Reviews

Press releases on trial

As the main means of communication between medical journals and the media, press releases have a pivotal role in conveying scientific accuracy. But can they do that at the same time as making the results attractive to lay journalists?

A study in last week's Journal of the American Medical Association (2002;287:2856-8)[Abstract/Free Full Text] found that press releases from medical journals often present data using formats that exaggerate the importance of the findings and do not routinely highlight study limitations or conflicts of interest.

The study authors, Steven Woloshin and Lisa M Schwartz, conducted telephone interviews with the press officers of nine leading medical journals, including the BMJ and the Lancet, and also looked back at press releases issued over the previous six months. Press releases were scored according to whether study results were quantified, whether limitations were mentioned, and whether sources of funding were included.

Seven of the nine journals routinely issued press releases and all these used the same basic approach. The press officer would select and write releases on research papers on the basis of their perceived newsworthiness. The involvement of journal editors in this process varied.

Only 23% (29 out of 127) press releases included information on study limitations. Although 23 of the studies were industry funded, only 22% of the corresponding releases acknowledged this.

Woloshin and Schwartz admit the limitations of their own studythat is, that the sample size was small, their criteria for assessing the quality of the releases were subjective, and they didn't look at the relationship between releases and media coverage. But they say that journals could do more to enhance the quality of medical reporting.

Commenting on the findings, Emma Wilkinson, press officer for the BMJ, said that studies were chosen for press release on the basis of public interest. Releases were approached as a "straightforward summary" and care was taken to avoid including anything not explicitly stated in the study.

Nigel Hawkes, health correspondent of the Times, said that in general he found medical journal press releases "good," and he understood that they could not go into every detail of the study. Both he and Daily Telegraph health correspondent Celia Hall said that they routinely read the original paper mentioned in the press release and sometimes spoke to the author. However, Nigel Hawkes did say that upon reading the original paper one sometimes found that the developments were not as positive as the release had indicated.

Woloshin and Schwartz suggest that ways of improving the quality of medical journal press releases would include adding a section putting results in context and providing more information on limitations and conflicts of interest.

Alex Vass, editorial registrar, 

BMJ

Claire McKenna, second year medical student. 

Queen's University, Belfast, and BMJ Clegg scholar