Case study

Platform to ingest patient data from external EHR systems

The regulatory-grade healthcare data generated by healthcare providers is rich in information and can be used for research purposes. With patient and provider consent, research organizations can securely access this data. Research organizations need significant amounts of healthcare data to run and derive dependable outcomes on their analytic models. As the data comes from multiple Electronic Health Record (EHR) systems, it might be in different formats. In this scenario, data integration becomes a bottleneck to research organizations and a major concern for the healthcare industry.

Our client is a US-based health intelligence company that applies advanced analytics on patient medical records to deliver personalized health insights. The company engages patients as partners in their research and develops digital tools based on the analytics’ outcomes. To improve research outcomes, they needed an interoperable system that can quickly onboard partnering practices and get patient health records from them.

Requirements

Build an interoperable platform to:

Get digitized health data with patient consent from multiple EHR systems
Extract clinical information, de-dupe, transform and store in custom schemas
Generate Continuity of Care Document (CCD) and share it with the patient
Create source data to perform analytics that can also be accessed through an API

Solution

Imaginea proposed to build a microservices-based architecture that is capable of supporting multiple applications and can quickly onboard new EHR systems. The solution begins with pluggable data sources based on HL7 FHIR to get healthcare data from multiple EHR systems.

Tech stack

How Our Solution Helped

We built a highly secure and scalable system capable of handling millions of patient records coming from disparate systems

Overall approach

The healthcare data from disparate systems is stored in a common unified relational DB schema to support any type of source specifications. The data from disparate sources is then denormalized, de-duped, and transformed before generating the CCD. The data is also transformed into OMOP CDM (Observational Medical Outcomes Partnership Common Data Model). This is utilized for analytics and other applications through a secured API.

This process is repeated continuously to stream the changes in the source EHR databases to the target database and to purge data of patients who have altered their consent to share their EHR data.

Google Cloud Platform and Common Unified Relational DB were used to support varying source specifications. The data is encrypted at storage, and rest. The security audit is approved by Rapid7 and West Monroe Partners.

Results

Pluggable data sources help in onboarding new participating providers quickly. This results in a rapid increase of source data for analytics.
Robust and reusable data mapping framework that helps map data from custom schema EHR vendors quickly.
Microservices architecture enables the reusability of components and usability of multiple technology stacks in the future.