by Angela McArthur and Joydeep Mukherjee, Contributing writers

How CareSource ditched its data silos

Jun 17, 2021
AnalyticsBusiness IntelligenceData Architecture

Taking a 'data fabric' approach in its shift from data silos to real-time access, the non-profit managed healthcare plan provider futureproofed its data platform, providing increased agility, interoperability, and accessibility for its members, network of doctors, and healthcare providers.

abstract data flows / data streams
Credit: Gonin / Getty Images

As companies re-evaluate current IT infrastructures and processes with the goal of creating more efficient, resilient, and intuitive enterprise systems, one thing has become very clear: traditional data warehousing architectures that separate data storage from usage are pretty much obsolete.  

The basic structure of current data platforms inhibits strategic outcomes by creating data silos and inconsistencies in dashboards and reports across an organization.  As a result, the quality of the data is often questionable since it is an amalgam of information from multiple sources. The design disadvantages and limitations of these platforms include:

  • Redundant layers in the architecture and multiple data marts that result in increased processing time and data latency.
  • Complexity in the data flow and the existing processing layers that make it cost-prohibitive to optimize or scale the performance of the hardware and software that comprise these systems.
  • The need for multiple data marts and independent data repositories due to the absence of a single user-based consumption layer.
  • An inability to support enabling data science technologies such as predictive modeling, AI, and machine learning that are future drivers of digital transformation.
  • Antiquated identity, security, and audit controls that escalate risks to the enterprise.

Since data is an anchor point for the digital transformation efforts at every company, it makes sense to create a modern data platform that can support real-time processing and enabling technologies like AI, while offering a future-proof architecture that can deliver actionable business intelligence to achieve an organization’s goals.

This is exactly what we did at CareSource, one of the nation’s largest Medicaid managed care plans, serving more than two million members across five states. Our goals included enabling seamless access to medical records; easily sharing medical information with members and health care providers throughout the network; and supporting new service offerings such as remote care.

We took it a step further in reimagining and re-architecting this platform by creating a tightly integrated data fabric that allows agile, efficient, and secure data movement within CareSource and with our partners to enable transparency and interoperability for our providers, members, community, and government organizations.  In effect, this integration creates a data supply chain that makes it easy to find, access, and consume data, thereby establishing a cost-effective single source of truth.

modern data platform CareSource

Diving into data architecture

CareSource’s modern data platform is a fully cloud-based solution that replaces an on-premises enterprise data warehouse and many other siloed reporting databases and is the central hub for all data needs across the organization. Once completed, the platform will host and manage 40TB of data with near real-time data movement supporting 700 to 1,000 users (data consumers, data analysts, and advanced analysts), and producing more than 1,500 prebuilt reports/dashboards.

Data is ingested in a raw format in the data lake and is then processed and refined through a data hub and enriched into a form ready for consumption in the data warehouse. Each layer of the architecture supports business-driven operational (transactional and regulatory) and analytical use cases.

Data governance was another key aspect in the design of this modern data platform, with principles and practices integrated into each layer of the architecture.  The objective was to create an overarching umbrella that would enable a high-level of trust in the data accessed by business users to make decisions. 

capture curate CareSource

An eye on outcomes

Successfully planning and developing a governance structure for a data architecture requires the active involvement of IT and business as co-owners and co-leaders of the effort. In our case, that collaboration consisted of participating in several development coordination groups, including a data governance executive steering committee, a data governance office, a data governance council, an executive data steering committee, and data stewardship, advisory, and consumer committees.

Some key points to consider in launching and guiding such an initiative:

  • Focus on enabling business outcomes. Although technology and IT obviously play key roles in driving and executing the project, the focus is not on IT, but on the overall objective to enable business outcomes and priorities. As such, the entire executive leadership, or C-suite, serves as the oversight and governing body of the program, as part of the executive data steering committee, in providing support and direction.
  • Keep business team members and leaders educated and informed. Scheduling demos with each iteration sprint gives these stakeholders an opportunity to provide feedback and recommendations.
  • Take a use case-driven approach. This will help clearly frame business objectives and avoid rabbit holes. At CareSource, we were also intentional in our definitions of the guiding principles around cost models, especially in applying new technologies and platforms, to avoid confusion around budget responsibilities and expenditures.
  • Make use of industry-standard data models related to speed of delivery and integration of new sources of data. We integrated such a model across different subject areas, allowing for improved ease in finding and consuming data, as well as creating and generating consolidated reporting.

Finally, remember that data governance needs to be metrics driven.  It is important to clearly lay out how the different data governance councils, in both IT and business, work with each other to get the work done and make decisions. Functional roles and responsibilities should be established and understood to create the capacity to engage and do the work.

In setting up the data governance program, we recommend the executive data steering committee also be made up of senior staff members who meet regularly. The functional roles for all data governance bodies are built into the roles and responsibilities, and capacity is created to successfully engage and do the work. 

Angela McArthur is Senior Vice President of IT Data and Enterprise Services, and Joy Mukherjee is Vice President, Enterprise Data Services at CareSource, a nonprofit, multi-state health plan recognized as a national leader in managed care.  CareSource is a CIO Executive Council member company.