Peer-Reviewed Articles and Public Health: Title and subTitle BreakThe Mad Cow Affair in Italian Newspapers FREE

JOURNALISTS consider peer-reviewed journals an important source of information on biomedical subjects¹ and there is evidence that articles published in peer-reviewed journals may have a significant impact on the lay press.² It has been suggested, however, that early announcements of research work to be published in peer-reviewed journals may greatly diminish newsworthiness of scientific articles.³

The aim of the present study was to analyze the impact on the news media, in terms of volume and prominence of coverage, of a scientific article published in peer-reviewed journals about issues with great relevance to public health compared with impact of preliminary release of information on the same issue. To this end we analyzed how the "mad cow affair" was reported in Italian newspapers in the spring of 1996.

Mad cow disease is the popular name for bovine spongiform encephalopathy (BSE), identified in 1986 among British herds.⁴ In 1989 concern was expressed on the possible risk of BSE transmission to humans, resulting in Creutzfeldt-Jakob disease (CJD).⁵ On March 20, 1996, the British secretary of state for health made a statement in the House of Commons announcing the identification of a new-variant CJD (v-CJD) in 10 young people, and stating that the most likely cause of these cases was exposure to BSE, without mention of clinical and neuropathological findings that led to the identification of v-CJD. Immediately all over Europe concerns were raised on eating British beef. The scientific article describing these cases in detail was published in The Lancet on April 6, 1996.⁶

We performed an analysis of press reports on CJD and BSE for the period March 20 to May 10, 1996, reviewing the 7 Italian newspapers with the highest nationwide circulations.⁷ Relevant articles were identified by hand search, reading each headline, subheading, and half title, and were classified according to date of publication, page number, and proportion of page occupied. To identify articles that contained more specific scientific information, we collected information on whether newspaper articles mentioned the following characteristics of v-CJD reported in The Lancet: number of cases, age, mean duration of disease, short incubation period, mention of a new variant, and neuropathology findings. If the article contained an interview with biomedical scientists, this was recorded. For articles published after April 6, we also analyzed whether the Lancet article was mentioned.

For the 2 newspapers with the highest circulations, the 6-month period preceding March 20, 1996, was also analyzed.

Overall, 535 articles on the mad cow affair were published during the study period in the newspapers considered; 62 (11.6%) of them appeared on the front page. A total of 5 articles on this subject appeared in the 2 newspapers with the highest circulations during the 6 months preceding the study period.

On March 21, 1996, the day following the first statement on v-CJD, 2 newspapers each had 1 article on the mad cow affair. In the days that followed, the overall number of articles, and the number of those appearing on the front page, rapidly increased. The peak of press coverage was recorded on March 26, 1996, with a total of 48 items; at least 5 items, including 1 on the front page, appeared in each of the newspapers studied. The newspaper attention to the subject decreased after March 31, 1996, and no further peak of press coverage was recorded after the Lancet publication (Figure 1). During the study period, 72% of all the articles and 89% of those appearing on the front page were published before the Lancet publication.

The median proportion of page occupied daily by the articles on the mad cow affair was 0.63 in the first week and 0.43 in the second week of our study. It dropped to 0.10 in the third week, and remained below this figure in the following 5 weeks.

Overall, 50 articles reporting at least 1 of the 6 characteristics of v-CJD were published in the 7 newspapers during the study period, and 46 (92%) of them appeared before April 6, 1996. After the Lancet publication, only 1 article reported 2 findings not reported before. Publication of the article on v-CJD in The Lancet was mentioned at least once by 3 of the 7 newspapers. Finally, 27 articles containing an interview with a scientist were identified; 21 (74%) of them were published before the Lancet article.

Our results show that the mad cow affair had a great impact on the press in Italy, and this finding is not surprising if its potential public health and economic importance is considered.⁸

However, this impact was concentrated during the 2 weeks that followed the first announcement on the emergence of v-CJD. The attention of the Italian press had already decreased by the time the Lancet article was published and no further peak of press coverage followed its publication. Moreover, 4 of the 7 Italian newspapers studied did not mention the Lancet article.

Journalists apply 2 tests to any piece of information in the field of science and medicine: is it genuine, and is it news?¹ The information presented in the peer-reviewed article should be expected to have a higher scientific credibility for journalists than the official announcement that started the affair. However, the time period between the first news release and the scientific article publication (slightly more than 2 weeks) could have been a long enough time span to determine that the information on CJD had already partially lost its newsworthiness by the time the Lancet article was published.

These data suggest that when a peer-reviewed scientific article on a health risk is published after this risk is disclosed to the public by other means, its impact on the media is low. However, some limitations of this study must be considered. First, we did not analyze coverage patterns of other news media such as television, although there is evidence in Italy that in other cases of emerging health risk, television coverage follows the same pattern as newspapers (Carlo Fido, oral communication, 1998). Second, the case we analyzed was characterized by the potential of significant and immediate implications for public health; therefore, our results cannot be generalized to the routine publication of peer-reviewed research.

The lack of synchronicity between announcement to the media and full publication of scientific data may negatively affect the quality of information conveyed to the public.

In the case we analyzed, scientists were asked by the press to give an expert opinion, most often during the period of maximum mass media attention, when most of the scientific information on v-CJD was conveyed by newspapers to their readership. Moreover, physicians may have been asked by their patients to give out more detailed information about a possible new health hazard, as reported in similar situations.⁹ However, the possibility for scientists and physicians to provide to the public a balanced view of this emerging problem could have been impaired by the fact that most of them did not have access to full scientific data eventually reported in the peer-reviewed article.¹⁰

To improve the communication to the public and within the scientific community in the case of emerging public health risks, 3 points should be considered. First, research work with potential public health implication should be promptly submitted to peer-reviewed journals, without delays because of political or economic considerations. Such delay in submission apparently occurred in the v-CJD article.¹¹

Second, scientific journals should expedite the peer review and publication process as much as possible in these cases, for example, by providing a fast track for articles with relevant public health implications.¹² Improved coordination between news release by public health authorities and scientific publication by peer-reviewed journals should also be pursued.

Third, peer-reviewed journal editors should consider placing articles with potential public health implications in Web sites. In an era in which information on health matters is disseminated rapidly by the media, circulation of information within the scientific community should be at least as fast, while preserving the quality and reliability of scientific journals.