and coordinate
the AIM-AHEAD Consortium

Assess, develop, and
implement data science
training curriculum

Data and Research Core

Infrastructure Core

Address research priorities and
needs to form a comprehensive
basis for AI/ML

Assess data, computing, and
software infrastructure to facilitate
AI/ML and health research

JK


NIH Leadership Team

Samson Gebreab, Ph.D. MSc.
Program Lead, AIM-AHEAD
Oﬃce of Data Science Strategy,
NIH

Shurjo K. Sen, Ph.D.
Program Director, Bridge2AI
Oﬃce of Genomic Data Science,
NIH

Haluk Resat Ph.D.
Program Lead, Bridge2AI
Oﬃce of Strategic Coordination
NIH

Dr. Emir Khatipov
Program Director, AIM-AHEAD
Oﬃce of Data Science Strategy,
NIH

Eva Lancaster, Ph.D.
Program Director, AIM-AHEAD
Oﬃce of Data Science Strategy,
NIH

JK


AIM-AHEAD  Leadership Team

Jamboor Vishwanatha, PhD
UNT Health Science Center
AIM-AHEAD PI

Toufeeq A. Syed, PhD
University of Texas Health
Science Center,
Houston, TX
AIM-AHEAD MPI

Nawar Shara, PhD
MedStar Health Research Institute
AIM-AHEAD DSTC MPI

JK


Bridge2AI CHoRUS Leadership Team

Eric Rosenthal
Massachusetts General
Hospital

Azra Bihorac
University of Florida

Xiaoqian Jiang
UTHealth Houston

Yulia Strekalova
University of Florida

Parisa Rashidi
University of Florida

Manlik Kwong
Tufts University

TS


Program Partnership

Strategic Collaboration: AIM-AHEAD and Bridge2AI collaborate to deliver specialized
AI/ML training for clinical care, leveraging shared resources and expertise.

Combined Expertise: AIM-AHEAD’s strength in trainee engagement and Bridge2AI’s AI data
and curriculum deliver a comprehensive training experience.

Partnership

Focus on Training: The partnership equips trainees with practical skills to apply AI/ML
methods to clinical care challenges using real-world data.

Goal: Build a skilled workforce prepared to advance AI/ML applications in health research.

TS


Program Purpose

The  AIM-AHEAD  Bridge2AI  for  Clinical  Care  Training  Program
expands access to Bridge2AI CHoRUS data through engagement, AI
training, and mentorship. It equips trainees to apply AI/ML methods
innovative  research  at  the
to  big  data  analysis  and  conduct
intersection  of  healthcare  challenges  and  AI/ML  using  multi-modal
datasets from a broad range of cohort participants.

TS


Bridge2AI for Clinical Care Dataset

Multicenter

Multimodal &
High-Resolution

ER


CHoRUS Dataset

• Retrospective data collection

• Controlled access

• As of August 2025, covers 14 different

hospitals with over 45K unique admissions
• OMOP and telemetry in enclave except:

• Clinical notes – stored locally except tokens

•

Imaging – currently 1000 images available
with de-id in process for larger cohort

• EEG – extraction in process at this point

• Datasets are being used for training

activities and publications

Data type

Data
standard

Demographics OMOP

OMOP

Controlled

Controlled

Access control

Metadata

Yes

Yes

Yes

Yes

OMOP

Controlled

OMOP

Controlled

Medication administration
(dosing time-stamped upon
each infusion change or
dose administration)

Procedures (documentation
by providers)

Nursing flowsheets
(high-frequency
documentation)

Diagnoses (documentation
by providers)

Clinical notes (extracted
and tokenized using OHNLP
toolkit)

OMOP

Controlled

Yes

Yes (OMOP schema)

OHNLP

Controlled

Planned

Published metadata
schema

Yes (OMOP schema)

Yes (OMOP schema)

Yes (OMOP schema)

Yes (OMOP schema with
extensions)

Yes (OHNLP open source
schema)

Yes (DICOM schema)

Yes (PhysioNet schema
extended)

Imaging (from PACS)

DICOM

Waveform telemetry
(bedside monitors,
gateway/middleware)

WFDB

Controlled

Controlled

Planned

Yes

Waveform EEG (hospital
database)

EDF+ and
Persyst

Controlled

Planned

Yes (open source EDF+ and
Persyst schema)

ER


Bridge2AI for Clinical Care Dataset

ER


CHoRUS Dataset

ER


CHoRUS Dataset

ER


CHoRUS Dataset

ER


Foundational Hands-On Training

Gain experience with AI/ML in the Bridge2AI CHoRUS ecosystem
Learn fundamental skills, tools, and design patterns for applying AI to clinical problems.

Asynchronous
Jupyter Notebooks

Synchronous
Office Hours

Hybrid Workshop:
AI in Clinical Care

Python and Version Control

Structured EHR Datasets

ML for Clinical Applications

Clinical Deep Learning

Ethics of Clinical AI

Virtual meetings
with AI experts

Guided coding exercises

Notebook review

Hands-on brainstorming

Open Q&A

Community team building

ER


Accessing the Data

Registration Process

Participants will ﬁll out a registration form with name, email, and institution

Once access is granted and compute is provisioned, an email will be sent

1

2

Licensing Agreement

All participants must sign a licensing agreement included in the registration form before gaining
access to the dataset

*This requirement is not a barrier to acceptance into the program. Program administrators will assist with this access if
needed.

ER


Training Overview

Trainees will receive hands-on training on the Bridge2AI AI/ML for Clinical Care Network
and leverage the data and tools to create practical use cases, putting their new skills
to work in real-life situations and innovative data-driven research. Training will include:

Workshops on using Jupyter
Notebooks

Didactics on generative AI
and speciﬁc use cases

Ongoing mentorship and support
using Collaborative Cloud
platforms

Workshops on using the OHDSI
tool stack

Instruction on creating
practical use cases

Workshops on the OHDSI/OMOP
common data model

Virtual live courses and
learning sessions

Hands-on training on the
Bridge2AI AI/ML for Clinical
Care Collaborative Cloud

ER


Curriculum Overview

Examples of Learning Sessions Provided:

Host

Lecturer(s)

Delivery

Approach

Topic and Description

MGH

Morteza
Zabihi

Live Online

Didactic

Machine Learning Basics - Intro to ML
methods for AI

Hosts: MGB, UF, UTH, Tufts

Delivery: Live Online,
Recorded, Asynchronous

Format: Didactic, Workshop,
Office Hours, Self-Directed

UF

Zhenhong Hu

Self-Directed

Python
Notebook

Intro to Python & Version Control

UTH

Debora
Simmons

Recorded

Lecture

Ethics of AI in Clinical Practice - Safety, risk,
and legal considerations

Tufts

Jared
Houghtaling

Live Online

Didactic

Working with EHR Data for Research

MGH

Siril Singa

Workshop

Collaborative

Data Schemas in Clinical Cloud

ER


Curriculum Overview

AI-LEARN Curriculum for Bridge2AI Broad Learning Communities

Courses to leverage & sync DSTC_MHRI Workshops

Curriculum
Offering

Target Audience /
Purpose

Key Topics

Format: Online, self-paced
with video lectures, case
studies, and exercises.

