A Lifeline for Medicaid Beneficiaries

Originally published 02/10/2020 on the Nuna blog

At Nuna, our Government Services Team has the privilege of working to improve the quality of healthcare for the 1 in 5 Americans that participate in Medicaid and the Children’s Health Insurance Program (CHIP). Together with our contract partners the government services team recently launched the first cross-agency API for Medicaid, connecting Medicaid eligibility data with the FCC’s Lifeline program. As of September, up to 60% of Lifeline beneficiaries can now be instantly verified for reduced cost mobile, internet, and broadband services via the Lifeline API, replacing a manual verification process.

The Lifeline API was developed in just five months. This short development time was possible due to years of work developing data warehousing and analytics platforms with our partners and the Centers for Medicare and Medicaid Services (CMS).

Beginning with an overview of these foundational platforms, we will discuss how bringing analysts onboard as partners and not merely data consumers enabled rapid development of the Lifeline ETL process. In addition, we will talk about our experiences leveraging Databricks notebooks and cloud-based serverless components in the development process. A technical deep dive into the Lifeline API will be the topic of an upcoming post.

Surfacing Medicaid Eligibility Data

The rapid development of the Lifeline API was enabled by three key elements:

  • Access to a current, nationwide database of Medicaid eligibility information
  • The ability to leverage serverless components
  • Partnering with analysts to develop Lifeline eligibility reports based on Medicaid eligibility

Nuna’s work developing The Transformed Medicaid Statistical Information System (T-MSIS) made the first two elements a reality. T-MSIS is a nationwide data warehouse for Medicaid data, a monumental undertaking that has been underway for several years. States are required to submit to on a monthly basis, ensuring that the data reflects current Medicaid enrollment.

T-MSIS is also the first cloud native platform at CMS, which required getting cloud-based technologies authorized for use in government systems. Following T-MSIS, the investment in cloud computing at CMS continued to grow. As a result, the Lifeline API was able to leverage cloud-based components from the very beginning of development.

As a data warehouse, T-MSIS is not designed to support diverse analytic workloads. The Analytics and Reporting Environment for Medicaid And CHIP (AREMAC) is built for this purpose. Using AREMAC, we partnered with analysts to determine Lifeline eligibility based on current Medicaid enrollment. In addition to T-MSIS data, AREMAC is designed to ingest data from other sources, providing analysts with the opportunity to create enriched reports and data products that can be published and shared on the platform.

The Nuna Government Services Team is extremely proud to have played a key part in designing, building and launching the T-MSIS and AREMAC projects over the last several years. These groundbreaking projects deserve blog posts in and of themselves, but hopefully the preceding primer provides some idea of the giants whose shoulders the Lifeline API stands on.

Connecting Analysts With The Data

Now that we have a national Medicaid database and analytics platform, how do we leverage them to build the Lifeline API?

Using T-MSIS data, our analytics partners developed SAS queries to generate the initial Lifeline eligibility report, identifying individuals eligible for Lifeline based on their Medicaid enrollment. The Lifeline data engineering team then created an ETL process by porting this report to Databricks Notebooks in AREMAC for further refinement and automation.

Notebooks enabled us to move quickly by:

  • Making the analysis itself part of the codebase, eliminating back and forth between analysts and engineers over requirements.
  • Providing a platform to quickly test and deploy changes without the overhead of traditional build and deploy processes.

The tight schedule for the Lifeline API required the engineering team to be executing while the business rules for determining eligibility were still in flux. As a result of using notebooks to anchor our ETL process, the team was able to quickly iterate with our partners at USAC as the business rules were refined.

The ETL process is depicted below. As business rules changed we were able to rerun the pipeline and provide USAC with sample output data to review. Once the ETL process was approved, the Notebook was scheduled to run as new T-MSIS data comes available.

Lifeline ETL Architecture

Powered by the cloud

Scalable, on-demand cloud computing services like AWS can minimize cost through charging only for the resources used, as opposed to the cost of running a data center where machines may sit idle. As an example, the Lifeline API costs just $500–600 a month to surface data on 70 million people, available 24/7 with a less than 3 second response time.

Cloud services are also a boon to developers, giving them access to out of the box solutions that previously would have required significant development effort to bring online, such as an API and credentials service. Because of this, developers can quickly iterate on potential designs and focus on building the platform. These efficiencies were crucial for getting the Lifeline API up and running within the five month timeframe. We will do a deep dive into these efficiencies in a future article.


The Lifeline API is a great example of technology the Nuna Government Services team seeks to build; a low cost, high availability platform that makes it easier for at-risk populations to access critical services.

In this post we introduced the Lifeline API and highlighted the key elements that contributed to its success:

  • Up to date Medicaid eligibility data provided by T-MSIS
  • Enabling data users to become co-creators via AREMAC
  • The ability to partner with analysts and quickly iterate using Notebooks
  • Cloud-based, serverless components enabling rapid development while minimizing cost.

Big data systems take a long time to develop and refine. We are now beginning to see the return on investment in T-MSIS and AREMAC, the Lifeline API being one of what will hopefully be many advancements thanks to these technologies.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.