AWS Public Sector Blog

Automating returned mail to keep members enrolled during Medicaid unwinding

Automating returned mail to members enrolled during Medicaid unwinding

As the COVID-19 public health emergency (PHE) begins to wind down, starting April 1, 2023, state Medicaid agencies (SMAs) will have one year to “unwind” temporary COVID-era changes and return to pre-pandemic ways of working. A major part of that process is to re-verify that all 91 million members are qualified to receive Medicaid benefits. The Kaiser Family Foundation (KFF) estimates indicate that 5 to 14 million individuals may lose their coverage across the nation due to this unwinding.

For nearly a year, Amazon Web Services (AWS) has supported SMAs with in-house Medicaid expertise to identify unwinding issues and develop solutions to address them. The top four concerns that SMAs have shared for unwinding are constituent outreach and engagement, staffing shortages, returned mail, and reporting capabilities. The blog post “4 ways AWS can help with Medicaid unwinding” provides an overview of how SMAs can use AWS to create solutions to meet these needs.

Now, we dive deeper into how AWS can help with automating returned mail to support members in keeping their coverage.

The challenge of returned mail for state Medicaid agencies

As a condition for states to receive increased federal funding, the Centers for Medicare & Medicaid Services (CMS) provided the following guidance:

  1. States must have a process in place to obtain up-to-date mailing addresses and additional contact information (i.e., telephone numbers, email addresses) for all enrollees; and
  2. Returned mail alone cannot result in terminating Medicaid benefits for enrollees. States must make a “good-faith effort” to reach out to enrollees in at least two different ways, using the most up-to-date contact information they have.

Failure to adhere to these two new requirements risks loss of federal funding for states, and more importantly, can negatively impact Medicaid members. If an SMA uses an out-of-date address to conduct outreach, and no follow-up occurs, individuals could lose Medicaid health coverage.

Handling returned mail is both laborious and expensive. It is estimated that returned mail costs $25 per piece. The average Medicaid enrollment for states is 1.8 million, and state Medicaid agencies estimate that 15% of letters sent are returned. This translates to an average cost of $6.75 million. In addition to these costs, SMAs typically do not have enough staff resources to manually scan, index, and conduct outreach for the large volume of returned mail. This means that high costs and manual effort may put Medicaid enrollees at risk of losing coverage because of returned mail backlogs.

AWS can help transform and automate the returned mail process through Amazon Textract and Amazon Pinpoint. Amazon Textract is a machine learning (ML) service that automatically extracts data from forms such as letters, applications, and renewal documents with a high degree of accuracy. Amazon Pinpoint can use that extracted data to use the contact information and send customized text and email messages to the right person at the right time. When that member responds to the message to update their address, the SMA can then send renewal forms to the correct address.

AWS solution framework for automating returned mail processing

We provide a three-part framework to address some of the key issues with returned mail:

  1. Reduce returned mail through proactive outreach to automatically update address and support data cleanup.
  2. Automate the handling and processing of returned mail that cannot be proactively fixed.
  3. Monitor, track, and report both proactive and automated returned mail process.

The following diagram presents a high-level architecture for the proposed solution framework, explained in more detail in the following sections.

Figure 1. Business level architecture for returned mail processing and constituent outreach.

Figure 1. Business level architecture for returned mail processing and constituent outreach

1. Reduce returned mail through data cleanup and proactive outreach to automatically update addresses

Proactive work to reduce the volume of returned mail can help reduce the costs and workload associated with returned mail. By prioritizing upfront automated address cleanup, states can significantly reduce the volume of returned mail they must manage.

Using Amazon Elastic Compute Cloud (Amazon EC2), AWS serverless computing, and AWS Batch, SMAs can retrieve and efficiently process weekly or monthly address update batch files from approved sources. These include the United States Postal Service (USPS) National Change of Address (NCOA), Managed Care Organizations (MCOs), sister state agencies such as Department of Motor Vehicles, or related eligibility systems including Supplemental Nutrition Assistance Program (SNAP) and Temporary Assistance for Needy Families (TANF). If the approved sources support APIs, SMAs can retrieve the updated address using API calls via AWS Lambda

Recent CMS guidance specifically provides SMAs with broad discretion to determine which other sources of contact information are reliable; however SMAs must document these sources and processes used in their unwinding operational plans. With a CMS waiver, SMAs can update case records with new contact information from reliable sources, and then use this information for any outreach or physical mail campaigns to enrollees.

States can also significantly reduce the volume of returned mail and manual processes with proactive outreach to enrollees. Amazon Pinpoint, an omni-channel communications service, can help in this scenario. It uses automated text messaging and email campaigns to help SMAs with outreach — including collecting address information to complete a Medicaid renewal. Learn more about creating a multi-lingual two-way messaging system and how to utilize two-way text messaging to get addresses from enrollees.

