ScHARe Platform Components

 

Datasets

STATUS: The ScHARe Datasets collection is accessible to all ScHARe-registered researchers. New datasets are being actively added.

ScHARe Datasets list

On ScHARe, researchers can access, link, analyze, and export a wealth of datasets relevant to research in health disparities and health care outcomes, including:

• Google-hosted Public Datasets: publicly accessible, federated, de-identified datasets hosted by Google through the Google Cloud Public Dataset Program
  • Example: American Community Survey (ACS)
• ScHARe-hosted Public Datasets: publicly accessible, de-identified datasets hosted by ScHARe
  • Example: Behavioral Risk Factor Surveillance System (BRFSS)
• ScHARe-hosted Project Datasets: publicly accessible and controlled-access, funded program/project datasets shared by NIH grantees and intramural investigators to comply with the NIH Data Sharing Policy
  • Example: Jackson Heart Study (JHS)

Datasets are grouped by these categories:

• Social Determinants of Health (Economic Stability, Education Access and Quality, Health Care Access and Quality, Neighborhood and Built Environment, Social and Community Context) and Health Behaviors
• Diseases and Conditions
• Clinical and EHR Data

Register for ScHARe and access the ScHARe Datasets collection in our Terra Workspace.

 

Data Repository

STATUS: The ScHARe Data Repository is currently under development.

The National Institute on Minority Health and Health Disparities (NIMHD) and the National Institute of Nursing Research (NINR) are developing the ScHARe Data Repository. When fully operational, the repository will allow for the required hosting, managing, and sharing of data from NIMHD- and NINR-funded research programs, including data on health disparities and health care outcomes. NIH-based researchers, external researchers, and health care organizations will be able to access data within the repository.

In parallel, NIMHD and NINR are developing core common data elements (CCDEs). CCDEs are standardized questions and responses that can be used across different studies to ensure consistent data collection and facilitate interoperability. These CCDEs will enable researchers to link data from different sources.

By enabling the uploading, collection, and sharing of data, and the linking of information across different sources, the ScHARe Data Repository will facilitate research into the mechanisms that foster health disparities and influence health care outcomes.

 

Collaborative Workspaces

STATUS: The ScHARe Collaborative Workspaces are available to all ScHARe-registered researchers.

ScHARe is powered by Terra , an open-source data analysis platform based on Google Cloud Platform. Terra was developed by the Broad Institute of MIT and Harvard in collaboration with Microsoft and Verily.

Using ScHARe’s Terra resources, researchers and their collaborators can access and cross-link the same publicly available or controlled-access data. They can also create secure online spaces for collaboratively running large-scale analyses and sharing reproducible results and resources.

ScHARe supports interactive analysis tools such as Jupyter notebooks. Jupyter notebooks are human-readable executable documents that can be run to perform advanced data analyses, including artificial intelligence and machine learning tasks, using coding languages such as Python and R. The platform also supports Dockstore as a repository for Docker-based analysis workflows that allow users to automate basic steps in their analyses.

• Register for ScHARe
• Explore the ScHARe Terra Workspace , or create your own workspace using the tutorials accessible from our Tutorials and Resources page.

 

Bias Mitigation Tools and Resources

STATUS: The ScHARe Bias Mitigation Tools are currently under development.

In computer science, algorithms are finite sequences of instructions used to solve specific problems or to perform automated reasoning and decision-making. Algorithms are widely used in healthcare- and policy-related decisions. However, many algorithms operate as “black boxes,” offering little opportunity for adequate testing to identify potential biases that can create unfair outcomes, such as privileging one category over another.

Biases can result from missing data and quality problems. Pre-existing social or cultural expectations and design limitations can also introduce biases into algorithms. If not identified, biased algorithms can result in healthcare decisions that lead to discrimination, unequitable healthcare, or unintentional arm to marginalized populations.

ScHARe will provide bioinformatics tools, best-practice workflows, and resources for collaboratively evaluating and mitigating biases associated with datasets and algorithms used to inform healthcare and policy decisions. Collaboratives will strategize on developing better tools to mitigate biases at the design, data, algorithm, model training, and implementation phases.

Page updated Feb. 9, 2024