Immunization Repository for a Large Healthcare Delivery System

Epic Integration

The Client 

Our client is one of the largest not-for-profit hospital and physician networks in the United States, with over 15,000 providers and over 2 million patients under care. 

Background 

Our client is engaged in the development of an enterprise-wide registry of immunization records for all patients under care within 11 hospitals and affiliated outpatient clinics in order to comply with government guidelines and improve continuity of care. Additionally, the organization is in the middle of a multi-year implementation and rollout process of EPIC EMR system across the entire delivery system, and is intended to import real-time and legacy immunization data into EPIC. 

The Challenge 

Hospitals and outpatient clinics in the clients’ network are running on multiple vendor and in-house EMR and ambulatory systems which are being used to store and manage immunization data. In total, the entire network has accumulated more than 40 million legacy immunization records over a number of years. The data is “dirty” – it has significant structural and content variances even within the same system. 

It is critical to preserve the historical sequence of data across all systems. Due to historical inconsistencies in data representation and ongoing edits of current and past records by clinicians, importing legacy and current immunization records concurrently presents a significant challenge. The organization decided to impose a temporary ban on edits of the current records until the legacy data import is complete. This decision imposes strict constrains on the timeline for importing extensive amount of legacy data. 

The Solution 

A joint team of First Line and client engineers, medical informaticians, analysts and hospital IT staff has worked together over a period of two years to design a sophisticated process and infrastructure to populate the immunization repository with historical and real-time data, and export this data to EPIC and government agencies. 

We have developed 13 interfaces from the client systems and outbound interfaces to EPIC and state vaccination registries. The interfaces used a variety of protocols including HL7, Web Services, flat file ingestion and proprietary APIs. We have also implemented complex data transformations that accounted for specifics of individual systems and variances in structure and content of data. 

In order to provide high throughput for legacy data import our engineers designed high performance parallel integration infrastructure with a routing logic to guarantee preservation of the historical sequence of records. 

Business User Interface 

Our team has developed a User Interface layer on top of the integration infrastructure specifically designed for clinical analysts to perform data validation, monitoring, and debugging of the workflow. 

The Result 

Using our troubleshooting and monitoring tools the analytical team on the project has been able to identify patterns and problems in the data during dry run imports of the data in the test environment. Based on this analysis we refined and improved data transformation scripts to achieve the desired level of fault tolerance and scalability in the interfaces. 

The integration pipeline designed and developed by the team has imported and processed 40 million immunization records into the immunization repository at a rate of 200 records per second with no loss of data, interface breakdowns, or reaching server capacity limits. 

The outbound interfaces have imported all records to EPIC and sent them to the state registry. The pipeline is now continuously processing real time records from all organizations participating in the immunization repository. 

Technologies 

Intersystems HealthConnect, CSP Web Services, Cache, Zen, DHTMLX Libraries

Want to know more details about this case study?

Get in touch

Related work

Interested in talking?

Whether you have a problem that needs solving or a great idea you’d like to explore, our team is always on hand to help you.