The Centers for Medicare & Medicaid Services (CMS) maintains the largest volume of health care data in the world. Its Integrated Data Repository (IDR) is a high-volume data warehouse integrating claim, beneficiary and provider data to support various Medicare and Medicaid programs. Access to this robust, integrated data supports analytics across CMS, including insights into medical trends, healthcare costs, and the prevention of fraud, waste and abuse.
Challenge: Build on Cloud Migration Successes to Offer Users More Capabilities
In 2023, under CMS’s leadership, GDIT successfully migrated the IDR to the cloud, evolving one of the largest public clouds in the entire federal government. Building on this foundation, GDIT continued working with CMS to enhance the capabilities of the IDR by giving CMS users access to advanced analytics, AI, and robust support solutions.
Approach: Continuous user-focused enhancements with zero downtime and minimal data movement
Over a span of two years, the team continuously rolled out new innovations to meet the overall objective of enhancing IDR user capabilities. These included:
Developing a Customer Analytics & AI Environment (CAE). The CAE streamlines users’ secure access to healthcare data. Built on AWS SageMaker AI and Amazon Bedrock, the CAE seamlessly integrates with IDR while offering the flexibility to connect to other CMS data sources. This enables CMS analysts and data scientists to harness enterprise-grade AI and machine learning capabilities without requiring extensive infrastructure expertise. Initially developed as a proof of concept in just 90 days, the CAE platform has demonstrated versatility by supporting diverse use cases, including document intelligence automation and predictive modeling.
Offering Instant Assistance via the IDR Support Bot. The team built, tested and deployed an AI-powered assistant capable of providing instant answers to IDR-related questions. Featuring an intuitive interface developed using React, the IDR Support Bot makes it easy for users to find the information they need efficiently. On the back end, it leverages Amazon Bedrock and large language models (LLMs), utilizing a multi-agent AI orchestration framework to deliver accurate and contextually relevant responses. Available to new users or experienced IDR analysts, this self-service capability advances CMS's strategic goal to streamline operations using AI while enhancing the overall user experience through prompt, reliable assistance.
Improving Program Integrity with Veterans Health Administration (VHA) Data. By incorporating VHA data into the IDR, the system now provides more robust insights that enable the detection of duplicate payments, reduce financial waste, and strengthen accountability. Within months of implementation, this effort helped to identify $106M in improper payments, advancing the Department of Health and Human Services (HHS) priorities for proactive detection strategies and ensuring the integrity of federal healthcare programs.
Developing the IDR Learning Platform and E-Learning Modules. Using the Moodle framework, GDIT built a custom Learning Management System to serve as the foundation of the IDR E-Learning platform. This platform provides CMS users with centralized access to bite-sized e-learning modules through the CMS Enterprise Portal. These modules enable users to learn about the IDR at their own pace, enhancing knowledge retention and making it easier to effectively leverage the IDR’s capabilities. This innovative approach has transformed how CMS users engage with the IDR, fostering a culture of continuous learning.
Transforming Data Access via IDR’s Enterprise Data Products (EDPs). Leveraging Snowflake’s cloud-native capabilities, including data shares and support of external tables, the EDP allows CMS users to access and analyze data across diverse systems without the need for data movement or duplication. This eliminates costly ETL processes and establishes a modern, scalable data federation architecture that supports enterprise-wide analytics, collaborative exploration, and decision-making. By onboarding multiple major data contributors and integrating additional users, the EDP delivers significant cost savings and improves data transparency and interoperability, aligning with HHS priorities for IT modernization.
Consolidating User Support and Request Processes with Jira Service Management (JSM). The team streamlined CMS's support processes by consolidating fragmented request systems into a centralized, automation-driven interface powered by Jira Service Management (JSM). This improvement has reduced ticket resolution turnaround time by 60%, enabling faster and more efficient handling of integration requests, cloud migration needs, data requests, and most recently all support requests. By standardizing intake workflows and automating manual processes, the JSM platform simplifies project enablement and eliminates confusion around accessing services, forms, or support. With this unified system, IDR users benefit from a more intuitive and reliable experience while leveraging the platform’s capabilities to support CMS's operational efficiency and IT modernization goals.
Impact: IDR Positioned as CMS's Enterprise Data Access Platform
These innovations have made the IDR more accessible, efficient, and user-friendly, further solidifying its role as CMS's enterprise data access platform. They also advance key HHS goals, including IT modernization, data transparency, and program integrity.
Today, users have access to faster insights, instant support, enhanced data integrity, easily accessible learning and seamless data access. Through close and constant collaboration with our customer and our strategic technology partners, like AWS and Snowflake, the GDIT team has empowered CMS users with the tools, support and knowledge they need to deliver on the agency’s mission and to do it in increasingly efficient and effective ways. As the IDR continues to evolve, these foundational innovations will ensure CMS users are well-equipped to leverage one of the federal government’s most comprehensive healthcare data resources.