Zobraziť v EDS

Topic modeling and sentiment analysis of Greek clinician-patient conversations in hematologic malignancies.

Uložené v:
Podrobná bibliografia
Názov: Topic modeling and sentiment analysis of Greek clinician-patient conversations in hematologic malignancies.
Autori: Chatzimina ME; Department of Electrical and Computer Engineering, Hellenic Mediterranean University, Heraklion 71004, Greece; Computer Science, Foundation for Research and Technology-Hellas (FORTH) Heraklion 700 13, Greece. Electronic address: hatzimin@hmu.gr., Papadaki HA; Department of Hematology, School of Medicine University of Crete, Heraklion 71003, Greece., Pontikoglou C; Department of Hematology, School of Medicine University of Crete, Heraklion 71003, Greece., Tsiknakis M; Department of Electrical and Computer Engineering, Hellenic Mediterranean University, Heraklion 71004, Greece; Computer Science, Foundation for Research and Technology-Hellas (FORTH) Heraklion 700 13, Greece.
Zdroj: International journal of medical informatics [Int J Med Inform] 2025 Dec; Vol. 204, pp. 106071. Date of Electronic Publication: 2025 Jul 30.
Spôsob vydávania: Journal Article
Jazyk: English
Informácie o časopise: Publisher: Elsevier Science Ireland Ltd Country of Publication: Ireland NLM ID: 9711057 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1872-8243 (Electronic) Linking ISSN: 13865056 NLM ISO Abbreviation: Int J Med Inform Subsets: MEDLINE
Imprint Name(s): Original Publication: Shannon, Co. Clare, Ireland : Elsevier Science Ireland Ltd., c1997-
Výrazy zo slovníka MeSH: Hematologic Neoplasms*/psychology , Hematologic Neoplasms*/therapy , Natural Language Processing* , Physician-Patient Relations* , Communication*, Humans ; Greece ; Emotions ; Male ; Female ; Palliative Care
Abstrakt: Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Background: Conversations in clinical settings, especially those involving hematologic cancers or palliative care are not only informational but also emotionally charged. Understanding how these conversations are structured could help develop AI systems that support not only information exchange but also personalized emotional care.
Objective: This study explored the themes discussed and emotional tones in real conversations between Greek speaking clinicians and patients. Our aim was to contribute toward building digital tools for low-resource languages such as Greek that are not only linguistically appropriate but also emotionally aware.
Methods: We analyzed over 52,000 anonymized utterances from real Greek clinical conversations using BERTopic for topic modeling and a domain-specific sentiment classifier fine-tuned on Greek medical data. Topics were labeled using both a large language model (Gemma-3) and extractive keyword methods (KeyBERT).
Results: The analysis revealed 35 thematic clusters, covering areas like symptoms, diagnosis, emotional states, treatment choices, and end-of-life planning. Sentiment analysis revealed that patients expressed more negative sentiment in utterances related to pain, uncertainty or personal loss. Clinicians often responded with a more neutral or even empathetic positive tone, especially when offering support or medical advice.
Conclusions: Our findings show that Natural Language Processing (NLP) can reveal both the content and emotional tone of conversations in medical settings. This approach could support the development of digital health tools such as conversational agents that are more emotionally aware in Greek and other low-resource languages, especially in emotionally charged domains like palliative care.
(Copyright © 2025 The Authors. Published by Elsevier B.V. All rights reserved.)
Contributed Indexing: Keywords: BERTopic; Clinical dialogues; Clinician patient communication; Hematologic malignancies; Large language models; Natural language processing; Palliative care; Sentiment analysis; Topic modeling
Entry Date(s): Date Created: 20250801 Date Completed: 20250907 Latest Revision: 20250907
Update Code: 20250908
DOI: 10.1016/j.ijmedinf.2025.106071
PMID: 40749353
Databáza: MEDLINE
Popis
Abstrakt:Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.<br />Background: Conversations in clinical settings, especially those involving hematologic cancers or palliative care are not only informational but also emotionally charged. Understanding how these conversations are structured could help develop AI systems that support not only information exchange but also personalized emotional care.<br />Objective: This study explored the themes discussed and emotional tones in real conversations between Greek speaking clinicians and patients. Our aim was to contribute toward building digital tools for low-resource languages such as Greek that are not only linguistically appropriate but also emotionally aware.<br />Methods: We analyzed over 52,000 anonymized utterances from real Greek clinical conversations using BERTopic for topic modeling and a domain-specific sentiment classifier fine-tuned on Greek medical data. Topics were labeled using both a large language model (Gemma-3) and extractive keyword methods (KeyBERT).<br />Results: The analysis revealed 35 thematic clusters, covering areas like symptoms, diagnosis, emotional states, treatment choices, and end-of-life planning. Sentiment analysis revealed that patients expressed more negative sentiment in utterances related to pain, uncertainty or personal loss. Clinicians often responded with a more neutral or even empathetic positive tone, especially when offering support or medical advice.<br />Conclusions: Our findings show that Natural Language Processing (NLP) can reveal both the content and emotional tone of conversations in medical settings. This approach could support the development of digital health tools such as conversational agents that are more emotionally aware in Greek and other low-resource languages, especially in emotionally charged domains like palliative care.<br /> (Copyright © 2025 The Authors. Published by Elsevier B.V. All rights reserved.)
ISSN:1872-8243
DOI:10.1016/j.ijmedinf.2025.106071