CMS tests AI for medical records with EPATHUSA (OCR, machine learning, NLP) — a federal contract (USAspending)

The Centers for Medicare and Medicaid Services (CMS) is the agency within the U.S. Department of Health and Human Services that runs Medicare, the public health insurance program for older Americans, and Medicaid, the program for people with low incomes. Operating programs of this scale involves handling vast amounts of documents, such as clinical records and billing paperwork, and making that reading and organizing more efficient has long been a challenge. This contract was put in place to assess whether AI can be brought to bear on that document processing.

What stands out is that this is a testing-stage contract, not a full system rollout. According to the source, the purpose is to test the viability of applying AI to medical records, including OCR (turning printed or scanned text into machine-readable data), machine learning (learning patterns from data), and natural language processing (having machines analyze human language). The source does not state what specific outcomes are expected, so this reads as an effort focused first on confirming whether these technologies are usable.

Viewed more broadly, the contract illustrates how a public agency can evaluate AI cautiously through a small-scale pilot rather than adopting it all at once. Whether AI becomes embedded in the operation of a large health program like CMS bears on future document-processing costs and administrative efficiency, and it offers one signal of where government AI adoption is heading. This article describes the facts of the contract and does not assert the medical effectiveness of any particular AI technology.