Certiﬁcation: Available upon
completion.

AI/ML Essentials for
Healthcare

No coding;
healthcare workers

Intro to AI/ML, ethics, patient engagement

Advanced
Decision-Making
Models

Focus on model
selection for
healthcare

Statistical modeling, decision trees, healthcare
use cases

Cutting Edge AI
Training Modules

Keeping up with AI
advancements

Integration with DSTC_MHRI workshops, latest
AI/ML trends

MS


Curriculum Overview

AI-LEARN Curriculum for Bridge2AI Broad Learning Communities

Courses to leverage & sync DSTC_MHRI Workshops

Titles

Key Topics

Format

Format: Sync workshops and
panel discussions hosted by
MHRI team

Navigating IRB, Data Compliance,
and Quality Assurance in AI/ML
Healthcare Research

- IRB protocol drafting
- HIPAA/GDPR compliance for OMOP/FHIR data
- QA audit frameworks

Hyperparameter Tuning, Model
Selection, and Deployment for
Healthcare AI

- Compare optimization techniques (e.g.,
Bayesian)
- Model interpretability vs. performance
- Docker deployment demo

Hands-on IRB drafting,
MedStar Program managers/
IRB Chair Q&A, Compliance
checklist toolkit

Live session with real-world
use cases

Bridging the Gap: Clinicians
and Data Scientists on
Methodological Challenges

- Align AI projects with clinical priorities
- Address EHR data limitations

Panel (MedStar clinicians, data
scientists), breakout consultations,
live Q&A

MS


Program Trainee Objectives

Objective 1

Exhibit advanced expertise in AI/ML principles as they are applied to clinical care.

Objective 2

Develop and present use cases suitable to apply in Bridge2AI Data Topics.

Objective 3

Participate  in  collaborative  research  on  the  Bridge2AI  Clinical  Care  Cloud  platform,  applying
program insights and engaging with BRIDGE Center expertise for responsible AI/ML.

Objective 4

Prepare a compelling poster presentation for the AIM-AHEAD and Bridge2AI Annual Meetings,
submit  an  abstract  for  a  health  informatics  conference,  and/or  develop  a  manuscript  for  a
peer-reviewed journal.

After completing the program, trainees will understand how to develop real-world use cases and how to address
concerns such as privacy in the responsible application of AI/ML. They will be equipped to engage in collaborative
research on the Bridge2AI platform and will connect with a community of professionals dedicated to advancing the
use of AI/ML in biomedical research.

Outcome

MS


Trainee Expectations

In order to successfully complete the program, selected trainees must:

Time Commitment: Be able to commit to 8 hours per week (on average) of coursework and synchronous
class sessions

Attendance: Attend one virtual, synchronous class session per week (day of the week and time TBD)

Assignments: Complete all assigned milestones and goals

Presentation of Work: Attend both the AIM-AHEAD Annual Meeting* (July 2026) and the
Bridge2AI Annual Meeting* (May 2026) and present a works-in-progress poster.

*These are both in-person events and a travel allowance will be given to each trainee for travel expenses.

TS


Program Benefits

Stipend

An $8,000 stipend upon successful
completion of trainee milestones

Travel allowances to attend both the
AIM-AHEAD Annual Meeting and the
Bridge2AI Conference in 2026

Support

Support and guidance
from an experienced
AIM-AHEAD mentor

Support from the
AIM-AHEAD Data
Science Training Core

Direct 1:1 guidance, virtual oﬃce
hours, helpdesk support, and
concierge services supporting R/
Python coding and the OHDSI tool
stack

Training

Training on:

● Introductory machine learning and feature engineering.
● The Bridge2AI AI/ML for Clinical Care Collaborative Cloud.
● Ethics and Policy issues in AI/ML.
● AI/ML for Clinical Care canonical Jupyter Notebooks.
● The OHDSI/OMOP common data model.
● Generative AI and speciﬁc use cases.
● Creating practical use cases during Bridge2AI topics.

TS


AIM-AHEAD Mentorship Process

Each  trainee  will  be  matched  with  a  mentor  who  will  provide  mentoring
and  support  throughout  the  training  program.  Mentors  are  matched  with
mentees using the Connect Platform.  Mentorship matches are made using:

AI Algorithm

Administrative
Matching

Mentor Pool
Search

TS


Applicant Eligibility

Citizenship
and Tax
Requirements

✓ Must be a U.S. Citizen, Permanent Resident, or Non-Citizen U.S. National
✓ Permanent Residency must be established by Sept. 26, 2025
✓ Temporary visa holders (F1, J1, H1, etc.) are not eligible
✓ Accepted candidates must be able to submit a W-9 form

✓ Current/former AIM-AHEAD program participants (awardees, fellows, trainees,

Participation
Restrictions

mentors, advisors, coaches) are ineligible

✓ Applicants may apply to more than one AIM-AHEAD training program but can only be

selected for one

Special Cases

✓ AIM-AHEAD Coordinating Center personnel and Federal employees may participate,

but will not receive stipend or travel allowance

Note: Please refer to the full CFA on AIM-AHEAD.net for all eligibility requirements.

TS


Education & Experience

Education
Requirements

✓ Minimum: Bachelor’s degree in physical sciences, life sciences, math, statistics, data

science, engineering, health sciences, or public health

✓ Eligible applicants include: post-baccalaureate and graduate students, postdocs, medical
students/residents, allied health trainees, early-career investigators, and early-career
professionals in non-academic institutions

Required Skills

✓ Prior programming experience
✓ Basic understanding of statistics
✓ Working command of English

Recommended
Skills

✓ Coursework in probability and statistics or higher-level math
✓ Coding experience in R or Python
✓ Experience with data manipulation and management through coursework or research

Note: Please refer to the full CFA on AIM-AHEAD.net for all eligibility requirements.

TS


Eligibility Requirements

Eligible Organizations

Higher education institutions

Local, state, and tribal
governments

Nonprofits
(with or without 501(c)(3) status)

For-profit businesses and
organizations

Other U.S.-based organizations

(e.g., school districts, housing
authorities, faith-based,
community-based, and regional
organizations)

Email Requirement

In order to gain access to
the dataset, you will need
to have a “.edu” email
address.*

*This requirement is not a barrier
to acceptance into the program.

Program administrators will
assist with this access if needed.

Note: Please refer to the full CFA on AIM-AHEAD.net for all eligibility requirements.

