Creation of descriptive metadata for digital objects tends to be a laborious process. Specifically, subject analysis that seeks to classify the intellectual content of digitized documents typically requires considerable time and effort to determine subject headings that best represent the substance of these documents. This project examines the use of topic modeling to streamline the workflow for assigning subject headings to the digital collection of New Mexico State University news releases issued between 1958 and 2020. The optimization of the workflow enables timely scholarly access to unique primary source documentation.