Sorcero GenAI conjures plain language from scientific papers

GenAI plain language
Siora Photography

Understanding complex scientific studies is often a challenge for patients as well as healthcare professionals, but one that may become easier thanks to artificial intelligence.

A new study has shown that a generative AI (GenAI) tool developed by life sciences AI specialist Sorcero, built on Google Cloud, can generate plain language summaries (PLS) of scientific literature that are much easier for patients and healthcare providers to understand and could help to transform medical affairs in the pharma sector.

The pilot study showed that the readability of medical literature could be improved from under 1% to 84% for both patients and providers using the Sorcero PLS software.

Abstracts made using the platform were significantly easier to read than those written by medical writers across all readability metrics, according to the study, and were generated in 40% less time.

Moreover, the AI-generated abstracts were ranked at higher quality than manually written ones by independent experts in the subject matter and also scored more highly by medical writers, patients, and patient advocates on aspects like compassion and empathy.

“Only 12% of adults have proficient health literacy, according to the National Assessment of Adult Literacy,” said Sorcero in a statement. “In other words, nearly nine out of 10 adults may lack the skills needed to manage their health and prevent disease. This creates barriers to understanding important research and can hinder informed decision-making about health.”

The new study was carried out by Belgian biopharma company UCB and medical, commercial, and regulatory affairs consultancy Lumanity, which in March partnered with Sorcero on the delivery of AI-powered medical strategy and communications, including the PLS technology. Under that alliance, the software generates the summaries, while Lumanity’s experts “validate, refine, and disseminate” them.

At the time, Sorcero chief commercial officer Richard Graves said that “communicating with stakeholders, such as patient communities, and using plain language that is tailored to them and their needs, is a key starting point for engagement and education.”

The findings have been presented via a scientific publication, poster, and a plenary presentation at a conference of the International Society for Medical Publication Professionals (ISMPP) conference.

Last year, the ISMPP published a position statement on AI, acknowledging that it will play a “pivotal role” in medical publications and communications, but also calling for measures to make sure it is used appropriately and with full disclosure.

UCB’s head of scientific communications excellence, Chris Rapier, said: “By piloting the generative AI solution from the Sorcero AI platform on Google Cloud to create patient-accessible summaries and by publishing on this novel patient-centric approach, we aspire to achieve transparency and empower patients throughout their healthcare journey.”

Photo by Siora Photography on Unsplash