DOES science need cleaning up? Some people think so. A panel convened by the US government鈥檚 National Institutes of Health is looking into tighter rules on collaborations between NIH scientists and outside companies. The move follows an article in the Los Angeles Times in December which alleged that NIH researchers had been collecting fees and stock options without publicly disclosing the ties to the companies concerned.
More recently in the UK, Andrew Wakefield, the doctor who first proposed a link between autism and the MMR vaccine, has faced charges that he failed to disclose a possible financial conflict of interest when he submitted the 1998 paper that triggered the scare. And last week, the Committee on Publication Ethics (COPE), which represents scores of leading medical journals, released summaries of cases brought to its attention in 2003. These included researchers submitting unethical or dubious research for peer review, and even attempts to bribe editors (see 鈥淢edical editors to take tougher line鈥�).
Whatever the truth of individual cases 鈥� and in many we do not have all the facts 鈥� they fuel a perception that financial interests and competitive career pressures are so straining the integrity of research and the peer-review system that more policing is needed. COPE鈥檚 draft code of conduct for editors is a step in the right direction. Editors who sign up to it will be obliged to vet papers more closely for undisclosed conflicts of interest. And ignoring dubious science or misconduct simply by rejecting the paper will no longer be an option.
Advertisement
But more needs to be done. In recent years, peer review has been so badly tarnished by high-profile retractions and outright fraud that nothing short of wholesale reform will do. And surprisingly, there is no code of conduct for journal referees, nor are their assessments made public. When controversial but crucial health research is published, shouldn鈥檛 the public have access to all the information?
Perhaps the most profound conflict of interest in science today is that drug companies can choose which clinical findings they publish and which stay secret. Despite evidence that this practice can cause real harm, governments continue to allow it (see 鈥淚n the dark鈥�).
Whatever shape reform takes, it would be a mistake to make peer review mandatory before findings are disseminated to the public. When the UK government announced in 1996 that its advisers had found a new type of CJD linked to mad cow disease, no paper had even been drafted let alone peer reviewed. Withholding the findings because they were unpublished would have been wrong. We want peer review to be reliable; not for it to be used as an instrument for censoring inconvenient findings.
We must also recognise that peer review will never be perfect. It is plain wrong to assume that a result or theory must be true because it has been peer reviewed. Witness the row earlier this year, following publication of a study reporting dioxin and PCB levels in salmon. Like it or not, disputes, uncertainties and honest errors will always be part of science. No magic wand 鈥� or truncheon 鈥� will make them go away.