As advances in data science, artificial intelligence (AI), and machine learning (ML) digitally transform our society, healthcare insurers increasingly use advanced analytics to improve health outcomes and lower cost of care. Our Client, a large payer group skilled at aggregating large data sets that feed transactional systems, found their complex array of repositories, warehouses, and consuming systems to be inefficient and ineffective for modern-day analytics. To address these growing limitations, the VP of Enterprise Data Solutions in partnership with the Chief Enterprise Architect asked Trexin to lead a data analytics platform assessment to design a modern data analytics platform and enterprise analytics capability.
Our approach followed Trexin’s Strategy, Assessment, and Roadmap methodology, applied over a 12-week timebox. Trexin began by conducting stakeholder interviews with 56 operational leaders and 4 health plans to establish a shared understanding of the business goal, strategy, and potential tactics related to emerging analytics business needs. Simultaneously, Trexin collected and reviewed prior material regarding data goals, plans, modernization projects, and current-state systems and capabilities. Working collaboratively with the stakeholders, Trexin then translated the selected tactics into a future-state capability vision, which included:
- A market scan of analytics workbench technologies and products for an integrated solution
- An analytics lifecycle process architecture including data sourcing, data ingestion, data validation, data integration, data wrangling, and data presentation
- A data analytics conceptual reference architecture
- A recommendation for aligning two large, existing data warehouses with the new platform
- An investment project roadmap outlining tasks, milestones, and an implementation timeline
Guided by a well-defined, consensus-driven set of 7 strategic tactics with example use cases, the Trexin-led team designed the future-state data analytics platform to have 4 primary components that could be built incrementally: 1) Data Lake & Ingestion, 2) Sandbox Data Access, 3) Analytic Workbenches, and 4) an AI/ML Algorithm Library. Implementation planning included options for re-building on premise, using a cloud-based platform, or leveraging a managed service provider. And the Analytic Workbench capabilities were framed in terms of 2 user categories with distinct user personas: Business Analytics (for “Leadership” and “Analyst” personas) and Data Science (for “Researcher” and “Data Scientist” personas).