TS



Application Requirements

Submission Deadline: September 26, 2025 by 11:59 PM EST

Required Application Elements

●
●

●
●
●

Proﬁle Information (basic applicant details in InfoReady portal)
Letters of Support & Recommendation (two letters)

○
○

Supervisor letter conﬁrming protected time
At least one faculty recommendation (additional letters optional)

Academic Transcript (undergraduate and, if applicable, graduate)
Biographical Sketch (NIH biosketch or CV, max 5 pages)
Statement of Rationale (≤2 pages) describing goals, need for training, relevant experience, and long-term plans

Important Note

●
●

Applicants may apply to more than one program but can only be selected for one
Applicants will rank program preferences in the application

*This is just an overview. Please see the CFA for the full list of application requirements

TS


Application Process

Applications must be submitted between September 02, 2025 and September 26, 2025 at 11:59 PM EST

Note: Please use Chrome, Firefox, or Edge browsers.

1

2

3

4

Familiarize
yourself with the
program
requirements
outlined in the call
for applications

Gather all of the
required
application
materials

Create an
account on
AIM-AHEAD
Connect  and
register as a
“mentee/learner”

Submit
application for
review using the
InfoReady
platform

Up to 30 trainees will be selected

TS


Key Program Dates

CFA Release Date

September 02, 2025

Application Deadline

September 26, 2025 by 11:59 PM EST

Notice of Award

November 10, 2025

Program Start Date

November 17, 2025

Bridge2AI Conference 2026

May 2026

AIM-AHEAD Annual Meeting 2026

July 2026

Program End Date

July 31,  2026

TS


Questions?

Please see the FAQ
document  linked
above and in the chat

Scan the QR code above
to access the
AIM-AHEAD Bridge2AI
for Clinical Care CFA.

TS


================================================================================

FILE: chorus4ai_org_row11.txt
PATH: data/preprocessed/individual/CHORUS/chorus4ai_org_row11.txt
SIZE: 2876 bytes
--------------------------------------------------------------------------------

SOURCE METADATA
Project: CHORUS
Source ID: project_documentation
Source type: documentation
Source URL: https://chorus4ai.org/
Raw file: data/raw/CHORUS/chorus4ai_org_row11.html
--------------------------------------------------------------------------------
CHoRUS – This repoitory is under review for potential modification in compliance with Administration directives.
Skip to content
CHoRUS
Homepage
Team
Project Overview
Tools
Data Acquisition
Standards
Community Engagement
Ethics
Dataset
Learning Resources & Opportunities
This repoitory is under review for potential modification in compliance with Administration directives.
AI for Clinical Care
Creating a patient-focused collaborative hospital repository uniting standards for clinical care AI
DataSet
Expanding Artificial Intelligence and Machine Learning to Improve Recovery from Acute Illness
The CHoRUS project is one data generation project of four in the National Institute of Health (NIH) funded Bridge2AI consortium. The goal is to work with projects to create ethically sourced datasets and tools around Artificial Intelligence and to create best practices for AI in health.
Learn more about Bridge2AI Consortium
Snapshot of the dataset
Anticipated Final Dataset
100,000
Patient admissions
9
Different data modalities
14
Data contributing hospitals
60+
CHoRUS consortium members across 20 different institutions
Current Released Dataset
50,000
Patient admissions from ICU, PICU, and NICU
1.6 Billion
Rows of EHR OMOP data
7,642
Admissions with Radiology Data
23 Tb
Waveform data
CHoRUS Project Components
Data
The CHoRUS project is the collection of multi-modal data across 14 different hospitals. The data utilizes ongoing and new standardization practices. The team creates tools to support the cloud enclave.
Ethics
The CHoRUS consortium evaluates community perspectives on clinical care AI to find ways to increase trustworthiness of provenance and privacy in AI. New approaches are being created as existing legal and regulatory landscape is analyzed.
People
Dedicated to engaging the community in CHoRUS and Bridge2AI work products, CHoRUS project is dedicated to training engineers, data scientists, and clinicians to solidify AI in Clinical Care Data Engineering Infrastructure.
Learning Opportunities
Learn of different opportunities to gain hands-on training using the CHoRUS dataset to support the development of transnational AI in clinical care.
Learn More
CHoRUS Consortium
This project is funded by the NIH under award number OT2OD032701. The content on this website is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Pages
Homepage
Team
Project Overview
Tools
Data Acquisition
Standards
Community Engagement
Ethics
Dataset
Learning Resources & Opportunities
Contact Us
Ciera McCrary, MGH, Program Manager
cmccrary@mgh.havard.edu


================================================================================

FILE: github_chorus_ai_overview_2025-11-14.txt
PATH: data/preprocessed/individual/CHORUS/github_chorus_ai_overview_2025-11-14.txt
SIZE: 9270 bytes
--------------------------------------------------------------------------------

SOURCE METADATA
Project: CHORUS
Source ID: github_organization_overview
Source type: historical documentation
Source URL: https://github.com/chorus-ai#table-of-contents
Raw file: data/raw/CHORUS/github_chorus_ai_overview_2025-11-14.pdf
--------------------------------------------------------------------------------
11/14/25, 5:25 PM

CHoRUS for Equitable AI

chorus-ai

Overview

Repositories

28

Discussions

Projects 3

Packages

Peopl

CHoRUS for Equitable AI

Follow

37 followers

United States of America

https://chorus4ai.org/

README.md

CHoRUS for Equitable AI

license MITMIT
license

Patient-Focused Collaborative Hospital Repository Uniting Standards (CHoRUS) for
Equitable AI

Table of Contents

About

Get Started

Data Site Managers

Clinical Collaborators

Project Managers

Software Developers

Community Users

License

Contact

https://github.com/chorus-ai#table-of-contents

1/8


11/14/25, 5:25 PM

About

CHoRUS for Equitable AI

The goal of the CHoRUS Network is to develop the most diverse, high-resolution, ethically

sourced, AI-ready data set to answer the grand challenge of improving recovery from acute

illness.

This collaboration spans 20 academic centers, of which 14 will contribute as Data

Acquisition centers.

Patient-focused efforts will determine the ethical and legal approaches to manage privacy

and bias, while accounting for Social Determinants of Health.

Unified standards will harmonize multi-modal EHR, waveform, imaging, and text data.

A visualization and annotation environment will label data with targets important for

prediction.

A comprehensive set of approaches will develop the skills and workforce for a next

generation of diverse academic and community AI scientists.

Federated access will enable sampling methods to ensure a balanced and diverse cohort.

Collaborating with Bridge2AI and 3 other data generation projects, the CHoRUS Network will

help us cross the Bridge2AI network together.

Get Started

The CHoRUS GitHub organization houses active repositories that provide:

1. Software and tooling to interact with and extract insight from clinical data in diverse

formats

2. Validated semantic mappings for connecting clinical data in various source formats to

international standards

3. Standard operating protocols (SOPs) to instruct data contributing sites about best

practices for curating and delivering interoperable datasets

4. Project management overviews to help track data delivery statuses and complex task

dependencies within CHoRUS

We've defined different groups of anticipated users of this GitHub in the sections below, and

direct those users to appropriate locations within the chorus-ai repository space.

(back to top)

Data Site Managers

The starting reference for data managers at data contributing sites is the Chorus_SOP page.