By implementing a two-way messaging system, SMAs can ask the enrollee to provide an updated address, or confirm an updated address obtained through a reliable source with the enrollee. This means less time and money spent on returned mail, and enrollees experience an improved renewal and redetermination process.

2. Automate the handling and processing of returned mail

While a proactive approach helps reduce returned mail, it will not completely eliminate it. For any mail that gets returned, automating the handling process is essential. These are the high-level steps for processing returned mail and automating the workflow to reduce burden on agency staff:

Step 1: Agency workers scan returned mail using physical scanners. Images can be indexed and tagged as “Returned Mail” and then uploaded to Amazon Simple Storage Service (Amazon S3) for storing the scanned documents, and Amazon DynamoDB to store the metadata. This step automates the population of a cloud-based database, which enables automation for the subsequent steps.

Step 2: Amazon Textract can automatically extract name and address information from the uploaded mail documents. If the confidence of extraction is low or otherwise needs further human review, Amazon Augmented AI supports human-in-the-loop oversight to correct the address.

Step 3: Using the name and address data extracted from the mail documents, this process automatically looks up the contact information (phone, email) in the current system corresponding to the name/address. Using this information, Amazon Pinpoint can send customized text and email messages to the enrollees. States can use these two outreach methods to collect the updated address information from the members, and satisfy the CMS requirement of two-modality outreach.

3. Monitoring, tracking, and reporting

State leaders need to monitor, track, and report on returned mail handling to get deeper insights into the program’s effectiveness and continued coverage for Medicaid enrollees. Proactive returned mail data obtained via Amazon Pinpoint and approved sources can be ingested into a data lake using Amazon S3 and AWS Lake Formation. Likewise, returned mail processing data from Amazon Textract and Amazon DynamoDB can be fed into the same Amazon S3 data lake. SMAs can then use Amazon QuickSight for business intelligence, visualization, and reporting with dashboards. Additionally, with Amazon QuickSight Q, program leaders can ask questions in natural language and receive accurate answers with relevant visualizations that help them gain deeper program insights from the data.

These dashboard metrics can include:

  • Total number of proactive address outreach calls/texts made
  • Total number of proactive address updates received by member, NCOA, MCO, or other source
  • Total number of redetermination letters returned, broken down by Nixie reason code
  • Total number of members notified via two-modality outreach events text/phone/email
  • Number of members who responded with updated address information
  • Number of members who have not responded via any modality
  • Number of members with outstanding invalid addresses

A dashboard with this data can help states gain insight into their current returned mail processes and help make data-driven decisions about further outreach.

Summary and next steps

AWS is ready to support Medicaid agencies as they transform to meet PHE unwinding efforts. Contact the AWS Public Sector Team to learn more. Health and Human Services agencies across the country are using the power of AWS to unlock their data, improve the citizen experience, and deliver better outcomes. See more Health and Human Services Cloud Resources here. Learn more about how governments use AWS to innovate for their constituents, design engaging constituent experiences, and more at the AWS Cloud for State and Local Governments hub.

Read related stories on the AWS Public Sector Blog:

Subscribe to the AWS Public Sector Blog newsletter to get the latest in AWS tools, solutions, and innovations from the public sector delivered to your inbox, or contact us.

Please take a few minutes to share insights regarding your experience with the AWS Public Sector Blog in this survey, and we’ll use feedback from the survey to create more content aligned with the preferences of our readers.

Srinath Godavarthi

Srinath Godavarthi

Srinath Godavarthi is a principal solutions architect at Amazon Web Services (AWS), based in the Washington, D.C. area. In that role, he helps public sector customers achieve their mission objectives with well architected solutions on AWS. Prior to AWS, he worked with global systems integrators for over 20 years serving the Food and Drug Administration, the Department of Veterans Affairs, and the Centers for Medicare and Medicaid Services. He focuses on innovative healthcare solutions using artificial intelligence (AI) and machine learning (ML).

Nick Aretakis

Nick Aretakis

Nick Aretakis is a Medicaid transformation leader for the Health and Human Services team at Amazon Web Services (AWS). He enables Medicaid agencies to improve the Medicaid experience through delivery of transformative technology. Prior to AWS, Nick authored and led delivery-focused federal policies and national oversight as a deputy director at the Centers for Medicare and Medicaid Services (CMS) Division of State Systems. At CMS, Nick spearheaded rescue efforts between states and, allowing 4.3 million people to gain access to care, and transformed the federal oversight process for Medicaid systems into an outcomes-based delivery-focused policy.