Tackling the Integration Challenge: A New Breed of Analytics in Healthcare

August 10, 2016 Jerry Hart, Product Director

Previously we outlined what tomorrow’s healthcare analytics platforms will need to resolve the challenges that healthcare organizations are facing. The first of these is a way to resolve the challenge presented by a lack of interoperability.

So how can healthcare organizations tackle the integration challenge?

The foundation of any meaningful healthcare analytics implementation is a repository of high-quality data. Despite significant efforts from standards organizations like HL7®, initiatives like FHIR®, and regulatory incentives for vendors and healthcare organizations to improve system interoperability and data sharing, inconsistencies and quality issues plague the information exchange interfaces of healthcare systems. 

As a result of these challenges, a majority of analytics initiatives fail to move past the task of implementing and populating the underlying data repository. Due to the large number of source systems that must typically be integrated to adequately profile and report on outcomes within a population, analytics platforms must be extremely cost effective at building high-quality integrations. This requires tools that: 

  • Natively understand the interoperability message formats of source systems
  • Support real-time message feeds and the specific challenges that event-based integration presents
  • Analyze message feeds to identify variants from message standards and other data quality issues
  • Provide tools to support efficient iterations of message loading and output analysis, along with robust monitoring and error-handling infrastructure

Orion Health Amadeus combines the depth of integration experience built into the Rhapsody integration engine for over 15 years and couples this with its next generation data engine. The raw data store component provides a highly functional implementation of the data lake enterprise integration pattern to store data of all shapes and sizes in a single repository. The data processing pipeline allows acquired data to be processed and mapped to evolving data models over time, avoiding costly re-acquisition efforts as usage requirements evolve. To effectively handle real-time data streams, it provides a sophisticated event-sourcing infrastructure that handles the common issue of message re-ordering and variable data update modes that message feeds inevitably present.

From a data modelling perspective, analytics platforms need a rich set of prebuilt, standards-based models that cover the core clinical, claims, and device-data domains. The HL7 FHIR resource set has rapidly evolved to provide a well-considered data model set to support the needs of healthcare integration, and it’s the obvious model for healthcare analytics platforms to adhere to.

However, the future requires these platforms to provide the flexibility to extend core models, create entirely new models to support evolving data acquisition needs, and—as social, behavioral, environmental, and genomic information becomes increasingly important in the effort to analyze and predict health outcomes—support the collection and utilization of a much broader range of data points.  Amadeus Data Spaces couple a rich library of HL7 FHIR-based data models that cover core clinical-, claims- and device-data domains with the ability to design and deploy custom data spaces through an intuitive user interface. This ensures that healthcare organizations can ingest and model all of the data required to advance their analytics requirements.


Download a white paper on Why Healthcare Analytics Needs a Digital Transformation.


Subscribe to our Blog!

First Name
Last Name
Thank you!
Error - something went wrong!