Composable Architecture and the Semantic Interoperability Union

By: Will Tesch

Composable Architecture explained

Last fall, Gartner hit the nail squarely in defining the future model of data-sharing across the enterprise:

“At a time of massive global disruption, the demands placed on application leaders are growing more complex and changing at a dizzying rate. Success and innovation require that the capabilities delivered by applications are modular, rapidly and safely assembled, disassembled and recomposed as business, customer and market needs shift.”1

Gartner’s concept of composable architecture aligns with the future we are enabling at HealthLX, both through our work with clients and our participation in HL7 FHIR Accelerator Programs like the Da Vinci Project and the Gravity Project. In the simplest terms, this is interoperability. It’s meaningful because data and transmission mechanisms are agreed upon across healthcare constituents.

Where this all goes is to a future where meaningful, real-time interoperability capabilities are the enablers of the real-time health system (RTHS). It is our mission at HealthLX to build a modular platform architecture to support capabilities that enable applications and interchange based on these standardized design patterns. Gartner refers to ‘Packaged Business Components (PBC)s’ as a piece of the overall composable architecture approach. Examples of these PBCs as applied to the HealthLX platform are:

• • • Ingestion design patterns for conversion to FHIR resources

    • Creation of prior authorization requests from traditional ADT messages

    • Workflow in and out of EMRs for CDA documents exchange

    • Patient Access API (PAAPI) to accommodate the CMS Final Rule

Today, these capabilities exist as packaged business components configurable for implementation with payers. These PBCs enable standardized workflow patterns configurable to various payer organization scenarios.

Semantic Clarity

Like two people speaking the same language having a discussion that is engaging and healthy, data sharing at its core implies the data requested is understood when it is received. The efficiency of this interchange, whether verbal or in data form, determines the semantic clarity of the interchange. If the words shared are understood as they were meant to be communicated, then both parties are effective in their understanding of each other. In the world of data interchange, the harmonization of meaning (aka ‘semantic clarity’) happens in a similar fashion. Using the FHIR standard as a point of reference the following characteristics apply in a data conversation:

• • • As the sender of data, were the words selected for my message correctly assembled to be understood accurately?

[Standard data format = FHIR Resources]

• • • As the listener to the data being sent, were the words correctly received from the message and with the meaning intended?

[Standard data format = FHIR Resources]

• • • For conversation to be interacted short-term memory is needed. Can the messages and discussion be saved to memory should either party need to refer to them in the immediate future for point of reference?

[Logging of interchange = message provenance]

• • • Applying the data interchange, are the sentences (data exchange) of value to the receiving party?

[data exchanged = value of the information, and can it be used to interoperate]

While there are other facets of communication that could be considered in this analogy, these cover the basics. As the analogy applies in the human to human discussion, the same applies to messages exchanged between two systems. If the messages are properly understood by the receiving system then downstream workflow gets realized and becomes actionable and valuable.

It is critical to continue forward on a path to viable and comprehensive data sharing across all healthcare stakeholders. Patient access to data mandated by the CMS Interoperability and Patient Access final rule is proving to be a significant hurdle for many healthcare organizations. While patient access to data is an important advance in industry-wide interoperability, it is only the beginning. Applying these configurable data flow design patterns and their nuanced use cases, our composable architecture approach ensures future steps in interoperability can not only be achieved, but will be easier to achieve with less cost and time invested.

Connecting the highways of healthcare data

The mission of HealthLX was to build a flexible data integration platform that we describe as “connecting the highways of healthcare data” to radically improve interoperability across the healthcare ecosystem. Our platform unifies multiple “packaged business capabilities” (PBCs in Gartner’s terms) to create workflows which enable data flow orchestration between enterprise applications. With our platform framework, every ‘PBC’ component inherits features from previously designed data flow patterns , enabling inherited data flow attributes which in turn enables configurable workflow. Our platform keeps the data flowing down the data highway while also capturing the state of the data on its journey, crucial for auditability and data provenance. This composable architecture approach to enterprise integration design enables us to use legacy assets where applicable, componentizing the interactions with those systems for reusability. The HealthLX platform provides a composable framework for organizations. It enables a new foundation for data-driven orchestration within and across healthcare organizations.

A viable strategy for meeting the upcoming deadlines

Gartner notes,

“Eventually, composable architecture approaches will change how we buy, build and deploy enterprise healthcare IT capabilities, much in the same way the cloud changed the way we buy, build and deploy IT systems and services now.”2

With the upcoming CMS and ONC Cures Act deadlines, the transition to a composable architecture approach will occur sooner rather than later, for several reasons:

• • • FHIR itself will broadly define the semantic layer for all healthcare data better because these data definitions become agreed upon for both the provider and payer’s traditional communication needs.

    • With FHIR as a semantic enabler, reusability of data definitions more closely aligns the underlying principles of Service-Oriented Architecture and thus can more quickly scale properly defined composable architecture concepts.

    • The realistic window for adding features or another layer to existing platforms is rapidly closing. It is also costly, and the result is often sub-par.

    • HealthLX’s FHIR Enterprise is already operational. Our clients are gaining interoperability of applications using prebuilt design patterns built with ‘PBCs’. By configuring the prebuilt workflows, we reduce the time and costs to implement new features.

    • At some point, data quality will need to be addressed. Our composable architecture makes it easy to integrate with data quality APIs provided by leading data quality vendors.

    • Composable architecture enables the flexibility for every healthcare organization to achieve interoperability in a way that delivers the greatest ROI for its business, whether that requires APIs developed internally, externally, or both.

The idea of modularity used to enable workflows is one we have embraced at HealthLX since our beginning. Composable architecture gives healthcare a more agile way to respond to regulatory requirements and create their best possible patient/member experience. We’ve seen it work. Our clients are innovating faster and can meet regulatory requirements with less cost. If you’d like to get on this road with us, contact us today. We believe interoperability is within reach.

1. Use Gartner’s Reference Model to Deliver Intelligent Composable Business Applications, 14 October 2020.
2. ibid.