Publication Bias and Public Trust - Poisoning the Well

Biologically Determined is always in search of a challenge, and so in this post we are going to try and get you animated about publication bias. We will also try to explain why so many scientists already are, and why they have formed movements such as All Trials that try to address this bias.

Scientists, paid for by private firms, carry out trials and studies of medicines and therapies and the results are published in peer-reviewed journals.  All of this is intended to ensure that a drug is safe and effective before it can be sold. For many drugs, the placebo effect is large relative to the actual effect, so multiple studies are necessary to show that a drug works.  If a firm can carry out many studies and publish only the positive results, it can market drugs that may not be safe or effective. The pharmaceutical industry has been doing just this.

The All Trials campaign is a heartening example of individual scientists standing up for integrity. The movement appears to be gathering momentum, with a committee recently assembled by the British House of Commons recently endorsing universal trial registration and describing the current lack of transparency as "unacceptable". 

The classic example of publication bias concerns the drug Lorcainide. Lorcainide was trialled in 1980 to treat arrhythmia in heart attack patients. Of the 50 people given the drug in the initial trial, 10 died, compared to only one in the control group – a huge increase in mortality.  Lorcainide was discontinued and the results were left unpublished. However other companies developed similar drugs over the following decade. These arrhythmia drugs, like Lorcainide, dramatically increased the mortality rate among heart attack patients. An estimated 100 000 unnecessary deaths occurred before these fatalities were linked to the arrhythmia drugs. If the results of the Lorcainide trial had been made public by law, many lives could have been saved. Publication bias in recent years has also fueled (possibly unnecessary) controversy over the efficacy of both the antidepressant Reboxetine and the antiviral Tamiflu, as the companies involved refuse to publish all of their data.

Despite their embarrassing track record, representatives of pharmaceutical companies claim that they are "firmly committed to enhancing public health" and are doing nothing wrong. The reality is that pharmaceutical companies are legally mandated to generate value for their shareholders. They are ‘institutionalised psychopaths’ - charged with maximising profit. Expecting them to act in the patients' interest, in the absence of regulation, is ridiculous. If pharmaceutical companies can exaggerate the efficacy of their drugs by cherry picking the results they publish, they will. Strict regulation in this case the only remaining option - other means have already failed to produce transparency. 

Considering this, news that pharmaceutical companies are exerting pressure on regulators to let them continue shouldn't come as a surprise. A leak published in a British newspaper, The Guardian, reveals some of the underhand measures undertaken by pharmaceutical companies to exert this pressure. They intend to 'mobilise' patient groups to oppose efforts in order to make all clinical trials public. Patient groups are concerned that efforts to make data public will violate their privacy as patients. This concern is a red herring - no campaign anywhere calls for individual patient data to be released - only anonymised data. Protests that companies' interests would be prohibitively damaged by the release of information are also false - GlaxoSmithKline has demonstrated this by consenting to publish even more than the All Trials campaign demands.

Opponents of the All Trials campaign would like to portray it as a further burden on an already over-regulated industry. It can be reasonably argued that the drugs market is already over-regulated, causing needless delays and expense. However bad regulation is not a solution to over-regulation.

Imagine a world in which pharmaceutical companies are required to publish all the evidence relevant to their customer’s health. In this world, how many people would want the system to be changed to allow companies to be able to hide some of this evidence? The current situation only seems reasonable because it is the status quo.

What the industry fails to understand or account for with their resistance to the All Trials campaign and other efforts is that trust, once lost, is difficult to regain. Cases like the Hindenburg disaster or Chernobyl illustrate what a powerful effect even a single incident can have on public trust in a technology. Not enough effort has been made to underline the importance of credibility in medicine and how pharmaceutical companies are gambling with it.