Healthy Skepticism Library item: 20466
Warning: This library includes all items relevant to health product marketing that we are aware of regardless of quality. Often we do not agree with all or part of the contents.
Publication type: Journal Article
Ed.
Betraying the public over nvCJD risk
The Lancet 1996 Dec 7; 348:(9041):1592
http://www.lancet.com/journals/lancet/article/PIIS0140-6736%2896%2921049-9/fulltext
Abstract:
Tuesday, Nov 26, contrasted the way in which risk information is brokered by scientists and the lay media. In the morning, the UK broadsheet newspaper The Independent had led its front page with an exclusive headlined “After the hype, the scientists’ verdict: CJD to kill hundreds”. In the afternoon, the Royal Statistical Society met in London to discuss the latest information about bovine spongiform encephalopathy (BSE) and its connection to Creutzfeldt-Jakob disease (CJD). The same data are in question: whether or not the 14 known cases of new-variant CJD (nvCJD) arise from eating BSE-infected cattle and are the vanguard of an epidemic.
The Independent reported Dr James Ironside from the CJD Surveillance Unit in Edinburgh, UK, as saying “It looks as though the total number of cases over the whole course of the disease will be in the hundreds, rather than the thousands”. The sharp eye could discern in a dark and unidentified accompanying photograph an abattoir hook and a cloven hoof, which invited reinforcement of the suspected but unproven link between eating beef products and nvCJD. Dr Ironside told The Lancet that he had been quoted incorrectly. At the press conference after the statisticians’ meeting, the expert panel was immediately pressed to predict the “size of the nvCJD epidemic”. The answer was given by Prof Adrian Smith, the President of the society: zero to millions.
What makes that answer any more defensible than the previous estimates fed to the public by the media, which ranged from millions last April, to hundreds of thousands in the summer, to the latest “hundreds”? Zero to millions is correct because it is the best estimate available with the known information. In truth, a useful prediction is impossible. Crucial parts of the equation remain unknown, in particular the incubation period in human beings and the minimum infective dose of the prion organism. The only way to estimate either is by continuing observation. There will be no quick answer. At the current rate of finding new cases, 2—3 further years of observation will be needed, unless there is a rapid increase in finding new cases. Other unknown factors include the extent of the interspecies barrier between cow and human being, and the effectiveness of the 1989 UK offal ban. A new finding, which will restrict the population at risk, is that all the known cases are homozygous for methionine at position 129 of the prion. Only 38% of the total population carry this homozygosity. At the statistical meeting, Prof Smith reaffirmed that no predictions about nvCJD are possible from what is currently known. The same line was taken by another expert in epidemic modelling, Prof Roy Anderson, also on that Tuesday morning at a press conference held by veterinary experts (see Lancet, Nov 30, p 1505).
What a newspaper such as The Independent has done is to assume that its readers cannot cope with the uncertainty of an estimate that runs from none to several millions. It was assumed that the public wants high precision. Yet this belief has led to needless confusion. If the “epidemic” is in the hundreds, the argument goes, the public will be reassured, even though no-one can say for certain which individuals will be affected. If the estimate is in the millions, then the chance of being affected is too high for comfort. Until there is good evidence that such assumptions are valid, we all must strive to balance prudent journalism with impartial accuracy.
The public rightly wishes to know about risks they take. The background in this case is that the UK Government, initially at least, denied any danger from eating contaminated beef products. What is less clear is what the public needs to know: absolute risk estimates or a range reflecting current uncertainties? When “don’t know” is the correct answer, then that is what should be printed. Anything else betrays people’s trust.
