Medicare claims processing modernization gains urgency at CMS

In 2025, the newly created U.S. Department of Government Efficiency (DOGE) was associated with the large-scale culling of staff and a proposed restructuring at HHS, along with substantial reductions in personnel at many other departments.

These days, in addition to releasing a massive file of provider-level Medicaid billing data, DOGE is placing its healthcare focus on modernizing programs such as Medicare via implementation of new technology.

To that end, the Medicare claims processing system is set for a long-overdue overhaul, says DOGE’s administrator.

The system was written during the 1970s using Common Business-Oriented Language (COBOL) data-processing technology, Amy Gleason, administrator, said during a session at the 2026 Virtual Value-Based Payment Summit.

Shortcomings today include an inability to support real-time data processing and a lack of compatibility with AI and modern analytics, Gleason noted.

“You can’t even find COBOL engineers, really, anymore,” she said.

How CMS seeks to modernize the Medicare claims processing system

Past attempts to upgrade the Medicare claims system have been slow and inefficient, according to Gleason.

For the current effort, CMS solicited input from financial institutions that have advanced beyond COBOL. And in January, a request for information (RFI) sought healthcare industry ideas to improve large-scale claims processing and adjudication via implementation of a real-time cloud-based platform.

Goals include enhancing beneficiary experience, reducing provider burden and improving administrative efficiency in Medicare. Criteria for vendor selection include the capacity to already support at least 2 million active members in a single platform and process more than 100,000 claims per day.

Processes that could improve under an augmented claims processing system include adjudication and status updates, along with fraud prevention. The hope is to achieve tighter integration among CMS, Medicare administrative contractors, providers and other stakeholders.

“It can’t take 10 more years to get off of [COBOL],” Gleason said. “We need to do it pretty quickly.”

CMS’s plans for a national provider directory

DOGE also is involved with the ongoing effort to develop a refurbished national provider directory that is scheduled to be unveiled March 31. CMS saw the need to develop a unified directory to replace the current fragmentation, with eight such directories having been posted in the agency’s own systems and an estimated 5,000 industrywide. The cost to keep information current amid the lack of cohesion is roughly $6 billion per year, Gleason said.

“Providers tell us that they get pinged 20 to 60 times a month to update information in all these different sources,” she said. “It’s a huge administrative burden, and even with that, it’s very difficult to figure out the right information about providers.”

A single directory can serve as a source of truth, Gleason said, with information initially set to include a provider’s affiliated hospital or medical group and how to connect with the provider for interoperability (e.g., through a particular health information exchange). In a second phase, information about the provider’s credentialing and their contracted insurance networks will be included.

The infrastructure will be shared to discourage other stakeholders from building their own distinct directories. Entities such as electronic health record vendors and payers will be able to update the new directory.

In addition to allowing for systemic improvements in routine scenarios, a viable directory will help the industry operate better during emergencies. Gleason recalled the shortcomings that emerged during the COVID-19 pandemic.

“You couldn’t figure out where all the hospitals were,” Gleason said. “We didn’t have a good listing of actually who the providers were in the country. And that was pretty horrifying to not actually know where we needed to work with them.”

Digital identity tools and trust in healthcare data exchange

The hope is that the focus on interoperability in the new directory will help apps and networks connect with providers for data exchange, while patients and other stakeholders will have an easier time sharing information with providers.

Lack of verification is a barrier to interoperability in situations where a provider cannot confirm that a request from another stakeholder is legitimate, Gleason said. The updated directory seeks to address that obstacle by embedding modern digital identity tools (e.g., ID Me, Clear), and such technology also will be implemented for beneficiaries at Medicare.gov.

“The main reason we hear things don’t happen in healthcare is coming down to a trust issue,” Gleason said. “How do I know that it’s actually Amy that’s requesting her information, or it’s actually Dr. Jones requesting that? By adding this identity piece, we think that we can help improve trust. I think that’s another example of where the government should weigh in.”

How the CMS Health Technology Ecosystem could expand patient tools

CMS’s Health Technology Ecosystem initiative has participation by providers, payers, patient advocacy groups, EHR vendors and Big Tech (e.g., OpenAI, Anthropic, Microsoft, Amazon, Apple, Google). The underlying goal is to give patents tools to help manage their own health, Gleason said.

“We’re really excited about this initiative because it’s relying also on the provider directory,” Gleason said. “We’re doing the provider directory to basically unlock the interoperability.”

Specific ecosystem initiatives include “Kill the clipboard,” an effort to give patients access to data networks where they can find their aggregated clinical information. Patients will be able to obtain their information in a private, secure app of their choosing, then use a QR code at the visit to share the information directly with their provider.

Another use case involves conversational AI technology that can help patients understand the information they access in their medical record.

“A lot of patients don’t remember their history or have the ability to really pull that in, or know how to prompt the AI, but this allows the AI to already have their information and be able to customize its recommendations,” Gleason said.

Plans are for diabetes and obesity prevention and treatment specifically to become more personalized in conjunction with a patient’s enhanced access to medical information, she added.