Visualizing Conversations in Health Care: Using Discursis to Compare Cantonese and English Data Sets

Abstract

Health care is shaped by often complex communication between multiple people such as doctors, nurses, patients and carers. Research has repeatedly shown that effective communication is key to safe and high-quality care yet improving communication remains a challenge across health systems. In recent years, the field of natural language processing has developed analytic tools to supplement the study of verbal communication through visual representation of analysis. To date, these tools have primarily been used on English data. This study used the software tool Discursis to compare visual representations of Cantonese conversational data that were analysed before and after English translation. Results indicate that some linguistic features of Cantonese that carry meaning may be lost in translation into English. Specific concerns relate to the multidimensional issues of equivalence, ranging from cultural and social associations to semantic, lexical and conceptual differences. These results highlight the importance of developing visual analytic tools that can be used on Cantonese data. Generating visual representations of such data contributes to local and international understandings about communication in health care.