Trials and Errors

Imagine an ordinary coin that lands on heads every time it is flipped. You know this, because you have seen evidence of it landing that way over and over again, hundreds if not thousands of times. The company that owns the coin has published evidence of it landing on heads every time. You buy into the coin, trusting it to land the way the evidence told you it would. But, it turns out, the evidence shown was only half the story – the company did not publish the videos of when, roughly half the time, it landed on tails.

Now what if it wasn’t a coin, but a treatment for your patients? What if you trusted the research body to tell you the truth, but then found out that they only published the clinical trials where it performed favourably, and sat on all the others? You would, justifiably, be outraged. Pharmaceutical companies surely cannot present their products in this way – selectively publishing data amounts to misleading evidence and should be treated as professional misconduct, and yet, it is not.

Around half of all clinical trials have not been published, leaving the medical community entirely in the dark as to their results. Many trials, in fact, have not even been registered, as is required by law. Companies can withhold results of studies even when asked for them, and investigations show that trials with negative results are twice as likely to be kept quiet. An analysis of trials conducted from 1999-2007 found that the results of 60 percent of industry-sponsored trials never made it to publication.1

This is the alarming situation that Ben Goldacre, author of Bad Science, and the AllTrials movement are hoping to rectify. They believe that, unless action is taken urgently, valuable information about what has been found in trials could be lost forever. They are calling on governments, regulators and research bodies to implement measures to require the publication of all trials. They call on universities, ethics committees and medical bodies to enact a change of culture, recognise that under-reporting amounts to misconduct, and to police their own members to ensure compliance.

Dr Richard Lehman explains why he has added his name to the campaign: “It is a scandal that doctors like myself often prescribe treatments without knowing their true benefits and harms, because research evidence from human trials has been withheld. Over my 25 years as a GP I have unintentionally spent large sums of NHS money on treatments that did not work, and some patients have suffered avoidable harm.”

Tracey Brown, Director of charity Sense About Science, explains that the reason this situation has been allowed to go on for so long may be because the medical community are used to not knowing everything. There are the ‘known unknowns’– the studies have not yet been carried out, that can be funded and completed in the coming decades, and the ‘unknown unknowns’ – limitations on our understanding, like the extent to which epigenetics might change the entire way drug treatments are administered. But when trials are run and the results kept a secret, this leads to something much more sinister – the ‘unknown knowns’ And neither doctor nor patient can benefit from knowledge they are unable to access.

In 2009 the UK spent over £500 million – that’s around five percent of the total NHS drugs budget – on Tamiflu to treat the H1N1 influenza virus. But the drug’s manufacturer, Roche, has published fewer than half of clinical trials it conducted on Tamiflu. It may, in fact, be no more effective than paracetamol.

There have been some attempts to fix this problem. In the US, the Food and Drug Administration (FDA) has, since 2008, required that the results of all trials be published within one year of completion. But in 2012 it was revealed that as many as 80 percent of trials still do not comply with the law, and not a single fine has ever been issued.
The AllTrials campaigners say that this is nowhere near good enough. Even if companies did comply with the FDA regulations, because most currently prescribed drugs came onto the market before 2008, trial results with the biggest impact on current medical practice would still not be released. For full disclosure, there must be retroactive law calling for publications of past trials.

The online petition created by AllTrials has so far collected over 50,000 signatures, as well as the backing of 200 medical organisations, including the British Medical Journal and the Centre for Evidence Based Medicine. The signatures come from doctors and patients, students and NHS professionals alike. Ross McDowell, a final year medical student at Newcastle University says he signed the petition because: “I had heard of publication bias over the course of my degree, but only recently have come to understand its implications. I find it absurd and indefensible. It may be a challenge to ensure trials are published, but when the stakes are so high, it is a challenge that surely should be accepted.”

Jennifer Durrant, an NHS Portfolio Support Manager in Leicester, signed the petition because of the strain put on the NHS by the current system: “The publication of results, good and bad, will prevent the replication of investment.” She goes on to say: “Participants who have been involved in a study should, as a matter of course, be informed of the results. After all, their time and contribution ensures that research happens at all.”
Ben Goldacre believes that the clock is ticking, and that if changes are not made soon, it may be too late. “Over the next two years,” he explains, “many drugs we commonly use will come to the end of their patent life. It may be even harder to get hold of the information then. Industry and academics must share their trial results. Without this information doctors are misled and patients are harmed.”

Writing for The Guardian in 2012, Goldacre makes the point that, if a company were to delete half the data points in its study to make a treatment look better, it would be misconduct. But failing to publish half the trials is not.

The good news is that the movement has already seen significant breakthroughs in the scientific community. GlaxoSmithKline, which employs 12,500 people in its research departments, has signed up to AllTrials and agreed to publish the results of every trial it has conducted since the company was formed.

But Goldacre is still concerned that unless the rules are changed quickly, less principled pharmaceutical giants will continue to move the goal posts and, ultimately, it will be patients who suffer. By selectively publishing, researchers are effectively changing their minds about what is the primary outcome of their trials after they have been completed. “You cannot change the rules after the game has started,” says Goldacre. But his petition, and the 50,000 names on it, are hoping they can change the rules by which the big boys play.


1. Ross JS, Mulvey GK, Hines EM, Nissen SE, Krumholz HM (2009) Trial Publication after Registration in ClinicalTrials.Gov: A Cross-Sectional Analysis. PLoS Med 6(9): e1000144. doi:10.1371/journal.pmed.1000144

Categories: ARTICLES