Intended for healthcare professionals


An amnesty for unpublished trials

BMJ 1998; 317 doi: (Published 19 September 1998) Cite this as: BMJ 1998;317:763

One year on, many trials are unregistered and the amnesty remains open

  1. Ian Roberts, Director. (Ian.Roberts{at}
  1. Child Health Monitoring Unit, Department of Epidemiology and Public Health, Institute of Child Health, London WC1N 1EH

    A year ago the editors of 50-100 medical journals called an amnesty for unpublished trials.1 Investigators with unreported trial data were invited to register their trials by completing an unreported trial registration form. The aim was to tap the silent subterranean pool of unpublished research and, by bringing these data to the surface, to increase the power of systematic reviews and reduce the effects of publication bias. In this respect the amnesty was a flop. To date, only 165 trials have been registered2—a drop in the ocean of unpublished research. But perhaps this is being too negative. Most of the 32 000 patients participating in these trials would have done so in the belief that they were contributing to medical knowledge, and, as a result of the amnesty, their contribution has been realised. At least one of the registered unpublished trials has been incorporated into a systematic review.3

    The idea for the amnesty did not win universal support. One editor felt that studies that were not peer reviewed could not possibly contribute to systematic reviews and said the amnesty would “encourage a vast gemische of junk.”4 Others objected to the use of the word amnesty, arguing that it implied serious wrongdoing. Oddly enough, if the amnesty made any contribution at all, it was because the implication was just that. To give amnesty means to give pardon for an offence. Withholding information about treatment effectiveness is bad for patients because it can lead those caring for patients to make inappropriate decisions.5 This is wrongdoing—and because so many journal editors said so the message got through.

    Probably the first to read the writing on the wall were those in the pharmaceutical industry. The Association of the British Pharmaceutical Industry responded immediately by acknowledging the need for industry to review its publication policy.6 Shortly afterwards GlaxoWellcome announced its intention to work with Dr Iain Chalmers of the UK Cochrane Centre to register clinical trials and to make available the results of these studies until they are published in full (GlaxoWellcome press release 12 November 1997). Planning is now underway to make this a reality.

    On the other hand, the government and drug regulatory agencies, notably the Medicines Control Agency, remain unconvinced about the need for freedom of drug information. They claim that data relating to clinical trials are commercially valuable and are exempt from disclosure unless there is an overriding risk to public health or safety. Their stance is surprising considering that the agency is responsible for “safeguarding public health by ensuring that all medicines for human use on the UK market meet appropriate standards of safety, quality and efficacy.” Secrecy has no place in this, and recent events have shown clearly that drug safety cannot be assured by regulatory agencies alone.7

    Prospective registration of clinical trials promises to make publication bias a thing of the past. But in the meantime we must make the best use of the information at hand. Finding out about unpublished studies is a critically important aspect of this. The amnesty remains open.


    View Abstract