Peters Joachim, Heckel Maria, Breindl Eva, Ostgathe Christoph
Department of Palliative care, University Hospital, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, Germany.
Chair of German Linguistics, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, Germany.
Palliat Med Rep. 2024 Dec 4;5(1):512-520. doi: 10.1089/pmr.2024.0057. eCollection 2024.
Little is known about the public perception of palliative care during and after the pandemic. Assuming that analyzing online language data has the potential to collect real-time public opinions, an analysis of large online datasets can be beneficial to guide future policymaking.
To identify long-term effects of the COVID-19 pandemic on the public perception of palliative care and palliative care-related misconceptions on the Internet (worldwide) through natural language processing (NLP).
Using large language model NLP analysis, we identified public attitudes, opinions, sentiment, and misconceptions about palliative care on the Internet, comparing a corpus of English-language web texts and X-posts ("tweets") (02/2020-02/2022) with similar samples before (02/2018-02/2020) and after the pandemic (03/2022-02/2024).
The study is a statistical analysis of website and social media data, conducted on six large language corpora.
Since the COVID-19 pandemic, palliative care situations are more often portrayed as frightening, uncertain, and stressful, misconceptions about the activities and aims of palliative care occur on average 44% more frequently, especially on the social media platform X.
The impact of the COVID-19 pandemic on public discussion on social media continues to persist even in 2024. Insights from online NLP analysis helped to determine the image of palliative care in the Internet discourse and can help find ways to react to certain trends such as the spread of negative attitudes and misconceptions.
关于大流行期间及之后公众对姑息治疗的看法,人们了解甚少。假设分析在线语言数据有潜力收集实时公众意见,那么对大型在线数据集进行分析可能有助于指导未来的政策制定。
通过自然语言处理(NLP)确定新冠疫情对全球互联网上公众对姑息治疗的看法以及与姑息治疗相关的误解的长期影响。
我们使用大语言模型NLP分析,确定互联网上公众对姑息治疗的态度、意见、情绪和误解,将一组英语网络文本和X帖子(“推文”)(2020年2月 - 2022年2月)与疫情前(2018年2月 - 2020年2月)和疫情后(2022年3月 - 2024年2月)的类似样本进行比较。
该研究是对六个大语言语料库的网站和社交媒体数据进行的统计分析。
自新冠疫情以来,姑息治疗情况更多地被描绘为可怕、不确定和有压力,对姑息治疗活动和目标的误解平均比以前多出现44%,尤其是在社交媒体平台X上。
即使在2024年,新冠疫情对社交媒体上公众讨论的影响仍然存在。在线NLP分析的见解有助于确定互联网话语中姑息治疗的形象,并有助于找到应对某些趋势的方法,如负面态度和误解的传播。