Observations Open Data Campaign

We are asking more questions about NICE appraisal of Tamiflu

BMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e8420 (Published 12 December 2012) Cite this as: BMJ 2012;345:e8420
  1. Michael Rawlins, chairman, National Institute for Health and Clinical Excellence

Dear Fiona

Thank you for your kind words as my tenure of office draws to a close. To have had the opportunity to chair NICE, since its inception, has been the most rewarding part of my professional career. Any success, however, has been a result of the commitment and expertise of the staff, the massive contributions of the professions, as well as (of course) support from the BMJ.

Your premise that NICE does not, in its technology appraisals programme, require access to all clinical trial data is wrong. On the contrary, we require full access to all the information that is available to marketing authorisation holders of medicines when they are subject to a NICE technology appraisal. The medical director of the company involved in making an evidence submission to NICE is asked to sign a declaration to confirm “that all relevant data pertinent to the technology appraisal has been disclosed to the Institute.” Pharmaceutical physicians are very well aware of their obligations, and if a medical director were knowingly to withhold relevant data, the institute would report the matter to the General Medical Council. You will recall that some years ago a pharmaceutical physician was erased from the medical register for failing to meet his obligations, albeit in a matter relating to an advertisement rather than a NICE appraisal. The institute’s staff also examines the European Medicines Agency’s “European Public Assessment Reports” to confirm that we have been provided with all relevant information.

I am not sure what additional measures we might take to be certain that we have access to all relevant data. The notion that legislation would help is, I am afraid, naive. From my experience as chair of the Committee on Safety of Medicines (during the 1990s), in those few instances of failure to disclose relevant data, fault lay not with the UK subsidiary but with the main company based outside the UK. As the Conservative MP Stephen Dorrell reminded us earlier this week, when Andrew Dillon and I gave evidence to the Health Select Committee, the British Parliament’s reach stops at the English Channel!

Nor can I accept your premise that we fail to make public the information on which our decisions are based. As you know, there is an expectation that the guidance we produce on new medicines is provided as close as possible to their launch in the UK. It is for that reason that we receive evidence submissions several months in advance of receipt of marketing authorisation. We therefore allow companies to designate some of the information as provided in confidence. This allows for subsequent publication in peer reviewed journals as well as for companies to provide data that is commercially sensitive.

The arrangements for the submission of confidential information are laid out in our process documents.1 We expect individual companies to follow these guidelines, and we spend a significant amount of time instructing companies to minimise the amount of information marked as confidential. When a submission concerns unpublished clinical data, the agreement specifically requires companies, as a minimum, to release what would normally be included in a CONSORT (or PRISMA) compliant abstract; and we state that this will be publicly disclosed. Our agreement further allows NICE to “quote publicly from either a full report or an abstract of unpublished trials, where the date of release, by NICE, of data from such reports/abstracts is not less than 12 months after the sign-off by the company of the trial report.”

In respect of your comments about our appraisals of oseltamivir, I am making further enquiries about the details of the assessment. As you know, the assessment report was prepared by the Centre for Reviews and Dissemination at the University of York, and I will write again when I have this additional information. It is clear, however, from the monograph2 underpinning TA 58 that rather than study M76001 having been completely excluded, data from the trial were used as part of an analysis of complication rates. The monograph3 that informed TA 168 includes the M76001 trial, citing an abstract by Treanor (2000b) and additional data provided by Roche as source of the information. The authors of the 2009 monograph indicate in their discussion that publication bias, and particularly time lag publication bias, cannot be ruled out, and they tested the impact of publication bias (excluding unpublished trials) on the results of the meta-analyses. I have not yet found reference to your suggestion that the “reviewers’ extraction sheets were filled in by Roche.”

Without full consideration of what you suggest has been uncovered by the Cochrane review team, it would be inappropriate for NICE to withdraw its guidance on amantadine, oseltamivir, and zanamivir for the treatment of influenza. The additional clinical trial data might, for example, support the use of oseltamivir under the circumstance we have already proposed. In that case, patients would be damaged by precipitate withdrawal of our guidance.

I can reassure you, though, that we will review all new information that comes to light, and consider whether it is necessary to reappraise oseltamivir, as well as its competing interventions, earlier than currently planned.


Cite this as: BMJ 2012;345:e8420