BBC conned by predatory publisher

Yesterday the BBC Radio 4 Today programme had a long item (click here) about a randomised trial of singing exercises to treat sleep apnoea and snoring.  Quote; “A study carried out by the Royal Devon and Exeter Hospital has proved that snoring can be reduced simply by singing”.  The paper is here, and trial registration here.

The trial ran between November 2005 and 2007 (not reported in the paper) but was first registered in March 2011 (late registration not reported either). This means we cannot be confident that the primary endpoint and sample size were predetermined. The reported primary endpoint, the Epworth score, differed from that reported in the pilot study (click here).

Participants were randomised to singing exercises or control after classification into four separate subgroups as follows: normal BMI snorer (n=50), raised BMI snorer (n=22), normal BMI obstructive sleep apnoea (OSA) (n= 30), and raised BMI OSA (n=25).   The sample size calculation suggests that the trial was powered to show a difference within each subgroup.

There’s nothing in principle wrong with this. The problem is that the only statement of results is this;

The Epworth scale improved significantly in the experimental group compared to the control group (difference −2.5 units; 95% CI −3.8 to −1.1; p = 0.000). 

That’s all. Nothing else. Zilch! No table of results. No raw data. No mean scores in either group. No indication of the data spread. No indication whether scores were similar at randomisation. No indication of how many patients contributed to this analysis. It’s not even clear whether it was on the whole sample or one of the four subgroups.  Presumably the former, but if so, none of the subgroups on which the trial sample size was based were reported at all. There are innumerable other problems.

The trial took six years to get published, presumably because all respectable journals turned it down, and finally ended up in an open access one. But why would even a low impact journal publish a paper with no data? Where were the editors and referees?

One reason is that the authors paid $500. Another that the editors and referees may not even exist. Scientific Research Publishing (SCIRP), who publish the International Journal of Otolaryngology and Head & Neck Surgery, are on Jeffrey Beall’s list of predatory publishers (click here). These are publishers who create pseudo-scientific, entirely web-based journals with non-existent editorial boards, and little or no peer-review, simply to harvest open access fees. To put it bluntly the authors fell for an academic version of the Nigerian “send us $500 to unlock the $5 billion we’re holding in your name” scam.

How did the BBC get suckered by this nonsense?

My guess is they thought it was complementary medicine. It’s not, but the website pushing the singing therapy (click here) is a bit “New Agey”, and in 2000 the pilot study was published in a journal called Complementary Therapy Medicine.  The BBC does to tend to suspend its critical faculties when complementary medicine comes up.

This’ll teach ’em!

Jim Thornton