Some of the most difficult healthcare conversations to have are the critical ones that happen around illness and end-of-life care. But machine learning tools might be able to analyze conversations between providers and patients about palliative care, resulting in improved communication.

That’s according to a study conducted at the University of Vermont’s (UVM) Conversation Lab, in which researchers have used machine learning and natural language processing to better understand what those conversations look like,

“We want to understand this complex thing called a conversation,” said Robert Gramling, director of the lab in UVM's Larner College of Medicine who led the study. “Our major goal is to scale up the measurement of conversations so we can re-engineer the healthcare system to communicate better.”

Grambling’s team used machine learning techniques to analyze 354 transcripts of palliative care conversations collected by the Palliative Care Communication Research Initiative, involving 231 patients in New York and California.

They broke each conversation into ten parts with an equal number of words in each, and examined how the frequency and distribution of words referring to time, illness terminology, sentiment, and words indicating possibility and desirability changed between each decile. According to the findings, which were published in the journal Patient Education and Counselling, conversations tended to progress from talking about the past to talking about the future, and from happier to sadder sentiments.

“We picked up some strong signals,” said Gramling. “There was quite a range, they went from pretty sad to pretty happy.”

Discussions also tended to shift from talking about symptoms at the beginning, to treatment options in the middle and prognosis at the end. Moreover, the team noted, the use of modal verbs like “can,” “will,” and “might,” that refer to probability and desirability also increased as conversations progressed.

The findings reveal the importance of stories in healthcare for patients, researchers noted.

“At the end there was more evaluation than description,” said Gramling. “What we found supports the importance of narrative in medicine.”

The team is now focused on using the machine learning algorithm to identify the different types of conversations that can occur in healthcare. This could help providers understand what might make a “good” conversation around palliative care, and how different conversations require different responses. Providers could then match patients to interventions they need the most.

“One type of conversation may lead to an ongoing need for information, while another may have an ongoing need for functional support,” said Gramling. “So one of the ways those types can help us is to identify what are the resources we are going to need for individual patients and families so that we're not just applying the same stuff to everybody.”