The volume of research published on COVID-19 was unprecedented and the volume and variety of COVID-related publications posed a major challenge for researchers and healthcare professionals who wanted to keep abreast of the latest developments. A major pharmaceutical company required an efficient and accurate continuous literature-monitoring process to select, highlight, and publicize the most relevant and high-quality peer reviewed publications.
The pharmaceutical company worked with IQVIA to use a blended set of Generative AI & NLP technologies as well as HEOR expertise to extract a broad set of data points from relevant publications. The team started with established linguistic rules-based NLP methods and ontologies, then expanding to use Generative AI, applying GPT, to produce an efficient process with high coverage of the target variables.
The automated NLP-based extraction process was 2.5 — 3.0 times faster than traditional manual methods and allowed over 150 data points in 25 major categories to be extracted from over 130 studies. The studies were presented and further summarized by medical experts and made available on two publicly available websites.
