Medicaid Waiver Automated Analysis Prototype “Hermes” (2019)

Goal:
To develop a web-based tool to automatically download, scan, and analyze a large number of publicly available Medicaid 1115 waiver documents to detect trends in certain topics of interest.

Summary:
CDC staff had to manually download hundreds of PDF documents, scan them, and compile coding tags in a spreadsheet. This process took 1 to 2 months to complete and could only be performed once or twice a year because of limited staff hours. As part of the prototype, the Hermes project leverages robotic process automation technology to scrape and process text data from the roughly 150 state Medicaid waivers available from the CMS website. The Hermes engine combs through each PDF document, ingesting tens of thousands of pages and analyzes the internal text to search for terms of interest and displays them in an interactive and intuitive web application that the Office of the Associate Director of Policy and Strategy (OAPDS) team can then use to perform rapid analysis. The newly automated system matches the manual results generated by the CDC staff with an accuracy of more than 96%. It is able to do this in an extremely short period of time (hours). Because of this improvement in speed, the textual analysis can be performed much more frequently, allowing the research team to quickly monitor any changes/updates in any of the hundreds of waivers.