Here, you will find an interactive workflow diagram describing the step-by-step process for

extracting and contributing clinical data to CHoRUS, with dynamic links to various SOP

documents that have undergone an internal validation and review process within the

Chorus_SOP repository.

https://github.com/chorus-ai#table-of-contents

2/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

The workflow diagram also includes links to a wealth of recordings and documentation

compiled by the various sub-teams (Standards, Data Acquisition, and Tooling) within the

CHoRUS DGP.

If you run into any issues in the creation or submission of your data extract, please feel free
to post them in the relevant context-specific discussion location:

Standards Discussions

Data Acquisition Discussions

Clinical Collaborators

(back to top)

Clinical expertise is invaluable to semantic mapping and validation within CHoRUS. We have

established a chorus-mapping repository with documentation and pooled tabular mappings

along with an associated clinical validation SOP for contributing to mapping efforts.

If you're interested in getting involved in downstream analytics on the mapped and
assembled dataset, please feel free to reach out using the contact details below

(back to top)

Project Managers

We've established two different modes of project management within the chorus-ai
organization:

1. Task tracking among CHoRUS members and across teams

2. Status tracking of data contributing sites to identify and resolve any blocking issues

Task Tracking

We are aggregating issues across various repositories into an overall task management

project. In this GitHub project, we assign users to tasks, create and track anticipated

delivery dates, and highlight dependencies between tasks and users. Several per-repo
projects are also active, but we are in the process of phasing those projects out and

migrating their contents to this central project space.

Status Tracking

We have asked data contributing sites to provide regular status updates with regard to their

progress in creating and curating a CHoRUS-specific clinical data extract. Sites can submit

updates either using the GitHub interface directly, or by submitting a Google Form (please

reach out to get a link to the form if you'd like to submit an update for your site). We've

created a GoogleScript that is triggered on each Google Form submission and makes calls
to the GitHub API to update status and issue information appropriately. These site statuses

end up in either the Standards Project or the Data Acquisition Project.

https://github.com/chorus-ai#table-of-contents

3/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

(back to top)

Software Developers

We have software contributors within CHoRUS who have a broad range of expertise, and

who have produced powerful open-source tooling for transforming and interacting with

clinical data. We have created a web guide for both contributors and users of the CHoRUS

software packages that compiles documentation in the chorus-developer repository.

You can check the versions, maintainers, and other metadata about CHoRUS packages

using our package status page.

(back to top)

Community Users

Welcome! Feel free to browse any of the resources listed above, or check out our public-

facing webpage for more information about the project, its progress, and high-level aims.

Thanks for stopping by!

License

(back to top)

This project is licensed under the MIT License. See the LICENSE file for more details.

(back to top)

Contact

For any inquiries or feedback, please feel free to reach out to us:

Request access: dbold@emory.edu or jared.houghtaling@tuftsmedicine.org

Website: www.bridge2ai.org/chorus

(back to top)

Popular repositories

chorus-container-apps

Public

Set of dockerized container applications and associated configurations to be deployed to support CHoRUS
services on Azure

JavaScript

 5

 1

https://github.com/chorus-ai#table-of-contents

4/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

privacy_scan_tool

A Privacy Scan tools for medical records

Python

 4

chorus_waveform

CHoRUS waveform documentation and various waveform conversion scripts

Jupyter Notebook

 3

 3

Chorus_SOP

ChoRUS centralized SOP documentation site

MDX

 3

 4

CADA

 2

data_acq_SOP

An online documentation SOP for Data Acquisition Module

JavaScript

 2

 2

 Repositories

Find a repository…

Type

Language

Sort

Public

Public

Public

Public

Public

chorus-container-apps

Public

Set of dockerized container applications and associated configurations to be deployed to

support CHoRUS services on Azure

https://github.com/chorus-ai#table-of-contents

5/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

JavaScript

 5

 1

 3

 0

Updated 11 hours ago

Chorus_SOP Public

ChoRUS centralized SOP documentation site

MDX

 3

Apache-2.0

 4

 2

 0

Updated yesterday

CTP-deid Public

0

0

 0

 0

Updated on Oct 3

UF-Geocoding Public

Forked from bihorac-LAB/Exposome

Open source code to geocode OMOP Location entities via DeGauss

Python

0

MIT

 3

 0

 0

Updated on Sep 25

chorus_waveform_resources

Public

Resources which support the preparation and release of waveforms

0

0

 0

 0

Updated on Sep 18

CHoRUSReports

Public

This repo is dedicated to producing characterization reports to return to sites following their
data submissions

R

0

0

 1

 0

Updated on Aug 8

cloud_public Public

https://github.com/chorus-ai#table-of-contents

6/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

0

0

 0

 0

Updated on Jun 16

chorus-extract-upload Public

Repository with tools to create and upload a CHoRUS Data Extract

Python

0

MIT

0

 1

 0

Updated on May 15

chorus-apps-internal

Public

CHoRUS internally built web applications.

0

0

 29

 0

Updated on May 14

chorus_waveform Public

CHoRUS waveform documentation and various waveform conversion scripts

Jupyter Notebook

 3

MIT

 3

 8

 1

Updated on Feb 21

View all repositories

Top discussions this past month

Discussions are for sharing announcements, creating conversation in your community, answering
questions, and more.

Start a new discussion

People

https://github.com/chorus-ai#table-of-contents

7/8


11/14/25, 5:25 PM

CHoRUS for Equitable AI

Top languages

JavaScript

Python

Jupyter Notebook

CSS

MDX

Report abuse

https://github.com/chorus-ai#table-of-contents

8/8


================================================================================
CRATE EVIDENCE
================================================================================

