Intended for healthcare professionals

Editor's Choice Editor's choice

Controversies revisited

BMJ 2007; 334 doi: (Published 15 February 2007) Cite this as: BMJ 2007;334:0
  1. Jane Smith, deputy editor
  1. jsmith{at}

    Last September Malcolm Potts and colleagues caused a small furore in our rapid responses by suggesting that some interventions were so obviously effective that they did not need evidence from randomised controlled trials (2006;333:701, doi: 10.1136/bmj.333.7570.701). Correspondents took them to task over their specific examples (misoprostol for postpartum haemorrhage and circumcision for preventing HIV transmission, but, interestingly, not for oral rehydration therapy) or simply for arguing that “obviousness” was a reasonable criterion.

    This week Paul Glasziou and his colleagues tease out some criteria for when observations may speak for themselves (doi: 10.1136/bmj.39070.527986.68). Their analogy is that of the signal to noise ratio: “how much difference between the treatment outcome (signal) and the natural outcome (noise) is enough? We know that confounding is common and often not obvious; indeed, this was the basis for inventing randomised controlled trials.” But there may be some cases where the signal is strong enough. They suggest that a relative rate between the effects and the natural outcome of over 10 is highly likely to reflect a real treatment effect. One of their examples is the “mother's kiss” technique for removing a foreign body from an infant's nostril: this, they show in a single case, has a rate ratio of 1440. Will they get a better hearing than Potts et al?

    Another controversy revisited in this week's issue is how healthcare organisations should conduct their relationships with the pharmaceutical industry (doi: 10.1136/bmj.39119.519664.BE). Michael Day seems to tell a simple story: the WHO can't take money from commercial organisations, but it can take it from patients' organisations. So when a WHO division seeks help in funding a report on neurological diseases it talks to a patient organisation, who in turn asks a drug company, which offers funds—which WHO then suggests would have to be seen to come from the patient organisation not the drug company. Except in this case the drug company didn't like the lack of transparency and withdrew the funds and the WHO official said he never meant to ask for the money. The picture is muddied because drug companies are sensitive about their relationships with patient organisations and WHO is cautious of the constraining influence of some relationships with industry.

    While some people worry about where money comes from, others worry more about how wisely it is spent. In this week's Head to Head Roger England piles up evidence that spending on HIV is disproportionate to the burden of disease it causes, to the proportion of deaths, to the costs per DALY averted (doi: 10.1136/bmj.39113.402361.94). And “because HIV interventions are not integrated into health services this excessive spending is not effective.” Paul de Lay and colleagues, however, point out that 31% of communicable, maternal, perinatal, and nutritional conditions were attributable to HIV and by 2030 it will be the third highest contributor of DALYs globally (doi: 10.1136/bmj.39113.539595.94) . They both agree on the need for stable funding for all international health and development, and England suggests that the Global Fund for AIDS should abandon disease specific support to become that fund.