A new computer model can automatically analyze conversations between seriously ill people, their families, and palliative care specialists in epidemiological studies.
The CODYM (COnversational DYnamics Model) developed by the University of Vermont (UVM)'s Laurence Clarfeld and Robert Gramling applies Markov Models to record the flow of data during conversations, based on patterns in speaker turn length.
The authors said a time-based definition of speaker turn length enables real-time automation and analysis of conversational dynamics without transcription or stored audio, which protects privacy.
The researchers wrote that they conducted analyses to validate the CODYM model, "identify normative patterns of information flow in serious illness conversations, and show how these patterns vary across narrative time and differ under expressions of anger, fear and sadness."
They also suggested CODYMs could compare "conversational dynamics across language and culture, with the prospect of identifying universal similarities and unique 'fingerprints' of information flow."
From University of Vermont Larner College of Medicine
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA