As healthcare costs continue to spiral upwards, regulators, doctors and patients more than ever rely on the integrity of clinical trials to assess the effectiveness of expensive novel therapies. But what if study results are skewed?
Governments around the world are stockpiling billions of dollars’ worth of Roche’s antiviral drug Tamiflu to fight a potential influenza epidemic. But doctors cannot be sure that the therapy works to prevent outbreaks because the company has for years refused to disclose study data.
The common industry tactic of sweeping undesirable results under the carpet is known as publication bias. If trial investigators manipulate their research, new treatments are not only riskier, they may also fail to offer patients any benefit, sceptical physicians warn.
Investigators may move threshold values to set easier-to-reach goals or use ambiguous statistical methods to emphasise positive effects and downplay side effects – and if the findings are unwelcome the files will likely end up in the researchers’ drawers, because nobody likes negative results.
“Doctors have a horror of negative results. They don’t publish them,” says Jean-Francois Cuttat, a Lausanne surgeon. He explains that investigators prefer to clean out data that doesn’t jibe, tweaking and streamlining the results.
“Companies don’t have any interest in publishing results about failures of certain blockbusters,” Hermann Amstad, general secretary at the Swiss Academy of Medical Sciences, seconds Cuttat. “And medical journals are also more keen to accept trials with positive findings for publication.”
About 70,000 clinical trials are conducted at any given time worldwide, but few of them bring new evidence, because many only serve marketing purposes, explains oncologist Reto Obrist, a counsellor at health regulator Swissmedic.
Today, only just over half of trials are published, a bias distorting results and complicating doctors’ choices, he said, adding that many trials end up not being published simply because the pharmaceutical industry doesn’t like them.
A new Swiss federal law regulating research on humans, which has been effective since the beginning of the year, aims to remedy some of those ills. All clinical trials conducted in Switzerland now have to be registered and investigators have to comply with additional ethical rules and fair trading practices to assure transparency and research quality.
For many years Roche refused to release detailed data on its influenza drug Tamiflu, which has been stockpiled around the world, arguing that laypeople would not be able to judge it correctly. Reviewers of independent research organisation Cochrane, who claimed that 60% of data had been withheld, received data from Roche in 2009, but still found evidence of bias.
Since 2013, Cochrane is able to review all studies in relation to Tamiflu, but Roche edited the material to “ensure patient confidentiality and protect commercial interests”. Roche today says it supports calls for more transparency. Upon request it provides researchers with trial data going back to 1998. An independent advisory group is analysing all data to identify any unanswered questions on the influenza drug.
Early this year Japan’s Health Ministry filed a criminal complaint against Novartis in Japan for allegedly misleading consumers through advertisements. The ministry said that some of the research conducted at Japanese universities was falsified to support the benefits of the blood pressure drug Diovan. Novartis said it is cooperating with the authorities and that it has implemented corrective measures to strengthen governance.
In early 2011, Actelion announced that it was halting the development of an experimental insomnia drug, almorexant, because of an undisclosed possible safety issue. According to an article in Forbes in December 2013, Actelion failed to make complete information available to other researchers and other companies developing similar drugs for several years.
“Today, the general recommendation is that companies should only do things that they don’t mind reading about in the papers the following day,” says Annette Magnin, head of the Swiss Clinical Trial Organisation. Misconduct is not only ethically reprehensible, it also is a financial burden and hurts the company’s reputation, she said.
The Swiss pharmaceutical industry, which accounts for 6% of the country’s gross domestic product, has in the past resisted calls for wider disclosure and openness, particularly when it comes to study results. It relies on a self-imposed code of conduct drafted by its trade association ScienceIndustries.
But many companies have started to voice their support for greater transparency over the past few years and have recently changed their policies.
Roche is now publishing trial results, releasing full clinical study reports via regulators and extending access to clinical trial data for third-party researchers. With regard to Tamiflu’s effectiveness, a group of independent researchers is reviewing the data, the “robustness and integrity” of which Roche says supports its “efficacy and safety”.
“We understand and support calls for our industry to be more transparent about clinical trial data with the aim of meeting the best interests of patients and medicine,” said Roche Pharma Chief Operating Officer Daniel O’Day in 2013. “At the same time, we firmly believe that health authorities need to remain the gatekeeper for drug assessment and approval.”
Drug makers stress that full disclosure also has its drawbacks. They remain concerned about the protection of patient data and intellectual property if all data were to be disclosed to everybody, said Thomas Cueni from Interpharma. For Cueni, self-regulation is sufficient.
“The pharmaceutical industry believes it is well regulated,” Cuttat said. “But it doesn’t bode well if once more the authorities condemn companies to disclose information and become more transparent. This sounds like something we recently heard from the banking industry.”
Pharma companies active in the US have had to publish payments to doctors under the Physician Payments Sunshine Act since September 2013. Companies have so far disclosed more than $2 billion worth of payments to health professionals. In 1997, the US congress passed a law requiring trial registration and publication, and in 2000 the US authorities made the clinical trials website clinicaltrials.gov publicly available.
From 2015, researchers in the European Union Europe will have to register clinical trials before they start and publish summary results within a year after a study ends. The European Parliament and the Council of Ministers still have to approve the new legislation.
According to the new Swiss law on human research, companies have had to register their trials on a portal of the Federal Office of Public Health since January 2014, but publication of results is not specified.
Swiss pharmaceutical companies are also encouraged to comply with ScienceIndustries’ pharmaceutical code. Unlike the US, where the data is stored centrally, drug makers in Switzerland will have to publish their conflicts of interest on their own websites from 2016, but there won’t be any sanctions because disclosure is not provided for by Swiss statute.
The Academy of Medical Sciences is one of the only major organisations in Switzerland supporting the alltrials initiative, which calls for companies to make all clinical trial data – including results – accessible, even for older trials.
Swiss companies’ ethical conduct has come under fire in recent years. Not only has the financial industry been battered by negative headlines about manipulations, corruption and secrecy, the drug industry has also lost some of its shine, Obrist said.
To be fair, global companies can’t afford to be discredited and already comply with international codes of conduct and legislation in the markets they’re active in, Hermann said. But they all seem to have to atone for past sins. All benchmark drug companies have had their share of negative headlines.
Companies, such as Roche in the case of Tamiflu and Novartis and its blood pressure drug Diovan (see infobox), have come under scrutiny for spinning their trials and omitting or delaying the publication of detrimental data. Obrist hopes that these recent scandals, which fuelled public calls for more transparency, will help promote overdue changes.
“Today, the pharmaceutical companies are only barely ahead of banks in the popularity scale,” said Obrist. “It will take a long time to mend their chipped image and win back trust.”
The United States is leading the way when it comes to disclosure and transparency in research, and Europe is lagging behind. In the US, drug trials have had to be registered since 1997 and drug companies must already disclose their conflicts of interest. Switzerland has now made a start to register trials and Europe will follow next year.
For sceptics, existing and planned legislation in Europe is late and doesn’t go far enough. Organisations like “No free lunch” and proponents of the AllTrials campaign call for full transparency, warning that because of “expensive pseudo innovations” without proven additional benefit, patients can no longer trust their doctors.
“Drug companies claim that patients are their top priority, but their financial interests cannot be ignored,” said David Klemperer from No free lunch Germany. “In order to sell products in a market worth $1.2 trillion in 2017, the industry systematically works the doctors, politicians and the media to lower awareness of bad data and spruce up their image.”
Transparency advocates echo Ben Goldacre, who writes in his book Bad Pharma that the evidence physicians use to decide what drugs to approve and prescribe is “hopelessly and systematically distorted” because companies withhold results from researchers, doctors, regulators and patients.
Time will tell whether the new rules will succeed in reducing some bias.