Algorithms are finding hidden human connections in end-of-life care

When patients and family members discuss end-of-life matters with professional caregivers, the silences between words can be as telling as the words themselves.

To better understand these communications, a palliative care specialist at the University of Vermont is leading an effort to categorize and code such silences. From these inputs, the team is training AI algorithms to identify meaningful moments of emotional connection that can inform palliative care going forward.

The work of Robert Gramling, MD, and his brother David, a linguistics professor at the University of Arizona, gets a thorough going-over in a long-form piece published by Mosaic.

One fruit of the Gramlings’ work is the Vermont Conversation Lab, the reader learns, which has amassed a database of more than 12,000 minutes and 1.2 million words of conversation involving 231 patients.

This database is set up to help find features of those conversations “that make patients and family members feel heard and understood,” explains article author Michael Erard, PhD, a language expert.

The piece also looks at similar work going on elsewhere. Among the sources Erard spoke with is the daughter of a terminal cancer patient whose diagnosis had come as a shock.

“I ask[ed] Judy’s daughter Kate what she thinks of using artificial intelligence to enrich human connections,” Erard writes.

Judy replied: “I wouldn’t worry about the technology. The more technology, the more sacred the conversation becomes.”

Read the whole thing: