Recent calls in the UK for pharmaceutical companies to make unpublished clinical trial data more widely available – and in particular to the National Institute for Health and Clinical Excellence (NICE) – look unlikely to prompt any drastic changes in policy from industry, the govern
ment or NICE.
The calls come on the back of extensive publicity in the UK over a Hull University study, published in the journal PLoS Medicine, that suggested widely used antidepressants were no better than placebos for the vast majority of patients.
The meta-analysis of both publish
ed and unpublished data from trials of four selective serotonin reuptake inhibitors (SSRIs) – Eli Lilly’s Prozac (fluoxetine), Wyeth’s Effexor (venlafaxine), GlaxoSmithKline’s Paxil/Seroxat (paroxetine) and Bristol-Myers Squibb’s drug Serzone (nefazodone) – showed that, compared with pla
cebo, antidepressants did not have clinically significant effects in mildly depressed patients or in most patients with very severe depression.
It was also claimed the apparent clinical effectiveness of these antidepressants in the small population of extremely depressed patients was more do
wn to a muted response to placebo rather than any notable improvement in response to antidepressants.
Critics of the industry have seized on the fact that the Hull University researchers used Freedom of Information legislation to obtain data on the SSRIs that was filed to regulators in approv
al dossiers but not subsequently made public. Much has been made of NICE’s lack of authority to demand access to unpublished trial data when it is evaluating the cost-effectiveness of medicines for use in the National Health Service.
Moreover, the UK government says it cannot force pharmace
utical companies to publish trial data as this would conflict with EU legislation. That is despite a pledge in the Labour Party’s 2005 election manifesto to “require registration of all clinical trials and publication of their findings for all trials of medicinal products with a marketing author
isation in the UK”, in line with a recommendation from the House of Commons Health Select Committee.
Nonetheless, there are signs of a gathering political impetus to make sure unpublished trial results are available when needed. Talking to The Observer newspaper about NICE’s cur
rent review of its clinical guideline on the management of depression, Ivan Lewis, the minister for care services, called on “any company in possession of any relevant evidence to make it available to NICE.
Failure to do so “would leave the inevitable impression that they have something
to hide”, he told The Observer.
We need access
NICE’s clinical and public health director, Professor Peter Littlejohns, commented: “In order for our guideline development groups to carry out their job effectively we need to have access to all available data. However, w
e cannot force organisations or companies to submit unpublished data.”
All the same, the Health Select Committee’s last report on NICE, published in January, had “highlighted the importance of unpublished data” to the Institute’s work. “The study by Hull University reinforces thi
s point,” Professor Littlejohns added.
“In the case of the depression guideline we asked for all published and unpublished data on SSRIs to be made available so we could accurately assess the role of these drugs in the treatment of depression,” he pointed out. “The update of the depression guideline is underway and we will be looking at the study carried out by Hull University as part of this work.”
The updated guideline is scheduled for publication in June 2009. As Professor Littlejohn noted, NICE’s existing depression guideline, published in 2004, does not recommend the use of SSRIs for the treatment of mild depression. It does, however, set out a role for them in the treatment of moderate and severe depression, as part of a wider package of care including psychological or ‘talking’ therapies.
As NICE explained, Professor Littlejohn’s comments do not signal any departure from current procedure on unpublished trial data. If NICE is aware of any relevant unpublished data, it will ask for them.
While acknowledging the impracticalities of sifting through a potentially “vast” pool of unpublished data for every appraisal, the Institute does want “everything we need to have”, it told PharmaTimes. But it prefers to do so through existing channels, noting that “we do have good working relationships with pharmaceutical companies”.
It was hard to say how often pharmaceutical companies actually did submit unpublished data on request, the NICE spokesperson added. Sometimes the issue was complicated by commercial confidentiality or the potential impact on share prices.
What it requires
As far as the Association of the British Pharmaceutical Industry (ABPI) is concerned, NICE “should get the information it requires”. This did not mean all data (e.g., toxicological studies) were relevant to the Institute’s appraisals, the ABPI said. But companies would be “shooting themselves in the foot” if they did not supply all the requested data, be it published or unpublished.
Ultimately, though, it was down to individual companies to decide what they wanted to release, the association stressed. The ABPI could not coerce them to do so, nor was it aware of any coercion from government to make unpublished data more widely available.
Study data from 2005 onwards were already “covered” under the UK industry’s commitment to the joint position on clinical trial disclosure announced in January of that year by the European Federation of Pharmaceutical Industries and Associations (EFPIA), the International Federation of Pharmaceutical Manufacturers and Associations (IFPMA), the Japanese Pharmaceutical Manufacturers Association (JPMA) and Pharmaceutical Research and Manufacturers of America (PhRMA).
Under this voluntary agreement, the results of all industry-sponsored clinical trials evaluating the safety and benefits of newly approved medicines should be made generally available through free, publicly accessible databases (including the ABPI’s own UK-centred website), regardless of outcome and within one year of approval or one year of completion for post-marketing studies.
This did leave a “lacuna” of unpublished results from before 2005, the association admitted, while insisting it was impractical to put every last chunk of historical raw data in the public domain.