================================================================================
FILE: CHORUS_crate_metadata_reduced.json
ROLE: crate JSON-LD with file inventories collapsed; the substantive evidence (rai:* fields, ethics, access, provenance)
SIZE: 19,691 characters
--------------------------------------------------------------------------------
{
 "@context": "https://w3id.org/ro/crate/1.2/context",
 "@graph": [
  {
   "@id": "ro-crate-metadata.json",
   "@type": "CreativeWork",
   "conformsTo": {
    "@id": "https://w3id.org/ro/crate/1.2"
   },
   "about": {
    "@id": "ark:59853/rocrate-chorus-ro-crate-package/"
   }
  },
  {
   "@id": "ark:59853/rocrate-chorus-ro-crate-package/",
   "@type": [
    "https://w3id.org/EVI#Dataset",
    "https://w3id.org/EVI#ROCrate"
   ],
   "name": "CHoRUS RO-Crate Package",
   "identifier": "https://doi.org/10.18130/V3/XNBOPG",
   "description": "The Collaborative Hospital Repository Uniting Standards (CHoRUS) for Clinical Care AI is a clinical data network developing a flagship multi-hospital, multimodal, high-resolution dataset under the NIH Bridge2AI program. CHoRUS harmonizes electronic health record (EHR) data, imaging, high-resolution physiologic signals (including telemetry data and EEG where available), and tokenized note and report text (via OHNLP toolkit) with raw text retained locally across participating institutions using standardized data models and governance frameworks.\nThe repository is designed to enable secure, auditable, and reproducible artificial intelligence (AI) research to improve recovery from acute illness. Data are curated within a NIST 800-53–aligned secure enclave architecture with role-based access controls, audit logging, and controlled-access governance review. CHoRUS supports AI model development, validation, benchmarking, and translational research while maintaining compliance with HIPAA and NIH data-sharing policies.\n",
   "keywords": [
    "Bridge2AI",
    "CHoRUS",
    "Electronic health records",
    "physiological data",
    "medical images"
   ],
   "version": "1.0 Beta",
   "datePublished": "2026-04-03",
   "isPartOf": [
    {
     "@id": "ark:59852/organization-bridge2ai-s6VouUf8Gkm"
    }
   ],
   "hasPart": [
    {
     "@id": "08cf7419-b94d-4508-8f64-c99c557351d7",
     "@type": [
      "https://w3id.org/EVI#Dataset",
      "https://w3id.org/EVI#ROCrate"
     ],
     "name": "CHoRUS RO-Crate EHR SubRoCrate"
    },
    {
     "@id": "b9b41c72-0895-4ec2-9e39-8de2a83abcd6",
     "@type": [
      "https://w3id.org/EVI#Dataset",
      "https://w3id.org/EVI#ROCrate"
     ],
     "name": "CHoRUS RO-Crate Waveforms SubRoCrate"
    }
   ],
   "author": "Eric S. Rosenthal(1), Rishikesan Kamaleswaran(2), Yulia Levites Strekalova(3), Andrew E. Williams(4), Ashley Cordes(5), Ishan C. Williams(6), Parisa Rashidi(3), Barbara J. Evans(3), J. Randall Moorman(6), Soojin Park(7), Edilberto Amorim(8), Paul M. Vespa(9), Xiaoqian Jiang(10), Hongfang Liu(10), Mihai V. Podgoreanu(2), Vitaly Herasevich(11), Mark Wainwright(12), Andrew J. Barros(6), Brynne A. Sullivan(6), Michael J. Young(1), Marta Alvarez(4), Polina Talapova(4), Jennifer A. Muszynski(13), Alasdair Gent(2), Delgersuren Bold(14), Manlik Kwong(4), Dileep Gunda(14), Jackie Le(1), India Loar(6), Brian J. Gow(15), Heidi Schmidt(1), Jared Houghtaling(4), Robert T. Miller(4), Timothy Clark (6), William B. Ashe(6), Tony C. Pan(14), Tom J. Pollard(15), Ciera McCrary(1), Azra Bihorac(3), Xiao Hu(14), Gilles Clermont(6) on behalf of the CHoRUS Investigators\n\n(1) Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA\n(2) Duke University, Durham, North Carolina, USA\n(3) University of Florida, Gainesville, Florida, USA\n(4) Tufts CTSI / Tufts Medical Center, Boston, Massachusetts, USA\n(5) University of Oregon, Eugene, Oregon, USA\n(6) University of Virginia, Charlottesville, Virginia, USA\n(7) Columbia University Irving Medical Center, New York, New York, USA\n(8) University of California, San Francisco, San Francisco, California, USA\n(9) David Geffen School of Medicine at UCLA, Los Angeles, California, USA\n(10) McWilliams School of Biomedical Informatics, UTHealth Houston, Houston, Texas, USA; Mayo Clinic, Rochester, Minnesota, USA\n(11) Mayo Clinic, Rochester, Minnesota, USA\n(12) Seattle Children’s Hospital, Seattle, Washington, USA\n(13) Nationwide Children’s Hospital / The Ohio State University, Columbus, Ohio, USA\n(14) Emory University, Atlanta, Georgia, USA\n(15) Massachusetts Institute of Technology, Cambridge, Massachusetts, USA",
   "publisher": "B2AI CHoRUS",
   "principalInvestigator": "Eric Rosenthal, EROSENTHAL@mgh.harvard.edu",
   "funder": "NIH Common Fund OT2OD032701",
   "contactEmail": "Ciera McCrary, cmccrary@mgh.harvard.edu",
   "citation": "The CHoRUS for Clinical Care AI Network. The Bridge2AI CHoRUS for Clinical Care AI Dataset: A Multi-Center, Multi-Modal, High-Resolution Critical Care Dataset, version 1.0 Beta. Harvard Dataverse, Apr. 2026.",
   "license": "Data Use Agreement available at 'https://chorus4ai.org/dataset/'",
   "conditionsOfAccess": "https://chorus4ai.org/wp-content/uploads/2025/10/Data-Agreement-9.30.2025.docx",
   "copyrightNotice": "© 2026 THE GENERAL HOSPITAL CORPORATION D/B/A MASSACHUSETTS GENERAL HOSPITAL. Copyrights are owned by each subaward site for curation performed locally. Copyrights are jointly held for software jointly developed. All rights reserved",
   "contentSize": "1.2 tb",
   "ethicalReview": "Eric S. Rosenthal, Michael J. Young, Ishan Williams, Ashley Cordes, and Barbara J. Evans in collaboration with the Mass General Brigham IRB with local context review by IRBs, HIPAA Privacy Officers, and institutional counsels at contributing medical centers.",
   "confidentialityLevel": "HL7:2V (very restricted)",
   "irb": {
    "@type": "Organization",
    "name": "Mass General Brigham Institutional Review Board (MGB IRB)",
    "contactPoint": {
     "@type": "ContactPoint",
     "contactType": "IRB Reliance and Compliance",
     "email": "irbreliance@mgb.org",
     "telephone": "+1-857-282-1900"
    },
    "address": {
     "@type": "PostalAddress",
     "streetAddress": "399 Revolution Drive, Suite 710",
     "addressLocality": "Somerville",
     "addressRegion": "MA",
     "postalCode": "02145",
     "addressCountry": "US"
    }
   },
   "irbProtocolId": "#2022P000707",
   "humanSubjectExemption": "HIPAA exemption 4 ((45 CFR 46.104(d)(4))",
   "fdaRegulated": true,
   "deidentified": true,
   "humanSubjectResearch": "Yes",
   "dataGovernanceCommittee": "Eric Rosenthal, EROSENTHAL@mgh.harvard.edu",
   "rai:dataLimitations": "•\tObservational data; and not randomized; collected as real-world clinical data, not through prospective research interactions with patient.\n•\tDocumentation quality dependent on routine clinical workflows and data elements being collected in clinical practice.\n•\tVariable sampling rates across real-world data from hospital waveform systems and middleware\n•\tInstitutional heterogeneity in coding and clinical data collection practices\n•\tLimited generalizability beyond participating hospitals\n•\tNot appropriate for re-identification, contact tracing, or patient-level intervention\n•\tHold-out splits are available for testing; training and validation in development splits require internal mechanisms to avoid overfitting\n",
   "rai:dataBiases": "•\tReferral bias from tertiary/quaternary academic centers\n•\tSocioeconomic and demographic distribution reflective of participating institutions\n•\tDocumentation bias in clinical notes\n•\tMissingness not at random (MNAR) in laboratory, waveform, and imaging acquisition\n•\tDifferential care pathways affecting label assignment (e.g., ICU vs non-ICU populations)\n•\tTemporal trends related to evolving clinical practice\nBias assessment is ongoing and documented in dataset release notes\n",
   "rai:dataReleaseMaintenancePlan": "\n•\tVersioned dataset releases e.g., CHoRUS vX.Y)\n•\tRelease notes documenting schema changes, harmonization updates, and known issues\n•\tGovernance committee oversight for structural modifications\n•\tData upload\n•\tDeprecation policy includes:\no\tAdvance notice of schema changes\no\tBackward compatibility window where feasible\no\tArchive of prior versions within enclave where possible or incremental change records\nPrimary maintainers: CHoRUS Data Pillar in collaboration with institutional data stewards.\n",
   "rai:dataCollection": "Data are derived from routine clinical care at participating hospitals under IRB oversight. Collection includes:\n•\tElectronic health record extraction and transformation to OMOP common data model\n•\tImaging metadata de-identification using RSNA Clinical Trial Processor and pixel-level de-identification using IbisWorks EICON. \n•\tHigh-resolution physiologic waveform capture extraction and transformation to the WFDB standard, \n•\tStandardized harmonization to common data model\nEthical safeguards include:\n•\tIRB approval or waiver as appropriate\n•\tCompliance with CHoRUS governance\n•\tLayered privacy protection including de-identification, data use agreement, controlled access data repository with adherence to NIST 800-53 principles, audit-level protection, registration, role-based access controls, protection against data exfiltration.\nData are not collected solely for research but repurposed from clinical workflows under ethical oversight.\n",
   "rai:dataCollectionMissingData": "Variable sampling rates across real-world data from hospital waveform systems and middleware",
   "rai:personalSensitiveInformation": [
    "CHoRUS operates within a secure enclave environment aligned with NIST 800-53 controls.",
    "De-identification consistent with HIPAA Safe Harbor",
    "Removal or tokenization of direct identifiers",
    "Date shifting or limited date fields when required",
    "Role-based access control (RBAC)",
    "Multi-factor authentication",
    "Full audit logging of data access and queries",
    "Prohibition of re-identification attempts",
    "Data export restrictions and output review procedures to reduce risk of re-identification",
    "Re-identification risk is periodically assessed and mitigation strategies updated as needed."
   ],
   "completeness": "Interim release with partial data. Not all patients in the CHoRUS full cohort are included. No DICOM images are included.",
   "contentUrl": "http://chorus4ai.org/dataset",
   "project_name": "CHoRUS",
   "rai:conditionsOfAccess": "Access to CHoRUS data is governed through a controlled-access framework and requires:\n•\tProviding institutional affiliation and credentials\n•\tApproved research proposal reviewed by CHoRUS governance committee\n•\tData Use Agreement (DUA) executed between requestor institution and CHoRUS\n•\tIRB approval or exemption documentation, as applicable\n•\tCompliance with NIH Bridge2AI OT terms and applicable federal regulations\n•\tAccess restricted to secure enclave environment; no raw data export permitted unless explicitly approved\nData use may not be used for commercial purposes unless specifically authorized through a separate agreement consistent with governing award and consortium terms.\n",
   "rai:potentialBiases": "•\tReferral bias from tertiary/quaternary academic centers\n•\tSocioeconomic and demographic distribution reflective of participating institutions\n•\tDocumentation bias in clinical notes\n•\tMissingness not at random (MNAR) in laboratory, waveform, and imaging acquisition\n•\tDifferential care pathways affecting label assignment (e.g., ICU vs non-ICU populations)\n•\tTemporal trends related to evolving clinical practice\nBias assessment is ongoing and documented in dataset release notes\n",
   "rai:intendedUseCases": "\n•\tAI model training within enclave\n•\tModel validation and benchmarking\n•\tAlgorithm development and feature engineering\n•\tRetrospective hypothesis testing\n•\tFine-tuning of multimodal models\n•\tComparative effectiveness research\n•\tReinforcement learning and development of expert AI\n•\tMethodological research in trustworthy AI\n•\tTrial emulation or powering\nNot recommended\n•\tDirect translation of trained algorithms to real-time clinical deployment without validation or regulatory clearance\n•\tUse as sole evidence for clinical decision-making\n•\tPopulation inference beyond participating institutions\nUsage must align with approved research scope and governance policies.\n",
   "rai:maintenancePlan": "\n•\tVersioned dataset releases e.g., CHoRUS vX.Y)\n•\tRelease notes documenting schema changes, harmonization updates, and known issues\n•\tGovernance committee oversight for structural modifications\n•\tData upload\n•\tDeprecation policy includes:\no\tAdvance notice of schema changes\no\tBackward compatibility window where feasible\no\tArchive of prior versions within enclave where possible or incremental change records\nPrimary maintainers: CHoRUS Data Pillar in collaboration with institutional data stewards.\n",
   "releaseDate": "03/04/2026",
   "ro-crate-metadata": "ro-crate-metadata.json"
  },
  {
   "@id": "b9b41c72-0895-4ec2-9e39-8de2a83abcd6",
   "@type": [
    "https://w3id.org/EVI#Dataset",
    "https://w3id.org/EVI#ROCrate"
   ],
   "name": "CHoRUS RO-Crate Waveforms SubRoCrate",
   "description": "Waveforms sub ro-crate for Bridge2AI CHoRUS RO-Crate Package",
   "keywords": [
    "Bridge2AI",
    "CHoRUS",
    "Electronic health records",
    "physiological data",
    "medical images"
   ],
   "version": "1.0 Beta",
   "datePublished": "2026-04-03",
   "isPartOf": [
    {
     "@id": "ark:59853/rocrate-chorus-ro-crate-package"
    }
   ],
   "hasPart": [],
   "author": "Eric S. Rosenthal(1), Rishikesan Kamaleswaran(2), Yulia Levites Strekalova(3), Andrew E. Williams(4), Ashley Cordes(5), Ishan C. Williams(6), Parisa Rashidi(3), Barbara J. Evans(3), J. Randall Moorman(6), Soojin Park(7), Edilberto Amorim(8), Paul M. Vespa(9), Xiaoqian Jiang(10), Hongfang Liu(10), Mihai V. Podgoreanu(2), Vitaly Herasevich(11), Mark Wainwright(12), Andrew J. Barros(6), Brynne A. Sullivan(6), Michael J. Young(1), Marta Alvarez(4), Polina Talapova(4), Jennifer A. Muszynski(13), Alasdair Gent(2), Delgersuren Bold(14), Manlik Kwong(4), Dileep Gunda(14), Jackie Le(1), India Loar(6), Brian J. Gow(15), Heidi Schmidt(1), Jared Houghtaling(4), Robert T. Miller(4), Timothy Clark (6), William B. Ashe(6), Tony C. Pan(14), Tom J. Pollard(15), Ciera McCrary(1), Azra Bihorac(3), Xiao Hu(14), Gilles Clermont(6) on behalf of the CHoRUS Investigators\n\n(1) Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA\n(2) Duke University, Durham, North Carolina, USA\n(3) University of Florida, Gainesville, Florida, USA\n(4) Tufts CTSI / Tufts Medical Center, Boston, Massachusetts, USA\n(5) University of Oregon, Eugene, Oregon, USA\n(6) University of Virginia, Charlottesville, Virginia, USA\n(7) Columbia University Irving Medical Center, New York, New York, USA\n(8) University of California, San Francisco, San Francisco, California, USA\n(9) David Geffen School of Medicine at UCLA, Los Angeles, California, USA\n(10) McWilliams School of Biomedical Informatics, UTHealth Houston, Houston, Texas, USA; Mayo Clinic, Rochester, Minnesota, USA\n(11) Mayo Clinic, Rochester, Minnesota, USA\n(12) Seattle Children’s Hospital, Seattle, Washington, USA\n(13) Nationwide Children’s Hospital / The Ohio State University, Columbus, Ohio, USA\n(14) Emory University, Atlanta, Georgia, USA\n(15) Massachusetts Institute of Technology, Cambridge, Massachusetts, USA",
   "principalInvestigator": "PI Eric Rosenthal  EROSENTHAL@mgh.harvard.edu",
   "funder": "NIH Common Fund OT2OD032701",
   "contactEmail": "Ciera McCrary, cmccrary@mgh.harvard.edu",
   "citation": "The CHoRUS for Clinical Care AI Network. The Bridge2AI CHoRUS for Clinical Care AI Dataset: A Multi-Center, Multi-Modal, High-Resolution Critical Care Dataset, version 1.0 Beta. Harvard Dataverse, Apr. 2026.",
   "license": "See Data Use Agreement",
   "conditionsOfAccess": "https://chorus4ai.org/wp-content/uploads/2025/10/Data-Agreement-9.30.2025.docx",
   "copyrightNotice": "© 2026 THE GENERAL HOSPITAL CORPORATION D/B/A MASSACHUSETTS GENERAL HOSPITAL. Copyrights are owned by each subaward site for curation performed locally. Copyrights are jointly held for software jointly developed. All rights reserved",
   "contentSize": "1.201567472832 tb",
   "contentUrl": "https://chorus4ai.org/dataset/",
   "ro-crate-metadata": "waveforms/ro-crate-metadata.json"
  },
  {
   "@id": "08cf7419-b94d-4508-8f64-c99c557351d7",
   "@type": [
    "https://w3id.org/EVI#Dataset",
    "https://w3id.org/EVI#ROCrate"
   ],
   "name": "CHoRUS RO-Crate EHR SubRoCrate",
   "description": "Electronic Health Records sub ro-crate for Bridge2AI CHoRUS RO-Crate Package",
   "keywords": [
    "Bridge2AI",
    "CHoRUS",
    "Electronic health records",
    "physiological data",
    "medical images"
   ],
   "version": "1.0 Beta",
   "datePublished": "03/04/2026",
   "isPartOf": [
    {
     "@id": "ark:59853/rocrate-chorus-ro-crate-package"
    }
   ],
   "hasPart": [],
   "author": "Eric S. Rosenthal(1), Rishikesan Kamaleswaran(2), Yulia Levites Strekalova(3), Andrew E. Williams(4), Ashley Cordes(5), Ishan C. Williams(6), Parisa Rashidi(3), Barbara J. Evans(3), J. Randall Moorman(6), Soojin Park(7), Edilberto Amorim(8), Paul M. Vespa(9), Xiaoqian Jiang(10), Hongfang Liu(10), Mihai V. Podgoreanu(2), Vitaly Herasevich(11), Mark Wainwright(12), Andrew J. Barros(6), Brynne A. Sullivan(6), Michael J. Young(1), Marta Alvarez(4), Polina Talapova(4), Jennifer A. Muszynski(13), Alasdair Gent(2), Delgersuren Bold(14), Manlik Kwong(4), Dileep Gunda(14), Jackie Le(1), India Loar(6), Brian J. Gow(15), Heidi Schmidt(1), Jared Houghtaling(4), Robert T. Miller(4), Timothy Clark (6), William B. Ashe(6), Tony C. Pan(14), Tom J. Pollard(15), Ciera McCrary(1), Azra Bihorac(3), Xiao Hu(14), Gilles Clermont(6) on behalf of the CHoRUS Investigators\n\n(1) Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA\n(2) Duke University, Durham, North Carolina, USA\n(3) University of Florida, Gainesville, Florida, USA\n(4) Tufts CTSI / Tufts Medical Center, Boston, Massachusetts, USA\n(5) University of Oregon, Eugene, Oregon, USA\n(6) University of Virginia, Charlottesville, Virginia, USA\n(7) Columbia University Irving Medical Center, New York, New York, USA\n(8) University of California, San Francisco, San Francisco, California, USA\n(9) David Geffen School of Medicine at UCLA, Los Angeles, California, USA\n(10) McWilliams School of Biomedical Informatics, UTHealth Houston, Houston, Texas, USA; Mayo Clinic, Rochester, Minnesota, USA\n(11) Mayo Clinic, Rochester, Minnesota, USA\n(12) Seattle Children’s Hospital, Seattle, Washington, USA\n(13) Nationwide Children’s Hospital / The Ohio State University, Columbus, Ohio, USA\n(14) Emory University, Atlanta, Georgia, USA\n(15) Massachusetts Institute of Technology, Cambridge, Massachusetts, USA",
   "principalInvestigator": "PI Eric Rosenthal  EROSENTHAL@mgh.harvard.edu",
   "funder": "NIH Common Fund OT2OD032701",
   "contactEmail": "Ciera McCrary, cmccrary@mgh.harvard.edu",
   "citation": "The CHoRUS for Clinical Care AI Network. The Bridge2AI CHoRUS for Clinical Care AI Dataset: A Multi-Center, Multi-Modal, High-Resolution Critical Care Dataset, version 1.0 Beta. Harvard Dataverse, Apr. 2026.",
   "license": "See Data Use Agreement",
   "conditionsOfAccess": "https://chorus4ai.org/wp-content/uploads/2025/10/Data-Agreement-9.30.2025.docx",
   "copyrightNotice": "© 2026 THE GENERAL HOSPITAL CORPORATION D/B/A MASSACHUSETTS GENERAL HOSPITAL. Copyrights are owned by each subaward site for curation performed locally. Copyrights are jointly held for software jointly developed. All rights reserved",
   "contentSize": "18.136671 mb",
   "contentUrl": "https://chorus4ai.org/dataset/",
   "ro-crate-metadata": "ehr/ro-crate-metadata.json"
  }
 ]
}

