Schat Joop, Bossema Francien G., Numans Mattijs E., Smeets Ionica, Burger Peter
Wetenschapsstagiair, LUMC, afd. Public Health en Eerstelijnsgeneeskunde, Leiden. (thans: anios, Alrijne Ziekenhuis, afd. Interne geneeskunde, Leiderdorp)
Onderzoeksgroep Science Communication and Society, Universiteit Leiden, Leiden.
Ned Tijdschr Geneeskd. 2018 Jan 2;162(1):D1936.
To determine how often press releases and news articles contain exaggeration and to locate its origin in the trajectory from research paper to news article.
Retrospective quantitative content analysis.
We analysed press releases on health-related research published by Dutch universities and university medical centres in 2015 (n = 129) as well as news media articles related to those press releases (n = 185).
20% of press releases and 29% of news articles exaggerated the conclusion or causal claim. Explicit health advice was, when present, exaggerated in 7% of press releases and 10% of news articles. When press releases exaggerated the conclusion or causal claim, 92% of associated news articles contained the same exaggeration. When the conclusion was not exaggerated in the press release, 6% of the news articles was exaggerated. The relative chance for exaggerated news associated with exaggerated press releases was 16.08 (95% CI: 7.35-35.18). Exaggerated press releases were associated with news articles more frequently. The relative chance for news articles to be associated with exaggerated press releases vs. a non-exaggerated press release was 1.45 (95% CI: 1.02-2.04).
Exaggeration in health-related news is strongly correlated with exaggeration in the original press release and occurs in more than 1 in 5 articles. Monitoring and, if necessary, improving the accuracy and correctness of academic press releases seem to be important measures to improve the quality of health related news.
Conflict of interest and financial support: I. Smeets is a columnist with one of the newspapers investigated. She was not involved in collecting, coding and analysing the relevant data. Additional potential conflicts of interest have been reported for this article. ICMJE forms provided by the authors are available online along with the full text of this article.
确定新闻稿和新闻文章中夸大内容出现的频率，并找出其在从研究论文到新闻文章的传播过程中的源头。
回顾性定量内容分析。
我们分析了荷兰各大学及大学医学中心在2015年发布的与健康相关研究的新闻稿（n = 129）以及与这些新闻稿相关的新闻媒体文章（n = 185）。
20%的新闻稿和29%的新闻文章夸大了结论或因果关系声明。若有明确的健康建议，7%的新闻稿和10%的新闻文章对其进行了夸大。当新闻稿夸大结论或因果关系声明时，92%的相关新闻文章也存在同样的夸大。当新闻稿中的结论未被夸大时，6%的新闻文章存在夸大。与夸大的新闻稿相关的新闻被夸大的相对概率为16.08（95%置信区间：7.35 - 35.18）。夸大的新闻稿与新闻文章的关联更为频繁。新闻文章与夸大的新闻稿而非未夸大的新闻稿相关联的相对概率为1.45（95%置信区间：1.02 - 2.04）。
健康相关新闻中的夸大与原始新闻稿中的夸大密切相关，超过五分之一的文章存在这种情况。监测并在必要时提高学术新闻稿的准确性和正确性，似乎是提高健康相关新闻质量的重要措施。
利益冲突和资金支持：I. Smeets是其中一家被调查报纸的专栏作家。她未参与相关数据的收集、编码和分析。本文已报告了其他潜在的利益冲突。作者提供的ICMJE表格可在网上与本文全文一同获取。
