Editor's Choice

The positive in negatives

BMJ 2003; 327 doi: https://doi.org/10.1136/bmj.327.7409.0-g (Published 31 July 2003) Cite this as: BMJ 2003;327:0-g
  1. Kamran Abbasi, deputy editor

    “Does the BMJ publish studies on American datasets?” asked a researcher at a meeting in Atlanta, Georgia, last week on reporting behavioural interventions. We do, of course, although it is worrying that a scientist from the Centers for Disease Control and Prevention thinks we may not. Inevitably, there is considerable international interest in the US healthcare system, just as there is in the NHS, and studies in one setting might generalise to others. But there are more reasons for us to be interested. Our website, bmj.com, attracts about 185 000 users a week, mostly from outside the UK, many from North America. BMJ USA, a monthly digest of the weekly BMJ reaches about 100 000 US family physicians.

    One method of shortening research papers, suggested a few journal editors at the same meeting, is to chop out negative findings. We do not—this is publication bias, where positive findings are reported and negative ones consigned to the dustbin of research history. Publication bias distorts the scientific record but many editors abet it unwittingly—and wittingly—in an attempt at “sexing up” their journals. Researchers and funding bodies—especially industries with an interest in product development—are reluctant about submitting negative studies for publication. Yet there are positives in negatives: they can shake our faith. Some of us, George W Bush for example, have more than our fair share of self belief. “I know what I believe,” he said in Rome two years ago. “I will continue to articulate what I believe and what I believe—I believe what I believe is right.” Well, often what we believe is not right.

    This week we publish several articles that generate more doubt than certainty. Enzyme potentiated desensitisation has been used to treat a range of allergic conditions for over 20 years, notably as a method of pollen desensitisation for hay fever. Previous randomised trials have suggested it works. Michael Radcliffe's team investigated this treatment's ability to reduce symptoms in severe hay fever and report that enzyme potentiated desensitisation is no better than placebo (p 251). Might changing smoking practices in the home, such as not smoking near children or airing a room after smoking, protect children from exposure to tobacco smoke? No, say Clare Blackburn and coworkers, although a complete ban on smoking in the home does have a small beneficial effect (p 257). Current guidelines in the UK and elsewhere recommend tuberculin testing before BCG vaccination for all children older than 3 months. Graham Bothamley and colleagues have reviewed the literature and conclude that tuberculin testing is unnecessary (p 243).

    Many bmj.com users have a Bush like faith in our website remaining fully free forever. From 2005, this will only be the case for users who are BMA members or from the world's poorest countries. Tony Delamothe and Richard Smith explain why (p 241).


    View Abstract