A recent article in the Medical Journal of Australia should have received more US media attention than it has.
On June 5, the MJA ran an article "Media reporting on research presented at scientific meetings: more caution needed" (HTML PDF ). Dartmouth Medical School professors Dr. Lisa Schwartz write that because "the public has a strong appetite for medical news and scientific meetings provide the media with an easy source of provocative material" oral presentations (and the news releases that accompany them) usually omit key disclaimers and key study facts. For example, one-third did not report the study size and 40% did not quantify the main result, only 6% of animal studies noted that the results might not apply to humans and only 2 or 175 stories admitted that the data were not (peer reviewed or) published.
It's easy to fault honest researchers and reporters for failure to disclose the limits of these studies. What is more bothersome, and unmentioned in the article, is the practice of research project sponsors like the National Heart, Lung and Blood Institute which, for example, put out a news release the day before its funded researchers supported the DASH-Sodium Study. NHLBI has stated that they have no access to the underlying data, yet their news release claimed the study showed every American would benefit by reducing dietary sodium. Reporters called me for reaction and, in fact, I did attend the presentation of the paper at the American Society of Hypertension meeting the next day, but what could I "react" to? The paper was not only not peer-reviewed or published, it hadn't even yet been presented! Not only that, but now that the data have dripped out over the past couple years, it shows that in six of eight subgroups (collectively including well over three-fourths of the population) there was no statistically significant association at all. So don't just blame resume-sensitive researchers or headline-hunting reporters.
