Earlier this year my colleagues and I carried out a Cochrane review of the effects of cholera vaccines. The review was large, but because of the data available we concentrated mainly on the older killed whole cell (KWC) variety of vaccines. At the beginning of the review we were aware that cholera KWC vaccines have been discouraged since the 1970s in either endemic or epidemic areas on the grounds that they had low efficacy and short duration (four to six months), required multiple doses, and were less effective in children under 5 than in adults. Important adverse effects were said to occur in 30% of recipients of injected KWC vaccines. Our Cochrane review re-examined the evidence for these statements through a systematic review of randomised and quasirandomised controlled cholera vaccine trials of over 2.5m participants conducted world wide since 1963. The overall efficacy of vaccine compared with placebo was 57% (95% confidence interval 50% to 64%) in the first seven months, 51% (41% to 59%) in the first year as a whole, and 47% (36% to 56%) in the second year. Most of the 17 trials in this comparison used only a single dose of vaccine. Vaccine efficacy in the first year was as great in children under 5 years as in older individuals, but declined to 22% (−7% to 43%) in the second year.
By and large, editors were unaware of any impact of the editorials
A similar pattern of protection was found for both injected and oral KWC vaccines. Short and booster schedules of injected KWC vaccination seemed to induce equivalent protection for two years. While local reactions were common for injected vaccines, systemic reaction rates were low (<13%) for all types of injected and oral vaccines. We concluded that the level and duration of efficacy of KWC cholera vaccines seemed to have been routinely underestimated in the literature and the incidence of adverse effects overestimated. This finding illustrated that anybody judging cholera KWC vaccines on the basis of current wisdom exemplified by editorials, viewpoints, and seminars (reviews) covering the topic in the past two decades would have been fed a diet of misinformation. This in turn made me curious to know the motives and effects of publishing reviews based on non-systematically collected evidence. I carried out a small survey among editors and authors of reviews of KWC vaccines. I identified five reviews on KWC vaccines published between 1991 and February 1998 in important international journals, wrote to editors and corresponding authors, and received 10 answers.
The answers showed that most editors wanted to update and educate their readers by commissioning the review from an author well known in the subject. By and large, editors were unaware of any impact of the editorials. Authors’ answers showed that the majority had written the piece because they had been asked to do so by editors and in assembling the evidence they had used what I call “desk drawer” methods. They had included the evidence that was known to them but had not assembled it or evaluated it systematically, a practice with which most busy authors will sympathise. Authors were also unaware of any impact of the editorials despite requests for reprints. One said that his had been used by the British Council to teach English to Italians.
Other researchers have drawn our attention to the danger of publishing in the columns of major journals authoritative sounding statements limited by the range of one’s knowledge and experience—for example, the paper by E M Antman and colleagues (A comparison of results of meta-analyses of randomised control trials and recommendations of clinical experts. JAMA 1992;268:240-8). In the past no alternative was available.
I believe that the rise and increased availability of knowledge base tools, such as the Cochrane Library, have made this practice increasingly difficult to justify. The example of cholera vaccines is perhaps not the most glaring, but the point is this: if editorials, viewpoints, and updates are written to update and educate then they should contain correct and trustworthy information. Nothing less will do, as misinformation in health care may cost lives. If they are written for other reasons, such as for the benefit of journals, authors, or other parties, then this should be made explicit in the text.
I challenge readers, editors, and authors who dissent with these sentiments to state their reasons. I invite those who agree with my views to put pressure on editors to be explicit about the aims and methods used to commission and write editorials. Our lives may depend on it.