Amanjot Singh Sachar, an alumnus of the UB School of Engineering and Applied Sciences, also co-authored the study.

In “Perceived Impact of COVID-19 in an Underserved Community: A Natural Language Processing Approach,” the researchers examined 50 transcribed interviews conducted from June 2021 to May 2022 with a purposively selected sample size of adults living in underserved communities in Western New York. Prior research indicates that the COVID-19 pandemic disproportionately impacted underserved populations because of socioeconomic factors and other social determinants of health, such as health care access.

The researchers combined two methodological approaches to analyze a subset of interviews with respondents from underserved communities: natural language processing (NLP), a type of artificial intelligence for understanding human language, and thematic analysis, an approach that identifies themes and patterns that emerge from collected data.

“I was surprised to see the complementary nature of natural language processing (NLP) and traditional thematic analysis,” Holmes said. “NLP uncovered new patterns and combinations of words while thematic analysis offered richness and detail.”

The two methods yielded unique insights about the negative, positive and highly individualized impacts of the COVID-19 pandemic within underserved communities. The study also significantly expands nursing research methodology.

“This study exemplifies the potential of combining traditional thematic analysis with innovative NLP techniques to gain a deeper understanding of complex, qualitative datasets … offering a promising direction for future interdisciplinary studies,” Chang said.

BY SHANNON O'SULLIVAN