Graduate Program in Complex Systems and Data Science, College of Engineering and Mathematical Sciences, University of Vermont, Burlington, Vermont, USA.
Department of Civil and Environmental Engineering, University of Vermont, Burlington, Vermont, USA.
J Palliat Med. 2023 Dec;26(12):1627-1633. doi: 10.1089/jpm.2023.0087. Epub 2023 Jul 13.
Developing scalable methods for conversation analytics is essential for health care communication science and quality improvement. To assess the feasibility of automating the identification of a conversational feature, which is associated with important patient outcomes. Using audio recordings from the Palliative Care Communication Research Initiative cohort study, we develop and test an automated measurement pipeline comprising three machine-learning (ML) tools-a random forest algorithm and a custom convolutional neural network that operate in parallel on audio recordings, and subsequently a natural language processing algorithm that uses brief excerpts of automated speech-to-text transcripts. Our ML pipeline identified with an overall sensitivity of 84% and specificity of 92%. For and subtypes, we observed sensitivities of 68% and 67%, and specificities of 95% and 97%, respectively. These findings support the capacity for coordinated and complementary ML methods to fully automate the identification of in natural hospital-based clinical conversations.
开发可扩展的对话分析方法对于医疗保健沟通科学和质量改进至关重要。为了评估自动化识别与重要患者结局相关的对话特征的可行性，我们使用姑息治疗沟通研究倡议队列研究的音频记录，开发并测试了一个自动化测量管道，该管道包括三个机器学习 (ML) 工具-随机森林算法和一个在音频记录上并行运行的自定义卷积神经网络，以及随后使用自动语音转文本摘录的自然语言处理算法。我们的 ML 管道的总体灵敏度为 84%，特异性为 92%。对于 和 亚型，我们分别观察到 68%和 67%的灵敏度以及 95%和 97%的特异性。这些发现支持协调和互补的 ML 方法有能力完全自动化识别自然医院临床对话中的 。
