75N992-25-SS-102 Need Summary.pdf

Sources Sought Notice – 75N992-25-SS-102 Advanced Research Projects Agency for Health (ARPA-H) Real World Data – Data Sources, Platforms, and Services

Needs Summary Document

This document is provided to summarize the Government’s needs with respect to real world healthcare data Sources, platforms and related services. These are draft requirements issued for market research purposes, and this document does not constitute the Government’s final requirements with respect to these need areas.

1. Data Sourcing: The contractor shall provide access to Real-World Data collections on a national and international scale with specific attention to underrepresented populations or those underserved by the current healthcare system. Data should be linkable across sources and deidentified to ensure the privacy of individuals. Where possible, data should comply with or be convertible to common data standards such as OMOP, CDSIC, ICD10, etc., and accessible via Application Programming Interface (APIs). Data should have flexible data use agreements to meet the needs of ARPA-H program performers in developing new technologies and internal program impact tracking.

a. Data sources:

i. United States data requirements:

The vendor shall provide access to the following types of data:

Data Description Population Refresh Rate Longitudinal (>5 years) Electronic Medical Record data

>100 million persons

Minimum of 90-day refresh

Longitudinal (>5 years) medical claims data

>100 million persons

Minimum of 90-day refresh

Longitudinal (>5 years) pharmacy claims data

>100 million persons

Minimum of 90-day refresh

Longitudinal (>5 years) laboratory and diagnostic data (genetic testing, diagnostic imaging, etc.)

>50 million persons

90-day refresh

Demographic data at the ZIP3 >100 million persons

N/A

Socio-economic data at the

ZIP3

>100 million persons

Minimum of 180-day refresh

Educational access at the ZIP3 >100 million persons

Minimum of 180-day refresh

Nutritional access at the ZIP3 >100 million persons

Of 180-day refresh

Pollution and/or environmental exposure data at the ZIP3 level

Entire United States

N/A

Data regarding a person’s access to healthcare ZIP3 level

Entire United States

N/A

Longitudinal weather and climate data at the ZIP3

Entire United States

Clinical trial participation data and results

>100 trials

Observational study data >1,000 participants per study

ii. International data requirements (Europe, Asia, South Asia, Central America, South America, Africa):

The vendor shall provide access to the following types of data:

Data Description Population Refresh Rate Longitudinal (>5 years) Electronic Medical Record data

N/A Minimum of 90-day refresh

Longitudinal (>5 years) medical claims data

N/A Minimum of 90-day refresh

Longitudinal (>5 years) pharmacy claims

N/A Minimum of 90-day refresh

Longitudinal (>5 years) laboratory and diagnostic data (genetic testing, diagnostic imaging, etc).

N/A Minimum of 90-day refresh

Demographic data N/A N/A Pollution and/or environmental exposure data

N/A N/A

Data regarding a person’s access to healthcare

N/A N/A

Longitudinal weather and climate data

N/A N/A

Observational study data >1,000 participants per study

N/A

b. Privacy preserving technology: The vendor shall be capable of implementing privacy preserving technology with the following attributes and capabilities:

i. The ability to deidentify individuals and datasets including geolocation information, metadata, and images

ii. Privacy-preserving record linking (PPRL) to connect individuals’ data across the data sources listed in section 1(a)

iii. The ability to bridge between multiple PPRL technologies to combine data sources from multiple vendors or brokers.

iv. Provide aggregated or summarized dataset information or dashboards (e.g. descriptive statistics about the data in section 1(a))

c. Data use and access: The Government has the following general requirements with respect to access and data use for the data provided under section 1(a)

i. The contractor shall create templated and reproducible data use agreements (DUA) for program execution and tracking (one DUA per program).

ii. Standardize data use agreements across programs to simplify ARPA-H tracking and management of DUAs.

iii. Provide common or standard APIs to connect data sources to analytic environments (FHIR, Smart on FHIR, GA4GH DRS, etc.)

iv. Ability to subset or assemble a cohort of data across data sources

v. Provide access to datasets within 30 days from contract award.