================================================================================
FILE: ai_ready_score.json
ROLE: AI-readiness self-assessment
SIZE: 7,654 characters
--------------------------------------------------------------------------------
{
  "name": "AI-Ready Score for CHoRUS RO-Crate Package",
  "fairness": {
    "findable": {
      "has_content": true,
      "details": "Dataset has DOI: https://doi.org/10.18130/V3/XNBOPG"
    },
    "accessible": {
      "has_content": true,
      "details": "The RO-Crate's JSON-LD metadata is machine-readable and publicly accessible by design."
    },
    "interoperable": {
      "has_content": true,
      "details": "The dataset uses the schema.org vocabulary within the RO-Crate framework and conforms to the Croissant RAI specification for interoperability."
    },
    "reusable": {
      "has_content": true,
      "details": "License: Data Use Agreement available at 'https://chorus4ai.org/dataset/'"
    }
  },
  "provenance": {
    "transparent": {
      "has_content": true,
      "details": "1468 dataset(s) documented"
    },
    "traceable": {
      "has_content": true,
      "details": "2 computation/experiment steps documented"
    },
    "interpretable": {
      "has_content": true,
      "details": "1 software instances documented"
    },
    "key_actors_identified": {
      "has_content": true,
      "details": "Author specified, Publisher: B2AI CHoRUS, PI: Eric Rosenthal, EROSENTHAL@mgh.harvard.edu"
    }
  },
  "characterization": {
    "semantics": {
      "has_content": true,
      "details": "Data is semantically described using the schema.org vocabulary within a machine-readable RO-Crate."
    },
    "statistics": {
      "has_content": false,
      "details": "No statistical characterization available. To add statistics, set 'contentSize' and/or 'hasSummaryStatistics' in Dataset/ROCrate entities"
    },
    "standards": {
      "has_content": true,
      "details": "44 schema(s) documented"
    },
    "potential_sources_of_bias": {
      "has_content": true,
      "details": "\u2022\tReferral bias from tertiary/quaternary academic centers\n\u2022\tSocioeconomic and demographic distribution reflective of participating institutions\n\u2022\tDocumentation bias in clinical notes\n\u2022\tMissingness not..."
    },
    "data_quality": {
      "has_content": true,
      "details": "Variable sampling rates across real-world data from hospital waveform systems and middleware"
    }
  },
  "pre_model_explainability": {
    "data_documentation_template": {
      "has_content": true,
      "details": "Documentation is provided via the RO-Crate's structured JSON-LD metadata, this HTML Datasheet, and Croissant RAI properties."
    },
    "fit_for_purpose": {
      "has_content": true,
      "details": "Limitations: \u2022\tObservational data; and not randomized; collected as real-world clinical data, not through prospective research interactions with patient.\n\u2022\tDocumentation quality dependent on routine clinical workflows and data elements being collected in clinical practice.\n\u2022\tVariable sampling rates across real-world data from hospital waveform systems and middleware\n\u2022\tInstitutional heterogeneity in coding and clinical data collection practices\n\u2022\tLimited generalizability beyond participating hospitals\n\u2022\tNot appropriate for re-identification, contact tracing, or patient-level intervention\n\u2022\tHold-out splits are available for testing; training and validation in development splits require internal mechanisms to avoid overfitting\n"
    },
    "verifiable": {
      "has_content": true,
      "details": "99% of files have checksums (1469/1477)"
    }
  },
  "ethics": {
    "ethically_acquired": {
      "has_content": true,
      "details": "Data collection: Data are derived from routine clinical care at participating hospitals under IRB oversight. Collection includes:\n\u2022\tElectronic health record extraction and transformation to OMOP common data model\n\u2022\tImaging metadata de-identification using RSNA Clinical Trial Processor and pixel-level de-identification using IbisWorks EICON. \n\u2022\tHigh-resolution physiologic waveform capture extraction and transformation to the WFDB standard, \n\u2022\tStandardized harmonization to common data model\nEthical safeguards include:\n\u2022\tIRB approval or waiver as appropriate\n\u2022\tCompliance with CHoRUS governance\n\u2022\tLayered privacy protection including de-identification, data use agreement, controlled access data repository with adherence to NIST 800-53 principles, audit-level protection, registration, role-based access controls, protection against data exfiltration.\nData are not collected solely for research but repurposed from clinical workflows under ethical oversight.\n, Human subject info: Yes"
    },
    "ethically_managed": {
      "has_content": true,
      "details": "Ethical review: Eric S. Rosenthal, Michael J. Young, Ishan Williams,\u00a0Ashley\u00a0Cordes, and Barbara J. Evans in collaboration with the Mass General Brigham IRB with\u00a0local context review by IRBs, HIPAA Privacy Officers, and institutional counsels at contributing medical centers., Governance: Eric Rosenthal, EROSENTHAL@mgh.harvard.edu"
    },
    "ethically_disseminated": {
      "has_content": true,
      "details": "License: Data Use Agreement available at 'https://chorus4ai.org/dataset/', Sensitive info: ['CHoRUS operates within a secure enclave environment aligned with NIST 800-53 controls.', 'De-identification consistent with HIPAA Safe Harbor', 'Removal or tokenization of direct identifiers', 'Date shifting or limited date fields when required', 'Role-based access control (RBAC)', 'Multi-factor authentication', 'Full audit logging of data access and queries', 'Prohibition of re-identification attempts', 'Data export restrictions and output review procedures to reduce risk of re-identification', 'Re-identification risk is periodically assessed and mitigation strategies updated as needed.']"
    },
    "secure": {
      "has_content": true,
      "details": "Confidentiality level: HL7:2V (very restricted)"
    }
  },
  "sustainability": {
    "persistent": {
      "has_content": true,
      "details": "Dataset has DOI: https://doi.org/10.18130/V3/XNBOPG"
    },
    "domain_appropriate": {
      "has_content": true,
      "details": "Maintenance plan: \n\u2022\tVersioned dataset releases e.g., CHoRUS vX.Y)\n\u2022\tRelease notes documenting schema changes, harmonization updates, and known issues\n\u2022\tGovernance committee oversight for structural modifications\n\u2022\tData upload\n\u2022\tDeprecation policy includes:\no\tAdvance notice of schema changes\no\tBackward compatibility window where feasible\no\tArchive of prior versions within enclave where possible or incremental change records\nPrimary maintainers: CHoRUS Data Pillar in collaboration with institutional data stewards.\n"
    },
    "well_governed": {
      "has_content": true,
      "details": "Governance committee: Eric Rosenthal, EROSENTHAL@mgh.harvard.edu"
    },
    "associated": {
      "has_content": true,
      "details": "All data, software, and computations are explicitly linked within the RO-Crate's provenance graph."
    }
  },
  "computability": {
    "standardized": {
      "has_content": true,
      "details": "Formats: .ipynb, text/tab-separated-values, wfdb"
    },
    "computationally_accessible": {
      "has_content": true,
      "details": "Publisher: B2AI CHoRUS"
    },
    "portable": {
      "has_content": true,
      "details": "The dataset is packaged as a self-contained RO-Crate, a standard designed for portability across systems."
    },
    "contextualized": {
      "has_content": true,
      "details": "Context is provided by the RO-Crate's graph structure and detailed in properties such as rai:dataLimitations."
    }
  }
}
