Johns Hopkins Hospitals – HERO
Johns Hopkins Hospitals wanted to build a brand new custom solution for their event reporting software. They had used a vendor in the past, but were determined to have their own in-house solution that had more features including new AI functionality to identify events. Each event cataloged needed to be sorted, identified for commonalities and […]
Client:
Nebulr
Date:
05/11/2020
Website:
https://www.hopkinsmedicine.org/institutional-review-board/guidelines-policies/guidelines/use-of-patient-safety-hero-and-sca-data
Category:
Artificial Intelligence
Technology
Johns Hopkins Hospitals wanted to build a brand new custom solution for their event reporting software. They had used a vendor in the past, but were determined to have their own in-house solution that had more features including new AI functionality to identify events. Each event cataloged needed to be sorted, identified for commonalities and sent to the appropriate administrators. Having to do this work was a pain, so including a new AI solution made sense. After two years of delays from the previous contractors we came in and delivered the solution.
- Backend API via java spring boot specifically with JHipster for Johns Hopkins Hospitals. Leveraged JPA and Hibernate for entity management with the database. Leveraged Azure Cognitive Search and custom-built AI Large Language Model (LLM) to “tag” new events with matching indicators (common tags) based on the description of the event submitted. Johns Hopkins needs to have this software as part of their federal compliance with event reporting in their 1822 hospitals and affiliated locations.
- Complete custom SSO solution built into backend
- Includes a graphql federated microservices architecture to send either data from azure search or a collection of MSSQL databases to a react web interface.