2. Data Analysis Platform: The contractor shall provide a cloud-based platform for the hosting and analysis of ARPA-H’s Real World Data requirements. ARPA-H will leverage this platform to perform data analysis for healthcare data, social determinants of health, environmental data, and research data.

a. ARPA-H has a core need for platforms capable of processing and analyzing the types of data identified in need area 1. There are two core areas to this need:

i. Specialized Platforms: The Government has a need for access to three types of specialized platforms:

1. Biostatistical: a platform that can support common, reproducible, statistical software packages (e.g. SAS, STATA, SPSS, R, etc.) used by clinical and biostatistical researchers for clinical trials, longitudinal and observational studies, and data visualization

2. Genomic: a platform that supports scalable and reproducible analysis and management of common –omics data (DNA sequencing (single-cell or bulk), RNA sequencing (single-cell or bulk), long-read sequencing, LC-MS/MS or other proteomic assays, metabolomics, DNA methylation analysis, DNA structural analysis (Hi-C, ATAC, etc.), etc.) and metadata

3. Imaging: a platform that supports biomedical image data management and analysis including radiographic (x-ray, CT, PET, MRI, etc.) and light-based (wide-field, fluorescence, confocal, NIR, etc.) with viewing and annotation capabilities plus scalable batch processing of DICOMs or other image stacks with metadata

ii. Notebook Platform: a cloud environment supporting writing, executing, and managing common notebook scripts such as JupyterLabs, R Studio, etc. Environments should support a variety of languages such as bash, Python, R, Julia, PERL, Java, SQL, etc. with the ability to install custom packages and call APIs

b. The provided platform(s) should have the following capabilities:

i. ARPA-H anticipates a need for 20 – 30 enclaves (short-term) and upwards of 200 program enclaves (long-term).

1. Program specific enclaves consisting of:

a. A program specific website or front door including descriptions of the program, the data available, the tools available, and how to access the resources

b. Storage and access to program specific data

c. Storage and access to program specific analytical tools

d. Flexibility for length of performance periods ranging from 6 months to 5 years

2. Ingestion (extraction, transformation, and loading) of program generated data to make discoverable to other programs and users

3. Ingestion of data resources (area 1) to make discoverable to users in a program specific enclave

ii. Interactive analysis applications such as R Shiny, DASH, etc.

iii. Low-code reproducible workflows based on program-specific needs

(e.g. CWL, WDL, SnakeMake, NextFlow, etc.)

iv. A sandbox or workspace for individuals to manage their data and analyses (as well as the ability to limit and share access to these workspaces)

v. The ability to download data from a workspace

vi. The ability to prohibit downloading of data from a workspace

vii. Native support for common biomedical data APIs (e.g. FHIR, GA4GH APIs (DRS, Passports, WES, etc.), etc.)

c. Additional Requirements:

i. FEDRAMP Certified Commercial Platform

ii. Section 508 Compliant

iii. Low or No Code solutions to enable users to develop their own tools

iv. Support for Artificial Intelligence and Machine Learning (AI/ML) workflows

v. Flexible and scalable compute and storage access such as AWS Elastic

Compute, Google Compute Engine, Azure Virtual Machines, etc.

including CPU, GPU, high-memory, etc. instances

vi. Ability to integrate with Global Alliance for Genomics and Health (GA4GH) APIs such as Passports, Data Repository Service, and Workflow Execution Service

vii. Ability to integrate with cloud streaming services such as Amazon Kinesis, Google Data Flow, Azure Stream Analytics

viii. Ability to integrate with data warehouse services such as AWS Redshift, Google BigQuery, Snowflake, Databricks, etc

ix. Ability to leverage distributed computing or clusters such as Apache SPARK, Kubernetes, etc.

3. Data Management and Curation Services

a. This task area is designed to provide the ability for a program manager to quickly acquire the data management and data curation services needed by their program and specific to the data types and data sets used by their programs. This includes real world and other data from data providers under this contract but may also include unique data being generated by the Performers on their program. Under this task, the contractor may be required to

i. Support data generators/stewards in preparing data for ingest into ARPA-H Platforms, in a consultant type role:

1. Support ingestion of varied data types, including, but not necessarily limited to, data from electronic health records, research clinic surveys and assessments, genotypes, molecular and laboratory assays, and imaging procedures.

2. Identify deidentification needs and recommend best practices

3. Alignment to data management, model, and practice standards

4. Examine quality of data and make recommendations to improve quality (such as additional sourcing, harmonization etc)

5. Support integration of data from multiple sources for the same data type

6. Make recommendations for combining acquired data with program generated data including metadata requirements, standardized/reproducible pipelines to combine data sources, and creating SOPs for future data integration

4. Program Management and Data Integration: the contractor(s) will provide program management and reporting to ARPA-H including regular updates on milestones, costs, and potential schedule changes. The program management contact for the contractor will be available for meetings with the ARPA-H team on a regular basis and support working groups as necessary.