Data & Technology Infrastructure
Learning enabled by data & technology
We built infrastructure within health system partners to make data usable to researchers and provide the technology needed to improve care through practice changes implementations. Our team primarily supports LHS researchers and CLHSS partners. But our data technology partnerships but our work also broadly benefits researchers who want to work with Fairview and our partners.
Our projects use FAIR principles for technology and data management which seek to improve the Findability, Accessibility, Interoperability, and Reuse of digital assets.
LHS Data Platform
Our Learning Health System (LHS) data platform is a collaborative framework that stores and processes real world data from the electronic health record to be used in LHS research and healthcare operations.
Key aspects of the data platform include:
- Data infrastructure compliant with data security regulations and research ethics that is based on the Academic Health Center–Information Exchange Infrastructure
- Datamarts (subsets of data) that make surgical and acute care data accessible and easier to utilize
- New data sources such as electronic health record (EHR) notes, logfiles, and other metadata
- Data management that makes data more standardized and reusable, including mapping data to the Observational Medical Outcomes Partnership (OMOP) common data model
- A Fast Healthcare Interoperability Resources (FHIR) façade that translates data stored in different formats into a standardized form
The data platform is developed with support from the Clinical and Translational Science Institute, the Center for Clinical Quality & Outcomes Discovery and Evaluation (CQODE), the Department of Surgery, investigators from the Academic Investment Clinical Program (AICP), and Fairview Health Services.
Federated Machine Learning Collaborative
Federated Machine Learning is an approach to artificial intelligence (AI)/machine learning (ML) development where copies of a model are trained in independent sessions using different sets of data. The trained copies of the model are uploaded to a server and integrated into one, centralized model. This allows models to be trained across organizations without sharing sensitive health data.
Together with M Health Fairview and other partners, CLHSS leads a Federated Computer Vision in Healthcare Collaborative with other partner institutions. In addition to computer vision, CLHSS applies federated learning to other AI/ML uses like Natural Language Processing (NLP).
Interoperable Clinical Decision Support
Clinical Decision Support (CDS) tools give patients and healthcare workers the right information at the right time to improve health outcomes.
CLHSS is collaborating with other institutions and an industry partner to develop CDS solutions that can be used with different data formats through the SCALED Venous Thomboembolism Prevention project.
Customer-Centric Design of Digital Solutions
In addition to our own Digital Technology Innovation (DTI) program, we collaborate with M Health Fairview to build innovative digital solutions that engage patients and care teams in their health.